Abstracts

The call for abstracts for poster presentations is now closed and the notifications of acceptance have been sent on 8th May. We received a great number of excellent submissions. We look forward to engaging discussions and the sharing of knowledge at the symposium.

Posters will be exhibited during 15^th– 17^th June at the Kuopio Music Centre (Kuopionlahdenkatu 23, 70100 Kuopio, Finland).  Each author is kindly requested to be present at their posters during the poster sessions I-III, if possible.

Schedule of poster sessions:

Monday, June 15, 2026 15.30–16.00 Poster Session I

Tuesday, June 16, 2026 10.30–11.00 Poster Session II

Wednesday, June 17, 2026 10.00–10.30 Poster Session III

The poster panels will be at the Music Centre on Monday 15^th June at 10 am. The posters should be placed between 10-12 am. If you will arrive later, please inform us in advance. The posters must be removed on Wednesday 17^th June by 16.30. 

If you have any questions, please contact [email protected].

Selected Abstracts

1. Bridging Scales: Adapting RAPSODI for High-Precision Rat Brain MRI-Histology Co-Registration  

Authors: Ebawak Wodajo ¹, Omar Narvaez ¹, Mirabela Rusu ², Jussi Tohka ¹, Alejandra Sierra ¹ 

Author affiliations: ¹ A.I. Virtanen Institute for Molecular Sciences, Faculty of Health Science, University of Eastern Finland, Kuopio, Finland. ² Department of Radiology, School of Medicine, Stanford University, Stanford, CA, USA.  

Rationale: Multimodal co-registration of MRI and histology is essential for linking macrostructural and microstructural information in brain research, but remains challenging due to differences in image contrast, scale, and tissue deformation. We adapted the RAPSODI framework, originally developed for prostate MRI-histopathology alignment, to rat brain histology and ex vivo MRI co-registration.   
Methods: The dataset included naïve, sham, and moderate traumatic brain injury (moTBI) subjects from a lateral fluid percussion model. Ex vivo T1-weighted MRI was acquired at 11.7T with 100µm isotropic resolution; histological sections stained for Nissl and Myelin were imaged at 0.1369µm × 0.1369µm in-plane resolution with 30µm section thickness. Preprocessing included identification of missing sections and tissue deformations, binary mask generation, and histology downsampling for computationally feasible slice-to-volume reconstruction. Affine registration aligned the reconstructed Nissl histology volume to the MRI volume, with accuracy evaluated using anatomical landmark pairs across cortex, ventricles, corpus callosum, and gray-white matter boundaries.   
Results: Qualitative assessment showed strong anatomical correspondence after co-registration. Quantitative evaluation on one sham subject using Euclidean centroid distance yielded minimum, mean, and median errors of 21.2µm, 191.7µm, and 146.2µm, with 73% of landmark pairs within a sub-voxel to low multiple-voxel error range relative to MRI resolution.   
Conclusions: Sub-voxel accuracy indicates that affine registration effectively captures gross anatomical alignment, while remaining errors likely reflect local deformations. These findings demonstrate that RAPSODI is a feasible, promising framework for preclinical neuroimaging. Future work will compare linear vs. non-linear registration and expand the framework to include more subjects, Myelin stains, and diffusion MRI contrasts.  

2. Association between structural changes in brain MRI and future frailty; FINGER sub-study  

Authors: Johanna Pöyhönen (1,2), Hanna-Maria Roitto (1,3,4), Jenni Lehtisalo (1,5), Esko Levälahti (1), Timo Strandberg (2,6,7), Miia Kivipelto (2,8,9,10,11), Riitta Antikainen (7,12), Hilkka Soininen (5,13), Jaakko Tuomilehto (14,15,16), Tiina Laatikainen (9), Alina Solomon (5,8,10), Ruth Stephen (5,8,17), Juha Rinne (18,19), Eric Westman (8,10,11), Tiia Ngandu (1,8,9)  

Author affiliations: 1 Department of Public Health, Finnish Institute for Health and Welfare (THL), Helsinki, Finland; 2 Department of Medicine, Clinicum, University of Helsinki, Helsinki, Finland; 3 Faculty of Medicine and Health Technology, University of Tampere, Tampere, Finland; 4 Department of Geriatrics, Wellbeing Services County of Pirkanmaa, Finland; 5 Institute of Clinical Medicine/Neurology, University of Eastern Finland, Kuopio, Finland; 6 Department of Geriatrics, Helsinki University Hospital, Helsinki, Finland; 7 Center for Life Course Health Research/Geriatrics, University of Oulu, Oulu, Finland; 8 Division of Clinical Geriatrics, Center for Alzheimer Research, Care Sciences and Society (NVS), Karolinska Institutet, Stockholm, Sweden; 9 Institute of Public Health and Clinical Nutrition, University of Eastern Finland, Kuopio, Finland; 10 Ageing Epidemiology Research Unit (AGE), School of Public Health, Imperial College London, London, United Kingdom; 11 Theme Inflammation and Aging, Karolinska University Hospital, Stockholm, Sweden; 12 Medical Research Center Oulu, Oulu University Hospital, Oulu, Finland; 13 Neurocenter, Department of Neurology, Kuopio University Hospital, Kuopio, Finland; 14 National School of Public Health, Madrid, Spain; 15 South Ostrobothnia Central Hospital, Seinäjoki, Finland; 16 Department of Public Health, University of Helsinki, Helsinki, Finland; 17 FINGERs Brain Health Institute, Stockholm, Sweden; 18 Turku University Hospital, Turku, Finland; 19 Turku PET Centre, University of Turku, Turku, Finland.  

Rationale: There are established brain magnetic resonance imaging (MRI) biomarkers related to dementia. However, the association between MRI findings and frailty, another major geriatric syndrome, remains unclear. We aimed to investigate whether brain MRI biomarkers associate with future pre-frailty or frailty over 11 years.   
Methods: Baseline MRI data were available for 120 participants, aged 60-77 years, in the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) trial. Frailty status (Fried phenotype) was measured at baseline, and at 2, 7, and 11 years. Assessed MRI biomarkers included total gray matter volume, left and right hippocampal volume, Alzheimer’s disease signature thickness, and periventricular and deep white matter hyperintensities (WMH) by Fazekas score. The risk of future pre-frailty/frailty was evaluated using a mixed effect logistic regression model.   
Results: MRI biomarkers were not associated with pre-frailty/frailty at baseline. Greater left hippocampal volume was associated with a lower risk of future pre-frailty/frailty at 2 (OR 0.45, 95% CI 0.21-0.97, p=0.042) and 7 years (OR 0.41, 95% CI 0.19-0.85, p=0.017). Higher periventricular WMH score was associated with higher risk of future pre-frailty/frailty at 2 years (OR 2.16, 95% CI 1.01-4.62, p=0.048). All associations were independent of baseline cognition.   
Conclusions: Smaller left hippocampal volume and higher periventricular WMH score may indicate risk of future frailty, in addition to dementia, highlighting the need for prevention strategies targeting both conditions. The predictive value of MRI biomarkers was independent of cognition, supporting possible shared pathophysiological mechanisms of frailty and cognitive impairment.  

3. Law Clinic: Integrating Legal and Clinical Perspectives on Legal Capacity and Rights in Dementia  

Authors: Kaisa Näkki 1; Jenni Niinimäki 2; Mervi Issakainen 1; Noora Suhonen 3, 4; Henna Nikumaa 1; Juho Kalapudas 2, 5; Kaijus Ervasti 1; Eino Solje 2, 5; Anna Mäki-Petäjä-Leinonen 1  

Author affiliations: 1 Law School, University of Eastern Finland, Joensuu; 2 Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio; 3 Neurocenter, Neurology, Oulu University Hospital, Oulu; 4 Research Unit of Clinical Medicine, Neurology, University of Oulu; 5 Neuro Center – Neurology, Kuopio University Hospital, Kuopio  

Rationale: Progressive dementias challenge patients’ overall legal position. Cognitive and neuropsychiatric symptoms may lead to legal problems eg. related to legal capacity, service access, family matters, and guardianship. Legal aspects remain poorly integrated into clinical research and service systems. This project develops an interdisciplinary Law Clinic model that combines medical, neuropsychological and legal expertise. The aim is to establish a scalable research and service platform for identifying legal needs among people with dementia and linking them with clinical findings and traits.   
Methods: A pilot Law Clinic was launched in 2023 at the Brain Research Unit of the UEF. Twenty patients with dementia and their adult relatives (n=40) participated. Patients underwent structured neuropsychological assessments. Relatives were interviewed regarding functional abilities and behavioral symptoms. Both groups participated in legally focused interviews mapping legal issues, service access and decision-making challenges. Clinical data, interview material and health records were integrated for analysis.   
Results: Medical–legal approach is feasible, ethically robust and meaningful for participants. The pilot revealed previously unrecognized legal needs linked to dementia related symptoms. Participants reported increased understanding of legal issues and felt empowered. The pilot supports expanding recruitment and refining standardized assessment and interview protocols.   
Conclusions: The Law Clinic offers a novel framework for studying and addressing legal capacity and rights in dementia. By linking clinical assessments with legal analysis, it helps identify legal vulnerabilities and supports policy and service development and legislative discussion. The project establishes a promising platform for future research and for improving participation, autonomy and rights realization.  

4. Frontotemporal dementia patient-derived iPSC neurons show synaptic dysfunction and DNA damage  

Authors: Nadine Huber1, Tomi Hietanen1, Sami Heikkinen2, Anastasia Shakirzyanova1, Dorit Hoffmann1, Hannah Rostalski1, Ashutosh Dhingra3, Salvador Rodriguez-Nieto3, Sari Kärkkäinen4, Marja Koskuvi5, Eila Korhonen1, Päivi Hartikainen6, Katri Pylkäs7, Johanna Krüger8,9,10, Tarja Malm1, Mari Takalo2, Mikko Hiltunen2, Jari Koistinaho11, Anne M. Portaankorva12, Eino Solje4,6, Annakaisa Haapasalo1  

Author affiliations: 1 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 2 Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland 3 German Center for Neurodegenerative Diseases (DZNE), Tübingen, Germany 4 Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 5 Neuroscience Center, University of Helsinki, Helsinki, Finland 6 Neuro center, Neurology, Kuopio University Hospital, Kuopio, Finland 7 Translational Medicine Research Unit, Medical Research Center Oulu and Biocenter Oulu, University of Oulu, and Northern Finland Laboratory Centre Nordlab, Oulu, Finland 8 Research Unit of Clinical Medicine, Neurology, University of Oulu, Oulu, Finland 9 Neurocenter, Neurology, Oulu University Hospital, Oulu, Finland 10 Medical Research Center, Oulu University Hospital and University of Oulu, Oulu, Finland 11 Helsinki Institute of Life Science, University of Helsinki, and Drug Research Program, Division of Pharmacology and Pharmacotherapy, University of Helsinki 12 Clinical Neurosciences, University of Helsinki, Helsinki, Finland  

Rationale: Synaptic dysfunction is suggested to play a major role in frontotemporal dementia (FTD). Some of the clinical symptoms can be associated to disturbances in the synaptic function. Here, we aim to study the underlying molecular mechanisms leading to synaptic disturbances in induced pluripotent stem cell (iPSC) –derived neurons from FTD patients carrying the C9orf72 repeat expansion (C9-HRE), the most common genetic etiology for FTD, and sporadic FTD patients in comparison to healthy controls.    
Methods: Human iPSCs were generated from C9-HRE-carrying and non-carrying FTD patients and healthy individuals from a well-characterized Finnish FTD cohort. The iPSCs were differentiated to cortical neurons using the dual SMAD inhibition approach.   
Results: iPSC neurons derived from C9-HRE carriers developed typical C9-HRE-associated hallmarks, including RNA foci and dipeptide repeat accumulation. All FTD neurons demonstrated increased TDP-43 cytosolic localization, p62/SQSTM1 accumulation, and changes in nuclear size and morphology. In addition, the FTD neurons displayed reduced number and altered morphology of dendritic spines together with significantly altered synaptic function. These structural and functional synaptic disturbances were accompanied by upregulation of genes related to synaptic function, including synaptic signaling and glutamatergic transmission, in FTD neurons as compared to control neurons. Pathways involved in DNA repair were significantly downregulated in FTD neurons. Only one gene was differentially expressed between the sporadic and C9-HRE-carrying FTD neurons.   
Conclusions: Our results show that the iPSC neurons recapitulate pathological changes found in FTD patient brain and strongly support the hypothesis that synaptic dysfunction is a crucial contributor to disease pathogenesis in FTD.  

5. Neuromodulation of hiPSC-Derived Neurons by Electrical Stimulation to Reduce Hyperexcitability in Frontotemporal Dementia  

Authors: Alberto Sesena-Rubfiaro1, Nadine Huber1, Viivi Pekkala1, Rebekka Fuchs1, Anssi Pelkonen1, Mireia Gómez-Budia1, Tarja Malm1, Eino Solje2,3 and Annakaisa Haapasalo1  

Author affiliations: 1. A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 2. Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 3. Neuro Center, Neurology, Kuopio University Hospital, Kuopio, Finland  

Rationale: Electrical brain stimulation is explored as a potential therapeutic strategy for frontotemporal dementia (FTD). For example, repetitive transcranial magnetic stimulation (rTMS) can modulate brain activity through repeated magnetic pulses that induce long-lasting synaptic changes. However, clinical outcomes remain inconsistent. We have developed an in vitro capacitive electrical stimulation platform mimicking the electric field profile of TMS in patients. This system enables controlled stimulation of human induced pluripotent stem cell (hiPSC)–derived neurons on microelectrode array (MEA) chips, enabling electrical stimulation and monitoring of neuronal network activity.   
Methods: hiPSC-derived neurons from healthy controls, sporadic and C9orf72 hexanucleotide repeat expansion (C9-HRE)-carrying FTD patients were stimulated between two parallel carbon electrodes delivering biphasic pulses at 10 Hz. Each stimulation train lasted 5 s, followed by a 25 s interval; 20 trains were applied/session. Stimulation was performed once daily, three times/week, for five weeks. MEA recordings were collected weekly to assess changes in neuronal network dynamics.   
Results: Upregulation of c-Fos demonstrated effective activation of hiPSC-derived neurons upon stimulation. MEA recordings showed that stimulation did not adversely affect neuronal viability, as indicated by sustained presence of single action potentials and burst activity. C9-HRE neurons showed higher neuronal network activity compared to sporadic or control neurons, but the activity decreased and synchronized in all neurons following stimulation.   
Conclusions: These findings suggest that capacitive electrical stimulation can modulate neuronal network activity and mitigate the hyperexcitability in FTD neurons. Our system provides a feasible platform to test the effects of different stimulation paradigms on patient-derived neurons.   

6. Cellular responses to traffic-related ultrafine particles and the role of NRF2 signaling as a mitigation strategy against toxicological effects  

Authors: Donya Behzadpour1*, Riikka Lampinen1, Laura Mussalo1, Laura Salo2, Mika Ihalainen3, Olli Sippula3, Topi Rönkkö2, Pasi Jalava3, Katja M. Kanninen1  

Author affiliations: 1. Department of Health Sciences, A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 2. Department of Engineering and Natural Sciences, Physics Unit, Aerosol Physics Laboratory, Tampere University, Tampere, Finland 3. Department of Environmental and Biological Sciences, Inhalation Toxicology Laboratory, University of Eastern Finland, Kuopio, Finland  

Air pollution is a global health concern with serious cellular consequences. Traffic emissions are a major source of ultrafine particles (UFPs; ≤100 nm), which are particularly harmful because their small size allows them to penetrate biological barriers and reach the central nervous system. Cellular redox balance is tightly regulated by signaling pathways, with nuclear factor erythroid 2-related factor 2 (NRF2) playing a central role in antioxidant and anti-inflammatory responses.   
  
Exposure to UFPs can trigger toxicity, inflammation, and oxidative stress. Although UFPs can activate NRF2 as a protective response, this defense is often insufficient. Therefore, pharmacological pre-activation of NRF2 may offer a strategy to reduce UFP-induced cellular damage. However, the relationship between air pollution and NRF2 activation is not well understood, particularly in the upper respiratory tract. Cells in the olfactory mucosa, located in the nasal cavity, are directly exposed to airborne pollutants and represent a relevant but understudied target.   
  
This study investigates how UFPs from different sources influence NRF2 signaling in primary human olfactory mucosa cells using air-liquid interface and submerged culture models. NRF2 activity will be measured using transcription factor assays, while expression of target genes and proteins will be analyzed by qPCR and Western blot. To assess protective effects, cells will be pretreated with NRF2 activators such as sulforaphane, dimethyl fumarate, and curcumin prior to UFP exposure.   
  
Overall, the study aims to evaluate NRF2 activation as a strategy to mitigate UFP-induced cellular damage and clarify its role in cellular defense mechanisms.   

7. Patient-based in vitro model of cortical iPSC-neurons for studying the effects of C9orf72 repeat expansion on synaptic dysfunction in frontotemporal dementia  

Authors: Viivi Pekkala1, Alberto Seseña Rubfiaro1, Dorit Hoffmann1, Tomi Hietanen1, Eino Solje2,3, Nadine Huber1, Annakaisa Haapasalo1  

Author affiliations: 1 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 2 Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland 3 Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland  

Rationale: The hexanucleotide repeat expansion of the C9orf72 gene (C9-HRE) is the most common genetic cause of frontotemporal dementia (FTD). The molecular mechanisms of FTD are still not fully understood. Thus, elucidating the underlying pathophysiology of C9-HRE-associated and non-genetic (sporadic) forms of FTD is essential to increase our understanding of disease mechanisms in the different forms of FTD. Our previous studies suggest deficiencies in neuronal function, including alterations in synaptic structures and neurotransmission, in FTD patient-derived neurons.   
Methods: Induced pluripotent stem cells (iPSCs) obtained from C9-HRE-carrying and sporadic FTD patients as well as healthy control individuals are differentiated into cortical neurons using overexpression of the NGN2 gene via viral transduction. Neuronal co-cultures with human primary astrocytes are used for studying changes in the synaptic functions, including expression of synaptic genes, number of synaptic puncta, and electrophysiological properties. RNA-sequencing will be performed for identifying altered biological pathways in FTD in comparison to healthy neurons.   
Results: The NGN2-based neuronal differentiation protocol has been validated and optimized for all used iPSC lines. C9-HRE-carrying neurons expressed typical hallmarks e.g., RNA foci and DPR proteins. All neurons expressed pre- (Syn1, Bassoon) and post-synaptic proteins (PSD95, Homer), and the number of synapses was found to stabilize at ~35 days in culture.   
Conclusions: Neurodegenerative disorders are an increasing concern in the aging population and need urgent therapeutic solutions. Although numerous studies have described synaptic dysfunction in FTD, the underlying molecular mechanisms remain insufficiently understood. Understanding disease mechanisms is essential for identifying novel treatment targets or biomarkers.  

8. Psychiatric Morbidity Before Early-Onset Dementia: Longitudinal Evidence from a Population-Based Study  

Authors: Kasper Katisko (MD,PhD)1,2 // Mikko Aaltonen (PhD)3 // Annakaisa Haapasalo (PhD)4// Päivi Hartikainen (MD,PhD)5// Johanna Krüger (MD,PhD)6,7,8 // Eino Solje (MD,PhD)1,5  

Author affiliations: 1=Institute of Clinical Medicine, Neurology, University of Eastern Finland, Kuopio, Finland. 2=Department of Psychiatry, Oulu University Hospital, Oulu, Finland. 3=Law School, University of Eastern Finland, Joensuu, Finland. 4=A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland. 5=Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 6=Research Unit of Clinical Medicine, Neurology, University of Oulu, Oulu, Finland. 7=MRC, Oulu University Hospital, Oulu, Finland. 8=Neurocenter, Neurology, Oulu University Hospital, Oulu, Finland.   

Rationale: Psychiatric conditions frequently intersect with neurodegenerative diseases, yet the long‑term psychiatric trajectories preceding early‑onset dementia (EOD) remain insufficiently characterized. Although associations between psychiatric disorders and dementia in general have been studied, evidence focusing specifically on EOD and its subtypes is limited. This study examined temporal patterns of psychiatric diagnoses before EOD diagnosis.   
Methods: We performed a population‑based longitudinal case–control study covering all new EOD diagnoses from 2010 to 2021 across two Finnish regions. Clinical records from dementia outpatient clinics were systematically reviewed (N=12,490), resulting in 794 validated EOD cases, including Alzheimer’s disease, frontotemporal dementia (FTD), alpha-synucleinopathies, vascular, and other EOD subtypes. For each case, 10 region-, age-, and sex‑matched controls without neurodegenerative diseases were selected from national registries (n=7,930). Psychiatric diagnoses (ICD‑10 F10–F98) were retrieved from the Finnish Care Register for Health Care to identify psychiatric morbidity within 1–15 years preceding the dementia diagnosis.   
Results: Across the 1–15-year pre‑diagnostic window, 28.3% of EOD cases had at least one psychiatric diagnosis, compared with 14.1% of matched controls (RR=2.01, p<0.001). Psychiatric history was most common among individuals with FTD (35.0%, RR=2.57, p<0.001) and vascular patients (42.2%, RR=2.76, p<0.001). Distinct longitudinal patterns emerged across different psychiatric and EOD diagnostic subgroups.   
Conclusions: Psychiatric disorders appear substantially more common in individuals who later develop EOD, with differences already detectable more than a decade prior to dementia diagnosis. These results underscore the importance of considering psychiatric symptoms as potential early indicators or risk factors for EOD and highlight subtype‑specific associations.  

9. Frequency-dependent diffusion tensor distribution imaging in the evaluation of ischemic stroke  

Authors: Sara Gröhn1, Angela Naranjo2, Omar Narvaez1, Maxime Yon3, Buse Buz-Yalug1, Santos Blanco2, Daniel Topgaard4, Esther Martinez-Lara2, Ma Ángeles Peinado2, Jussi Tohka1, Alejandra Sierra1  

Author affiliations: 1A.I. Virtanen Institute for Molecular Sciences; University of Eastern Finland, Finland. 2Department of Experimental Biology, University of Jaén, Spain. 3University of Rennes, LTSI, Inserm-U1099, Rennes, France. 4Department of Chemistry, Lund University, Sweden  

Rationale   
Early detection of microstructural tissue alterations is essential for understanding the pathogenesis of neurodegenerative diseases, such as stroke or epilepsy. While conventional MRI is central in many diagnostic workflows, it often fails to capture subtle, subvoxel microstructural changes or reliably assess tissue viability. Frequency dependent diffusion tensor distribution imaging (ωDTD) leverages tensor valued diffusion encoding to probe sub voxel microstructure.    
Methods   
We performed ex vivo ωDTD (11.7 T scanner, 2D multislice spin echo sequence; b = 700–8000 s/mm²; normalized anisotropy bΔ = −0.5, 0, 0.5, 1; TR = 800 ms; TE = 30 ms; multiple orientations (θ, φ); centroid frequency ωcent/2π = 34–115 Hz; in plane voxel size 80×80 μm², 250 μm slice thickness) on ischemic (transient middle cerebral artery occlusion) and sham operated rat brains 24 h after reperfusion, followed by histology. Random forest models were trained to predict quantitative histological metrics directly from ωDTD.   
Results   
Leave one animal out cross validation indicated that ωDTD reflected changes in cellularity and cellular morphology. Permutation testing confirmed that ωDTD outperformed standard DTI accros perfomance metrics (R, Q², MAE and MSE). Feature importance estimation showed edema, disrupted cell processes, and loss of microstructural organization as major contributors to predictions.   
Conclusion   
Our findings demonstrate that ωDTD based predictive modeling can detect microstructural alterations associated with ischemic stroke, outperforming standard DTI metrics. Although the present study used ischemic stroke, MRI-based predictive modeling shows promise as a generalizable, biopsy free MRI biomarker approach for quantifying tissue pathology for a variety of disorders.    

10.  Electrical stimulation of human primary astrocytes and iPSC-neurons in frontotemporal dementia  

Authors: Rebekka Fuchs1, Viivi Pekkala1, Tomi Hietanen1, Dorit Hoffmann1, Eino Solje2,3, Nadine Huber1, Alberto Sesena-Rubfiaro1, and Annakaisa Haapasalo1 

Author affiliations: 1. A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 2. Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 3. Neuro Center, Neurology, Kuopio University Hospital, Kuopio, Finland  

Rationale. Electrical stimulation (ES) has been shown to improve cognitive performance in individuals suffering from neurocognitive diseases, including frontotemporal dementia (FTD). However, the underlying physiological mechanisms remain poorly understood and need further investigation. Thus, modeling repetitive transcranial magnetic stimulation (rTMS) in vitro could provide novel insights into the cellular and molecular processes involved.   
Methods. Human induced pluripotent stem cell (iPSC)-derived neurons from healthy control individuals and sporadic as well as C9orf72 hexanucleotide repeat expansion (C9-HRE)-carrying FTD patients were co-cultured with human primary astrocytes. The cultures were electrically stimulated starting from day-in-vitro (DIV) 11 three times weekly for 10 min (5 s stimulus on, 25 s off) with biphasic pulses at 10 Hz and a voltage of 1.28 V, 8 V or 30 V until DIV 32. To assess whether ES influences cell viability, MTT assay was performed at DIV 9 and DIV 30. Also, immunocytochemical stainings (ICC) for c-Fos and synaptic markers were assessed.   
Results. The MTT assay indicated that the astrocytes’ viability was not adversely affected by ES. c-Fos expression was upregulated in astrocytes at 8 V and 30 V, suggesting activation in response to stimulation. ICC indicated that expression of the postsynaptic protein Homer1 decreased in neurons after ES when compared to non-stimulated neurons, suggesting alterations in synaptic density.    
Conclusions. These results suggest that ES does not compromise cell viability but activates the neuronal co-cultures at 8 V and 30 V. Furthermore, ES may contribute to synaptic remodeling as indicated by changes in postsynaptic protein expression.   

11.  Association between the blood-based Alzheimer’s disease biomarkers and diet  

Authors: Jenni Lehtisalo 1,2, Anna Rosenberg 2,3, Esko Levälahti 2,  Anna Matton 3,4, Francesca Mangialasche 3,4,5, Alina Solomon 3,6,7, Henrik Zetterberg 8, Tiia Ngandu 1,2, Miia Kivipelto 1,3,4,5,7    

Author affiliations: 1.Institute of Public Health and Clinical Nutrition, University of Eastern Finland, Kuopio, Finland 2. Department of Public Health, Lifestyles and Living Environments, Finnish Institute for Health and Welfare, Helsinki, Finland 3. Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden 4. FINGERS Brain Health Institute, Stockholm, Sweden 5. Theme Inflammation and Aging, Karolinska University Hospital, Sweden  6. Department of Neurology, Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland 7. The Ageing Epidemiology Research Unit, School of Public Health, Imperial College London, London, UK 8.Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, The Sahlgrenska Academy, University of Gothenburg, Mölndal, Sweden  

Rationale: Blood-based Alzheimer’s disease (AD) biomarkers are gaining interest as tools for optimising prevention strategies, but understanding their associations with preventive approaches like lifestyles in cognitively unimpaired cohorts is lacking. We analysed overall diet quality, energy intake, and macronutrients in relation to AD biomarkers in an at-risk cohort.   
  
Methods: Data from the FINGER study (n=1259, 60-77 yrs) included diet records (baseline n=1252, 2y=1080) and plasma biomarkers (baseline n=1225, 2y=1105). For diet we calculated quality indices (FINGER adherence and healthy Nordic diet) and macronutrient intake. Phosphorylated Tau protein (ptau; 231 and 181), glial fibrillary acidic protein (GFAP), and neurofilament light (NfL) were analysed using the Alamar Biosciences NULISAseq (CNS Disease panel 120); and ptau217 with Fujirebio Lumipulse.    
  
Results: Diet quality indices were not associated with the blood biomarkers. Cross-sectionally higher carbohydrate intake was related to higher levels of ptau217, ptau231, ptau181, and NfL (p<0.05 for each) at baseline, and higher total fat was related to lower NfL(p=0.002). Longitudinally higher carbohydrate predicted increase in ptau181 (p=0.018) and GFAP (p=0.008). Also, higher baseline ptau181 predicted less improvement in FINGER index (p=0.023) and higher GFAP increase in carbohydrate (p=0.041) but decline in energy (p=0.038).   
    
Conclusions: The most consistent finding was the cross-sectional association between higher carbohydrate intake and higher level of AD biomarkers. Longitudinal associations between diet and AD biomarkers appear multidimensional and we cannot distinguish if already very early disease results in changes in diet or if diet contributes to development of AD pathology.    

12. Non-Motor Symptoms Questionnaire-Based profiles in Early-Stage Dementias: A prospective Cohort Study  

Authors: Mika Kinnunen1, Emma Vitikka2, Johanna Krüger2,3,4, Dorit Hoffmann5, Annakaisa Haapasalo5, Alberto Lleó6,7, Olivia Belbin6,7, Noora-Maria Suhonen2,3,4, Barbara Borroni8,9, Anne M. Portaankorva10, Laura Säisänen1,11, Kasper Katisko1, Eino Solje1,11  

Author affiliations:  1Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 2Research Unit of Clinical Medicine, Neurology, University of Oulu, Oulu, Finland 3Medical Research Center, Oulu University Hospital, Oulu, Finland 4Neurocenter, Neurology, Oulu University Hospital, Oulu, Finland 5A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 6Sant Pau Memory Unit, IR SANT PAU Hospital de la Santa Creu i Sant Pau, Barcelona, Spain 7Centro de Investigación Biomédica en Red en Enfermedades Neurodegenerativas (CIBERNED), Madrid, Spain 8Department of Clinical and Experimental Sciences, University of Brescia, Brescia, Italy 9Molecular Markers Laboratory, IRCCS Istituto Centro San Giovanni di Dio Fatebenefratelli, Brescia, Italy 10Clinical Neurosciences, University of Helsinki, Helsinki, Finland 11Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland  

Background Current understanding of the nature and prevalence of non-motor and autonomic symptoms across dementia subtypes remains incomplete. Although Alzheimer’s Disease (AD), behavioral variant frontotemporal dementia (bvFTD), and dementia with Lewy Bodies (DLB) are primarily defined by cognitive and motor impairments, non-motor symptoms and autonomic dysfunction are increasingly recognized across these disorders. The aim of this study was to assess non-motor symptoms, reflecting autonomic dysfunction in early-stage dementia, and characterize symptom profiles across different neurodegenerative dementias.   
  
Methods A total of 64 early-stage AD, 38 bvFTD, and 17 DLB patients, along with 63 cognitively healthy controls (HC) were prospectively recruited from Kuopio and Oulu University Hospitals. Non-motor symptoms were assessed using the Non-Motor Symptoms Questionnaire (NMSQ), with defined in categories of gastrointestinal, neuropsychiatric, cognitive, sleep-related, and sensory motor domains across three predefined time periods prior to the baseline evaluation.   
  
Results DLB patients showed the highest prevalence of non-motor symptoms across all assessed domains. Neither the bvFTD nor the AD group exhibited distinct disease-specific symptom profiles. However, apathy and paranoid thoughts emerged earlier in bvFTD than in AD or DLB. Diplopia was reported by 14 % of DLB patients one year prior to the diagnosis (Chi-Square, p = 0.035). Neuropsychiatric and cognitive symptoms were more common in all dementia groups compared to HC (Chi-Square, p < 0.01), although no significant differences between the dementia groups were observed.    
  
Conclusions Non-motor symptom profiles vary across dementia subtypes already in early disease stages, indicating their potential value as early disease-specific markers.   

13. BMAL1-Dependent Circadian Regulation of the TREM2–PLCγ2 Signaling Pathway in Human Microglia  

Authors: Emine Erdag1, Mette Heiskanen1, Heli Jeskanen1, Nadine Huber2, Taisia Rolova3, Petra Mäkinen1, Inka Kervinen1, Jessica Rosa1, Annakaisa Haapasalo2, Henna Martiskainen1, Mari Takalo1, Mikko Hiltunen1  

Author affiliations: 1. Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland  2. A.I. Virtanen Institute of Molecular Sciences, University of Eastern Finland, Kuopio, Finland  3.Neuroscience Center, Helsinki Institute of Life Science (HiLIFE), University of Helsinki, Helsinki, Finland.  

Rationale:   
Although circadian disruption is a common feature of Alzheimer’s disease (AD), the role of the core clock regulator BMAL1 in linking circadian dysregulation to microglial dysfunction through TREM2–PLCγ2 signaling is not examined in human microglial models. Moreover, APOE4, the strongest common genetic risk factor for sporadic AD, is associated with impaired microglial responses and altered TREM2 signaling, whereas the protective PLCG2-P522R variant enhances microglial function and reduces AD risk. Together, these observations suggest that circadian regulation may interact with the APOE–TREM2–PLCγ2 axis in AD. This study aims to determine whether circadian disruption alters microglial immunometabolism through the BMAL1–TREM2–PLCγ2 pathway and whether melatonin can restore microglial homeostasis.   
Methods:   
Human iPSC-derived microglia carrying wild-type or protective PLCG2-P522R, on APOE3/3 or APOE4/4 backgrounds, will be used. Circadian disruption will be induced by BMAL1 silencing using RNA interference or CRISPR, followed by assessment of TREM2–PLCγ2 expression and downstream signaling. Cells will also be treated with melatonin to evaluate restoration of circadian and microglial homeostasis. Functional outcomes will include Seahorse analysis, Aβ phagocytosis, reactive oxygen species, lipid droplet accumulation, cytokine expression, and RNA sequencing with pathway enrichment.    
Results:   
We hypothesize that BMAL1 disruption impairs microglial metabolism and Aβ clearance through dysregulated TREM2–PLCγ2 signaling, with greater effects in APOE4/4 and wild-type PLCG2 cells than in PLCG2-P522R cells. Melatonin is expected to restore metabolic homeostasis and enhance Aβ clearance.   
Conclusions:   
This study will define the BMAL1–TREM2–PLCγ2 axis in AD microglia and may identify novel targets for chronotherapeutic intervention.   

14. Generation of heterozygous Tyrobp knockout BV2 microglia by CRISPR-Cas9 genome editing  

Authors: Mette Heiskanen¹, Päivi Harju¹, Roosa-Maria Willman¹, Mikko Hiltunen¹, Henna Martiskainen¹  

Author affiliation: ¹ Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland  

Rationale   
TYROBP gene encodes DAP12, a transmembrane signaling adapter protein expressed in the brain by microglia. DAP12 mediates intracellular signaling from cell surface receptors such as TREM2, which regulate microglial activity. Homozygous loss-of-function variants in TYROBP cause Nasu-Hakola disease, a rare disease characterized by bone cysts and early dementia. Recently, we identified heterozygous TYROBP deletion as a novel risk factor for Alzheimer’s disease. To investigate the effects of partial loss of DAP12 on downstream signaling and microglial activity, we used CRISPR-Cas9 genome editing to create mouse-derived microglial cell lines with heterozygous knockout (KO) of Tyrobp.   
  
Methods   
RNP complex targeting the exon 3 of Tyrobp was delivered to BV2 microglial cells by lipofection. Monoclonal cultures were created by FACS, and clones carrying either heterozygous (Tyrobp+/-) or homozygous (Tyrobp-/-) KOs of Tyrobp were identified by Sanger sequencing and ICE analysis. Selected lines were validated by Western blotting for DAP12.   
  
Results   
Edits were detected in 84% (62/74) of sequenced colonies, yielding 19 candidate Tyrobp+/- and 18 candidate Tyrobp-/- monoclonal lines. We performed Western blot screening on 19 Tyrobp+/- and 12 Tyrobp-/- lines. Tyrobp-/- lines did not express DAP12, while Tyrobp+/- lines exhibited DAP12 levels ranging from 2–107% (median 30%) relative to control cells. Based on sequencing results, DAP12 levels, and growth rates, we selected 4 Tyrobp+/- and 3 Tyrobp-/-lines for further experiments.   
  
Conclusions   
We generated BV2 microglial cell lines carrying heterozygous KO of Tyrobp. In addition, we generated homozygous Tyrobp KO lines for comparison. Cell lines were used in further experiments.  

15. Alzheimer’s Disease Plasma Biomarkers in the FINGER Multidomain Lifestyle Trial: Associations with Cognition and Intervention Response and Changes Over 2 Years  

Authors: Anna Rosenberg (1,2), Anna Matton (2,3,4), Burak Arslan (5), Andrea L. Benedet (5), Makrina Daniilidou (2,3), Jenni Lehtisalo (1,6), Esko Levälahti (1,6), Francesca Mangialasche (2,3,7), Ilaria Pola (5), Juha Rinne (8,9), Alina Solomon (4,10), Kübra Tan (5), Henrik Zetterberg (5), Tiia Ngandu (1,2,6), Miia Kivipelto (2,3,4,6,7)   

Author affiliations: 1) Department of Public Health, Lifestyles and Living Environments, Finnish Institute for Health and Welfare, Helsinki, Finland; 2) Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm Sweden; 3) FINGERS Brain Health Institute, Stockholm, Sweden; 4) The Ageing Epidemiology Research Unit, School of Public Health, Imperial College London, London, UK; 5) Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, The Sahlgrenska Academy, University of Gothenburg, Mölndal, Sweden; 6) Institute of Public Health and Clinical Nutrition, University of Eastern Finland, Kuopio, Finland; 7) Theme Inflammation and Aging, Karolinska University Hospital, Sweden; 8) Turku PET Centre, University of Turku, Turku, Finland; 9) Division of Clinical Neurosciences, Turku University Hospital, Turku, Finland; 10) Department of Neurology, Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland  

RATIONALE: Plasma Alzheimer’s disease biomarkers can enable early disease detection and intervention, but studies are scarce in at-risk cohorts and non-pharmacological trials. We investigated plasma biomarkers in the FINGER trial.   
METHODS: FINGER included 1260 older adults without substantial cognitive impairment, randomized to 2-year multidomain lifestyle intervention vs. general health advice. Biomarkers were analyzed in all available baseline (n=1225) and 2-year (n=1105) samples, with Lumipulse (ptau217) and NULISAseq™ CNS Disease Panel120 (Aβ42, 40, ptau181, ptau231, NFL, GFAP). Cognition was assessed with Neuropsychological Test Battery.    
RESULTS: Majority (85.4%) had normal p-tau217 (<0.22 pg/ml) at baseline (8.0% intermediate, 6.6% abnormal >0.34 pg/ml). AUCs for prediction of amyloid-PET positivity were modest (e.g., ptau217: 0.72, 95% CI 0.56–0.88). Higher ptau217, GFAP, and Aβ40 were associated with poorer baseline cognition. Higher levels of all biomarkers apart from Aβ42 predicted less cognitive improvement over 2 years. Individuals with abnormal ptau217 had increased risk of memory decline (OR 1.91, 95% CI 1.12–3.28, vs. normal). Biomarker levels did not modify response to intervention, i.e., beneficial effects were observed both in those with higher and lower levels. Over 2 years, biomarkers increased cohort-wide, yet changes were subtle, with most participants remaining within normal ptau217 range (82.1%). Biomarker changes were similar in the intervention and control groups.    
CONCLUSIONS: New plasma AD biomarkers, including ptau217, associate with cognitive changes and capture subtle accumulation of neuropathology already in at-risk stages prior to onset of cognitive impairment. The FINGER lifestyle intervention can benefit cognition also in individuals with underlying AD neuropathology.   

16. Monoallelic TYROBP deletion affects TREM2 levels and disrupts myelin phagocytosis in microglial BV2 cells  

Authors: Linnea Luhtaniemi1, Rosa Sinisalo1, Mette Heiskanen1, Päivi Harju1, Roosa-Maria Willman1, Annakaisa Haapasalo2, Mikko Hiltunen1 and Henna Martiskainen1   

Author affiliations: 1 School of Medicine, Institute of Biomedicine, University of Eastern Finland 2 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland  

Activation of disease-associated microglia (DAM) requires TREM2 signaling through its adaptor protein DAP12, encoded by TYROBP. Complete loss of TYROBP causes Nasu-Hakola disease, whereas partial loss (monoallelity) increases Alzheimer’s disease (AD) risk. However, the mechanisms by which reduced DAP12 availability disrupt TREM2 signaling and microglial function are still unclear.    
To address this, we used BV2 mouse microglial lines generated by CRISPR-Cas9 editing, including cells lacking Tyrobp (Tyrobp-/-) or carrying a heterozygous deletion (Tyrobp+/-), along with untargeted controls. Membrane fractions were isolated by biotinylation to assess surface and intracellular TREM2 and evaluate TREM2 expression by Western blot. Soluble TREM2 (sTREM2) levels were measured by ELISA. Microglial phagocytic function was quantified using Incucyte live-cell imaging and uptake of pHrodo-labelled Zymosan and myelin to distinguish substrate- and pathway-specific effects.   
Both full-length TREM2 and its C-terminal fragment (TREM2-CTF) were reduced in Tyrobp-/- and Tyrobp+/- cells, with markedly diminished membrane-associated TREM2-CTF and similar reductions in cytosolic fractions. Consistent with reduced availability of functional receptor at the cell surface, sTREM2 level was decreased across DAP12-deficient genotypes. Phagocytic assays showed substrate-specific alterations: myelin uptake—dependent on TREM2—was significantly impaired, whereas Zymosan phagocytosis was increased, consistent with a shift toward TLR4-mediated phagocytic pathways when DAP12-dependent signaling is compromised.   
Together, our findings indicate that partial or complete DAP12 loss reduces functional TREM2 at the cell surface, lowers sTREM2 production, and disrupts TREM2-dependent phagocytosis while enhancing TLR4-driven uptake. These alterations highlight impaired microglial homeostasis as a potential mechanism by which monoallelic TYROBP deletion increases AD risk.   

17. Associations between chronic CNS-active medication use and paired-pulse Transcranial Magnetic Stimulation (ppTMS) 

Authors: Meri Juslin ¹, Laura Säisänen ² ³, Sara Määttä ¹ ², Matti Tyrväinen ⁴, Jelena Hyppönen ¹ ⁴, Eino Solje ² ³  

Author affiliations: 1 Department of Clinical Neurophysiology, Kuopio University Hospital, Kuopio, Finland. 2 Institute of Clinical Medicine, Neurology, University of Eastern Finland, Kuopio, Finland. 3 Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland. 4 Institute of Clinical Medicine, Clinical Neurophysiology, University of Eastern Finland, Kuopio, Finland. 

Rationale: Paired-pulse Transcranial Magnetic Stimulation (ppTMS) has been proposed as a potential diagnostic tool in early-stage dementing diseases. However, use of central nervous system (CNS) active medication is often listed as an exclusion criterion in ppTMS studies. For clinical applicability of ppTMS, it’s essential to understand how long-term and concurrent use of CNS-active medication influence its outcomes. Most prior pharmaco-TMS studies have been conducted in a setting, where the drug is administered to a previously unmedicated subject, with ppTMS measured relatively shortly after.  
 
Methods: The study is based on an observational study design. We evaluated 200 subjects (17 Frontotemporal dementia, 33 Alzheimer’s disease, 13 Dementia with Lewy bodies, 17 parkinsonian, 15 Mild cognitive impairment, 41 Normal pressure hydrocephalus, 64 other) and 36 unmedicated healthy controls, with TMS performed at baseline and after 1 year. Subjects continued their regular medications to reflect a real-life clinical implementation. The medications were divided into 13 groups based on their primary pharmacological mechanism of action. We measured resting motor threshold (RMT) and three measures of cortical excitability: SICI-ICF, LICI and SAI; accounting for glutamatergic, GABAergic and cholinergic functions of the brain.  
 
Results: Antipsychotic use was associated with a higher RMT (independent samples t-test, p < 0.05). Memantine use could only be evaluated at baseline and was associated with weaker ICF at 7 ms (independent samples t-test, p < 0.001).  
 
Conclusions: The continuous use of CNS-active drugs may influence ppTMS findings and should therefore be further investigated and considered when interpreting ppTMS results in clinical settings.  

18. PREDICTFTD: Accelerating the Validation of Predictive Liquid Biomarkers for Frontotemporal Dementia Diagnosis and Subclassification 

Authors: Abbe Ullgren1,2*, Noora-Natalia Veikkolainen1, Mervi Könönen3, Annakaisa Haapasalo4, Harro Seelar5, Eino Solje1,6, on behalf of the PREDICTFTD Consortium *Presenting author 

Author affiliations: 1 Institution of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 2 Department of Neurobiology, Care Sciences and Society, Division of Neurogeriatrics, Karolinska Institutet, Solna, Sweden 3 Department of Clinical Radiology, Kuopio University Hospital, Kuopio, Finland. 4 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 5 Department of Neurology and Alzheimer Center Erasmus MC, Erasmus MC University Medical Center, Rotterdam, the Netherlands. 6 Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 

Rationale  
PREDICTFTD is a Horizon Europe project aiming to transform early and accurate diagnosis of frontotemporal dementia (FTD), particularly sporadic FTD, by validating predictive liquid biomarkers and enabling AI-supported, multimodal diagnostic tools. Aside from genetic testing, FTD lacks definitive diagnostic biomarkers and is often misdiagnosed, contributing to multi-year delays to correct diagnosis and missed opportunities for timely care and trial recruitment.  
Methods  
The prospective arm of PREDICTFTD aims to build a harmonized, multicenter observational cohort with baseline and annual follow-up visits, including clinical phenotyping, neuroimaging, and biospecimen collected using standardized protocols. The data collected will be used for biomarker validation and lay the foundation for developing new diagnostic tools.  
Results  
Within the first year, PREDICTFTD has established unified study protocols, set up a new study database, conducted cross-site training, and is on the verge of starting recruitment. Recruitment targets include ~195 FTD, 120 Alzheimer’s disease, 55 non-neurodegenerative controls, and 130 healthy controls.  
Conclusions  
By delivering harmonised longitudinal multimodal datasets and biosamples, PREDICTFTD provides the foundation to validate minimally invasive biomarkers and train multimodal diagnostic AI tools to aid in differential diagnostic, support molecular subtyping, and enable earlier identification of disease onset.  

19. Understanding the Factors Contributing to Alzheimer’s Disease Misdiagnosis 

Authors: Ave Kivisild1, Salome Ervasti2, Per Johansson3,4, Maria Landqvist Waldö3,4, Kasper Katisko1, Johanna Krüger2,5,6, Eino Solje1,7  

Author affiliations: 1 = Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 2 = Research Unit of Clinical Medicine, Neurology, University of Oulu, Oulu, Finland 3= Center of Cognitive Medicine, Skåne University Hospital, Ängelholm, Sweden 4= Clinical Sciences Helsingborg, Department of Clinical Sciences Lund, Lund University, Lund, Sweden 5 = MRC, Oulu University Hospital, Oulu, Finland 6 = Neurocenter, Neurology, Oulu University Hospital, Oulu, Finland 7 = Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 

Rationale  
Early and accurate diagnosis of Alzheimer’s disease (AD) is essential in the era of disease-modifying therapies but remains challenging due to nonspecific symptoms, phenotypic heterogeneity, and diagnostic overlap in early disease stages. We investigated factors associated with AD misdiagnosis and diagnostic revision following longitudinal reassessment.  
Methods  
This retrospective case-control study included patients from three university hospitals in Sweden (Ängelholm/Lund) and Finland (Kuopio and Oulu). Misdiagnosis cases (n = 87) comprised patients initially diagnosed with AD whose diagnoses were later revised based on longitudinal follow up, genetic testing, or neuropathology. The misdiagnosis group was compared with biomarker confirmed AD patients (n = 177), verified using cerebrospinal fluid amyloid biomarkers, amyloid PET imaging, or neuropathology.  
Results  
AD misdiagnosis was significantly less frequent among patients with complete cerebrospinal fluid biomarker profiles (OR 0.13, 95% CI 0.052–0.32, p < 0.001), availability of CSF biomarker assessment (OR 0.09, 95% CI 0.037–0.233, p < 0.001), hippocampal atrophy of Scheltens grade ≥ 2 (OR 0.44, 95% CI 0.240–0.812, p = 0.018), and amnestic cognitive presentation (OR 0.02, 95% CI 0.005–0.10, p < 0.001). Lower educational attainment was associated with an increased likelihood of misdiagnosis (OR 2.49, 95% CI 1.13–5.50, p = 0.024). Thirty-one percent of misdiagnosed patients showed no evidence of a progressive neurodegenerative disease. Frontotemporal dementia was the most frequent revised diagnosis (22%).  
Conclusions  
AD misdiagnosis primarily occurred in patients lacking AD-typical clinical and biomarker features, emphasizing the need for careful diagnostic confirmation and longitudinal reassessment in biologically targeted treatment settings. 

20. Mid-to-late life subjective memory impairment and edentulism may jointly identify increased dementia risk 

Authors: Sam Asher^1,2, Alina Solomon^2,3,4, & Anna Liisa Suominen^1,5,6 

^{Author affiliations:} 

¹Institute of Dentistry, University of Eastern Finland, Kuopio, Finland   

²Division of Clinical Geriatrics, Karolinska Institute, Stockholm, Sweden 

³Institute of Clinical Medicine, Neurology, University of Eastern Finland, Kuopio, Finland ⁴Ageing Epidemiology Research Unit, School of Public Health, Imperial College London, London, United Kingdom 

⁵Department of Public Health, Finnish Institute for Health and Welfare (THL), Helsinki, Finland 

⁶Oral and Maxillofacial Teaching Unit, Primary Oral Health Care Services, Wellbeing Services County of North-Savo, Kuopio, Finland  

Rationale: Literature links subjective memory impairment and severe tooth loss with incident dementia, however, synergistic effects of these markers on incident dementia risk has not been investigated. 

Methods: Data on participants (N=1834) without baseline dementia, depression, or functional impairment, was retrieved from Finnish Health 2000 Survey, supplemented with dementia diagnosis until 2015 from health registers. Baseline cognitive assessment included objective assessment via neuropsychological cognitive test battery and inquiries about subjective memory: experiencing recent memory decline (no/yes), and memory changes adversely affecting daily life (no/yes). We defined subjective memory impairment (SMI) as experiencing both memory decline and its adverse functional impact. Analysis included Cox proportional regression, tested interaction between SMI and tooth count, accounted for relevant covariates including objective compositive cognitive score, reverse causality bias, and competing mortality.  

Results: After accounting for relevant confounders, potential reverse causality bias, and competing risk, SMI was associated with increased dementia risk. The association was independent of baseline objective cognition and tooth count, but no SMI-tooth count interaction was observed. However, significant positive interaction between SMI and edentulism (i.e., complete tooth loss) was found (p=0.047) including among participants aged ≥55years (p=0.037). In analysis stratified by edentulism, SMI increased incident dementia risk among edentate participants (HR range=3.70-3.80), but not dentate participants (having ≥1 teeth).      

Conclusions: Subjectively meaningful memory impairment and edentulism may in conjunction identify at-risk individuals. Natural tooth preservation is important for healthy cognition, and integration of oral health considerations in geriatric healthcare may facilitate early identification and risk reduction. 

21. The Cerebrospinal Fluid and Plasma Proteomics of Idiopathic Normal Pressure Hydrocephalus 

Authors: Mikolaj Wojnicki 1,2, Anssi Lipponen 2, Sami Heikkinen 2, Tony Wyss-Coray 3, Tarja Malm 4, Mikko Hiltunen 2, Ville Leinonen 1,5 

Author affiliations: 1. Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland, 2. Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland, 3. Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, USA, 4. A.I. Virtanen Institute, University of Eastern Finland, Kuopio, Finland, 5. Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland 

Rationale  
Idiopathic normal pressure hydrocephalus (iNPH) is a potentially reversible cause of dementia characterized by impaired gait and cognition and by urinary incontinence. However, limited understanding of its molecular pathophysiology and its subtypes contributes to varied treatment outcomes. To predict disease severity and progression, individual CSF biomarkers such as amyloid-beta and tau have been used. Recently, alterations in CSF and plasma proteome have been associated with cognitive impairment, but the proteomics of iNPH remain largely unexplored. To address this, we investigated how the CSF and plasma proteome is associated with iNPH, its severity and treatment response.  
Methods  
The study cohort consists of 324 iNPH patients and 106 non-iNPH individuals treated in Kuopio. Paired plasma and lumbar CSF samples underwent SomaScan aptamer-based proteomics (>11,000 proteins quantified). Additionally, iNPH patients have brain biopsy pathology as well as cognitive and gait test results. To discover how iNPH is manifested in the human proteome, we ran linear model differential expression analyses using limma and SomaDataIO R packages on iNPH diagnosis status, severity and treatment response. The biological mechanisms underlying the associated proteins were interpreted with the help of pathway analysis and by finding tissues that each protein originates from.  
Results  
Preliminary results revealed 138 CSF and 975 plasma proteins significantly altered in iNPH vs non-iNPH group (adj.p<0.05, |log2FC|>1).   
Conclusions  
We conclude that iNPH leaves a significant mark on both CSF and plasma proteome. Further analysis should advance our understanding of iNPH physiology and help identify novel disease mechanisms. 

22. AMZ1/GNA12 and SLCO1A2 Variants Are Associated with Altered Ventricular Volume in Normal Pressure Hydrocephalus 

Authors: Umay Atalay1, Toni Tuovinen1, Nikita Keskinen1,2, Mikolaj Wojnicki1, Alli Pesola1, Joel Räsänen1,3, Sami Heikkinen1, Anssi Lipponen1, Ville Leinonen2,3 

Author affiliations: 1 Institute of Biomedicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland (UEF); 2 Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, UEF; 3 Kuopio University Hospital. 

Normal Pressure Hydrocephalus (NPH) is a chronic neurological syndrome in older adults, characterized by gait disturbance, cognitive impairment, and urinary dysfunction. The condition is marked by abnormal accumulation of cerebrospinal fluid (CSF) in the cerebral ventricles. Although symptoms can be alleviated by draining excess CSF to abdominal cavity via shunting, many patients will continue to experience long-term impairments highlighting the need for a better understanding of disease mechanisms. A recent GWAS of NPH in FinnGen identified eight risk loci (SLCO1A2, AMZ1/GNA12, MLLT10, CDCA2, C16orf95, PLEKHG1, ARHGEF12, and CSNK1E; Räsänen et al., 2024). In this study, we investigated whether lead variants at these loci are associated with characteristic neuroanatomical features of NPH. Specifically, GWAS lead variants were tested for association with ventricular and CSF volumes derived from brain MRI and cNeuro software. Associated loci were further studied by identifying variants in linkage disequilibrium with the lead variants which were subjected to functional annotation. In total, 234 common variants (GWAS , minor allele frequency > 0.05) were annotated using Ensembl Variant Effect Predictor and brain expression quantitative trait loci data, followed by secondary phenotype–variant association analyses. Preliminary results suggest that variants near SLCO1A2, including rs2199687 ( ), were associated with reduced ventricular and CSF volumes, whereas rs2793354 near MLLT10 ( ) showed evidence of increased ventricular volume. Variants near AMZ1/GNA12 and CDCA2 demonstrated more mixed effects across volumetric measures. Overall, these findings link genetic risk for NPH to disease-relevant neuroanatomical phenotypes and provide basis for future studies investigating disease mechanisms and progression. 

23. Integrating Genetics, Proteomics, and in vitro Modeling to Elucidate SLCO1A2-Mediated CSF Clearance in Idiopathic Normal Pressure Hydrocephalus 

Authors: Alli Pesola¹,²; Anssi Lipponen²; Sami Heikkinen²; Joel Räsänen¹,³; Tony Wyss-Coray⁴; Evan Macosko⁵; Tarja Malm⁶; Mikko Hiltunen²; Ville Leinonen¹,³  

Author affiliations: ¹Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland; ²Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland; ³Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland; 4Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, USA; 5Broad Institute of MIT and Harvard, Boston, USA; 6A.I. Virtanen Institute, University of Eastern Finland, Kuopio, Finland 

In idiopathic normal pressure hydrocephalus (iNPH) cerebrospinal fluid (CSF) clearance is impaired leading to metabolic waste accumulating in the brain. Co-occurring iNPH and Alzheimer’s disease (AD) pathologies are common, showing altered CSF levels of amyloid β-42 (Aβ42), total Tau, and phosphorylated Tau (pTau), which suggests shared mechanisms. We identified four iNPH-associated loci with a FinnGen genome wide association study, including SLCO1A2, which showed two independent signals suggesting previously unknown haplotype structure. SLCO1A2 encodes organic anion transporting polypeptide 1A2 (OATP1A2) regulating key processes in barrier-mediated clearance. We aim to identify novel molecular mechanisms influencing OATP1A2-mediated CSF clearance with proteomics, genetics and biomarker studies, and validate the mechanisms with an in vitro blood-CSF barrier model.  
Whole genome sequencing data from 317 Kuopio iNPH cohort patients and 718 imputed genotypes were used to define SLCO1A2 haplotypes. We will assess SLCO1A2 haplotypes by associating them with phenotypes within the Kuopio iNPH cohort, including cortical Aβ42 and Tau immunohistochemistry, CSF proteomics, and functional and cognitive data. We will validate the results on Aβ42 and Tau clearance testing cellular uptake by in vitro model of blood–CSF barrier.  
Our preliminary findings suggest that SLCO1A2 haplotypes influence the onset age of iNPH and AD, and are associated with altered CSF Tau, Tau/Aβ42, and pTau/Aβ42 ratios.  
We conclude that these results support a functional role for OATP1A2 in blood-CSF barrier-mediated clearance of neurodegeneration-related proteins. Our integrative genotype-phenotype approach enables identification of CSF clearance-related genetic mechanisms and highlights SLCO1A2 as a potential contributor to regulate CSF clearance.  

24. Mapping cognitive function assessment in the World-Wide FINGERS Network to optimise evidence for dementia risk-reduction: the WW-FINGERS Cognitive Outcome International Working Group. 

Authors: Laboni Akter¹, Mariagnese Barbera^1,2, Celeste de Jager², Tiia Ngandu^3,4,5, Francesca Mangialasche^5,6,7, Alina Solomon^1,2,5, Miia Kivipelto^2,4,5,6,7, on behalf of the WW-FINGERS Network and the CO-IWG Team. 

Author affiliations:  

1. Department of Neurology, Institute of Clinical Medicine, University of Eastern Finland, Yliopistonranta 1C, 70211, Kuopio, Finland 

2. The Ageing Epidemiology Research Unit, School of Public Health, Imperial College London, Charing Cross Hospital, St Dunstan’s Road, London, W6 8RP, UK 

3. Department of Public Health, Lifestyles and Living Environments, Finnish Institute for Health and Welfare, Mannerheimintie 166, Helsinki, 00300, Finland 

4. Institute of Public Health and Clinical Nutrition, University of Eastern Finland, Yliopistonranta 1C, 70211, Kuopio, Finland. 

5. Division of Clinical Geriatrics, Center for Alzheimer Research, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Karolinska Vägen 37A, 171 64, Solna, Sweden 

6. Theme Inflammation and Aging, Medical Unit Aging, Karolinska University Hospital, Karolinska Vägen 37A, 171 76, Solna, Sweden 

7. FINGERS Brain Health Institute, Karolinska Vägen 37A, Solna, 171 64, Sweden 

Rationale: World-Wide FINGERS (WW-FINGERS) is a global scientific initiative (73 countries as of January 2026) supporting the generation of robust evidence for dementia risk-reduction through aligned methodologies, data harmonisation, and joint analysis, in multidomain intervention trials. Within WW-FINGERS, the Cognitive Outcomes International Working Group (CO-IWG) is developing a framework for cognitive outcomes harmonisation. The aim of this study was to map the available cognitive assessments/measures in the WW-FINGERS trials.   

Methods: In March 2025, a survey gathering basic information on study design and detailed data on cognitive assessments was completed for all CO-IWG trials. A descriptive analysis was conducted to synthesise trials’ general characteristics (e.g., status, target population), and map available cognitive tests/measures and domains assessments. 

Results: 38 trials (N=23,999) at different stages (e.g., planning, ongoing, completed) were included across multiple countries. Most focused on cognitively healthy/at-risk populations (84%). Cognitive assessments were mostly in-person (86%) and on paper (61%) only. Most trials (86%) assessed all FINGER cognitive domains (memory, executive function, processing speed). Additionally, executive function, processing speed, and working and episodic memory domains could be assessed in ≥ 24% of the trials using ≥3 measures. Digit-span, Trail-Making-Test, and CERAD verbal fluency were the most common tests, used in 73%, 70%, and 57% of the included trials, respectively.  

Conclusions: Mapping of the cognitive assessments across the WW-FINGERS trials showed that joint analyses to assess the effectiveness of multidomain interventions on cognitive function is possible. More work is needed to optimise the data and variables harmonisation which will enable pooling of data. 

25. EXPLORING THE NEUROPROTECTIVE ROLE OF THE PLCG2 VARIANT IN C9ORF72 REPEAT EXPANSION-ASSOCIATED FRONTOTEMPORAL DEMENTIA 

Authors: Vili Hakosalo1, Nadine Huber1, Eino Solje2,3, Mikko Hiltunen4, Juzoh Umemori1, Mari Takalo4, Annakaisa Haapasalo1 

Author affiliations: 1 – A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland.  2 – Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland.  3 – Neuro Center, Neurology, Kuopio University Hospital, Kuopio, Finland.  4 – Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland 

Rationale  
Frontotemporal dementia (FTD) is the second most common cause of early‑onset dementia, characterized by progressive neuronal degeneration leading to synaptic dysfunction and imbalanced neuronal activity. The most common genetic cause underlying FTD is a hexanucleotide repeat expansion (HRE) in the C9orf72 gene (C9‑HRE), with a particularly high prevalence in Finnish patients. The complex HRE region poses major challenges for disease screening and genetic correction, slowing both diagnostics and research. Recent discoveries have highlighted a rare PLCG2 P522R variant with neuroprotective gain‑of‑function effects against Alzheimer’s disease (AD) and FTD, raising the possibility that PLCG2‑mediated mechanisms might also counteract C9‑HRE‑associated pathologies. This idea is now tested using human induced pluripotent stem cell (hiPSC)–derived neuronal models from C9-HRE-carrying FTD patients.  
Methods  
We generated hiPSC lines from three C9‑HRE-carrying patients. Isogenic control hiPSC lines were generated with CRISPR/Cas9. In addition, the protective PLCG2 P522R variant was introduced into both C9-HRE-carrying and isogenic control backgrounds. RNA fluorescent in situ hybridization (FISH) was used to assess pathological RNA foci across all lines.   
Results  
To date, three isogenic C9‑HRE‑corrected lines and one PLCG2 P522R knock‑in line have been successfully established. FISH analysis confirmed that the C9-HRE correction abolished the pathological RNA foci in the isogenic iPSCs.  
Conclusions   
Our studies represent an important step toward understanding the molecular mechanisms underlying C9‑HRE‑related pathogenesis and offer options for testing the potential PLCG2‑mediated neuroprotection beyond AD in FTD neurons. Next, the iPSCs will be differentiated into cortical neurons and their functionality, viability, and structural integrity will be assessed.  

26. Structural correlates of paired‑pulse TMS in cognitive impairment 

Authors: Miika Korhonen1,2; Laura Säisänen1,2; Abbe Ullgren1,3; Jelena Hyppönen4,5; Sara Määttä4,5; Johanna Krüger6,7,8;Annakaisa Haapasalo9; Barbara Borroni10; Ville Leinonen11,12; Mikke Niinistö13;Juhana Hakumäki13;Eino Solje1,2 

Author affiliations: 1 Institute of Clinical Medicine – Neurology, University of Eastern Finland, 70211, Kuopio, Finland; 2 Neuro Center, Neurology, Kuopio University Hospital, 70211, Kuopio, Finland; 3 Department of Neurobiology, Care Sciences and Society, Division of Neurogeriatrics, Karolinska Institutet, 17177, Solna, Sweden; 4 Department of Clinical Neurophysiology, Kuopio University Hospital, 70211, Kuopio, Finland; 5 Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland, 70211, Kuopio, Finland; 6 Research Unit of Clinical Neuroscience, Neurology, University of Oulu, 90220, Oulu, Finland; 7 Neurocenter, Neurology, Oulu University Hospital, 90220, Oulu, Finland; 8 Medical Research Center, Oulu University Hospital, 90220, Oulu, Finland ; 9 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, 70211, Kuopio, Finland; 10 Molecular Markers Laboratory, IRCCS Istituto Centro San Giovanni di Dio Fatebenefratelli, 25125, Brescia, Italy; Department of Clinical and Experimental Sciences, University of Brescia, 25121, Brescia, Italy: 11 Institute of Clinical Medicine – Neurosurgery, University of Eastern Finland, 70211, Kuopio, Finland; 12 Neuro Center – Neurosurgery, Kuopio University Hospital, 70211, Kuopio, Finland; 13 Department of Clinical Radiology, Kuopio University Hospital, 70211, Kuopio, Finland 

Rationale  
Volumetric magnetic resonance imaging (MRI) and paired-pulse transcranial magnetic stimulation (TMS) are used to differentiate neurocognitive disorders, but their interrelationship remains unclear. We investigated neuroanatomical correlates of TMS parameters reflecting cholinergic, GABA-mediated, and glutamatergic dysfunction.  
 
Methods  
The study included 134 patients with cognitive impairment: progressive neurocognitive diseases (n=65), mild cognitive impairment (n=20), subjective cognitive impairment (n=36), and idiopathic normal pressure hydrocephalus (n=13). All underwent paired-pulse TMS and MRI with voxel-based morphometry using the cNeuro® artificial intelligence (AI) tool. Regional brain volumes were analyzed in relation to TMS parameters in the total cohort and by diagnosis.  
 
Results  
In the overall cohort and in progressive dementia, short-interval intracortical inhibition (interstimulus interval, ISI = 2 ms) correlated with volumes of the pallidum, ventral diencephalon, brainstem, and cerebellum. In subjective cognitive impairment, short-latency afferent inhibition (ISI 24 ms) was associated with volumes of the ventral diencephalon, thalamus, brainstem, midbrain, and inferior frontal operculum. In normal pressure hydrocephalus, intracortical facilitation (ISI 7 ms) showed a stronger correlation with medial precentral gyrus volume than in subjective cognitive impairment. Mild cognitive impairment showed a stronger inverse association between long-interval intracortical inhibition (ISI 100 ms) and medial postcentral gyrus volume.  
 
Conclusions  
This study demonstrates anatomically plausible, partly diagnosis-specific associations between AI-based volumetric MRI and paired-pulse TMS markers of cholinergic, GABA-mediated, and glutamatergic dysfunction in cognitive impairment. The findings support paired-pulse TMS as a non-invasive marker of biologically relevant circuit-level dysfunction linked to structural brain changes. 

27. Optically enhanced plasticity in interneuron clears amyloid beta plaques in Alzheimer’s disease 

Authors: Mariia Ivanova, Lotta Esselström, Delight Degu, Lisa Rampazzo, Rashid Giniatullin, Riikka Martikainen, Heikki Tanila, Tarja Malm, Juzoh Umemori 

Author affiliations: A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Alzheimer’s disease (AD) is a progressive neurodegenerative condition and increasing evidence implicates dysfunction of parvalbumin-positive interneurons (PV-INs) as key regulators for network hyperexcitability and disease progression. In parallel, deficits in activity-dependent neuronal plasticity are believed to play a central role in AD-related cognitive impairment.  
To selectively enhance plasticity within PV-INs, we developed an optogenetic strategy of cell-type-specific expression of an optically activatable TrkB receptor (optoTrkB), the receptor for brain-derived neurotrophic factor (BDNF).   
Activation of optoTrkB in PV-INs in vitro reduced intrinsic excitability of PV-INs and attenuated glutamate-evoked calcium responses in surrounding pyramidal neurons, consistent with dampening of hyperactive network states. In vivo activation of optoTrkB in cortical PV-INs at an advanced stage of AD pathology led to a marked decrease in both the number and size of amyloid-beta (Aβ) plaques, accompanied by elevated pCREB levels, a downstream marker of BDNF/TrkB signaling, across a broad cortical region. Notably, overall microglial abundance remain unchanged. However, a modest increase in CD68 expression was detected in IBA1-positive microglia surrounding plaque cores, suggesting focal microglial activation without widespread proliferation.  
These findings indicate that selective enhancement of TrkB signaling in PV-INs restores aspects of circuit plasticity and is sufficient to reduce amyloid pathology even at late disease stages. The observed plaque reduction likely arises from neuron-driven mechanisms, possibly complemented by localized microglial responses. Overall, this work highlights PV-IN–specific optoTrkB activation as a promising disease-modifying strategy. 

28. SomaNet: Weakly Supervised Instance-Level Soma Segmentation in 3D Electron Microscopy with Partial Annotations 

Authors: Mohammad Khateri, Morteza Ghahremani, Jussi Tohka, Alejandra Sierra 

Author affiliations: Mohammad Khateri, Jussi Tohka, Alejandra Sierra are with A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Finland; Morteza Ghahremani is with Munich Center for Machine Learning, Technical University of Munich, Germany 

Rationale. The neuronal cell body, or soma, is a central landmark for studying brain organization, classifying cell types, and reconstructing neural circuits. Three-dimensional electron microscopy (3D EM) now resolves somas at nanometer scale, but identifying and outlining each individual soma across large volumes remains a major bottleneck. Manual annotation is extremely time-consuming, and in practice only a subset of somas is typically labeled, creating a need for methods that can learn from partial annotations.  
Methods. We introduce SomaNet, a deep learning framework that learns to segment every soma in a 3D EM volume using only partial annotations. SomaNet uses a teacher–student training strategy: a teacher model trained on the labeled somas generates reliable predictions on the unlabeled regions, which a student model uses to progressively learn from the entire volume. To separate closely touching somas, the network learns similarity relationships between neighboring voxels, producing sharp boundaries even in densely packed tissue. A 2.5D design combines a 2D network that captures the full extent of large somas within a section with a lightweight 3D component that ensures consistency across adjacent sections.   
Results. Evaluated on 3D EM volumes of the adult Drosophila brain, SomaNet accurately segmented individual somas in dense, morphologically diverse regions and outperformed existing methods across standard segmentation metrics.  
Conclusions. SomaNet substantially reduces annotation effort for soma segmentation in 3D EM, enabling scalable cell-type mapping, morphological analysis, and connectomic reconstruction. The framework generalizes to other 3D EM datasets, including mammalian and human tissues such as clinical epilepsy and tumor samples. 

29. Multimodal electrophysiological characterization of in vitro seizure-induction paradigms in resected human brain slices 

Authors: Vera Lezhneva1, Polina Abushik1, Mireia Gomez Budia1, Anssi Pelkonen1, Nataliia Novosolova1, Liudmila Saveleva1, Henri Eronen2, Omar Narvaez1, Mastaneh Torkamani-Azar1, Tuomas Rauramaa2,4, Arto Immonen5, Ville Leinonen2,5, Antti Huotarinen2, Kuopio Epilepsy Center Epilepsy Surgery Group3, Leena Jutila3, Jussi Tohka1, Olli Gröhn1, Reetta Kälviäinen2,3, Alejandra Sierra1, Tarja Malm1  

Author affiliations: 1A.I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland 2Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland 3Kuopio Epilepsy Center, Kuopio University Hospital, Member of ERN EpiCARE, Kuopio, Finland 4Department of Clinical Pathology, Kuopio University Hospital, Kuopio, Finland 5NeuroCenter Neurosurgery, Kuopio University Hospital, Kuopio, Finland 

To investigate the cellular drivers of human epileptogenesis, direct analysis of pathological network dynamics within ex vivo tissues is essential. Slices of resected human cortex serve as a vital experimental framework to evaluate intrinsic neuronal traits and disrupted circuit functions in a disease-relevant setting. The present research utilized a multimodal approach incorporating extracellular microelectrode arrays (MEA) and whole-cell patch-clamp recordings to analyze cortical samples from patients with focal, drug-resistant epilepsy (Kyyriäinen et al., 2025).  
The efficacy of two standard in vitro induction protocols, 100 μM 4-aminopyridine (4-AP) and high K⁺/low Mg²⁺, was assessed for their capacity to trigger epileptiform activity. Although these paradigms are standard in rodent models, their specific effects on the reorganized synaptic circuits and unique intrinsic properties of the human epileptogenic cortex remain poorly understood. Our results indicate markedly different network outcomes between the two models. Sustained, synchronized firing patterns resembling ictal events were consistently driven by a significant depolarizing shift in resting membrane potential under high K⁺/low Mg²⁺ conditions. In contrast, 4-AP application failed to induce comparable tonic depolarization, resulting instead in interictal-like discharges and sporadic burst firing without achieving sustained transitions.  
This study suggests that global ionic manipulation, rather than isolated potassium-channel blockade, generates a hyper-excitable state that more accurately reflects seizure activity in human tissue. These insights refine the translational utility of classical in vitro models, aiding the selection of appropriate methodologies for future human-centered epilepsy research. 

30. Intellectual Disability and Biomarkers of Alzheimer’s Disease and Neurodegeneration in the Evaluation of Suspected Dementia 

Authors: Jaana Laasanen, M.D, M.Sc, PhD, Maija Kantele, M.D., Eino Solje, Assoc. Chief Phys., Assoc. Prof., Head of the Brain Research Unit 

Author affiliations: Eino Solje, Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio ja Neuro Center – Neurology, Kuopio University Hospital, Kuopio. Jaana Laasanen, Kuopio University Hospital (the pediatric neurology clinic) and Center of expertise for disability services (Wellbeing services county of North Savo). Maija Kantele, Center of expertise for disability services (Wellbeing services county of North Savo)

Rationale  
People with intellectual disability (ID) have an elevated risk of developing dementia which often presents at a younger age and with atypical symptoms. The objective is to determine whether the altered behavior or function disability observed in an aging individual with ID is attributable to a neurodegenerative memory disorder or to other causes by using memory/neurodegenerative-related blood biomarker analyses.    
To obtain information on how memory/neurodegenerative biomarkers function in different etiologies of ID. More information is needed specifically about the aging of people with mild ID.  
    
Methods  
To collect blood samples and analyze memory/neurodegenerative biomarkers like pTau217, NfL etc. as a follow-up study.  
 
Results   
Memory-related blood biomarker analyses yield results categorized as normal, borderline, or abnormal.    
The purpose is to investigate whether the same biomarker thresholds used to support clinical decision-making and diagnostics in neurotypical older adults with early-stage dementia can be applied to individuals with IDs, or whether population-specific reference values are required for this patient group.   
 
Conclusion   
Specialised intellectual disability memory services and evaluation using reliable biomarkers of neurodegeneration are needed to improve diagnostic and prognostic in Down syndrome and other intellectual disabilities.    
Early diagnosis of Alzheimer’s disease in individuals with IDs improves prognosis, increases quality-adjusted life years, and enables more independent living by facilitating the timely initiation of pharmacological treatment. 

31. Barrier-mediated Clearance of Cerebrospinal Fluid in Idiopathic Normal Pressure Hydrocephalus 

Authors: Anssi Lipponen[1], Sami Heikkinen[1], Joel Räsänen [2,3], Tony Wyss-Coray[4], Evan Macosko[5], Tarja Malm[6], Mikko Hiltunen[1], Ville Leinonen[2,3] 

Author affiliations: [1] Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland. [2] Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland. [3] Institute of Clinical Medicine-Neurosurgery, University of Eastern Finland, Kuopio. [4] Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Stanford, CA, USA. [5] Broad Institute of MIT and Harvard, Boston, USA. [6] A.I. Virtanen Institute, University of Eastern Finland, Kuopio, Finland. 

Rationale  
In idiopathic normal pressure hydrocephalus (iNPH), cerebrospinal fluid (CSF) clearance is impaired, which can contribute to the accumulation of metabolic waste and disease-relevant proteins, making iNPH an excellent model for studying brain clearance mechanisms. Half of iNPH patients have Alzheimer’s disease (AD), and both show altered CSF levels of amyloid-β42 (Aβ42) and Tau. Our recent GWAS in FinnGen identified SLCO1A2, which showed two independent signals suggestive of a haplotype structure. SLCO1A2 encodes OATP1A2, a transporter involved in barrier mediated clearance, supporting its potential functional role in brain clearance.  
 
Methods  
Whole genome sequencing data from 317 patients in the Kuopio iNPH cohort were used to define SLCO1A2 haplotypes. Associations between haplotypes and iNPH phenotype were assessed using brain biopsy Aβ and Tau status, CSF Aβ42, Tau, and phosphorylated Tau (pTau) levels, and cognitive measures. To assess haplotype-specific barrier-mediated clearance, CSF/plasma proteomic ratios measured using SomaScan will be analyzed. FinnGen will be used to evaluate haplotype effects on age at onset of iNPH and AD.  
 
Results  
Preliminary results indicate that SLCO1A2 haplotypes are associated with altered age at onset of iNPH and AD. Haplotypes are also associated with altered CSF Tau levels and with Tau/Aβ42 and pTau/Aβ42 ratios in iNPH-patients.  
 
Conclusions  
These preliminary findings support a functional concept in which genotype-phenotype integration can identify haplotypes related to brain clearance. Based on the observed associations with Tau/Aβ42 ratios, we aim to validate haplotype-specific effects on Aβ42, Tau, and selected SomaScan-identified proteins using a cell model of the blood-CSF barrier.  

32. Polypharmacy and Comorbidity Burden in Adult Epilepsy Patients: A National Study from Kazakhstan 

Authors: Dina Kalinina, Temirgali Aimyshev, Alimzhan Muxunov, Abduzhappar Gaipov, Antonio Sarria-Santamera 

Author affiliations: Department of Medicine, School of Medicine, Nazarbayev University, Astana, Kazakhstan 

Rationale:  
Epilepsy often coexists with psychiatric and somatic conditions, leading to complex treatment regimens. In settings with limited access to newer antiseizure medications (ASMs), patients may face increased exposure to polypharmacy, raising concerns about safety and drug–drug interactions. Evidence from Central Asia remains scarce.  
 
Methods:  
A retrospective analysis of nationwide healthcare data was conducted, including over 54,000 adult patients with epilepsy. Prescription records were evaluated to identify ASMs and concomitant medications, classified using the Anatomical Therapeutic Chemical system. Patterns of monotherapy, polytherapy, and co-prescription across major drug classes were assessed.  
 
Results:  
ASM monotherapy was observed in approximately 62% of patients; however, a substantial proportion received multiple additional medications. Drugs acting on the nervous system were most frequently co-prescribed, with psycholeptics comprising nearly half of this group. Cardiovascular medications accounted for over one-third of non-ASM prescriptions. These findings indicate a high burden of psychiatric and cardiometabolic comorbidities. Despite gradual adoption of newer ASMs, prescribing patterns remain dominated by older agents, contributing to complex treatment combinations and potential interaction risks.  
 
Conclusions:  
Epilepsy care in Kazakhstan is characterized by significant multimorbidity and polypharmacy. Management strategies should address comorbid conditions alongside seizure control, with attention to medication safety and rational prescribing. Nationwide data provide valuable insight into real-world treatment complexity and may support efforts to improve epilepsy care in similar healthcare settings. 

33. The performance of plasma pTau181 andpTau217 in distinguishing Alzheimer’sdisease from various neurodegenerativedisorders, psychiatric disorders, andcognitively unimpaired controls ^a 

Authors: Juho-Antti Rissanen 1,2, Sari Kärkkäinen 1, Kasper Katisko 1, Aleksi Vanninen 3,4, Antti J. Luikku 3,4, Tuomas Rauramaa 5,6, Tadeusz Musialowicz 7, Merja Kokki 7,8, Valtteri Julkunen 1,2,  Anne M. Portaankorva 9, Annakaisa Haapasalo 10, Eino Solje 1,2, Päivi Hartikainen 2, Ville Leinonen 3,4, Tarja Kokkola 1 and Sanna-Kaisa Herukka 1,2  

Author affiliations: 1 Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland; 2 Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland; 3 Institute of Clinical Medicine – Neurosurgery, University of Eastern Finland, Kuopio, Finland; 4 Neuro Center – Neurosurgery, Kuopio University Hospital, Kuopio, Finland; 5 Institute of Clinical Medicine, Pathology, University of Eastern Finland, Kuopio, Finland; 6 Department of Pathology, Kuopio University Hospital, Kuopio, Finland; 7 Anaesthesiology and Intensive Care, Kuopio University Hospital, Kuopio, Finland; 8 School of Medicine, University of Eastern Finland, Kuopio Finland; 9 Clinical Neurosciences, Faculty of Medicine, University of Helsinki, Helsinki, Finland ;10 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale ^b: Plasma phosphorylated tau isoforms 181 (pTau181) and 217 (pTau217) are promising Alzheimer’s disease (AD) biomarkers.  
Objective: We evaluated the performance of pTau181 and pTau217 in the differential diagnostics between AD, other neurodegenerative diseases and non-neurodegenerative participants.  
Methods: We included 104 patients with neurodegenerative diseases (37 with AD, 21 with synucleinopathies [SYNU], 24 with frontotemporal dementia [FTD] and 22 with idiopathic normal-pressure hydrocephalus [iNPH]) and 50 participants without neurodegenerative disorders (33 individuals undergoing knee arthroplasty and 17 with psychiatric diagnoses).  
pTau181 and pTau217 were measured via single-molecule array.  
Results: pTau181 differentiated AD patients from psychiatric patients with an area under the curve (AUC) of 0.879 and AD patients from all other participants with an AUC of 0.685. pTau181 was higher in patients with AD compared to FTD, iNPH,  
and non-neurodegenerative (ND) patients. pTau217 differentiated AD patients from psychiatric patients, with an AUC of 0.998, and AD patients from all other groups, with an AUC of 0.835. pTau217 was higher in AD patients compared to ND, FTD, and SYNU patients, but it did not differ between AD and iNPH patients without adjustment for age as a covariate.  
Conclusions: Our prospective cohort data indicate that pTau217 differentiates AD patients from psychiatric patients, with an excellent AUC value in receiver operating characteristic analysis. Our study supports the use of pTau217 rather than pTau181 as a minimally invasive tool to differentiate AD from non-neurodegenerative diseases (e.g., psychiatric disorders). Further studies are needed to determine the nature of pTau217 in iNPH. 

^a Shared in accordance with CC BY 4.0. This abstract has been previously published as part of a journal article in the Journal of Alzheimer’s disease (DOI 10.1177/13872877261431800) 

^b  In the original publication the used term is “Background” 

34. The Neuropsychological Memory Questionnaire in Detecting Early-Onset Dementia in Memory Clinic Setting 

Authors: Aino Pirkkala[1], Noora-Maria Suhonen[1,2,3], Anna-Leena Heikkinen[1,2,3], Christer Hublin[4], Tuomo Hänninen[5], Anne M. Koivisto[5,6,7], Toni T. Saari[8], Anne M. Portaankorva[7], *Teemu I. Paajanen[4] & *Johanna Krüger[1,2,3]. *Shared last authors. 

Author affiliations: [1] Research Unit of Clinical Medicine, Neurology, University of Oulu, Finland. [2] Neurocenter, Neurology, Oulu University Hospital, Finland. [3] Medical Research Center Oulu, Oulu University Hospital, Finland. [4] Occupational Health Unit, Occupational Medicine, Finnish Institute of Occupational Health, Helsinki, Finland. [5] Neurocenter, Neurology, Kuopio University Hospital, Finland. [6] Department of Geriatrics and Neurological Memory Clinic, Helsinki University Hospital, Finland. [7] Clinical Neurosciences, University of Helsinki, Finland. [8] Institute for Molecular Medicine Finland (FIMM), University of Helsinki, Finland. 

Rationale: Early-onset dementia (EOD) poses a diagnostic challenge, as clinical presentations are often atypical. Informant reports may support the diagnostic process by capturing subtle cognitive and behavioral changes not readily detected during clinical assessment. The Neuropsychological Memory Questionnaire (NMQ) enables evaluation of informant- and patient-reported symptoms and their comparison, but its utility in identifying EOD has received little research attention.  
 
Methods: In a memory clinic population with symptom onset before age 65, we analyzed informant NMQ (NMQ-i) data from 174 participants, of whom 164 also completed the patient self-report NMQ (NMQ-p). Patients were categorized into four groups: EOD (n = 45), mild cognitive impairment due to neurological (MCI-n, n = 27) and other (MCI-o, n = 89) etiologies, and subjective cognitive decline (SCD, n = 13). Group differences in NMQ total and domain scores were assessed using the Kruskal–Wallis test. The diagnostic accuracy of the NMQ-i for EOD was evaluated using receiver operating characteristic analysis.  
 
Results: NMQ-i total scores differed between groups (p = .003), whereas NMQ-p scores did not (p = .556). Pairwise NMQ-i domain comparisons showed greater memory and visuo-practical impairments in EOD than in MCI-o (both p ≤ 0.005) and MCI-n (both p < 0.05). NMQ-i domains showed modest discrimination for EOD versus other groups (area under the curve, AUC = .60–.70).  
 
Conclusions: Informant NMQ ratings outperform patient self-reports in identifying EOD. The NMQ-i provides a structured tool for symptom assessment and monitoring, despite limited etiological differentiation.  

35. Unraveling disease mechanisms of PSP using an iPSC-derived astrocyte model 

Authors: Giacomo De Luca1, Dorit Hoffmann1, Sami Heikkinen2,3, Eino Solje2,3, Annakaisa Haapasalo1 

Author affiliations: 1. A.I. Virtanen Institute for Molecular Sciences, UEF    2. Institute of Clinical Medicine – Neurology, UEF  3. Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 

Rational: Astrocytic dysfunction and formation of 4R-Tau-positive (tufted) astrocytes are two main characteristics of progressive supranuclear palsy (PSP). Although PSP is the second most prominent parkinsonian disorder, the molecular and cellular pathophysiological mechanisms are not well understood. In this study, we aim to unravel the mechanisms that contribute to astrocytic dysfunction in PSP as well as the impact of PSP-related astrocytic dysfunction on the crosstalk between astrocytes and microglia.  
Methods: Skin fibroblasts from PSP patients and healthy individuals were reprogrammed into induced pluripotent stem cells (iPSCs) and differentiated into neural progenitor cells (NPCs) and further to astrocytes. Immunocytochemistry (ICC) was performed to assess the expression of NPC markers, such as Nestin and Sox. Global RNA sequencing, ICC focusing on astrocytic markers, such as GFAP, S100β, and AQP4, as well as functional analyses, including glutamate uptake, ATP-induced calcium signaling, lactate release, and cytokine-driven inflammatory responses will be performed to confirm the successful differentiation into astrocyte-like cells and to determine the phenotypes of PSP astrocytes. PSP astrocytes will be co-cultured with iPSC-derived microglia to gain insights into the potentially changed interaction between the two cell types.   
Results: ICC results confirmed the presence of NPC markers in the differentiated NPCs. NPCs are currently going through the last stages of differentiation into astrocytes. Next, ICC and other analyses of the differentiated astrocytes will be undertaken.  
Conclusions: Understanding the mechanisms that determine astrocytic dysfunction is a crucial step to unravel disease etiology in PSP, providing a foundation for the development of novel therapies. 

36. Donanemab in Early Symptomatic Alzheimer’s Disease: Additional Insights from the TRAILBLAZER‐ALZ 2 Long-Term Extension 

Authors: Jennifer A. Zimmer1, Cynthia Evans1, Emel Serap Monkul Nery1, Hong Wang1, Christophe Sapin1, Lars Lau Raket1, Marie-Ange Paget1, Ivelina Gueorguieva1, Rashna Khanna1, Scott W. Andersen1, Shoichiro Sato1, Dawn A. Brooks1, John R. Sims1, Mark Mintun1, Minna Korolainen (Non-author Presenter)2 

Author affiliations: 1 Eli Lilly and Company, Indianapolis, IN, USA; 2 Oy Eli Lilly Finland AB, Helsinki, Finland 

Introduction: TRAILBLAZER‐ALZ 2 (NCT04437511) was a multicenter, randomized, double-blind Phase-3 trial (76-week placebo-controlled (PC) and 78-week long-term extension (LTE) period).   
OBJECTIVES: During PC period, donanemab slowed clinical progression in early symptomatic Alzheimer’s disease (AD). This study describes the long-term efficacy and safety of donanemab in participants who completed treatment by 52-weeks.  
Methods: Randomised participants underwent amyloid PET scans every 24 weeks. Treatment completion criteria required amyloid plaque level <11 Centiloids (CL) on 1 PET scan or <25 CL on 2 consecutive scans. Participants meeting completion criteria received blinded placebo infusions. Efficacy was assessed by change in Clinical Dementia Rating–Sum of Boxes (CDR-SB), compared to an external control group from AD Neuroimaging Initiative (ADNI). Safety was assessed through MRI and adverse event reporting.  
Results: Of 860 participants, 325 completed treatment criteria within a year. At 52-weeks, donanemab-treated participants declined -0.6 CDR-SB points less than external controls, with benefits increasing to -1.1 and -1.3 points after 102 and 154 weeks; and had a mean amyloid plaque level of 10.99 at end of LTE. After participants completed donanemab treatment and switched to placebo, amyloid-related imaging abnormalities (ARIA) event rates were comparable to placebo.   
 
Conclusion: Participants with early symptomatic AD completing treatment by 52 weeks demonstrated sustained clinical benefits and maintained amyloid reduction, with ARIA rates similar to placebo after completing treatment. 

37. Smell function testing and alpha-synuclein seed amplification assay for inclusion and stratification in a clinical trial of exidavnemab in Parkinson’s disease and Multiple System Atrophy 

Authors: Ebba Amandius, Tomas Odergren, Teresa Sandvall, Petra Domeij, Olga Björklund, Linda Söderberg, Paulina Appelkvist, Patrik Nygren, Susanne Fabre, Gunilla Osswald, Gabrielle Åhlberg Hillert and Johanna Fälting 

Author affiliations: BioArctic AB 

Rationale: For a successful outcome in clinical trials, inclusion of patients with the targeted pathology is critical. In the ongoing EXIST, Phase 2a study with exidavnemab, targeting aggregated a-synuclein, we evaluated two new methods for enrichment of a homogenous patient population.   
Methods: To enrich the ongoing trial for exidavnemab in Parkinson’s disease (PD) and multiple system atrophy (MSA) for patients likely to be positive on the CSF alpha-synuclein seed amplification assay (SAA), we implemented a screening approach for PD subjects using reduced smell on the University of Pennsylvania Smell Identification Test (UPSIT) as inclusion criteria.   
Results: In PD Cohort 1, 13 of 14 patients subjected to the UPSIT screening were positive for reduced smell and proceeded to lumbar puncture. All 13 individuals had a positive SAA status. In the second PD cohort, all 13 participants subjected to the UPSIT were positive for reduced smell, and 10 of 13 were SAA positive. The full screening, including CSF SAA testing, was completed within the 6-week screening period for all subjects. Participants in the MSA cohort were not subjected to the UPSIT, although 11 of 12 randomized patients were SAA positive.   
Conclusions: The study outcome shows that pre-screening on the UPSIT followed by CSF SAA testing can be successfully applied in clinical trial practice with reasonable screening timelines. Where CSF SAA positivity is applied as an inclusion criterion, the UPSIT remains an applicable tool for prescreening that can possibly reduce the number of lumbar punctures performed to screen out SAA negative PD participants.  

38. Lecanemab demonstrates highly selective binding to Aβ protofibrils isolated from Alzheimer’s disease brains 

Authors: Linda Söderberg, Malin Johannesson, Nicolas Fritz, Eleni Gkanatsiou, Patrik Nygren, Gunilla Osswald, Lars Lannfelt, Christer Möller    

Author affiliations: BioArctic AB, Stockholm, Sweden 

RATIONALE   
Immunotherapy against amyloid β (Aβ) is a promising approach for treating Alzheimer’s disease (AD). In a phase 3 trial in participants with early AD, lecanemab demonstrated disease-modifying effects on clinical endpoints and promoted clearance of Aβ plaques. Lecanemab has been approved for the treatment of AD in several countries, including the USA and Japan. Lecanemab is a humanized IgG1 monoclonal antibody that selectively targets Aβ protofibrils, soluble Aβ aggregates are considered more toxic than monomers or insoluble fibrils. However, anti-Aβ antibodies can cause amyloid-related imaging abnormalities with edema (ARIA-E), potentially linked to binding of vascular Aβ fibrils, particularly in cerebral amyloid angiopathy (CAA). Here, we evaluated lecanemab binding to Aβ species isolated from post-mortem AD brain tissue.  
 
METHODS   
Protofibrils were extracted from human post-mortem cortical tissue, and CAA-associated fibrils were isolated from leptomeningeal tissue from AD patients. Aβ levels and the distribution of Aβ species across fractions were quantified using Aβ-specific immunoassays.  
 
RESULTS  
Protofibril levels were markedly higher in AD subjects than in controls. Lecanemab bound Aβ42-rich protofibrils across stages of disease severity, with minimal binding to Aβ40-rich fibrils associated with CAA.  
 
CONCLUSION  
Lecanemab preferentially targets soluble, aggregated Aβ protofibrils and shows limited binding to Aβ40-enriched fibrils in the cerebral vasculature. This limited binding to CAA-associated fibrillar structures may help explain the low frequency of ARIA-E reported for lecanemab. 

39. Real world data of causes of death in early- and late-onset neurodegenerative diseases in Finland 

Authors: Annemari Kilpeläinen1,2, Mikko Aaltonen3, Kasper Katisko1, Kalle Aho1, Sami Heikkinen1, Ave Kivisild1, Adolfina Lehtonen4, Laura Leppänen4, Iina Rinnankoski4, Helmi Soppela1, Laura Tervonen4,5,6, Päivi Hartikainen2, Annakaisa Haapasalo7, Johanna Krüger4,5,6 Eino Solje1,2 

Author affiliations: 1 = Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 2 = Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 3 = Law School, University of Eastern Finland, Joensuu, Finland  4 = Research Unit of Clinical Medicine, Neurology, University of Oulu, Oulu, Finland 5 = Medical Research Center, Oulu University Hospital, Oulu, Finland 6 = Neurocenter, Neurology, Oulu University Hospital, Oulu, Finland, Member of ERN-RND FTD 7 = A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale: Progressive neurodegenerative diseases are major cause of mortality yet often results from secondary complications rather than neurological disease itself. Comprehensive data on immediate causes of death (CoD), especially across the full clinical spectrum and in early-onset cases, remain scarce.  
Methods: In this study, we examined the causes of death associated with progressive neurodegenerative diseases by including all patients with a confirmed diagnosis from two defined regions in Finland. The cohort encompassed all major subtypes of neurodegenerative diseases (NDD), as well as both early-onset and late-onset cases. Patient data were collected between January 1, 2010, and December 31, 2021. We assessed early onset NDDs (EO-NDD) and late onset NDDs (LO-NDD) as separate groups.  
Results: Among 1,743 deceased patients, respiratory diseases were the leading immediate CoD in both groups (EO-NDD 56.7%, LO-NDD 45.7%). In EO-NDD, this was followed by neurological (5.0%) and cardiovascular diseases (4.3%), whereas in LO-NDD, cardiovascular (16.7%) and urinary system diseases (4.7%) were more prominent. Neurological diseases were a rare immediate CoD in LO-NDD (0.7%). As underlying causes, neurodegenerative diseases accounted for the most of deaths (EO-NDD 67.4%, LOD 52.9%).  
Conclusions: Neurodegenerative diseases predominantly contribute to death indirectly, with respiratory complications, particularly pneumonia, being the leading immediate cause across both early- and late-onset populations. Cardiovascular causes are more prominent in late-onset disease. These findings highlight the importance of managing systemic complications in addition to neurological decline.  

40. The protective PLCG2 variants delay Alzheimer’s disease onset age in APOE ε4 carriers 

Authors: Heli Jeskanen1, Sami Heikkinen1, Inka Kervinen1, Jenni Lehtisalo2,3, Tiia Ngandu2,4, Roosa-Maria Willman1, Jessica Rosa1, Dorit Hoffmann5, FinnGen6, Ville Leinonen7,8, Annakaisa Haapasalo5, Mari Takalo1†, Henna Martiskainen1†, Mikko Hiltunen1†  

Author affiliations:  † These authors contributed equally to this work 1 Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland 2 Department of Public Health, Lifestyles and Living Environments Unit, Finnish Institute for Health and Welfare (THL), Helsinki, Finland. 3 Institute of Clinical Medicine/Neurology, University of Eastern Finland, Kuopio, Finland. 4 Division of Clinical Geriatrics, Center for Alzheimer Research, Care Sciences and Society (NVS), Karolinska Institutet, Stockholm, Sweden. 5A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland; Kuopio, Finland 6 A full list of authors and their affiliations at the supplementary file of the manuscript 7Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland 8Institute of Clinical Medicine – Neurosurgery, University of Eastern Finland, Kuopio, Finland 

Rationale:  
The aim in this study was to investigate the impact of protective PLCG2-P522R and PLCG2-3’UTR variants, and the risk-associated TREM2-R62H variant, on the onset age of Alzheimer’s disease (AD), particularly in relation to APOE ε4 status. Additionally, plasma-based inflammatory and metabolic biomarkers of these genetic variants were explored.  
Methods:  
To evaluate the timing and risk of disease onset, we performed Kaplan-Meier survival and Cox proportional hazard analyses using genotype and clinical endpoint data from the FinnGen cohort, focusing on AD, anxiety, and type 2 diabetes (T2D). Plasma biomarkers related to metabolism and inflammation were measured in 145 participants from the FINGER cohort, along with body mass index (BMI), waist circumference, and waist-to-height ratio.  
Results:  
The PLCG2-P522R and PLCG2-3’UTR variants associated with delayed AD onset, including among APOE ε4 carriers. PLCG2-3’UTR reduced the risk of AD, especially in homozygous females carrying APOE ε4. Additionally, PLCG2-P522R carriers exhibited elevated plasma ghrelin levels compared to non-carriers. However, there were no differences in BMI, waist circumference or waist-to-height ratio between the genotypes. In contrast to the PLCG2 variants, TREM2-R62H was linked to an earlier onset of AD in males carrying APOE ε4.  
Conclusions:  
The PLCG2-P522R and PLCG2-3’UTR variants may mitigate APOE ε4-related AD risk, the PLCG2-P522R potentially through ghrelin-mediated neuroprotective mechanisms. Notably, PLCG2-P522R and PLCG2-3’UTR may exert sex-dependent effects, which could influence their protective role in AD. These findings highlight the importance of further investigation of plasma biomarkers and molecular pathways associated with protective genetic variants, especially in the context of APOE ε4.  

41. Challenging diagnosis of atypical central neurocytoma – a case report 

Authors: Mari Valkonen MD PhD, Paula Bendel MD PhD, Tuomas Rauramaa MD PhD, Susanna Rantala MD PhD 

Author affiliations: Department of Clinical Pathology, Kuopio University Hospital, Kuopio, Finland. Unit of Pathology, Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland. Department of Radiology, Kuopio University Hospital, Kuopio, Finland.  Neurosurgery KUH NeuroCenter, Institute of Clinical Medicine, Kuopio University Hospital, University of Eastern Finland, Kuopio, Finland   

Rationale  
We report a case of challenging diagnosis of a rare, atypical central neurocytoma causing obstructive hydrocephalus. Despite the advancements of molecular diagnostic methods, these tumors lack definite molecular features. Our case demonstrates unusually anaplastic histological features in this rare central nervous system (CNS) tumor.  
Methods  
Brain magnetic resonance imaging (MRI) showed a tumor in left frontal horn of previously healthy patient. Previously patient had suffered from headaches and vomiting and due to obstructive hydrocephalus and midline swift. The tumor was swiftly resected in Neurosurgery unit of Kuopio University Hospital. Histologically this tumor demonstrated anaplastic features including necrosis and increased mitotic activity. Methylation classifier and molecular analyses failed to classify the tumor.  
Results  
After deliberation the diagnosis of atypical central neurocytoma was made. This tumor type still lacks official diagnostic criteria in the current WHO classification of CNS tumors. In the literature these tumors are reported to have increased mitotic activity, histological atypia and necrosis. These features are uncommon in WHO grade 2 central neurocytomas which generally have favorable prognosis. Due to these features close follow-up with MRI was recommended for our patient.   
Conclusions  
Despite the progress of molecular diagnostic methods, the diagnosis of rare CNS tumors is still challenging and depends on multidisciplinary approach. Furthermore, the current WHO classification of CNS tumors lacks the definite and official criteria for atypical central neurocytomas. Especially, increased mitotic counts have been associated with the reports of increased aggressive behavior compared to traditional WHO grade 2 central neurocytomas.  

42. A multimodal MRI–TMS study of cortico-thalamo-cerebellar dysfunction in EPM1: dissociation between diffusion metrics and network organization 

Authors: Yawu Liu, Katri Silvennoinen, Janina Gunnar, Henri Eronen, Mervi Könönen, Juhana Hakumäki, Reetta Kälviäinen 

Author affiliations: Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland; Imaging Center, Kuopio University Hospital, Kuopio, Finland; Kuopio Epilepsy Center, Kuopio University Hospital, Kuopio, Finland. 

Rationale  
Unverricht–Lundborg disease (EPM1) is a progressive myoclonus epilepsy marked by cortical hyperexcitability and motor dysfunction. Although diffusion tensor imaging (DTI) studies have shown widespread white matter abnormalities, their functional relevance remains unclear. Structural covariance networks (SCNs) may better capture system-level changes linked to neurophysiological dysfunction. We investigated whether DTI-based microstructural measures or SCN organization better explains cortical excitability in EPM1.  
Methods  
Ninety-two patients with EPM1 underwent multimodal assessment including structural MRI (n=85), DTI (n=69), and transcranial magnetic stimulation (TMS; n=64). Fractional anisotropy and mean diffusivity were extracted from cortico-thalamo-cerebellar tracts. SCNs were constructed from regional brain volumes focusing on the motor cortex, thalamus, cerebellum, and putamen. Principal component analysis generated summary measures for DTI and SCN features. Linear regression tested associations of these measures with TMS-derived motor threshold and cortical silent period, adjusting for age, sex, disease duration, and scalp–cortex distance, with false discovery rate correction.  
Results  
DTI measures were not associated with SCN metrics or cortical excitability (all p>0.1). In contrast, SCN measures showed associations with motor threshold (node strength: β=−1.56, p=0.020; edge strength: β=−1.80, p=0.049), with trend-level significance after correction. Greater SCN connectivity was linked to lower motor threshold, indicating higher cortical excitability. No associations were found with cortical silent period.  
Conclusions  
In EPM1, network organization rather than diffusion-based microstructural integrity relates to cortical excitability, supporting network-based biomarkers for progressive myoclonus epilepsy.  

43. Environmental Nanoplastics as Emerging Drivers of Parkinson’s Disease Pathology 

Authors: Demet Sinem Guden-Yildirim¹, Sara Kälvälä¹, Kelvin C. Luk², Katrina Albert¹, Šárka Lehtonen¹  

Author affiliations: ¹A.I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, 70211 Kuopio, Finland  ²Center for Neurodegenerative Disease Research, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Perelman School of Medicine, Philadelphia, PA 19104-4238, USA 

Rationale: Parkinson’s disease (PD) is currently one of the fastest-growing neurological disorders worldwide, with environmental factors increasingly implicated in its pathogenesis. Micro- and nanoplastics (MNPs), pervasive environmental pollutants, have recently been detected in human tissues, including the brain, raising concerns about their long-term biological effects. However, their role in neurodegeneration remains unclear.  
 
Methods: Here, we investigated the impact of polystyrene nanoplastics (PS-NPs) on α-synuclein (αSyn) pathology using human induced pluripotent stem cell-derived dopaminergic neurons and midbrain organoids. Cells were exposed to PS-NPs with different surface charges and concentrations, followed by treatment with αSyn pre-formed fibrils (PFFs). Nanoparticle uptake, cell viability, and αSyn accumulation were assessed via live-cell imaging and immunofluorescence, alongside analyses of neuronal morphology and organoid growth and function.  
 
Results: We found that anionic PS-NPs (10 μg/ml) were efficiently internalized and increased total αSyn levels without inducing overt cytotoxicity, in contrast to higher concentrations. Co-exposure to anionic PS-NPs and αSyn PFFs resulted in reduced neurite length and branching complexity. In midbrain organoids, PS-NP exposure altered organoid size and impaired neuronal function.  
 
Conclusions: These findings demonstrate that NPs can modulate αSyn homeostasis and exacerbate neuronal dysfunction at sub-cytotoxic levels. Our results provide mechanistic evidence supporting the role of MNPs as potential environmental modifiers of PD pathology and highlight their relevance to neurodegenerative disease risk.  

44. Lithium chloride alters Tau phosphorylation, kinase activity, and Rho GTPase signaling in cell models 

Authors: Dorit Hoffmann1, Virpi Ahola2, Nadine Huber1, Teemu Natunen2, Stina Leskelä1, Mari Takalo2, Henna Martiskainen2, Stephanie Ballweg3, Egor Vorontsov3, Stefan Selzer3, Pekka Kallio4, Ian Pike5, Jouni Sirviö6, Annakaisa Haapasalo1, Mikko Hiltunen2 

Author affiliations: 1 A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, 70211 Kuopio, Finland,  2 Institute of Biomedicine, University of Eastern Finland, 70211 Kuopio, Finland; 3 Proteome Sciences R&D GmbH & Co. KG, Germany, 4 Finnish Drug Discovery Center, 5 Proteome Sciences plc. Hamilton House, London, United Kingdom 6 Oy Sauloner Ltd, Kuopio, Finland 

Rationale  
Hyperphosphorylation and intracellular aggregation of Tau are pathological hallmarks of Alzheimer’s disease (AD). Modulation of Tau phosphorylation using lithium salts has yielded mixed results in clinical trials for AD patients. However, a recent study showed that lithium can be sequestered into Aβ plaques and that using Aβ-evading salts could improve clinical outcomes, making them attractive therapeutic candidates. Here, we aimed to gain an in-depth understanding of the biological effects of lithium chloride (LiCl), which has previously been shown to reduce Tau phosphorylation.  
Methods   
We assessed the effects of LiCl treatment using two different cell-based models: A co-culture model of mouse embryonic primary cortical neurons and BV-2 microglial cells under neuroinflammation and a U2OS cell line overexpressing human triple mutant Tau. We assessed changes in Tau phosphorylation using Western blot and phosphoproteomics analyses and utilized pathway and kinase-substrate enrichment analyses to assess the impact on different kinases and other AD-relevant pathways.  
Results   
Treatment with LiCl reduced Tau phosphorylation depending on the concentration, the examined phosphosites, and the cell model. Phosphoproteomics data also suggested that LiCl affects different kinases and the phosphorylation status of several proteins belonging to different Rho GTPase cycles.  
Conclusions   
We identified several new AD-relevant Tau phosphorylation sites affected by LiCl treatment. Moreover, our data suggest that LiCl may affect other kinases beyond GSK-3β, a kinase commonly phosphorylating Tau in AD, and also Rho GTPases, known to play a role in AD pathogenesis. The effects of LiCl may extend beyond Tau phosphorylation and thus warrant further examination.  

45. The Emergence of a Purpose-Oriented Health Research Network: Sustaining Shared Purpose and Collaboration Over Time 

Authors: Hannaneh Moradi[1], Tero Montonen[2]

Author affiliations:  

1. Doctoral Researcher, University of Eastern Finland, UEF Business School, Kuopio, Finland 

2. Professor, University of Eastern Finland, UEF Business School, Kuopio, Finland 

Purpose-oriented research networks are increasingly addressing complex health challenges that no single organisation can tackle alone. Yet how such networks actually unfold over time remains poorly understood. Questions of how a network’s role shifts, how its purpose is sustained, and how its identity is renegotiated as it grows are largely absent from existing research. Understanding network evolvement is also practically relevant; those working within and leading health research networks benefit from knowing how such networks take shape and change over time.  
A qualitative, longitudinal case study was conducted on an international health research network over approximately ten years. Empirical material consisted of semi-structured interviews with nine core managers and founding members, supported by longitudinal analysis of the network’s public website across nine years. Data were analysed using Braun and Clarke’s reflexive thematic analysis. Identity-related themes emerged inductively from the material rather than being imposed as a predetermined lens.  
The network evolved from informal scientific collaboration toward a more structured and internationally recognised organisation. Throughout this evolution, leaders continually returned to questions of who the network is, what it exists to do, and how its purpose should be understood as membership and geographic reach widened. Identity evolvement and organisational evolvement were found to be deeply intertwined.  
A longitudinal perspective reveals how purpose-oriented brain health research networks are not static structures but continuously reworked through ongoing narration of role, purpose, and meaning with practical implications for how such networks are built and sustained. 

46. Behavioral Changes Associate with Cortical Circuit Function Independent of Neurocognitive Disorder 

Authors: Matti Tyrväinen¹, Laura Säisänen² ⁷, Jelena Hyppönen¹ ³, Sara Määttä¹ ³, Kaarlo Ylimäki², Meri Juslin³, Mika Kinnunen², Kasper Katisko², Johanna Kruger⁴ ⁵ ⁶, Noora-Maria Suhonen⁴ ⁵ ⁶, Päivi Hartikainen⁷, Barbara Borroni⁸, Annakaisa Haapasalo⁹, Esa Mervaala¹ ³, Eino Solje² ⁷  

Author affiliations: 1. Institute of Clinical Medicine – Clinical Neurophysiology, University of Eastern Finland (UEF), Finland; 2. Institute of Clinical Medicine – Neurology, UEF, Finland; 3. Department of Clinical Neurophysiology, Kuopio University Hospital (KUH), Finland; 4. Research Unit of Clinical Medicine, Neurology, University of Oulu, Finland; 5. MRC, Oulu University Hospital, Finland; 6. Neurocenter, Neurology, Oulu University Hospital, Finland; 7. Neuro Center – Neurology, KUH, Finland; 8. Neurology Unit, Centre for Neurodegenerative Disorders, Department of Clinical and Experimental Sciences, University of Brescia, Italy; 9. A. I. Virtanen Institute for Molecular Sciences – UEF, Finland 

Rationale  
Behavioral changes are characteristic for frontotemporal dementia and other neurocognitive disorders. Transcranial magnetic stimulation (TMS) can be used to non-invasively assess disease-specific changes in inhibitory and excitatory processes. We aimed to determine the association of TMS variables with personality and behavioral changes captured using the Modified Frontal Behavior Inventory (FBI-mod) questionnaire, independently from diagnostic classification.  
 
Methods  
In total, 130 patients (37 % female, mean age 68.1 years) presented to the neurology outpatient clinic with cognitive or memory complaints were included. For TMS, motor threshold and paired-pulse paradigms for short interval intracortical inhibition (SICI), intracortical facilitation (ICF), long-interval cortical inhibition (LICI) and short-latency afferent inhibition (SAI) were measured. The 22 items in FBI-mod were filled by a relative of the patient. For each negative and positive symptom, patients were grouped according to its presence, and TMS variables were compared between groups using the Mann-Whitney U test.  
 
Results  
Multiple associations were found in the symptom-wise analysis. Most notably, weaker SICI was associated with disorganization, inattention, and logopenia (Z = -3.04, p < .01; Z = -2.7, p < .01; Z = -2.48, p = .013, respectively), whereas weaker SAI was associated with aspontaneity, inattention, and poor judgement[LS3.1]. (Z = -2.61, p < .01; Z = -2.57, p = .01; Z = -2.26, p = .024, respectively).  
 
Conclusions  
Changes observed in the GABAergic (SICI), glutamatergic (ICF) and acetylcholinergic (SAI) systems were found to correlate with specific behavioral symptoms, independently from the underlying diagnostic entity, paving way for novel treatment targets. 

47. Sleep-time subclinical Epileptiform Discharges (EDs) in Alzheimer’s disease – how harmful is it? 

Authors: Häkli Sara1, Koivisto Henna1, Gureviciene Irina1, Schimmel Jonas2, Castell Caubet Eric3, Rozas Casero Maria4, Beinert Diana2, Malathi Keerthana5, Jin Nanxiang1 and Tanila Heikki1 

Author affiliations: 1 A. I. Virtanen Institute, University of Eastern Finland, Finland, 2 Hasselt University, Belgium, 3 Universitat de Barcelona, Spain, 4 Universidad Autonoma de Madrid, Spain, 5 Indian institute of science education, India 

Objectives: Sleep-time subclinical Epileptiform Discharges (ED have been found to correlate with faster cognitive decline and memory impairment in human Alzheimer’s disease (AD) patients, compared to AD patients who do not demonstrate EDs. The systems consolidation theory suggests that during sleep, memories are transferred from short-lasting working memory in the hippocampus to long-term storage in the cortex, which requires synchronization of hippocampal, cortical and thalamo-cortical oscillations. We set a study to investigate how memory consolidation is altered when the amount of interictal spikes is manipulated with drug treatments.   
 
Methods: APP/PS1 transgenic (n=15) and wild-type (n=9) male mice at 3.5-5.5 months of age were implanted with cortical, hippocampal and olfactory bulb EEG electrodes. Each mouse underwent a weekly test session consisting of 1) memory training (Barnes circular platform/Novel Object Recognition (NOR)), 2) drug administration and 3) 3h video-EEG recorded after the memory training. The amount of interictal spiking during the post-learning sleep period was manipulated with either Lamotrigine (increased spiking frequency), Levetiracetam (decreased spiking) or vehicle (baseline). On the next day, the recalling ability of the previous day’s training was tested.  
 
Results: The final results will be revealed during the symposium.  

48. Establishing a CRISPR-screen based platform to study schizophrenia-associated genetic risk factors in human brain organoids 

Authors: Ohtonen Sohvi¹*, Zhou Yan¹*, Chaudhary Mridul¹, Rigamonti Alessandra¹, Samudyata¹, Sellgren Carl¹ 

Author affiliations: 1. Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden 

Rationale  
Schizophrenia (SCZ) is a severe psychiatric disorder with a strong genetic component, yet its underlying disease mechanisms remain poorly understood. It is suggested that disruptions during early neurodevelopmental contribute to the disease onset in late adolescence. Although numerous genetic risk loci and genes have been identified, the functional role of these factors in neurodevelopment is not well established. Here, we combine CRISPR/Cas9-based genetic screening with human induced pluripotent stem cell (iPSC)-derived brain organoids to investigate the impact of SCZ-associated genes in a human-specific context.  
Methods  
Candidate genes for the CRISPR-screen were selected from recent genome-wide association studies and validated for expression in forebrain organoids using an integrated transcriptomic atlas of human neural organoids. A doxycycline-inducible Cas9 (iCas9) iPSC-line was generated. A lentiviral guide RNA (gRNA) library targeting prioritized genes is introduced into iCas9-iPSCs, followed by enrichment of gRNA-containing cells are via fluorescent reporter -based sorting. Sorted iPSCs are differentiated into dorsal forebrain organoids, and genetic perturbations are induced by doxycycline treatment. The effects of genetic perturbations in cellular survival and differentiation will be assessed using gRNA enrichment analysis as well as single-cell sequencing.  
Results  
A set of schizophrenia-associated genes was defined and confirmed to be expressed in human forebrain organoids. A Cas9-expressing iPSC line was successfully generated. Optimization of gRNA delivery, enrichment of gRNA-expressing cells, and dorsal forebrain organoid differentiation is ongoing.  
Conclusion  
This approach enables systematic analysis of SCZ-associated genetic risk factors in human brain organoids, yielding disease-relevant insight into neurodevelopment and broader applicability to other neurodevelopmental disorders. 

49. INCIDENCE TRENDS OF YOUNG-ONSET FIRST-EVER STROKE IN NORTH KARELIANS 2010-2024 

Author: Jussi Sipilä  

Author affiliations: University of Eastern Finland 

Rationale: Young-onset (<50 years of age) ischemic stroke incidence is increasing around the world but data are insufficiently available from Finland. Stroke risk is high in the easternmost part of the country, North Karelia, where possible trends should be readily apparent.  
Methods: All patients 16-50 years of age when diagnosed with first-ever stroke were identified from the patient records of North Karelia Central hospital, which served a population of 66,298 (2010) 62,006 (2024) of this age.  
Results: Altogether 336 patients (37.2% women) with a first-ever stroke were identified. Ischaemic stroke [temporary ischemic attack (TIA) or cerebral infarction (CI)] constituted 75.6 % of these, cerebral venous thrombosis (CVT) 1.8% and the rest were haemorrhagic [61% of these subarachnoid haemorrhages (SAH)]. In 2010-2018 annual overall incidence was never <30/100,000 person-years and >40/100,000 in five years (201, 2013, 2015, 2017-2018) while in 2019-2024 it was constantly <30/100,000 person-years save for 2024 (35.1/100,000). Overall incidence changed from 37.9 (2010-2014) via 39.4 (2015-2019) to 27.6 (2020-2024) per 100,000 person-years (nonlinear regression slope -1.08, 95% CI -2.10 to -0.07). The incidences of SAH, TIA and CI declined during the study period while that of intracerebral haemorrhage (ICH) remained stable. These trends were clearer in men. Six CTSs were observed in 2010-2019 and zero in 2020-2024.   
Conclusions: Young-onset first-ever stroke incidence declined in north karelians. Further studies are needed to elucidate the causes for this and investigate if the trend will continue. It is unclear why ICH rates have not declined like those of other strokes. 

50. Early Alzheimer’s disease pathology alters ultrastructure of synapses and synapse-glia interaction 

Authors: Henna Jäntti1, Zewen Zhuo1, Polina Abushik1, Anssi Pelkonen1, Mireia Gómez-Budia1, Mohammad Rezaie1, Annajulia Järvelä1, Graham Knott2, Ilya Belevich3, Ali Abdollahzadeh1, Mikko Hiltunen4, Tuomas Rauramaa5,6, Ville Leinonen7,8, Alejandra Sierra Lopez1, Jussi Tohka1, Tarja Malm1  

Author affiliations: 1A.I.Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland.  2 Bioelectron Microscopy Core Facility,  Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland. 3 Electron Microscopy Unit, Institute of Biotechnology, Helsinki Institute of Life Science, University of Helsinki, Helsinki, Finland. 4 Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland. 5 Institute of Clinical Medicine, Pathology, University of Eastern Finland, Kuopio, Finland. 6 Department of Pathology, Kuopio University Hospital, Kuopio, Finland. 7 Institute of Clinical Medicine, Neurosurgery, University of Eastern Finland, Kuopio, Finland. 8 Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland. 

Synaptic dysfunction is a hallmark of Alzheimer’s disease (AD), yet its early ultrastructural correlates in the human brain remain poorly defined. Here, we investigated AD pathology-associated changes in human cortical brain biopsies from donors with and without AD-related pathology.  
We analysed 17 Serial Block-Face Scanning Electron Microscopy (SB-EM) datasets from layer III of Brodmann area 9 obtained from patients with idiopathic normal pressure hydrocephalus (iNPH; 63–86 years). These rare live-biopsy samples enabled the study of preclinical and early-stage pathology across four neuropathological groups: control (C) with no amyloid or tau, amyloid beta (Ab), amyloid beta and tau (A+T), and tau only (T). Based on earlier electrophysiological and transcriptomic evidence of regional hyperexcitability, we imaged a 33×25×18 μm block at 8×8×62.5 nm voxel resolution from cortical layer III, validated using Nissl staining, IBA1 and WO2 immunohistochemistry, and MEA recordings from consecutive sections.  
Using WebKnossos and Voxelytics AI-based 3D instance segmentation, we performed automated detection and volumetric segmentation of mitochondria and excitatory (asymmetric) and inhibitory (symmetric) synaptic clefts, with manual reconstruction of microglia, dendrites, and pre- and postsynaptic boutons. Preliminary analyses treating each donor as one biological replicate revealed pathology-associated alterations in microglial morphology, mitochondrial number and volume, excitatory and inhibitory synapse density, synaptic cleft volume, microglia–synapse interactions, and novel presynaptic inclusions. Together, we establish SB-EM as a powerful approach for resolving synapse pathology in human brain disease and identify novel early ultrastructural correlations of human AD pathology aligning with prior functional and transcriptional evidence of hyperactivity. 

51. SYSTEMIC-BRAIN METABOLIC COUPLING IN AGING: EFFECTS OF THE PROTECTIVE P522R VARIANT IN THE PLCγ2 GENE 

Authors: Jessica Rosa1, Inka Kervinen1, Heli Jeskanen1, Emine Erdag1, Petra Mäkinen1, Mari Taipale2, Janne Capra1, Sami Heikkinen1, Annakaisa Haapasalo2, Heikki Tanila2, Mikko Hiltunen1 & Mari Takalo1 

Author affiliations: 1 Institute of Biomedicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland; 2 A. I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale  
Aging, a major risk factor for neurodegeneration, is associated with systemic metabolic alterations and impaired cellular homeostasis in the brain. The P522R variant in the PLCγ2 gene, which encodes the PLCγ2 (phospholipase C gamma 2) enzyme, has been associated with a reduced risk of Alzheimer´s disease and other neurodegenerative diseases. PLCγ2-P522R has been shown to modulate microglial responses and enhance cellular energy production. This study aims to determine whether the neuroprotection associated with the PLCγ2-P522R variant in aging is mediated by modulation of systemic-brain metabolic interactions and whether these effects are sex-dependent.   
Methods  
We are conducting a comparative study in aged wild-type (WT) and PLCγ2-P522R knock-in (KI) mice of both sexes. A protocol for metabolic phenotyping was established using an automated home-cage system that continuously records indirect calorimetry data via non-invasive sensors. Estrous cyclicity is determined both before and after metabolic phenotyping. Systemic metabolic status is also assessed via circulating hormones. Brain cellular and metabolic states are evaluated using cutting-edge techniques, including spatial transcriptomics, lipidomics, and molecular imaging, with a focus on mitochondrial and lysosomal pathways.  
Results  
Data collection and analysis are currently ongoing. Preliminary results will be presented at the symposium.  
Conclusions  
By integrating hormonal, metabolic, molecular, and behavioral data across sexes in aged mice, this study will provide insights into the mechanisms of genetic resilience mediated by the PLCγ2-P522R variant, with potential clinical applications to promote healthy aging and address neurodegenerative diseases. 

52. Impact of Alzheimer’s and cardiovascular genetic risk scores on AD blood biomarkers in the FINGER multidomain lifestyle randomized controlled trial 

Authors: Gazi Saadmaan Hossain, MBBS, MPH1; Anna Rosenberg, PhD2,3; Alfredo Ramirez, MD, PhD4,5,6,7,8; Anna Matton3,9,10; Andrea L. Benedet11; Esko Levälahti, MSc2,12;  Francesca Mangialasche, MD, PhD3,9,13; Henrik Zetterberg11; Hilkka Soininen, MD, PhD1; Jenni Lehtisalo, PhD2,12; Maria Carolina Dalmasso, PhD14; Mikko Hiltunen, PhD15; Minna U. Kaikkonen, PhD16; Tiia Ngandu, MD, PhD2,3,12; Miia Kivipelto, MD, PhD3,9,10,12,13; Alina Solomon, MD, PhD1,3,10 

Author affiliations: 1. Department of Neurology, Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland. 2. Department of Public Health, Lifestyles and Living Environments, Finnish Institute for Health and Welfare, Helsinki, Finland. 3. Division of Clinical Geriatrics, Center for Alzheimer Research, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden. 4. Division of Neurogenetics and Molecular Psychiatry, Department of Psychiatry and Psychotherapy, University of Cologne, Medical Faculty, Cologne, Germany. 5. Department of Neurodegenerative Diseases and Geriatric Psychiatry, University Hospital Bonn, Bonn, Germany. 6. German Center for Neurodegenerative Diseases (DZNE Bonn), Bonn, Germany. 7. Glenn Biggs Institute for Alzheimer’s & Neurodegenerative Diseases, University of Texas Health Sciences Center, San Antonio, TX, USA. 8. Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases (CECAD), University of Cologne, Cologne, Germany. 9. FINGERS Brain Health Institute, Stockholm, Sweden. 10. The Ageing Epidemiology Research Unit, School of Public Health, Imperial College London, London, UK. 11. Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, The Sahlgrenska Academy, University of Gothenburg, Mölndal, Sweden. 12. Institute of Public Health and Clinical Nutrition, University of Eastern Finland, Kuopio, Finland. 13. Theme Inflammation and Aging, Karolinska University Hospital, Sweden. 14. Studies in Neuroscience and Complex Systems Unit (ENyS-CONICET-HEC-UNAJ), Florencio Varela, Argentina. 15. Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland. 16. A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Background: Alzheimer’s disease (AD) plasma biomarkers are minimally invasive, cost-effective ways to assess pathology. Little is known about the associations between emerging AD plasma biomarkers of amyloid, tau, and neuroinflammation and genetic risk scores (GRSs), among at-risk older adults without substantial cognitive impairment. Therefore, we investigated associations between AD plasma biomarkers and GRSs for AD and coronary artery disease (CAD) in the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER)  
Method: FINGER randomized controlled trial included 60-77 years, at-risk individuals without dementia, randomized to 2-year multidomain lifestyle-based intervention or regular health advice (N=1260). Plasma biomarkers were analyzed in all available banked baseline (n=1225) and 2-year (n=1105) samples, using the NULISAseq™ CNS Disease Panel120 (Aβ42/40, ptau181, ptau217, ptau231, NfL, GFAP) and Lumipulse (ptau217). GRSs (n= 1177) were based on a recent comprehensive AD-GRS (83 genome-wide significant variants, Nat Genet 2022;54(4):412-436) and a validated CAD-GRS (6,338,802 variants, Nat Genet 2018;50(9):1219-1224).  
Result: This study included 1149 participants (mean age=68.8 years; 527[45.9%] female). At baseline, N=80(6.6%) had abnormal Lumipulse ptau217 (>0.34 pg/ml). Preliminary baseline analyses suggest that higher CAD-GRS was associated with higher GFAP and NfL (p<.05) after adjusting for age, sex, and APOE. No significant correlation between other AD plasma biomarkers and GRSs were observed. Results will also be shown from ongoing analysis investigating the impact of GRSs on biomarker changes over time.  
Conclusion: Results demonstrate associations between CAD-GRS and markers of inflammation and neurodegeneration. Findings from these exploratory analyses need to be further tested across several multimodal lifestyle trials. 

53. THE MECHANISMS OF TMEM106B PROTECTIVE VARIANT IN CELLULAR AND PATHOLOGICAL PROCESSES OF NEURODEGENERATION 

Authors: Sini-Pauliina Juopperi1, Susanna Kemppainen1, Anssi Lipponen1, Mikolaj Wojnicki1,2, Roosa-Maria Willman1, Sami Heikkinen1, Henna Martiskainen1, Nadine Huber3, Annakaisa Haapasalo3, FinnGen, Tuomas Rauramaa2,4, Ville Leinonen2,5, Tarja Malm3, Mikko Hiltunen1 

Author affiliations: 1 Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland 2 Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland 3 A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland, 4 Department of Pathology, Kuopio University Hospital, Kuopio, Finland 5 Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland 

Rationale: Transmembrane protein 106B (TMEM106B) is a lysosomal membrane protein expressed in neurons and glia. Its C-terminal domain forms amyloid filaments in the brain in neurodegenerative diseases (NDDs) and aging, but its role in disease pathogenesis remains unclear. Genetic variants in TMEM106B modulate risk for frontotemporal dementia and Alzheimer’s disease (AD). A threonine-to-serine substitution at position 185 (TMEM106B Thr185Ser; rs3173615 C>G) has been associated with protection, although the underlying mechanisms are unknown.  
Methods: TMEM106B haplotype variants described in Salazar et al. (2025) were identified in FinnGen and analyzed for association with AD in a Finnish genome-wide association study. Effects on AD onset age were assessed. Cellular mechanisms of rs3173615 were investigated in postmortem brain samples of NDD patients.  
Results: 185-Ser associated haploblock in the TMEM106B coding region reached genome-wide significance in the Finnish AD cohort, while haploblock in the upstream regulatory region did not. The 185-Thr-associated haplotype showed increased AD risk (HR 1.22, p=3.6×10⁻¹⁰) compared to 185-Ser-associated haplotype. 185-Ser/Thr (HR 0.89, p=2.71×10⁻⁹) and 185-Ser/Ser (HR 0.84, p=3.82×10⁻¹⁰) genotypes showed reduced risk and delayed onset.  
TMEM106B C-terminal staining in inferior temporal cortex increased in advanced Braak stages, while 185-Ser had no effect. pTDP-43 pathology was absent in homozygous 185-Ser/Ser carriers.  
Conclusions: 185-Ser-associated haploblock in the TMEM106B coding region, but not upstream regulatory variants, modifies AD risk and onset in the Finnish population. Ongoing studies in postmortem brain samples, brain biopsies of idiopathic normal pressure hydrocephalus patients, and human induced pluripotent stem cell -derived microglia aim to clarify rs3173615-associated mechanisms in neurodegeneration.  

54. xACSON: Myelinated axon segmentation in X-ray holographic nano-tomography 

Authors: Reyhaneh Aghayousefi¹, Michela Fratini², Artem Vorobyev³, Sofia Boccanera², Omar Narvaez¹, Ilya Belevich⁴, Eija Jokitalo⁴, Alejandra Sierra¹, Alexandra Pacureanu³, Ali Abdollahzadeh¹ 

Author affiliations: 1 A.I. Virtanen Institute, University of Eastern Finland, Kuopio, Finland.  2 CNR-Nanotec (Institute of Nanotechnology), Rome, Italy. 3 European Synchrotron Radiation Facility, The European Synchrotron, Grenoble, France. 4 Electron Microscopy Unit, Institute of Biotechnology, University of Helsinki, Helsinki, Finland. 

Introduction: X-ray holographic nano-tomography (XNH) enables non-destructive, isotropic 3D imaging of densely stained neural tissue over large volumes ∼1mm³ at sub-100 nm resolution. We developed an automated pipeline for segmentation and morphology analysis of white matter XNH tomograms, xACSON. xACSON segments myelin and intra- and the extra-axonal spaces, tracing the entirety of individual axon instances, yet extracting local axonal morphologies such as axonal diameter along their trajectories.  
Method: We acquired an XNH tomogram from the splenium of the corpus callosum of a sham-operated rat, at the European Synchrotron Radiation Facility beamline ID16A. A multi-position acquisition at 33 keV yielded a reconstructed volume of 128.6×128.6×100µm³ with an isotropic voxel size of 40 nm. After denoising using a deep-learning-based self-supervised approach, we trained a U-Net on a few annotated XNH images for the semantic segmentation of white matter compartments. We developed a bottom-up instance segmentation technique to trace myelinated axons, hierarchically merging oversegmented superpixels.  
Results: The automated xACSON pipeline segmented ∼13,800 myelinated axons, excluding axons smaller than 10µm, along with myelin and the extra-axonal space. On a skeleton-based 3D morphometry, we measured a mean diameter of 0.44±0.07µm for myelinated axons—that is excitingly consistent with our previous measurements of rats’ corpus callosum using high-resolution 3D electron microscopy. We quantified the volume fractions (φ) for the intra-axonal space φₐₓ=17.20%, myelin φₘ=31.52%, and extra-axonal space φₑₓ=51.28%.  
Conclusions: xACSON allows a scalable framework for axonal segmentation and morphometry in XNH, enabling characterization of white matter microstructure in ultra-large fields of view.  

55. Human forebrain assembloids to study microglia-neuron interactions during neurodevelopment 

Authors: Susanne Michels, Nataliia Novosolova, Mireia Gómez-Budia, Mohammadsaeed Nabati Saravani, Reyhaneh Aghayousefi, Mohammad Rezaie, Polina Abushik, Anssi Pelkonen, Anastasia Shakirzyanova, Minna-Mari Tervo, Nelli-Noora Välimäki, Sohvi Ohtonen, Akash Mali, Vera Lezhneva, Ali Abdollahzadeh, Luca Giudice, Paula Korhonen, Tarja Malm 

Author affiliations: A. I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Neulaniementie 2, 70211 Kuopio, Finland 

Rationale  
Microglia invade the brain during early embryogenesis, where they continuously survey their micro-environment, secrete signaling factors, are capable of phagocytosis, and regulate synapse stabilization and pruning. Due to the close microglia-neuron interaction during neuronal development and maturation, microglia may also play a supportive role in the migration and functional integration of interneurons in the developing cerebral cortex and could thus be involved in the regulation and control of the brain excitatory/inhibitory (E/I) balance.  
 
Methods  
We fused iPSC-derived region-specific organoids to model early human brain development and interneuron migration from the ventral to the dorsal forebrain in vitro. Microglia were incorporated as erythromyeloid progenitors and the immunocompetent assembloids were maintained as air-liquid interface slice cultures. At specific developmental stages, the assembloids were characterized by immunohistochemical stainings. Spatial transcriptomics were performed on a 10x Xenium device. The 4- to 6-month-old assembloids were assessed electrophysiologically via whole cell patch clamp and 3D microelectrode array recordings.  
 
Results  
We found that the presence of microglia promoted the frequency of spontaneous excitatory postsynaptic currents in the migrated interneurons. Furthermore, the interneurons functionally integrated into the dorsal assembloid part, as shown by firing rate changes in response to the GABA-A receptor agonist isoguvacine, the NMDA receptor agonist N-methyl-D-aspartate, and the GABA-A receptor antagonist picrotoxin.  
 
Conclusions  
Thus, microglia-containing human forebrain assembloids contribute to the continuous improvement of human brain model systems and could prove particularly valuable for studying microglia-/immune-related effects on synaptic plasticity, neuronal circuit formation, and E/I balance in the context of neurodevelopmental and psychiatric disorders. 

56. IFNγ Modulates α-Synuclein-Induced Microglial Responses and Dopaminergic Neuron Function in a Human Model 

Authors: Jonna Niskanen1), Vili Hakosalo1), Marko Lehtonen2), Mervi T. Hyvönen2), Sanni Peltonen1), Sohvi Ohtonen1), Yinjia Zhang3), Kelvin C. Luk4), Jari Koistinaho5),6), Wilhelmiina Hämäläinen3), Gundars Goldsteins1), Katrina Albert1), Šárka Lehtonen1) 

Author affiliations: 1) A.I.Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland;  2) School of Pharmacy, Faculty of Health Sciences, University of Eastern Finland;   3) Department of Computer Science, Aalto University, Helsinki, Finland;  4) Center for Neurodegenerative Disease Research, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Perelman School of Medicine, Philadelphia, USA;  5) Helsinki Institute of Life Science, University of Helsinki; 6) Drug Research Program, Division of Pharmacology and Pharmacotherapy, University of Helsinki 

Rationale:  
Microglia are the brain’s resident immune cells that regulate immune surveillance, inflammatory responses, and proteostasis. In Parkinson’s disease (PD), alpha-synuclein (αSyn) misfolds and aggregates, forming Lewy bodies in dopaminergic (DA) neurons. Over time, PD progression leads to degeneration and death of DA neurons and is frequently accompanied by sustained chronic neuroinflammation, in which microglia are key mediators. 

Methods:  
Human induced pluripotent stem cells were used to generate microglia and DA neurons. To model a PD-relevant inflammatory environment, microglia were exposed to interferon gamma (IFNγ) to induce an inflammatory state and αSyn preformed fibrils (PFFs) to mimic protein burden. Following stimulation, microglial conditioned media (CM), representing the secretome, were collected and applied to cultured DA neurons to assess downstream effects.  
Results:  
Exposure of microglia to αSyn PFFs induced aberrant autophagy and increased proinflammatory mediators, such as iNOS. IFNγ stimulation induced a strong activation phenotype characterized by reduced phagocytosis and mobility. Prolonged combined exposure to IFNγ and αSyn PFFs further induced expression of TGM2 and TGFβ1, alongside neuroprotective metabolites like MTA and pantothenic acid. The microglial secretome significantly modulated DA neuron architecture, mitophagy, and activity. Notably, neurotoxic effects of αSyn PFFs-CM exposure were partially reversed by CM from IFNγ/αSyn PFFs-exposed microglia.   
Conclusions:  
Exposure to IFNγ/αSyn PFFs reprograms human microglia, reducing αSyn uptake and shifting protein processing toward a more resolving secretome profile. This state is associated with beneficial support for DA neuron function compared to exposure to αSyn PFFs alone. These findings highlight context-dependent microglial plasticity in PD-relevant conditions.  

57. Association between accumulation of chronic diseases and cognition among older adults at risk of dementia 

Authors: Konsta Valkonen1,2 , Anette Hall2, Anna Rosenberg1, Johanna Pöyhönen1,5,Riitta Antikainen7,8, Tiina Laatikainen4, Hilkka Soininen2, Timo Strandberg5,6,7, Jaakko Tuomilehto9,10, Amaia Calderon Larrañaga14, Miia Kivipelto4,3,11,12,13, Alina Solomon2,11, Tiia Ngandu1,4 

Author affiliations: 1) Department of Public Health, Lifestyles and Living Environments,  Finnish Institute for Health and Welfare, Helsinki, Finland; 2) Institute of Clinical Medicine/Neurology, University of Eastern Finland, Kuopio, Finland; 3) Division of Clinical Geriatrics, Center for Alzheimer Research, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Sweden; 4) Institute of Public Health and Clinical Nutrition, University of Eastern Finland, Kuopio, Finland; 5) Department of Medicine, Clinicum, University of Helsinki, Helsinki, Finland; 6) Department of Geriatrics, Helsinki University Hospital, Helsinki, Finland; 7) Center for Life Course Health Research/Geriatrics, University of Oulu, Oulu, Finland; 8) Medical Research Center Oulu, Oulu University Hospital, Oulu, Finland; 9) South Ostrobothnia Central Hospital, Seinäjoki, Finland; 10) Department of Public Health, University of Helsinki, Helsinki, Finland;  11) The Ageing Epidemiology Research Unit, School of Public Health, Imperial College London, London, United Kingdom; 12) FINGERS Brain Health Institute, Sweden; 13) Theme Inflammation and Aging, Medical Unit Aging, Karolinska University Hospital, Sweden; 14) Aging Research Center, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Sweden 

Rationale: Multimorbidity, defined as co-existence of two or more chronic diseases, is associated with adverse outcomes, including dementia. Less is known about the association of chronic disease accumulation and cognition in older individuals without substantial cognitive impairment. In this study, we investigated chronic disease accumulation among older adults and associations with cognition in The Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER).   
Methods: FINGER included 1260 participants (aged 60–77, at risk of dementia), randomized 1:1 to multimodal lifestyle intervention or control. Cognition was assessed using a Neuropsychological Test Battery. Information on diagnoses and drug purchases from 2000 until FINGER baseline (2009-2012) were collected from Finnish Prescription and Care Registers, to assess the prevalence of chronic diseases (conditions included in a previously defined list of 60 disease categories). Linear regression models were used in the analyses.  
Results: Approximately half of all participants (n=635, 50%) had two or more chronic diseases at baseline. The most common diseases were osteoarthritis and other degenerative joint diseases, diabetes, and ischemic heart disease. The number of chronic diseases did not differ significantly by randomization group (p>0.05) or sex (p>0.05). Higher chronic disease accumulation was associated with worse global cognition, processing speed, and memory (p<0.05), but not executive functioning.   
Conclusions: Multimorbidity was common in this study population of older at-risk adults, and higher chronic disease accumulation was associated with worse cognition. The next analyses will clarify associations between the FINGER intervention and the rate of chronic disease accumulation over a follow-up period.  

58. Organic Anion Transporting Polypeptide 2B1 (OATP2B1)-utilizing prodrugs in improving intra-brain drug delivery of neuroprotective agents 

Authors: Janne Tampio 1, Seyed Hamed Maljaei 1, Unna Sinivuori 1, Riina Nieminen 1, Adela Králová 1, Arun Kumar Tonduru 1, Petri Turhanen 1, Aaro Jalkanen 1, Kristiina M. Huttunen 1 

Author affiliations: 1) School of Pharmacy, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale  
Transmembrane transporters have a major regulatory role for pharmacokinetics by their expression profiles. In central nervous system, insufficient drug delivery across the blood-brain barrier (BBB) often limits therapeutic efficacy despite adequate systemic pharmacokinetics. While transporters at BBB have been studied, individual brain cell types and regions have been less explored. In the present study, we hypothesized that utilization of organic anion transporting polypeptide 2B1 (OATP2B1) could enhance targeted brain drug delivery.  
 
Methods  
We quantified protein expression levels of several OATP transporters with targeted liquid chromatography-mass spectrometry from human and mouse microglial and astrocyte cell lines. Simultaneously, we evaluated a series of OATP2B1-utilizing prodrugs of ketoprofen for their ability to improve cellular uptake. Subsequently, distribution of the most promising prodrug in mouse brain was assessed.  
 
Results  
We determined that human astrocytes and microglia have differences in their OATP expression profiles, allowing cell-type-preference in drug delivery. The OATP2B1-utilizing prodrugs were able to improve the drug uptake in microglial cells, where OATP2B1 was more abundant. On the in vivo level, the intrabrain prodrug distribution differed from the ketoprofen distribution, showing that utilization of OATPs could alter the delivery of ketoprofen in the brain.  
 
Conclusions  
Since OATP subtypes have different expression profiles in brain cells and areas, transporter utilization in delivery holds a strong potential for improved targeted drug delivery. Beyond the ketoprofen-derived prodrugs assessed in the present study, this approach opens new possibilities for the development of neurodegenerative therapeutics with site-selective intra-brain delivery to increase efficacy and safety. 

59. Music as an Alternative Probe of Emotion Recognition in Cognitive Disorders 

Authors: Laura Säisänen 1,2, Kasper Katisko 1, Miika Korhonen 1,2, Walter Werzowa 3, Eino Solje 1,2 

Author affiliations: 1 Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland 2 Neuro Center, Neurology, Kuopio University Hospital, Kuopio, Finland 3 Institut für Kompositionsstudien, Ton- and Musikproduktion, University of Music and Performing Arts, Vienna, Austria 

Rationale: Emotion recognition deficits are common in neurodegenerative disorders and are traditionally assessed using facial expression recognition tasks. Recent work suggests that it can also be evaluated using alternative modalities. We examined whether recognition of emotional states in music differentiates individuals with cognitive impairment from healthy controls (HC), distinguishes between diagnostic groups, and relates to cortical neurotransmitter functioning assessed by transcranial magnetic stimulation (TMS).   
Methods: Within 149 participants 45 were HC, 31 were diagnosed with Alzheimer’s disease (AD), 13 with frontotemporal dementia (FTD), 19 with synucleinopathy (SYNU), 11 with mild cognitive impairment (MCI), 23 with subjective cognitive impairment (SCI), and 7 with other or undefined cognitive impairment. The task included free description of emotions elicited by three original musical works and a forced-choice task pairing six excerpts with basic emotions: anger, fear, disgust, happiness, sadness and surprise. TMS measures included motor threshold and paired pulse paradigms assessing cortical inhibition and facilitation.   
Results: HC performed significantly better than participants with AD, FTD, SYNU, and MCI, but not SCI. Notably, the free description task differentiated AD from FTD. Anger, surprise, and disgust were the most difficult emotions to recognize (22-26% in patients, 42-62% in HC). No correlations were observed between music emotion recognition performance and TMS measures.   
Conclusions: Emotion recognition in music distinguishes cognitively impaired individuals from HC, partly paralleling findings from facial emotion recognition. The free description component shows promise as a diagnostic biomarker, particularly in differentiating AD and FTD. While no neurophysiological correlates were identified, the findings support further exploration 

60. Circadian and Seasonal Modulation of Intracortical Inhibition and Facilitation at a High Northern Latitude 

Authors: K.Mäntyvaara, L.Säisänen, J.Hyppönen, S.Määttä, E.Solje 

Author affiliations: Department of Clinical Neurophysiology, Kuopio University Hospital, Kuopio, Finland.  Institute of Clinical Medicine, Neurology, University of Eastern Finland, Kuopio, Finland.  Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 

Rationale  
Transcranial magnetic stimulation (TMS) assesses cortical excitation and inhibition and has been proposed as a biomarker in neurodegenerative disorders. However, substantial variability limits its clinical applicability. Biological rhythms, including circadian timing and seasonal photoperiod, as well as age-related changes, may contribute to this variability.  
Methods  
We examined circadian, seasonal, and age-related influences on intracortical excitability in 42 healthy adults (mean age 59.5 ± 14.5) in Finland. Short-interval intracortical inhibition (SICI), long-interval intracortical inhibition (LICI), and intracortical facilitation (ICF) were analysed using multivariable linear regression including, age, time of day, daylight duration, and rate of change in daylight.  
Results  
Earlier time of day was associated with stronger short-interval intracortical inhibition at 2–3 ms interstimulus intervals (SICI2 p=0.011 and SICI3 p=0.005). Seasonal photoperiod dynamics (p=0.068) and longer daylight duration (p=0.061) showed trend-level associations with SICI2 and LICI150, respectively, indicating stronger SICI2 inhibition during increasing daylight and weaker LICI150 inhibition with longer daylight duration. SICI2 inhibition weakened with increasing age (p=0.011), whereas LICI50 inhibition strengthened (p=0.001). Higher motor threshold was associated with greater ICF7 and ICF10 facilitation (p=0.037 and p=0.033) and weaker SICI1 inhibition (p=0.021). No consistent circadian or seasonal effects were observed for other parameters. Statistical significance was set at p < 0.05.  
Conclusion  
Intracortical inhibition appears to be modulated by circadian timing and age. Short-interval inhibition may be particularly sensitive to seasonal changes, while age shows varying effects on different inhibitory circuits. Accounting for age, circadian, seasonal, and methodological factors may improve the reliability of TMS metrics.  

61. Underlying Factors, Clinical Characteristics and Diagnostic Findings of Subjective Cognitive Impairment: A Multimodal Memory Clinic Study 

Authors: P. Saarinen¹, E. Vitikka², K. Katisko¹, J. Krüger^{2, 4, 5}, E. Solje^{1, 3} 

Affiliations: 
1= Institute of Clinical Medicine, Department of Neurology, University of Eastern Finland, Kuopio, Finland 

2= Research Unit of Clinical Medicine, Neurology, University of Oulu, Oulu, Finland 

3= Neuro Center-Neurology, Kuopio University Hospital, Kuopio, Finland 

4= Medical Research Center Oulu, Oulu University Hospital, Oulu, Finland 

5= Neurocenter, Neurology, Oulu University Hospital, Oulu, Finland 

Rationale: Subjective cognitive impairment (SCI) refers to self-experienced cognitive decline without objective deficits on testing. Although SCI may represent an early stage of a neurodegenerative disease in some individuals, the majority do not progress to dementia. 

Progression risk likely reflects a complex interplay of biomarkers, neuroimaging findings and clinical, psychological, and social factors. However, current evidence is inconsistent and mostly based on isolated variables, highlighting the need for multimodal approaches.  

Methods: The study cohort includes 135 participants: 33 eventually classified as SCI on clinical evaluation, 43 with Alzheimer’s disease, and 59 healthy controls without subjective or objective cognitive symptoms. The data includes clinical, neuropsychological, psychosocial, and lifestyle variables, as well as biomarkers (e.g., CSF-biomarkers beta-amyloid42, fostau and tau and plasma biomarkers pTau217 and NfLc) and neuroimaging when applicable. Analyses will include descriptive statistics, group comparisons, and multivariate models to identify factors associated with SCI and explore potential SCI subgroups. 

Results: Data collection has been completed and preprocessing is ongoing. Planned analyses will focus on identifying clinical, biological, and psychosocial patterns associated with SCI, as well as distinguishing features compared to Alzheimer’s disease patients and healthy controls. 

Conclusions: This study will provide a comprehensive characterization of SCI as a heterogeneous clinical condition. A multimodal approach may enable earlier risk identification and support the development of targeted and individualized interventions. 

62. Pericytes as Converging Targets of Genetic Risk and Environmental Exposure in Parkinson’s Disease 

Authors: Sanni Peltonen [1], Tuuli-Maria Sonninen [1], Demet Sinem Guden Yildirim [1], Kelvin C. Luk [2], Jari Koistinaho [3,4], Marika Ruponen [5], Šárka Lehtonen [1]  

Author affiliations: [1] A.I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland  [2] Center for Neurodegenerative Disease Research, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Perelman School of Medicine, Philadelphia, PA, USA   [3] Drug Research Program, Division of Pharmacology and Pharmacotherapy, University of Helsinki, Helsinki, Finland  [4] Helsinki Institute of Life Science, University of Helsinki, Helsinki, Finland  [5] Finnish Medicines Agency, Kuopio, Finland 

Rationale  
Parkinson’s disease (PD) is the fastest-growing neurological disorder in the world. As in other neurodegenerative diseases, blood-brain barrier (BBB) dysfunction and chronic neuroinflammation are well-recognized features in PD. Pericytes, mural cells that surround microvessels, play a key role in regulating BBB integrity and neuroinflammatory responses in the central nervous system; however, their contribution to neurodegenerative diseases remains insufficiently understood. Environmental exposures, including nanoplastics, have recently been implicated in exacerbating PD-related pathology, such as alpha-synuclein (aSyn) aggregation and BBB disruption. Due to their location, pericytes are among the first brain cell types exposed to nanoplastics that cross the BBB.  
 
Methods  
We investigated the impact of the PD-causing LRRK2 G2019S mutation in induced pluripotent stem cell-derived pericyte-like cells. In parallel, we examined how co-exposure to aSyn and polystyrene nanoplastics affects pericyte-like cell function.  
 
Results  
The LRRK2 G2019S mutation impaired the migratory capacity of pericyte-like cells and altered the secretion of inflammatory mediators. Preliminary exposure studies showed that aSyn alone modestly reduced cell motility. Notably, anionic nanoplastics further exacerbated aSyn-induced impairment, although nanoplastics alone increased cell motility.  
 
Conclusions  
These findings support a role for pericytes in mediating inflammation and vascular alterations in PD. Moreover, nanoplastics can modulate pericyte function and amplify the detrimental effect of aSyn. Overall, this work highlights pericytes as a critical interface between genetic susceptibility, and environmental exposure in PD pathogenesis.  

63. Healthcare Use as an Early Indicator of Parkinson’s Disease 

Authors: Aino-Lotta I. Alahäivälä [1], Saara Rissanen [2,3], Paulus Carpelan [2], Jaana M. Hartikainen [4,5], Anna Sormunen [4], Valtteri Julkunen [6,7], Jouni Ihalainen [1] 

Author affiliations: [1] VTT Technical Research Centre of Finland, Microkatu 1, 70211, Kuopio, Finland; [2] Adamant Health Oy, Microkatu 1, 70211, Kuopio, Finland; [3] University of Eastern Finland, P.O. Box 1627, 70211, Kuopio, Finland; [4] Institute of Clinical Medicine, Clinical Pathology and Forensic Medicine, University of Eastern Finland, 70211, Kuopio, Finland; [5] Institute of Clinical Medicine, Genome Center of Eastern Finland, University of Eastern Finland, 70211, Kuopio, Finland; [6] Department of Neurology, Neurocenter, Kuopio University Hospital, 70210, Kuopio, Finland; [7] Institute of Clinical Medicine, Department of Neurology, University of Eastern Finland, 70211, Kuopio, Finland 

Rationale  
Parkinson’s disease is typically diagnosed after the onset of motor symptoms, although subtle changes often emerge years earlier. Less is known about how these changes manifest in healthcare utilization at the population level. Routinely collected primary healthcare data may capture early, non-specific signals preceding diagnosis. This study examines whether longitudinal service-use patterns can identify early indicators of Parkinson’s disease.  
Methods  
This study uses longitudinal primary healthcare data from adults aged ≥65 in the Kuopio region, Finland. Individuals with Parkinson’s disease were compared with age- and sex-matched controls. A four-year pre-diagnostic period was examined. The data include detailed information on healthcare contacts (e.g., contact types and resource groups) as well as clinical content linked to these contacts, such as diagnoses, medications, and laboratory tests.  
Results  
Healthcare service use increased over time in both groups, with group differences emerging about 1.5 years before Parkinson’s disease diagnosis. The most pronounced increase was seen in home care services, in both contacts and resource use. In addition, the linked clinical information provides a basis for exploring risk modelling and early identification of Parkinson’s disease.  
Conclusions  
Longitudinal primary healthcare data reveal early changes in service use before Parkinson’s disease diagnosis. These findings suggest that routinely collected healthcare data may support earlier identification of the disease and future data-driven risk models, contributing to modern care of neurodegenerative diseases. This study is part of the TÄRY project funded by the European Regional Development Fund (ERDF).  

64. C9ORF72 REPEAT EXPANSION-CARRYING FTD PATIENT-DERIVED IPSC-MICROGLIA SHOW INCREASED PHAGOCYTIC ACTIVITY AND FEWER LYSOSOMAL VESICLES 

Authors: Tomi Hietanen 1; Hannah Rostalski 1,2; Sami Heikkinen 3; Nadine Huber 1; Ashutosh Dhingra 4; Salvador Rodriguez-Nieto 4; Viivi Pekkala 1; Henna Martiskainen 3; Mari Takalo 3; Mikko Hiltunen 3; Eino Solje 5,6; Dorit Hoffmann 1; Annakaisa Haapasalo 1 

Author affiliations: 1 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland;  (T.H.); (H.R.);  (D.H.); (N.H.); (V.P.); (A.H.) 2 Biotech Research & Innovation Center (BRIC), University of Copenhagen, Copenhagen, Denmark; (H.R.) 3 Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland; (S.H.); (H.M.); (M.T.); (M.H.) 4 German Center for Neurodegenerative Diseases (DZNE), Tübingen, Germany; (A.D.);  (S.R.) 5 Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland; (E.S.) 6 Department of Neurology, Neuro Center, Neurology, Kuopio University Hospital, Kuopio, Finland; (E.S.) 

Rationale. The C9orf72 hexanucleotide repeat expansion (HRE) is the most common genetic cause of frontotemporal dementia (FTD). C9orf72 has been shown to be involved in the autophagic pathway and is highly expressed in microglia, but little is known about how the HRE affects human microglial function. We aimed to study the effects of the C9orf72 HRE in FTD patient-derived microglia, with specific attention to phagocytosis and the autophagic pathway.  
Methods. We produced human induced pluripotent stem cell-derived microglia from healthy controls, sporadic FTD patients, and FTD patients carrying the C9orf72 HRE (C9FTD). The microglia were characterized for the common FTD-related hallmarks and by global RNA sequencing. Immunofluorescence imaging and Western blotting were used to quantify the levels of p62/SQSTM1, Lamp2a, TDP-43 and LC3B to examine the autophagic pathway. Phagocytic activity was assessed by a bead-based phagocytosis assay.  
Results. RNA foci and poly-GP dipeptide repeat proteins, but not C9orf72 haploinsufficiency, were present in the C9FTD microglia. RNA sequencing revealed alterations in pathways related to phagocytosis, inflammatory response, and energy metabolism in FTD microglia. The sporadic FTD microglia showed the largest gene expression changes. LC3B-I to II conversion occurred similarly in all microglia. Decreased number of Lamp2a-positive vesicles and enhanced phagocytic activity were observed in the C9FTD microglia.  
Conclusions. Some of the typical C9orf72 HRE-related hallmarks were found present in C9FTD microglia. The decreased number of lysosomal vesicles together with increased phagocytic activity might indicate enhanced function of the phagocytic pathway specifically in the C9FTD microglia. 

65. Microglia increase electrophysiological activity and maturation of excitatory neurons in a long-term cortical organoid model upregulating pathways involved in synaptic transmission and axon extension and creating a niche for cell-cell contact and synaptic remodelling 

Authors: Korhonen Paula1, Novosolova Nataliia1, Gomez Budia Mireia1, Ghasemi Fateme1, Giniatullina Raisa1, Shakirzyanova Anastasia1, Ohtonen Sohvi1, Tervo Minna-Mari1, Rezaie Mohammad1, Saveleva Liudmila1, Gribchenko Emma1, Nikupeteri Jutta1, Michels Susanne1, Mali Akash1, Della Pietra Adriana1, Lezhneva Vera1, Pelkonen Anssi1, Lehtonen Šarka1, Abushik Polina1, Dougalis Antonios1, Jäntti Henna1, Giudice Luca1, Malm Tarja1 

Author affiliations: 1 A. I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Neulaniementie 2, 70211 Kuopio, Finland  

Rationale  
Microglia are sculptors of neuronal circuits, yet their specific contribution to human neurodevelopment remain poorly defined. Recent advances in human induced pluripotent stem cell (iPSC) -derived cortical organoids (COs) and microglia-like cells enable direct investigation of microglia-neuron interactions in experimentally tractable human tissue context. Prior studies show that microglia promote neuronal maturation, however, the long-term mechanisms underlying these effects are not well understood.   
Methods  
We generated microglia-containing COs (MG-CO) and performed longitudinal electrophysiological characterization over 7-months culture period. Neuronal differentiation was directed using small molecules, COs were cut into 500 µm thick slices and grown in air-liquid interphase on inserts. Microglial progenitors were incorporated onto slices and neuronal activity was assessed using patch clamp recordings and multi-electrode arrays. Immunostainings and Xenium transcriptomics were used to spatially map microglia-neuron interactions and associated transcriptional programs.   
Results  
MG-COs developed both excitatory and inhibitory neurons and astrocytes with microglia persisting throughout the 7 months. Microglia increased neuronal baseline firing rates, spontaneous excitatory postsynaptic currents, neuron size and dendritic spine density. Physical contact with microglia increased presynaptic size. Spatial transcriptomics revealed that excitatory neurons in MG-COs upregulated genes involved in excitatory synaptic transmission and axon extension, and downregulated apoptotic pathways. Microglia-proximal niches showed signatures of cell-cell contact and synaptic remodelling and pruning.  
Conclusion  
Microglia enhance excitatory drive and structural maturation of human neurons, shaping circuit formation at functional and transcriptional levels. MG-CO platform enables mechanistic investigation of microglia-dependent circuit development and provides a human system for studying genetic and environmental modulators of microglia-neuron interaction.  

66. Tau pathology -driven alterations in the brain microvessel membrane proteome 

Authors: Jooseppi Hyvärinen, Sabrina Petralla, Mikko Gynther, Alexey Afonin, Marko Lehtonen, Seppo Auriola, Laura Mussalo, Gert Fricker, Katja M. Kanninen, Elena Puris 

Author affilations: A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland. Institute of Pharmacy and Molecular Biotechnology, Ruprecht-Karls-University, Heidelberg, Germany. School of Pharmacy, University of Eastern Finland, Kuopio, Finland 

The blood–brain barrier (BBB), formed by brain microvascular endothelial cells, maintains brain homeostasis by regulating molecular transport and protecting against blood-borne toxins and pathogens. In Alzheimer’s disease and other tauopathies, abnormal aggregation of tau protein is observed in neurons, and BBB integrity is often compromised. Recent in vitro reports link BBB disturbance to tau species propagating to the brain vessel endothelial cells. However, tau-associated BBB disturbances remain poorly understood.  
We examined tau-linked BBB alterations by analysing the whole membrane proteome of isolated brain microvessels from male and female JNPL3 transgenic mice, compared with wild-type (WT) controls. JNPL3-model recapitulates the features of human tauopathy, including BBB leakage. Membrane proteome isolation concentrates proteins essential to BBB function, such as transporters.   
Data-independent acquisition proteomics analysis revealed altered expression of 379 proteins in JNPL3 male mice and 219 in JNPL3 female mice compared to the WT controls. Between male and female groups, major differences were observed; only 51 proteins were similarly altered between the sexes. The changes were associated with several BBB-relevant pathways, including clathrin-mediated transcytosis and solute carrier (SLC)-mediated transport in females. Male model showed changes in pathways linked to cell cycle, immune system and mitochondrial function when compared to the sex-matched WT mice. Additionally, we observed differences between sexes in microvessels of WT animals.  
The analysis highlights novel tau-associated mechanisms of BBB disturbance and identifies significant sex-specific differences. The findings reveal potential therapeutic targets and help facilitate drug delivery to the central nervous system in sex-specific way.  

67. From OR to wet lab – sample handling protocol for multimodal and multicenter assessments 

Authors: Jenni Kyyriäinen1, Adriana Della Pietra1, Mastaneh Torkamani-Azar1, Mireia Gomez Budia1, Polina Abushik1, Nataliia Novosolova1, Henri Eronen1,2, Omar Narvaez1, Ekaterina Paasonen1,3, Vera Lezhneva1, Anssi Pelkonen1, Liudmila Saveleva1, Antti Huotarinen2,4, Ilya Belevich5, Eija Jokitalo5, Kuopio Epilepsy Center Surgery Group6, Leena Jutila6, Tuomas Rauramaa2,7, Arto Immonen4, Ville Leinonen2,4, Jussi Tohka1, Olli Gröhn1, Reetta Kälviäinen2,6, Tarja Malm1, Alejandra Sierra1 

Author affiliations: 1 A. I. V. Institute, University of Eastern Finland (UEF), Kuopio, Finland  2 Institute of Clinical Medicine, School of Medicine, UEF, Kuopio, Finland  3 Neurocenter, Kuopio University Hospital (KUH), Kuopio, Finland   4 NeuroCenter Neurosurgery, KUH, Kuopio, Finland   5 Electron Microscopy Unit, Institute of Biotechnology, Helsinki Institute of Life Science, University of Helsinki, Helsinki, Finland 6 Kuopio Epilepsy Center, KUH, Member of ERN EpiCARE, Kuopio, Finland. 7 Department of Pathology, KUH, UEF, Kuopio Finland 

Rationale. We have developed sample handling protocol for multiscale characterization of the resected epileptic tissue at structural, functional, and molecular level.   
Methods. Patients with drug-resistant epilepsy are recruited from Epilepsy Center at Kuopio University Hospital. Day before the resection surgery patients undergo magnetic resonance imaging (MRI) where additional research sequence protocol is applied. The entire resected tissue is placed under oxygenation into artificial cerebral spinal fluid solution. The tissue is immediately transported to A. I. Virtanen Institute, 800 m from the hospital, where it is sliced by a neuropathologist to extract tissue slabs for standard diagnostics and subsequent research. Two adjacent tissue slabs are provided for research and studied using microelectrode arrays, patch‐clamp recordings, organotypic cultures, single‐nuclei sequencing, spatial transcriptomics, proteomics, electron microscopy, ex vivo MRI, and histology.   
Results. The developed multimodal and multicenter protocol offers an opportunity to test emerging imaging techniques in combination with cutting-edge technologies from the same samples. Final goal is to build a 3D anatomical model of the resected tissue to allow predictive modeling of the underlying pathology from MRI images.   
Conclusions. Our approach provides a novel protocol to study microstructure, cellular heterogenicity, and electrophysiological properties of epileptogenic tissue for more precise characterization of epileptogenic region. 

68. Benchmarking Registration Methods for Cross-Modal Rat Brain Images with Mild TBI 

Authors: Zewen Zhuo ¹, Claude Lepage ², Thomas Funck ³ ⁴, Paule-Joanne Toussaint ², Alejandra Sierra Lopez ¹, Alan Evans ², Jussi Tohka ¹ 

Author affiliations: 1. A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland  2. Montreal Neurological Institute, McGill University, Montreal, Canada 3. Child Mind Institute, New York, USA   4. Institute of Neuroscience and Medicine (INM-1), Research Centre Ju¨lich, Ju¨lich, Germany 

Magnetic Resonance Imaging (MRI) is a widely used brain imaging modality thanks to its non-invasive nature. It provides mesoscale-to-macroscale anatomical and functional information, offering insight into tissue organisation, connectivity, and regional functional patterns relevant to disease and biology. However, MRI alone is insufficient for capturing cellular-level structures or revealing microscopic mechanisms underlying pathological brain processes. Histology complements this limitation by providing high-resolution structural information at the microscale.  
 
Different from MRI, histological data are typically acquired as serial 2D sections which are reconstructed into a coherent 3D representation to enable volumetric analysis and cross-scale interpretation. This reconstruction process is challenging due to section-to-section variability, tissue tearing, shrinkage, staining inconsistencies, missing slices, and distortions introduced during preparation. Such artefacts disrupt structural continuity across adjacent sections, making accurate 3D reconstruction difficult.  
 
Image registration plays a central role in establishing spatial correspondence across serial histology sections and between reconstructed histology volumes and MRI. Compared with MRI, histology images are substantially larger, acquired at higher resolutions, and vary across staining protocols, contrasts, and microscopy techniques, complicating multimodal alignment.  
 
In this research, we benchmarked multiple image registration strategies for 3D histology reconstruction of histological data and MRI-guided alignment, including optimisation-based methods such as ANTs, iterative approaches inspired by the BigBrain project, deep-learning techniques, and segmentation-assisted frameworks. Our goal is to identify a robust in-house pipeline for whole-rat-brain 3D histology reconstruction and multimodal registration. Initial results indicate that semantic segmentation masks improve tissue continuity during alignment, while iterative methods achieve strong registration performance. 

69. Stroke in working age and job accommodation in facilitating transition back to work (SWIFT) – protocol description and preliminary baseline clinical, cognitive questionnaire and occupational data analyses 

Authors: Teemu I. Paajanen¹, Karoliina Kiviranta²,³, Pyry Silomaa⁴, Laura Mannismäki⁵, Henrietta Törmänen⁵, Nicolas Martinez-Majander⁵, Karoliina Aarnio⁵, Aleksi Sihvonen⁵, Mohamed Elmegiri⁶, Juha Martola⁶, Christer Hublin¹, Nina Nevala⁷, Silja Palokangas², Hanna Jokinen²,³, Jukka Putaala⁵ 

Author affiliations: ¹Occupational Health Unit, Occupational Medicine, Finnish Institute of Occupational Health, Helsinki, Finland; ²Division of Neuropsychology, HUS Neurocenter, Helsinki University Hospital, Helsinki, Finland; ³Department of Psychology, Faculty of Medicine, University of Helsinki, Helsinki, Finland; 4ICT and Digital Services Unit, Finnish Institute of Occupational Health, Helsinki, Finland; 5Department of Neurology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland; 6Department of Radiology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland; 7Work ability and working careers Unit, Finnish Institute of Occupational Health, Helsinki, Finland 

Rationale: The incidence of young-onset stroke increases, with approximately 3,000 working-age Finns affected annually. Previous studies indicate that 50-90% stroke patients suffer from cognitive impairment in the subacute phase, with nearly half having longer-lasting cognitive or emotional symptoms affecting return and performance at work. Cognitive job accommodations may support recovery. We designed a study focusing on cognitive and occupational factors of working-age stroke patients.  
Methods: The SWIFT study enrolls 18-68-year-old patients treated at Helsinki University Hospital due to ischemic stroke. The study comprises four work packages (WP): medical records, clinical and web-based assessments (WP1); a randomized controlled trial (WP2); case studies (WP3); and national registries (WP4). We analyzed clinical and cognitive-emotional questionnaire data from the first WP1 patients recruited between 11/2024 and 2/2026.  
Results: Mean age for analyzed patients (n=200) was 53.3±10.7 years (range 24-68 years), 67% were male, and 76% were working prior to stroke. In the acute phase, the NIHSS score was 0–1 in 50% and 0–4 in 81% of patients. After 2 months post-stroke, subjective work ability score was 7.1±2.3 of ten, when 10% had returned to work. Of all participants, 51% reported more than two at least mild subjective cognitive and 83% more than two emotional/fatigue symptoms.  
Conclusions: Preliminary results indicate that working-age ischemic stroke patients may have milder acute neurological symptoms than previously described. However, mild cognitive, emotional, and fatigue complaints remain very common after two months post-stroke. Further follow-up and neuropsychological assessments are necessary to understand their impact on cognitive work ability. 

70. FUNCTIONAL CHARACTERIZATION OF A NEURODEGENERATION-ASSOCIATED GRN RISK VARIANT IN IPSC-DERIVED MICROGLIA 

Authors: Roosa-Maria Willman1, Elena Ppali2, Heli Jeskanen1, Sami Heikkinen1, Sini-Pauliina Juopperi1, Susanna Kemppainen1, Annakaisa Haapasalo2, Mari Takalo1, Ville Leinonen3,4, Tarja Malm2, Henna Martiskainen1, Mikko Hiltunen1 

Author affiliations: 1 Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland 2 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 3 Unit of Neurosurgery, Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland 4 Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland 

Rationale: Progranulin is a lysosomal protein expressed in microglia and neurons in the brain. A common variant, rs5848 (C>T), in the progranulin gene (GRN) has been associated with an increased risk for neurodegenerative diseases, including limbic-predominant age-related TDP-43 encephalopathy (LATE) and Alzheimer’s disease. However, the cellular mechanisms underlying this risk-modifying effect remain poorly understood. Here, the impact of the GRN rs5848 variant was studied in human induced pluripotent stem cell (iPSC) -derived microglia (iMG).  
Methods: Human iPSCs from individuals homozygous for the risk allele (TT) and non-carriers (CC) were differentiated into iMGs. The iMG cells were treated with lipopolysaccharide (LPS) or myelin and subsequently analyzed using omics-based approaches, functional assays and fluorescence staining techniques.  
Results: Bulk mRNA sequencing revealed that iMG cells carrying the risk allele (TT) showed an upregulation of genes involved in lipid metabolism. Consistently, lipid droplet accumulation was elevated in TT-carrying cells compared to control cells without any treatment and following LPS stimulation. Interestingly, the phagocytic uptake of pHrodo-labeled myelin and zymosan did not differ between the genotypes. Further functional assays are underway to assess the potential alterations in lysosomal function.   
Conclusions: Studying these cellular effects, along with changes in progranulin levels or processing, may help elucidate the mechanisms by which GRN influences neuronal health and disease susceptibility.  

71. Processing with presbycusis: Examining the interplay between auditory, linguistic, and other cognitive processes in older Finnish adults 

Authors: Melissa Gunning[1,2], Alexandre Nikolaev[3], Anette Hall[2], Laura Ihalainen[4,5], Pia Linder[4,5], Aarno Dietz[4,5], Alina Solomon[2] 

Author affiliations: 1: European Master’s in Clinical Linguistics; 2: Department of Neurology, Institute of Clinical Medicine, University of Eastern Finland; 3: School of Humanities, University of Eastern Finland; 4: Institute of Clinical Medicine, University of Eastern Finland; 5: Department of Otorhinolaryngology, Kuopio University Hospital 

Rationale: Age-related hearing loss (ARHL) affects millions of older adults and is often accompanied by declines in cognitive performance. However, it is not well understood how auditory, language-related, and cognitive functions are connected or how they are relevant during speech recognition.  
Methods: Neuropsychological assessments, speech-in-noise tasks, and self-reported questionnaires from 149 older Finnish adults with untreated ARHL in the Healthy heAring for Healthy Ageing (HAHA) trial were analyzed. Bootstrapped network analysis and community detection were applied to characterize functional relationships. Two random forest (RF) models ranked variable importance for the speech recognition thresholds (SRTs) from the Finnish Matrix Sentence Test (FMST) and the Digits-in-Noise (DIN).  
Results: Results showed that ranked variable importance varied by speech-in-noise task at the individual and community levels, with the variables better predicting the FMST than the DIN. The network analysis yielded low density, meaning fewer connections between variables. The communities detected were composed of mixed variables, indicating limited functional segregation.  
Conclusions: Different auditory, language-related, and cognitive processes dynamically reorganize in their importance as speech-in-noise task demand changes. Older Finnish adults with ARHL exhibit less clearly segregated communities of functions, which may suggest compensatory strategies at work. More complex speech-in-noise tasks may be preferred clinically because of the provided wider view of auditory, language-related, and cognitive performance in aging populations. 

72. Everyday legal problems of people with epilepsy in Finland and Sweden 

Authors: Kujansuu, E., Teerikangas, M., Issakainen, M., Lönnroth, M., Mäki-Petäjä-Leinonen, A., Kälviäinen, R., & Ervasti, K. 

Author affiliations: Center of Law and Welfare, Law School, University of Eastern Finland; Institute of Clinical Medicine, School of Medicine, University of Eastern Finland 

Rationale. The study examines the everyday legal problems experienced by people with epilepsy in Finland and Sweden. Additionally, we investigate which sociodemographic, socioeconomic, and health-related factors explain variation in perceived barriers across different domains of daily life, including working life, mobility, health, partnership, financial stability, and independence. Methods. The data consists of cross-sectional survey data (N=354) of people with epilepsy from both countries. Descriptive statistics were performed to summarize the key results concerning legal problems and linear regression analyses were used to examine associations between epilepsy-related barriers across domains of daily life and socioeconomic position and health status. Results. The greatest number of challenges in everyday life in both countries were reported in relation to work. The study revealed only a few differences between the countries, with problems related to goods and services being more common in Finland. Being unemployed or otherwise outside the labour market and experiencing seizures were significant predictors of perceived restrictions across all models. Conclusions. Experiencing seizures and being unemployed or otherwise outside the labour market emerged as key factors associated with epilepsy being perceived as a barrier in both Finland and Sweden. 

73. Prevalence of obstructive sleep apnea in patients with intracranial aneurysms 

Authors: Riina Pirinen ^1,4, Salla Ylä-Herttuala ¹, Antti Lindgren ^2,4, Nils Danner ^3,4, Timo Koivisto ^3,4, Juha E. Jääskeläinen ^3,4, Anu Muraja ¹, Esa Mervaala ^1,4 

Affiliations: 
¹Department of Clinical Neurophysiology, Kuopio University Hospital, Kuopio, Finland 

² Department of Radiology, Kuopio University Hospital, Kuopio, Finland 

³ Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland  

⁴ Institute of Clinical Medicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland                                                                                                                                                  

Purpose: Obstructive sleep apnea (OSA) is associated with various cardiovascular and cerebrovascular conditions; however, its prevalence among patients with intracranial aneurysms (IAs) remains insufficiently characterized. Emerging evidence suggests that OSA may contribute to IA formation, progression, and outcomes after rupture. 

Methods: This case–control study included 244 participants: 84 IA patients without prior clinical suspicion of OSA and 160 OSA-symptomatic controls. Overnight respiratory polygraphy (RP) was used to assess apnea–hypopnea index (AHI) and presence of OSA. Group differences were analyzed using models adjusted for age, BMI, and sex. 

Results: Among IA patients, 50.0% had findings consistent with OSA on RP (AHI ≥ 5 events per hour), of whom 27.4% were classified as mild, 15.5% as moderate, and 7.1% as severe. The corresponding proportions in the OSA-symptomatic control group were 71.3% for overall OSA, and 36.3%, 18.1%, and 16.9%, for mild, moderate, and severe, respectively. Log-transformed AHI in IA patients did not differ significantly from that in controls (F(1,238) = 1.88, p = 0.171, partial η² = 0.008).  

Conclusion: RP findings demonstrated a high and largely unrecognized burden of OSA in IA patients. These findings support more systematic and proactive screening for OSA in IA patients, particularly among those with established OSA risk factors. 

74. Reducing neuroinflammation through microglia targeting: ketoprofen OATP1C1 prodrug approach 

Authors: Adéla Králová¹ , Ville Honkanen¹ , Seyed Hamed Maljaei¹ , Janne Tampio¹ , Arun Kumar  Tonduru¹ , Aaro Jalkanen¹ , Jaana Rysä¹ , Kristiina M. Huttunen¹  

Author affiliations: ¹ School of Pharmacy, Faculty of Health Sciences, University of Eastern Finland 

Rationale  
Organic anion-transporting polypeptide 1C1 (OATP1C1) is a solute carrier transporter responsible for brain uptake of thyroid hormones.¹ In the human brain, OATP1C1 is expressed in glial cells, mainly in astrocytes and microglial cells, and endothelial cells of the blood-brain barrier.²ˉ³ Therefore, OATP1C1 might be considered as a promising target for transporter-utilizing prodrugs to enhance glial drug delivery while limiting peripheral side effects.   
 
Methods   
OATP1C1 expression in mouse microglial cells, astrocytes, and neurons was quantified using LC–MS proteomics. We designed and evaluated an OATP1C1-utilizing prodrug of ketoprofen to improve the delivery of an anti-inflammatory drug into immortalized mouse cells (BV2). Cellular accumulation of the prodrug was compared with ketoprofen. Anti‑inflammatory effects were assessed in lipopolysaccharide (LPS)‑stimulated BV2 cells by quantifying cyclooxygenase‑2 (COX2) and tumor necrosis factor‑α (TNFα). Brain distribution of the prodrug and ketoprofen was evaluated in a pharmacokinetic study with mice.   
 
Results   
OATP1C1 expression was significantly higher in BV2 microglial cells than in neurons and astrocytes. The prodrug increased cellular ketoprofen accumulation in BV2 cells. In LPS‑stimulated cells, the prodrug reduced COX2 and TNFα levels by 39.04% and 40.30%, respectively, compared with the LPS‑treated group. Prodrug accumulated effectively in the brain and exhibited higher and prolonged brain exposure compared to ketoprofen.   
 
Conclusions   
OATP1C1 enables targeted microglial drug delivery. The ketoprofen prodrug improved brain uptake and reduced inflammatory markers, supporting OATP1C1 as a promising transporter for microglia‑focused therapeutics. 

75. Retinal glial cells and glial scar formation 

Author: Niina Harju  

Author affiliation: University of Eastern Finland 

Rationale: The retina is both developmentally and anatomically an extension of the CNS [1]. It contains neurons with their synapses, the axons of ganglion cells form the optic nerve, and photoreceptors transmit light signals to the brain [1]. As in the CNS, damage to the retina (e.g. injury or disease) can lead to a neurotoxic environment and scarring [1–3]. There are three types of glial cells in the retina: Müller cells, astrocytes, and microglia [2–4]. During injury, glial cells activate gliosis, which aims to restore homeostasis, but can also cause pathological conditions [3]. In gliosis, glial cells hypertrophy, proliferate, migrate, and produce intermediate filaments (e.g. vimentin) [3].   
Methods: The scar tissue was fixed, blocked, and stained with Hoechst, anti-vimentin primary antibody, and Alexa Fluor anti-rabbit secondary antibody. The membrane was photographed under a microscope.   
Results: The membrane appeared thin and translucent. It contained occasional cells and abundant ECM (e.g. vimentin).    
Conclusions: Scarring disrupts the structure of the retina and impairs vision, which is why protective treatments are needed [5]. On broader level, the role of different cells in the formation of retinal scarring needs to be elucidated.   

76. Establishing in vitro iPSC model via CRISPR engineering for functional analysis of Schizophrenia risk variant 

Authors: Saana Elay 1, Maria Ilola1, Tarja Malm 1, Henna Jäntti 1 

Author affiliations: 1 A.I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale  
Schizophrenia (SCZ) is a severe neuropsychiatric disorder affecting ~1% of the population and remains inadequately treated, particularly regarding negative symptoms and cognitive impairment. Genome-wide association studies have identified hundreds of common SCZ risk variants, most of which are non-coding and predicted to act in a cell-type- and developmental-stage-specific manner. Functional dissection of these variants has been limited by the lack of experimentally tractable human models.  
 
Methods  
CRISPR-Cas9 genome editing is utilised to establish a human iPSC-based in vitro model to investigate a common, SCZ-associated, human-specific risk variant rs10503253 located in an intronic region of the complement-regulating gene CSMD1. Six iPSC lines were genotyped by Sanger sequencing, and a homozygous wild‑type KOLF2.1J-neurogenin‑2 (NGN2) line was selected for introduction of the risk variant. Edited clones were identified by restriction enzyme analysis and Sanger sequencing. CRISPR-editing was also utilized to integrate fluorescent reporter proteins into a genomic safe harbour of KolfJ2.1-NGN2 line to enable live imaging. For neuronal differentiation, NGN2 was integrated into a safe harbour locus of four iPSC lines. The edited clones were selected by PCR screening and Sanger sequencing.  
 
Results  
We successfully differentiated polyclonal NGN2-neurons, and immunostainings for dendritic and synaptic markers demonstrated robust neuronal differentiation and synapse formation, supporting the suitability of this system for downstream analysis.  
 
Conclusions  
This experimentally tractable human in vitro system with genome editing, and live imaging applications enables tracking of developmental trajectories and cell-type-specific effects of a common non‑coding SCZ risk variant at molecular and functional synaptic levels. 

77. THE ALZHEIMER’S DISEASE PROTECTIVE PLCG2-P522R VARIANT RESTORES MICROGLIAL PHENOTYPE DISRUPTED BY TREM2 DEFICIENCY 

Authors: Inka Kervinen¹, Heli Jeskanen¹, Jessica Rosa¹, Sami Heikkinen¹, Petra Mäkinen¹, Meri Kervinen¹, Kimmo Jokivarsi², Dorit Hoffmann², Annakaisa Haapasalo², Mikko Hiltunen¹, Mari Takalo¹ 

Author affiliations: ¹Institute of Biomedicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland ²A. I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland 

A rare coding variant in the PLCG2 (phospholipase C gamma 2) gene, PLCG2-P522R, reduces the risk of Alzheimer’s disease (AD) and other neurodegenerative diseases, promotes healthy aging, and slows cognitive decline. This variant enhances immune responses, metabolic activity, and viability of microglial cells. In contrast, multiple loss-of-function variants in the TREM2 (Triggering receptor expressed on myeloid cells 2) gene are associated with increased AD risk. TREM2 activates PLCγ2 upon binding to ligands, including beta-amyloid. The aim is to provide mechanistic insights into the protective features of this pathway, which remain poorly understood.  
RNA sequencing and lipidomics were performed in acutely isolated CD11b+ microglia from 13-month-old wildtype (WT), PLCG2-P522R knock-in (KI), TREM2 knock-out (KO) and crossbred PLCG2-P522R KI x TREM2 KO mice. Glucose metabolism in various tissues was assessed in 15-month-old mice using FDG-PET imaging. Different cell models will be used for functional assays to discover the molecular mechanisms of PLCG2-P522R upon TREM2 KO.  
PLCG2-P522R variant restored several biological pathways downregulated by TREM2 KO, including those related to cell motility, TFG-beta signaling and M2 polarization. PLCG2-P522R also reversed lipid profile alterations caused by TREM2 KO. Ongoing analyses are investigating the effects of PLCG2-P522R on TREM2 KO background on mitochondrial metabolism and lipid accumulation in cell models under AD-relevant stress conditions.  
The effects of PLCG2-P522R under TREM2 deficiency need further research to pinpoint the key mechanisms underlying its protective role. These findings provide insight into how the protective PLCG2-P522R variant may mitigate the detrimental consequences of impaired TREM2 function. 

78. Evaluation of two chronic stress exposure protocols for inducing depression-like phenotypes in female C57BL/6J mice 

Authors: Ari Suhonen1, Riikka Immonen1, Mina Yacoub1, David Delotterie1, Johanna Uhari-Väänänen1, Tuulia Huhtala1, Taneli Heikkinen1, Kimmo Lehtimäki1, Jussi Rytkönen1, Anna McGrath2, Susanne Bäck1  

Author affiliations: 1 Charles River Discovery Research Services, Kuopio, Finland;  2 Rare and Neurologic Dis. Res. Therapeut. Area, Sanofi, Cambridge, MA 

Rationale  
Depressive disorders reduce quality of life, and a substantial proportion of patients do not respond to existing treatments. There is therefore a need for improved therapeutic strategies, supported by translational animal models. Chronic stress models aim to mimic the longterm behavioral and neurobiological effects of stress that underlie mood disorders such as depression.  
 
Methods  
Female C57BL/6J mice were assigned to three groups (n=18/group): control, chronic restraint stress (CRS), or chronic unpredictable stress (CUS). Stress exposed mice were single housed one week prior to and throughout the 5week stress period. CRS involved restricted movement up to 4 hours daily, whereas CUS consisted of daily exposure to unpredictable stressors. Behavioral assessments included open field, elevated plus maze, sucrose preference, contextual fear conditioning, forced swim, and tail suspension tests. After stress exposure, brain function was assessed using dynamic functional ultrasound to evaluate whisker evoked activation and vascular reactivity.   
 
Results  
Both CRS and CUS induced anhedonia. However, CRS produced more robust depression like features, including reduced exploratory behavior. No effects were detected in forced swim or tail suspension tests. Functional ultrasound revealed marked hemodynamic alterations following CRS, including increased vascular reactivity consistent with elevated basal vasoconstriction. CRS mice also showed enhanced whisker evoked hemodynamic responses, hyperemia, and impaired perfusion recovery, suggesting disrupted vascular homeostasis.  
 
Conclusions  
Chronic restraint stress induces stronger behavioral and hemodynamic depression like phenotypes than chronic unpredictable stress in mice. While both models capture aspects of stress induced anhedonia, CRS appears particularly sensitive for studying neurobehavioral and neurovascular mechanisms.   

79. Assessment of amyloid related imaging abnormalities (ARIA) in the 5xFAD mouse model following anti amyloid antibody 3D6 treatment 

Authors: Riikka Immonen1, Taneli Heikkinen1, Johanna Uhari-Väänänen1, Tuulia Huhtala1, Mina Yacoub1, Tuukka Miettinen1, Hua Long2, Maya Sokolovsky2 

Author affiliations: 1 Charles River Discovery Services, Kuopio, Finland; 2 Alector LLC, South San Francisco, CA, USA 

Rationale  
Amyloid related imaging abnormalities (ARIA) are magnetic resonance imaging (MRI) findings associated with anti amyloid beta immunotherapies for Alzheimer’s disease. ARIA are broadly classified into edema and effusion (ARIA E), and microhemorrhages or superficial siderosis (ARIA H). This study assessed ARIA like events in the 5xFAD mouse model of Alzheimer’s disease following treatment with the anti amyloid beta antibody 3D6.  
 
Methods  
Approximately 9 month old 5xFAD mice received either 3D6 (n=15) or an isotype control antibody (n=8) for 11 weeks. ARIA like events were evaluated using longitudinal T2 and T2star MRI at baseline and at 3, 7, 9, and 11 weeks after treatment initiation.  
 
Results  
At baseline, mice exhibited few microbleeds (10±5) and no ARIA E or ARIA H lesions. ARIA E, ARIA H, and increased microbleed burden were observed only in 3D6 treated mice, starting at 3 weeks from treatment onset. ARIA E lesions were transient, with partial resolution over time. ARIA H incidence increased at 7–9 weeks and persisted throughout the follow up starting from 3D6 treatment onset. Also the microbleed burden increased progressively starting from 3 weeks. After 11 weeks, ARIA E incidence was 77%, ARIA H 69%, and the number of microbleeds 52±38.  
 
Conclusions  
Treatment with the murine anti amyloid beta antibody 3D6 induced ARIA E, ARIA H, and increased microbleed burden in 5xFAD mice. Longitudinal MRI enables sensitive detection of ARIA like events and supports development of improved anti amyloid treatment strategies.  

80. Non-Neurodegenerative Comorbidities and Healthcare Use Prior to Early-Onset Alzheimer’s Disease: A Population-Based Case-Control Study 

Authors: Laura Leppänen1, Adolfina Lehtonen1; Mikko Aaltonen2; Kalle Aho3; Ave Kivisild3; Abbe Ullgren3,4; Iina Rinnankoski1; Helmi Soppela3; Annakaisa Haapasalo5; Anne M Portaankorva6; Päivi Hartikainen7; Kasper Katisko3; Eino Solje3,7; Johanna Krüger1,8,9 

Author affiliations: 1 Research Unit of Clinical Medicine, Neurology, University of Oulu, Oulu, Finland 2 Law School, University of Eastern Finland, Joensuu, Finland  3 Institute of Clinical Medicine – Neurology, University of Eastern Finland, Kuopio, Finland  4 Department of Neurobiology, Care Sciences and Society, Division of Neurogeriatrics, Karolinska Institutet, Solna, Sweden 5 A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland  6 Clinical Neurosciences, University of Helsinki, Helsinki, Finland  7 Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 8 Neurocenter, Neurology, Oulu University Hospital, Oulu, Finland 9 Medical Research Center, Oulu University Hospital, Oulu, Finland  

Rationale  
To date, only a limited number of studies have investigated the association of previous diseases with early-onset Alzheimer’s disease (EOAD). We examined whether non-neurodegenerative comorbidities and healthcare utilization differ between individuals later diagnosed with EOAD and matched controls during 15 years preceding diagnosis.  
 
Methods  
We compared the prevalence of non-neurodegenerative diseases 1-15 years before EOAD diagnosis between EOAD patients (n=407) from two Finnish referral centers and age-, sex-, and geographically-matched controls without neurodegenerative diseases (n=4067). We also compared the number of healthcare visits between the groups during this period. EOAD diagnoses were made between January 1, 2010, and December 31, 2021, and all the diagnoses were manually re-checked.   
 
Results  
EOAD patients had higher prevalence of depression (5.7% vs. 2.3%), brain injuries (2.9% vs. 0.9%), alcohol-related disorders (3.4% vs. 1.4%), thyroid disorders (2.5% vs. 0.9%), epilepsy (2.2% vs. 0.8%), strokes (4.4% vs. 2.0%), and otorhinolaryngologic diseases (5.2% vs. 2.9%) 1–5 years before diagnosis. Only otorhinolaryngologic diseases differed 6–10 years prior, whereas no differences were detected 11–15 years before diagnosis. Cardiovascular risk factors (hypertension or dyslipidemia) showed no between-group differences. The mean number of health care visits was significantly higher in EOAD patients one year before and at the year of diagnosis, but not earlier.  
 
Conclusions  
Several conditions, such as epilepsy, thyroid disorders, depression, and brain injuries, were significantly more common in EOAD patients during the 1-5 years preceding diagnosis, while cardiovascular risk factors were not overpresented. These findings may aid earlier recognition and understanding of EOAD. 

81. Human cortical long-term potentiation: A dual electrophysiology-transcriptomics view of how Alzheimer’s related pathology alters synaptic plasticity 

Authors: Mireia Gómez-Budia1, Mohammadamin Beheshti Dehkordi 1, Luca Giudice1, Polina Abushik1, Anssi Pelkonen1, Antonios Dougalis1, Valeriia Sitnikova1, Martina Merga1, Nelli-Noora Välimäki1, Mikko Hiltunen2, Tuomas Rauramaa3, Ville Leinonen4,5, Tarja Malm1  

Author affiliations: 1 A. I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland.  2 Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland.  3 Department of Pathology, Kuopio University Hospital, University of Eastern Finland, Kuopio, Finland.  4 Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland.  5 Neurosurgery, Institute of Clinical Medicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland 

Alzheimer’s disease (AD) is a complex disease with beta-amyloid (Aβ) deposits and tau pathology that lead memory impairment. They are suggested to be due to deficits in synaptic plasticity, which mechanistically is mediated by long term potentiation (LTP). Although often studied in hippocampi, long-term memory also resides in the six-layered neocortex. To be able to understand the cellular and molecular mechanisms leading to altered LTP in AD, we use cortical living biopsies from idiopathic normal pressure hydrocephalus living patients. These patients exhibit AD-related pathological hallmarks (Aβ and Tau deposits), which makes them a unique model to study in a human context LTP and AD.  
Pairing Training Protocol (PTP) is used to induce synaptic strengthening in whole cell patch clamp recordings. Moreover, Theta Burst Stimulation (TBS) protocol is used on 60-electrode Microelectrode Arrays (MEA). We have established a novel computational pipeline to extract time and potentiation features from the evoked field post synaptic potentials waveforms. These are used to train a supervised machine-learning classifier to compare patients with different AD-related pathologies. In addition, Xenium spatial transcriptomic technology is used to find the differentially regulated transcripts when LTP is induced.  
Our data shows that iNPH cortical biopsies elicit synaptic plasticity with both PTP and TBS. Moreover, after TBS LTP-related transcripts and pathways are differentially altered. With our novel pipeline we can validate that AD-related pathology impairs synaptic plasticity.  
These unique samples can be a physiologically relevant model to study synaptic function in human cortex and unravel its correlation with AD-related pathology. 

82. Generative Tractography for White-Matter Mapping: Opportunities and Limitations for Clinical Neuroimaging 

Authors: Mahdad Esmaeili, Jussi Tohka, Dogu Baran Aydogan 

Author affiliations: A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale: Tractography reconstructs white-matter pathways from diffusion MRI (dMRI) and underpins connectivity analyses in neurodegenerative disease, stroke recovery, and surgical planning. However, it is fundamentally ill-posed: local fiber orientation distributions (FODs) do not uniquely determine global streamline configurations, leading to false positives and false negatives. Conventional pipelines rely on seeding, stepwise propagation, stopping rules, and post-hoc filtering. Recent learning-based methods suggest tractography can be reframed as data-driven generation, yet ambiguity in mapping local diffusion evidence to global trajectories persists. We investigated seedless generative tractography in a controlled synthetic setting and characterized its opportunities and intrinsic limits. 

Methods: We generated 20 thousand paired tractogram–FOD samples through clustering, purification, affine augmentation, and recombination of bundle-derived streamlines, voxelized into spherical-harmonic FOD volumes (MNI152, 2 mm, lmax=8, 45 coefficients) with fixed cardinality K=1000 and P=8 resampled points per streamline. Streamlines were direction-canonicalized and ordered using a 3D Morton (Z-order) key to provide a permutation- and reversal-invariant target. Non-zero voxels were tokenized (48-D: 45 SH coefficients + normalized coordinates) and processed by a Transformer encoder–decoder that maps sparse FOD tokens to fixed-cardinality streamline sets in a single inference pass. Training combined exact-token negative log-likelihood, neighborhood mass loss, and smooth-L1 coordinate loss with epoch-dependent weighting. 

Results: Training and test losses converged stably. The model preserved global bundle layout, spatial extent, and curvature, matching iFOD2 and PTT performances in compactness and structural regularity, with Hungarian-matched MDF histograms shifted toward lower distances. Crossing, kissing, and bottleneck phantoms recovered dominant global geometry; streamline-level accuracy was highest for crossings and weakest for kissing and bottleneck configurations. 

Conclusions: Generative tractography reconstructs coherent large-scale white-matter geometry in a single pass, offering a promising direction for connectivity analyses in neurodegenerative and cerebrovascular disease, while local ambiguity highlights intrinsic limits requiring caution in clinical translation. 

83. Subject-Specific White-Matter Mapping for Neuromodulation Targeting in Parkinson’s Disease, Epilepsy, and Treatment-Resistant Depression 

Authors: Simona Leserri, Kalle Saarinen, Dogu Baran Aydogan 

Author affiliations: A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale: Deep brain stimulation (DBS) for Parkinson’s disease, drug-resistant epilepsy, and treatment-resistant depression, as well as transcranial magnetic stimulation, act by engaging white-matter connections. Suboptimal targeting remains a major cause of variable clinical outcomes. Current targeting relies on anatomical landmarks, normative regions of interest in MNI space, or structural MRI, none of which capture subject-specific white-matter variability or the underlying organization of structural connectivity. We aimed to provide individualized, voxel-wise maps of spatial connectivity change to support precise neuromodulation targeting and white-matter research. 

Methods: We developed the directional derivative of structural connectivity method, which combines a tractogram with a brain surface to quantify how connectivity changes across space at each voxel. Diffusion MRI data from four Human Connectome Project subjects were used; whole-brain 100-million-streamline tractograms were generated with parallel transport tractography under a modified anatomically constrained tractography framework. Derivatives were computed over 512 directions with 0.5 mm voxels and 0.00125 mm shift, in weighted-magnitude mode. Two outputs were analyzed: the maximal derivative image (amount and direction of maximal connectivity change per voxel; RGB-coded along left–right, anterior–posterior, superior–inferior axes) and the sum image (aggregated changes across directions). Analyses focused on DBS targets for obsessive-compulsive disorder (anterior limb of the internal capsule) and major depressive disorder (subgenual anterior cingulate cortex). 

Results: Maximal derivative images revealed data-driven topographic boundaries aligned with known anatomy, providing information beyond Track Density Imaging. Connectivity-change patterns were consistent yet subject-specific, particularly near the MDD DBS target, where heterogeneous, multi-bundle organization was evident. 

Conclusions: Directional derivatives of connectivity offer a novel, subject-specific contrast supporting personalized DBS planning across neurodegenerative, epileptic, and psychiatric indications, pending anatomical and clinical validation. 

84. Aimstim: An In-House Neuronavigation Platform for Real-Time MRI Motion Correction Using External Optical Tracking 

Authors: Veikka Leppänen¹, Joonas Laurinoja¹², Dogu Baran Aydogan¹² 

Author affiliations: 

¹ A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

² Department of Neuroscience and Biomedical Engineering, Aalto University, Espoo, Finland 

Rationale: Real-time MRI motion correction requires low-latency, image-independent pose estimation that can be tightly integrated with scanner acquisition. Image-based registration tools (e.g., SPM, FSL FLIRT) are slow and can be unreliable for online use, particularly under large or low-contrast motion. We present Aimstim, our in-house C++/Qt neuronavigation software, extended into a real-time MRI motion correction platform that couples external optical tracking with on-the-fly volume resampling. Aimstim was developed for image-guided brain stimulation with specific features developed for applications inside an MRI scanner. 

Methods: Aimstim integrates four components: (1) Optical tracking module — 6-DOF head pose acquired continuously from an NDI Polaris Vega VT tracker (60 Hz, 17 ms latency, <0.3 mm RMS accuracy) via a custom forehead-fixed subject tracker and reference rigid bodies. (2) Subject registration module — point-based registration using ≥3 anatomical landmarks solved in closed form (Procrustes/SVD), exposed through Aimstim’s interactive GUI with real-time fiducial registration error (FRE) feedback (achieved FRE <3 mm). (3) Synchronization module — tracker poses and MRI volume triggers jointly logged by the EEG recording system under a shared clock, ensuring deterministic alignment between pose samples and EPI volumes. (4) Real-time correction module — for each EPI volume, the tracker-derived rigid transform relative to the first volume is inverted and applied via trilinear resampling using the in-house developed nibrary C++ library, producing motion-corrected volumes in a common reference frame. Experiments were performed on a 3T Siemens MAGNETOM Skyra. Validation included tracker noise-floor assessment, cued motion, and realistic motion paradigms; NDI-derived motion parameters were compared to SPM Realign & Unwarp, and image quality was assessed via normalized mutual information (NMI) to T1 and tSNR. 

Results: Aimstim’s real-time performance was benchmarked across 601 volumes on a Dell Precision 5680 workstation (RT Linux). Mean per-volume processing time was 37.00±1.78 ms, with a 99th percentile of 39.10 ms — well below the acquisition TR of 1 s. The processing breakdown was: image write 26.55 ms (72%), transform and resampling 7.35 ms (20%), and image read 3.11 ms (8%), confirming I/O as the dominant cost and identifying clear targets for further optimization. 

Conclusions: Aimstim provides a modular software platform for real-time, MRI motion correction driven by external optical tracking. Its low latency, deterministic synchronization, and extensible C++/Qt architecture support integration with a broad range of MRI experiments, complementing image-based registration approaches. 

85. Anatomically Refined Whole-Brain Tractography for Clinical Brain Connectivity Analysis: Reducing False Positives in Neurodegeneration, Epilepsy, and Stroke 

Authors: Simona Leserri, Dogu Baran Aydogan 

Author affiliations: A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale: Diffusion MRI tractography has potential for brain connectivity-based biomarkers in neurodegenerative disease, epilepsy, and stroke, but the technical problem is fundamentally ill-posed:  multiple fiber configurations explain the same diffusion signal, producing false-positive and false-negative streamlines. Existing constraint strategies—geometric thresholds, Anatomically Constrained Tractography (ACT), ExTractorFlow, and post-hoc filtering (SIFT2, COMMIT2)—reduce but do not eliminate implausible streamlines, and detailed neuroanatomical knowledge remains fragmented and unevenly distributed. We introduce Hierarchical Anatomical Refinement of Pathways (HARP), an extensible framework that injects anatomical constraints at progressively finer levels in parallel with increasingly granular brain segmentations. 

Methods: HARP organizes regions into nested levels: Level 0 (brain mask), Level 1 (≡ ACT: GM, WM, subcortex, brainstem, CSF, abnormalities), and Level 2 (hemisphere- and cerebellum-specific subdivisions). Level 2 adds ten trajectory rules enforcing categorical separation into association, commissural, and projection fibers, prohibiting configurations such as associations crossing the contralateral hemisphere or brainstem and commissural fibers crossing the brainstem. Implemented in our in-house fiber tracking software, Trekker (harp branch), using high-resolution surface meshes to avoid partial-volume effects, each streamline is evaluated against 27 criteria and classified by 13 rules with a defined priority order. We applied HARP to the ISMRM 2015 Challenge phantom and to TractoInferno (six clinical cohorts) and HCP datasets (98 subjects total), reconstructing 10M streamlines per subject using SD_STREAM, iFOD2, Default PTT, and Modified PTT, with linear mixed-effects analyses and harmonized comparison against SIFT2/COMMIT2. 

Results: On the ISMRM phantom, L2 flagged 2.46% of streamlines as implausible, none corresponding to ground-truth pathways. L1 removed 55.9–93.5% of unconstrained streamlines; L2 additionally removed 0.85–9.18%. PTT methods produced fewer implausible streamlines than iFOD2/SD_STREAM. Implausible streamlines spanned the full range of SIFT2/COMMIT2 weights, with no threshold separating them from plausible ones. 

Conclusions: HARP extends ACT with hierarchical anatomical priors, removing systematic false positives that post-hoc filtering cannot identify. HARP and SIFT2/COMMIT2 are complementary, jointly improving anatomical specificity of tractograms across clinical cohorts and supporting more reliable brain connectivity-based biomarkers for several brain disorders including neurodegeneration, epilepsy and stroke. 

86. Anomap: White-Matter Deformation Mapping Software for Surgical Planning and Tumor-Related Epilepsy 

Authors: Kalle Saarinen, Dogu Baran Aydogan 

Author affiliations: A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

Rationale: Gliomas are highly infiltrative tumors that alter white-matter architecture through infiltration and mass effect, inducing complex deformations linked to tumor aggressiveness, neurological dysfunction, and tumor-related epilepsy, which is prevalent in low-grade gliomas.  Diffusion MRI tractography is sensitive to these alterations, but whole-brain quantification of white-matter deformation remains limited owing to the scale and geometric complexity of tractography data. This work presents Anomap, our deep-learning–based software tool that converts whole-brain tractograms into voxel-wise anomaly maps of white-matter deformation. 

Methods: Anomap is structured as a three-module pipeline. (1) Streamline embedding module: a convolutional autoencoder with a multi-layer encoder–decoder architecture and a 64-dimensional latent space, trained on 100 million synthetic streamlines generated in MNI space (L1 loss, Adam optimizer) to produce geometry-preserving embeddings. (2) Normative reference module: tractograms from 91 Human Connectome Project subjects are embedded and clustered using a distance-based strategy to define a population-level reference in latent space. (3) Anomaly mapping module: patient tractograms are projected into the latent space, and voxel-wise anomaly scores are computed as Euclidean distances to cluster centers, producing a scalar 3D anomaly image co-registered with standard neuroimaging outputs. 

Results: Validation of the embedding module showed strong correlation between latent-space and geometric distances for synthetic (R2=0.88) and in vivo (R2=0.78) streamlines, confirming geometry preservation. Applied to glioma patients, Anomap was shown to produced anomaly maps that highlighted peritumoral white-matter deformation, with anomaly magnitude scaling with tumor size and grade (P1: 15.24±1.85; P2: 14.58±1.66; P3: 15.50±2.38). 

Conclusions: Anomap provides a reproducible, whole-brain software framework for generating voxel-wise white-matter deformation contrasts from tractography data. Its modular design enables integration into existing diffusion MRI pipelines and supports applications in glioma surgical planning and tumor-related epilepsy research. 

87. Early Amyloid Spine Response in Human Cortex: Structural and Functional Synaptic Alterations in Alzheimer’s Disease Revealed through iNPH Biopsies 

Authors: Polina Abushik¹⁺, Antonios Dougalis¹⁺, Mireia Gómez-Budia¹, Mohammad Rezaie¹, Luca Giudice¹, Saeed Saravani¹, Valeriia Sitnikiova¹, Dilyara Nurkhametova¹, Raisa Giniatullina¹, Anssi Pelkonen¹, Nataliia Novosolova¹, Vera Lezhneva¹, Fateme Ghasemi¹, Anastasia Shakirzyanova¹, Akash Mali¹, Henna Jäntti¹, Paula Korhonen¹, Eline Mertens², Huib Mansvelder², Hilde Lavreysen³, Nachiket Kashikar³, Maren Engelhardt⁴, Christian de Kock², Tuomas Rauramaa^5,6, Beth Stevens^7,8,9, Evan Macosko⁷, Mikko Hiltunen¹⁰, Ville Leinonen^5,11, Tarja Malm^1,12,* 

 Author affiliations: 

¹ A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland  

² Department of Integrative Neurophysiology, CNCR, Vrije Universiteit Amsterdam, The Netherlands 

³ Janssen Research and Development Beerse, Neuroscience Therapeutic Area 

⁴ Johannes Kepler University Linz, Institute of Anatomy and Cell Biology, Linz, Austria 

⁵ Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland 

⁶ Department of Pathology, Kuopio University Hospital, Kuopio, Finland 

⁷ Broad Institute of MIT and Harvard, Cambridge, MA, USA 

⁸ F.M. Kirby Neurobiology Center, Boston Children’s Hospital, Boston, MA, USA 

⁹ Howard Hughes Medical Institute (HHMI), Boston, MA, USA 

¹⁰ Institute of Biomedicine, University of Eastern Finland, Kuopio, Finland 

¹¹ Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland 

¹² Lead contact 

(+) Equal contribution 

* Corresponding author 

Understanding how early Alzheimer’s disease (AD) pathology alters human cortical circuit function has been limited by the inaccessibility of viable brain tissue. To address this limitation, we studied the acute frontal cortex biopsies from idiopathic normal pressure hydrocephalus (iNPH) patients—a unique cohort in which ~40% exhibit early amyloid‑β (Aβ) pathology and ~10% show additional tau—to directly probe early AD–related circuit dysfunction. Using an integrated approach combining single‑cell and network‑level electrophysiology, neuronal morphology, single‑nucleus RNA sequencing, and spatial transcriptomics, we studied structural, functional, and molecular adaptations in supragranular cortical circuits. 

Aβ deposition induces coordinated remodeling of supragranular pyramidal neurons, marked by reduced layer 1–mediated inhibition, NMDA‑dependent hyperexcitability, increased bursting, and a increase in dendritic spine density. In cases with combined Aβ and tau pathology, we found the increase of sodium and potassium current densities, enhanced action potential amplitudes, and strengthened afterhyperpolarization. 

Single‑nucleus and spatial transcriptomic analyses identify RORB⁺ layer 3 pyramidal neurons as selectively vulnerable, exhibiting proximity‑dependent suppression of excitatory synaptic gene programs, altered ion channel expression, and disrupted neuron–microglia ligand–receptor signaling. Machine‑learning classification reveals that intrinsic biophysical features, together with cognitive scores, robustly predict the underlying pathological state of the biopsy. 

Together, these results provide the first direct functional evidence that early Aβ and tau deposition restructure human supragranular cortical circuits at intrinsic, synaptic, and molecular levels, establishing a hyperexcitable cortical phenotype at the earliest stages of Alzheimer’s disease. 

88. Multimodal and Multiscale Data Integration for Focal Epilepsies: From Clinical Acquisition to Reproducible Analysis Pipelines 

Authors: Mastaneh Torkamani-Azar¹, Henri Eronen², Ekaterina Paasonen^1,3, Omar Narvaez¹, Jenni Kyyriäinen¹, Anssi Pelkonen¹, Mireia Gomez Budia¹, Adriana Della Pietra¹, Vera Lezhneva¹, Polina Abushik¹, Nataliia Novosolova¹, Liudmila Saveleva¹, Yawu Liu², Elias Ylä-Herttuala^1,4, Juhana Hakumäki^4,5, Antti Huotarinen^2,6, Leena Jutila⁷, Tuomas Rauramaa^2,8, Arto Immonen⁶, Tarja Malm¹, Ville Leinonen^2,6, Alejandra Sierra¹, Olli Gröhn¹, Reetta Kälviäinen^2,7, KEC Epilepsy Surgery Group⁷, Jussi Tohka¹   

Author affiliations:  
¹University of Eastern Finland, A.I. Virtanen Institute for Molecular Sciences, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland. 
²Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland. 
³NeuroCenter, Kuopio University Hospital, Kuopio, Finland. 

⁴Imaging Center, Kuopio University Hospital, Kuopio, Finland. 

⁵School of Medicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland. 
⁶NeuroCenter Neurosurgery, Kuopio University Hospital, Kuopio, Finland. 

⁷Kuopio Epilepsy Center, Full Member of ERN EpiCARE, Kuopio University Hospital, Finland. 

⁸Department of Pathology, Kuopio University Hospital, Kuopio, Finland. 

Rationale: Precise localization of epileptogenic zones remains a major clinical challenge in focal epilepsies. While addressing this challenge requires integration of structural, functional, and molecular data across scales, current datasets remain fragmented due to clinical heterogeneity, regulatory constraints, and lack of standardized workflows. We present a clinically embedded, GDPR-compliant framework for harmonized multimodal data integration to support reproducible epilepsy research and care. 

Methods: Building on a standardized operating procedure for preservation of resected tissue from the operating room to research laboratories (Kyyriäinen et al. 2025), we implemented a clinical-research framework spanning core clinical variables, intraoperative neuronavigation, and in vivo and ex vivo neuroimaging, electrophysiology, microscopy, and omics data. Data are pseudonymized at source and organized using BIDS with novel customization to modality-specific extensions. Storage and processing are conducted within secure UEF Bioinformatic Center infrastructures, with pipelines supporting metadata harmonization, quality control, and reproducible analyses aligned with FAIR principles. 

Results: Implemented in a cohort of 30+ healthy volunteers and 30+ patients with focal drug-resistant epilepsies, the framework enables systematic linkage of neuroimaging, electrophysiology, histopathology, and molecular data. Integrated analyses demonstrate feasibility for mapping structural and functional abnormalities across scales. Electrophysiological case studies with hippocampal recordings illustrate network-level characterization of epileptogenic tissue (Pelkonen et al. 2026). 

Conclusions: This framework establishes a secure and reproducible foundation for epilepsy research. By integrating robust data management with clinically grounded workflows, it contributes to emerging efforts in cross-center harmonization and data-driven stratification, enabling scalable analyses and translation of micro-level mechanisms into clinical insights for neurological disorders. 

89. Grey Matter Volume of Fronto–Striatal and Occipital Regions correlates with Visuospatial andGlobal Cognition in older adults 

Authors: Akansha Singh 1 , Gauri Shankar Kaloiya 2, Nasreen Akhtar 3  

Author affliations: 1 Department of Physiology, All india Institute of Medical sciences, New Delhi, 2 Department of Psychiatry, All india Institute of Medical sciences, New Delhi, 3 Department of Physiology, All india Institute of Medical sciences, New Delhi 

Rationale: Visuospatial memory and global cognition relies on the integrity of grey matter volume (GMV) in fronto-striatal and occipital structures. This study examines the association between GMV of different regions with visuospatial memory and global cognition in healthy older adults. This can help detect early neural vulnerability and prepare strategies for timely intervention to promote healthy cognitive ageing  
Methods: 58 older adults were recruited for this cross-sectional study (43 male). Indian Council of medical research-NeuroCognitive Toolbox was used for assessing visuospatial memory with Modified Taylor Complex Figure Test and TNI spatial recall. MoCA was used to assess global cognition. 3T T1-weighted structural MRI data was processed with CAT-12 toolbox in MATLAB to extract regional grey matter volume.  
Results: Mean age of the participants was 66±7 years. GMV in left medial frontal cortex and middle occipital gyrus both showed significant positive correlations with global cognition (r= 0.26, p=0.046; r= 0.27, p=0.037) and delayed spatial recall scores (r= 0.32, p=0.013; r= 0.37, p=0.004). Larger caudate nucleus and putamen GMV were significantly related to better visuospatial memory (MTCF immediate recall (r= 0.44, p=0.0007; r= 0.27, p=0.044 respectively). Caudate nucleus GMV was significantly associated with MoCA score (r= 0.30, p=0.018).  
Conclusions:  
In healthy older adults, greater GMV in frontal, occipital, and striatal regions was associated with better visuospatial memory and global cognition, highlighting the role of preserved fronto-occipital and cortico-striatal networks in successful cognitive ageing. This may represent structural markers of resilience to age-related cognitive decline. 

90. BCG Artifact Suppression in EEG–fMRI Using a Kalman Filter: A Validation Study Using a Realistic Agar Phantom 

Authors: Joonas Laurinoja¹², Veikka Leppänen¹, Joonas Ryssy³, Matilda Makkonen², Tuomas Mutanen², Risto Ilmoniemi², Dogu Baran Aydogan¹ 

Author affiliations: 

¹ A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland 

² Department of Neuroscience and Biomedical Engineering, Aalto University, Espoo, Finland 

³ Department of Applied Physics, Aalto University, Espoo, Finland 

Rationale: Simultaneous EEG–fMRI combines high temporal and spatial resolution but is compromised by ballistocardiogram (BCG) artifacts arising from pulsatile motion, which obscure neural signals. Existing correction methods have not been systematically benchmarked under controlled conditions. We aimed to (i) establish a controlled benchmarking framework using an MRI-compatible agar phantom that isolates motion-induced artifacts from neural activity, and (ii) introduce and evaluate a novel carbon-wire loop (CWL)-based Kalman filter for real-time BCG correction. 

Methods: A 4% agar, 0.5% NaCl gel layer mimicking skin conductivity was placed between the scalp of a healthy volunteer and a 64-channel EEG cap (NeurOne Tesla; 3T Siemens MAGNETOM Skyra), electrically isolating neural signals so recordings reflected only MRI-induced artifacts. Seven CWLs and one ECG channel were used for artifact suppression. The 5-minute artifact-only recording was combined with an artifact-free resting-state EEG to construct a benchmarking test signal. Three correction methods were compared: ECG-based Optimal Basis Set (OBS), CWL-based linear least-squares fitting (VDM), and the proposed CWL-based Kalman filter. Performance was evaluated against the reference using power spectral density, mu-band (8–12 Hz) topographies, Hilbert-based circular phase correlation, and time-domain RMSD and interquartile range. 

Results: The Kalman filter achieved the lowest RMSD in the 8–12 Hz band, the highest phase circular correlation in both full-band and mu-band analyses, and the smallest time-domain RMSD and IQR across channels. VDM produced competitive results, while OBS was least effective. 

Conclusions: The proposed Kalman filter outperforms established BCG correction methods and operates sample-wise, enabling real-time EEG–fMRI applications such as neurofeedback and adaptive stimulation. 

91. Geometric Moduli-Space Framework for Alzheimer’s Disease EEG Analysis 

Authors: R. Murat Demirer¹, Sultan Tarlacı²  ¹ Biomedical Engineering Program, Faculty of Engineering and Natural Sciences, Işık University, Istanbul, Türkiye ² Department of Neuroscience, Üsküdar University, Istanbul, Türkiye 

Author affiliations: 1. Biomedical Engineering Program, Faculty of Engineering and Natural Sciences, Isik University, Istanbul, Türkiye; 2. Faculty of Medicine, Department of Neurology, Üsküdar University, Istanbul, Türkiye 

Rationale. Neurodegeneration is usually described by symptoms or molecular pathology, but these descriptions do not explain how disease reshapes cortical dynamics over time. We model multichannel EEG as covariance trajectories on the symmetric positive-definite (SPD) manifold and define a moduli space after removing non-informative symmetries. We hypothesize that Alzheimer’s disease (AD) produces a directional contraction of this space, reflecting reduced dynamical degrees of freedom.  
Methods. Longitudinal EEG recordings from 27 AD patients (20 female, 7 male; mean age 69.7 ± 8.7 years; median follow-up 214 days) were analyzed. Covariance trajectories were computed with the affine-invariant Pennec-Fletcher-Ayache metric. Symmetries related to time reparameterization, montage/reference changes, source mixing and window-phase shifts were quotiented out. Three invariant descriptors were evaluated: moduli volume, length-spectrum exponent and Shannon entropy of the geodesic-current measure.  
Results. In patients with three or more EEG visits, preliminary estimates showed positive moduli-volume contraction over follow-up. This finding is consistent with loss of recurrent dynamical structure and supports a geometric interpretation of Alzheimer-related cortical decline.  
Conclusions. EEG covariance dynamics can be represented as trajectories in a moduli space whose contraction provides a falsifiable and biophysically interpretable marker of neurodegenerative progression. The framework links longitudinal EEG changes to invariant geometric descriptors and may support future manifold-based biomarkers in clinical neuroscience.  
Keywords: moduli space; SPD manifold; affine-invariant geometry; covariance trajectory; cortical dynamics; neurodegeneration; longitudinal EEG; Alzheimer’s disease.  

92. Effects of the AD- and FTD-associated SORT1 genetic risk variant on human microglia. 

Authors: Elena Ppali 1, Tomi Hietanen 1, Nadine Huber 1, Giacomo De Luca 1, Dorit Hoffmann 1, Roosa-Maria Willman2, Eino Solje3,5, Ville Leinonen 4, Anssi Lipponen 2, Sami Heikkinen 2, Mikko Hiltunen 2, Annakaisa Haapasalo 1  

Author affiliations: 1. A.I. Virtanen Institute for Molecular Sciences, UEF  2. Institute of Biomedicine, UEF 3. Institute of Clinical Medicine – Neurology, UEF  4. Institute of Clinical Medicine – Neurosurgery and Neuro Center, Neurosurgery, Kuopio University Hospital 5. Neuro Center – Neurology, Kuopio University Hospital, Kuopio, Finland 

Aims: Several variants of the SORT1 gene, encoding sortilin, an endo-lysosomal system receptor, have been associated with an increased risk of Alzheimer’s disease and frontotemporal dementia. However, the underlying molecular mechanisms of the risk effect are still unclear. In this study, we aimed to gain insights into how the SORT1 rs141749679 risk variant affects the phenotype, function, and gene expression in human microglia, the brain’s key immune cells.  
Methods: Skin fibroblasts from SORT1 rs141749679-C risk variant carriers and non-carrying individuals (rs141749679-T; controls) were reprogrammed into induced pluripotent stem cells (iPSCs) and differentiated into microglial cells (iMG). Functional analyses, immunofluorescence (IF) stainings, and global RNA sequencing (RNAseq) were carried out to assess the effects of the variant on iMG treated with proinflammatory stimuli or a vehicle.  
Results: RNAseq identified several rs141749679-C-specific gene expression changes in the iMG, including those involved in IFN responses, lysosome function, and trafficking between Golgi and the endoplasmic reticulum. IF results confirmed the co-localization of sortilin with TGN38, the Golgi marker, and the lysosomal protein Lamp2a in the iMGs. However, in the rs141749679-C iMG, sortilin localization appeared more widespread within the intracellular compartments compared to controls. Studies related to the iMG responses to proinflammatory stimuli are ongoing.   
Conclusions: Our data suggest potentially altered immune responses and sortilin localization in SORT1 rs141749679-C-carrying iMG. Understanding how sortilin dysfunction contributes to disease offers important insight, necessary for the development of novel treatments for neurodegenerative diseases.   

93. Unsupervised mapping of myeloarchitecture in glioblastoma 

Authors: Melina Estela¹, Raimo A. Salo¹, Minna Niittykoski¹, Tuomas Rauramaa², Ville Leinonen³, Jussi Tohka¹, Susanna Rantala³, Alejandra Sierra¹ 

Author affiliations: ¹A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland ²Department of Pathology, Kuopio University Hospital, Kuopio, Finland and Unit of Pathology, Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland ³Neurosurgery KUH NeuroCenter, Kuopio University Hospital, Faculty of Health Sciences, School of Medicine, Institute of Clinical Medicine, University of Eastern Finland, Kuopio, Finland 

Rationale  
Myelin integrity and axonal architecture are altered in brain tumor tissue and surrounding areas, yet the extent of this disruption remains poorly understood. In glioblastoma, one of the most aggressive brain tumors, understanding how tumor infiltration affects myeloarchitecture holds direct clinical significance. Myelin-stained glioblastoma sections offer a window into myeloarchitecture across tumor and peritumoral areas, yet unsupervised characterisation of these patterns remains challenging. Previously, we developed an autoencoder-based approach trained on myelin-stained rat brain sections; however, it fails to generalize to human tumor tissue, motivating exploration of alternative approaches.  
 
Methods  
Vision Transformers (ViT) pretrained on large natural image datasets learn rich, general-purpose representations capturing fundamental visual patterns across diverse domains. We hypothesise these may generalise to histological sections, which share the same underlying natural properties. We propose testing different ViT architectures to extract features from myelin-stained glioblastoma sections without task-specific training. Feature quality and attention maps will be evaluated, and the feature space analysed to assess separation between tumoral and peritumoral regions, or patterns of tumor infiltration into myelinated tissue.  
 
Results  
Preliminary results on ViT-based feature extraction and unsupervised tissue characterisation of glioblastoma sections will be presented.  
 
Conclusions  
Characterising how myeloarchitecture is altered across glioblastoma and peritumoral tissue is key to understanding the structural consequences of tumor infiltration into white matter, with implications for patient prognosis and surgical decisions. ViT architectures, with their promising generalisation across histological domains, may help understand how glioblastoma utilises white matter to migrate, invade and spread through the brain. 

94. Improved Alzheimer’s disease model with an hPSCderivedtri-culture system for drug screening 

Authors: Ropafadzo Mzezewa, Josefine R Christiansen, Alison Salvador, Nina Lydia Kazakou, Chiara Kavelar, Benjamin Schmid, Jonas Viswalingam Bagge, Agnete Kirkeby 

Author affiliations: 1. Novo Nordisk Foundation Center for Stem Cell Medicine (reNEW), Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark 2. Bioneer A/S, Denmark 3. Center for Stem Cell-based Disease Modeling and Drug Screening (StemScreen), University of Copenhagen, Denmark 

Alzheimer’s disease (AD) is the leading cause of age-related neurodegeneration, clinically characterized by progressive cognitive decline and dementia. Neuropathologically, it is hallmarked by the accumulation of extracellular β-amyloid (Aβ) plaques and intracellular neurofibrillary tangles containing tau. Accumulation of aggregate proteins results in massive loss of synapses and ultimately neuronal degeneration. Among the most vulnerable neuronal populations are basal forebrain cholinergic neurons (BFCNs), which innervate the cortex and hippocampus, and undergo severe degeneration in AD. As such, common symptomatic treatments target cholinesterase inhibitors. However, despite continued efforts of intense drug development, no effective disease-modifying therapies exist, partly due to the lack of physiologically relevant human models for drug development. Human iPSCs have become increasingly effective tools in modeling human diseases and hold potential for translational research in drug development  
The project aims to establish a disease-relevant human Alzheimer’s disease (AD) model enabling the development of a high-throughput phenotypic drug screen platform. The project utilizes hiPSC-derived neurons carrying an APOE4/APPmut genetic background co-cultured with glial cell types (i.e tri-culture system).   
We show that mutant and wild-type differentiated neuronal cultures express LHX8+/ISLT+/CHAT + markers, confirming BFCN identity. Furthermore, we observe endogenous population of GFAP+ astrocytes within the neuronal cultures. We apply key pathological insults to the model that induce tau fibril aggregation/phosphorylation, cellular ageing, and disrupted kinase and lysosomal machinery. Inducing cellular ageing in the cultures by inhibition of the neddylation pathway increased DNA damage, cell death, senescence marker expression, and acute aberrant neuronal activity detected through microelectrode 

95. Biobank of Eastern Finland provides high-quality samples and longitudinal clinical data for neuroscience research 

Authors: Ella Ikonen¹ ², Hanna Kujala¹, Emma Kaartinen¹, Timo Laitinen¹, Kimmo Savinainen¹, Arto Mannermaa¹ ² 

Author affiliations: ¹Biobank of Eastern Finland, Kuopio University Hospital, Finland. ²Institute of Clinical Medicine, Pathology and Forensic Medicine, Multidisciplinary Cancer research community RC Cancer, University of Eastern Finland. 

Biobank of Eastern Finland is a hospital biobank that enables access to biological samples and clinical data from a genetically unique population. The population base of Eastern Finland exceeds 550 000 individuals. Biobank samples are obtained during clinical or research procedures from individuals who have given voluntary biobank consent. Neurological diseases, including various neurodegenerative disorders, dementias, and dementia-like syndromes such as normal pressure hydrocephalus, are among our key focus areas. Sample donors can be recontacted for additional biomedical research, such as targeted sample collections, lifestyle questionnaires, or participation in clinical trials.  
 
Our sample collections include a wide range of sample types from neurological patients, such as blood, cerebrospinal fluid, and fibroblasts. We also have the capability to isolate PBMCs from patient blood samples. In addition to high-quality biological samples, we provide comprehensive longitudinal clinical data, including follow-up information. This is complemented by multimodal imaging data, including CT and MRI, as well as image analysis metrics such as volumes of brain structures. Furthermore, our biobank projects generate an increasing amount of analytical data, including GWAS, HLA imputation, sequencing, and proteomics. Integrating genotype and phenotype data opens new opportunities for personalized medicine in neurological diseases.