[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-470011-105":59,"doc-detail-470011-en":134},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":127,"head_meta":129,"extra_data":131,"updated_unix":133},105,"en","cortical-graph-neural-networks-to-predict-dementia-risk-based-on-mri-derived-cortical-surface-morphonology-poster-presentation","Cortical Graph Neural Networks to Predict Dementia Risk Based on MRI-Derived Cortical Surface Morphonology - Poster Presentation","","Alzheimer’s Disease (AD) is the leading cause of dementia, and cortical atrophy patterns from T1 MRI are sensitive neuroimaging biomarkers, yet early diagnosis at the prodromal stage remains difficult. This work develops an explainable cortical graph convolutional network (GCN) using cortical surface morphonology to represent early atrophy as graph-based features and identify individuals with elevated AD risk. T1 MRI baseline data from ADNI were processed with FreeSurfer, and five-fold models predicted future dementia risk from sMCI vs pMCI. The approach achieves balanced accuracy for AD vs CN and for sMCI versus pMCI risk stratification, supporting early prediction of future dementia onset.",{"@graph":69,"@context":126},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":35,"@type":76,"position":81},"https://docshare.wps.com/document/healthcare/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/cortical-graph-neural-networks-to-predict-dementia-risk-based-on-mri-derived-cortical-surface-morphonology-poster-presentation/470011/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/cortical-graph-neural-networks-to-predict-dementia-risk-based-on-mri-derived-cortical-surface-morphonology-poster-presentation/470011.png","ImageObject",300,407,{"name":92,"@type":93},"moss~","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-04","2026-09-30",true,{"@type":102,"interactionType":103,"userInteractionCount":29},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118,122],{"name":109,"@type":110,"acceptedAnswer":111},"What is the main goal of this poster presentation?","Question",{"text":112,"@type":113},"To develop an explainable cortical graph convolutional network that predicts elevated dementia risk by learning early cortical atrophy patterns from MRI-derived cortical surface morphonology.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How were the MRI data processed and models evaluated?",{"text":117,"@type":113},"T1 MRI baseline data from ADNI were processed using FreeSurfer (v7.4). Five-fold crossvalidated models used 90% for training and 10% for independent testing, with stratification by gender and clinical diagnosis.",{"name":119,"@type":110,"acceptedAnswer":120},"What performance was reported for dementia and progression risk prediction?",{"text":121,"@type":113},"The cortical GCN achieved a balanced accuracy of 0.736 for differentiating dementia (AD) versus cognitively normal (CN), and an average mean balanced accuracy of 0.644 for predicting stable MCI (sMCI) versus progressive MCI (pMCI).",{"name":123,"@type":110,"acceptedAnswer":124},"How will the study be extended in future work?",{"text":125,"@type":113},"Future work will validate the model on independent NACC/ADRC data to assess generalizability and improve explainability using methods such as Grad-CAM and integrated gradients.","https://schema.org",{"og:url":83,"og:type":128,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":130,"canonical":83},"index,follow",{"doc_id":132,"site_id":62},470011,1790776557,{"code":4,"msg":5,"data":135},{"doc_id":132,"user_id":136,"nickname":92,"user_avatar":137,"doc_module":4,"category_id":34,"category_name":35,"doc_title":65,"doc_description":67,"doc_content":138,"file_id":139,"file_url":140,"file_type":141,"file_size":142,"view_count":29,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":14,"language":143,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":144,"faqs":145,"seo_title":146,"seo_description":67,"update_tm":133,"read_time":24,"preview_page_count":4},962090895452,"https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd","DOI: 10. 1002/alz70856_107666  \nBIOMARKERS  \nPOSTER PRESENTATION  \nNEUROIMAGING  \nCortical Graph Neural Networks to Predict Dementia Risk Based on MRI-Derived Cortical Surface Morphonology  \nGuanlin Guo 1, 2  Harinishree Sathu3  Marc D. Rudolph4  Trey R. Bateman4   \nTimothy M. Hughes4  Suzanne Craft4  Metin Nafi Gurcan4  Sheng Luo2  Da Ma4  \n1 University of Texas Health Science Center at Houston, Houston, TX, USA  \n2 Duke University School of Medicine, Durham, NC, USA  \n3 Georgia Institute of Technology, Atlanta, GA, USA  \n4Wake Forest University School of Medicine, Winston-Salem, NC, USA  \nCorrespondence  \nDa Ma, Wake Forest University School of Medicine, Winston-Salem, NC, USA. [Email:](Email: dma@wakehealth.edu)[ dma@wakehealth.edu](Email: dma@wakehealth.edu)  \nAbstract  \nBackground: Alzheimer’s Disease (AD) is the leading cause of dementia. Cortical atrophy patterns derived from T1 MRI are sensitive neuroimaging biomarkers to detect early signs of AD-related neurodegeneration. However, early diagnosis at theprodromal stage is challenging duetothe subtle and heterogeneous neuropathological and neurodegeneration patterns over the cortical surface, which cannot be captured by conventional deep learning methods natively. In this study, we developed an explainable cortical graph convolutional network (GCN) that captures the early signs of atrophy patterns in the cortical surface as graph-based features to identify subjects with elevated AD risk.  \nMethod: T1 MRI data from the baseline visit of Alzheimer’s Disease Neuroimaging Initiative (ADNI; 1645 subjects, 902 Male, 743 Female; stable NC [sNC]: 523, stable [AD]:339, MCI:783)dataset was processed through FreeSurfer(v7.4). Five-fold crossvalidated models were built using 90% of the CN+AD data as training (stratified by gender and clinical diagnosis), with 10% reserved for independent testing. The dementia risk models were then applied to predict stable mild cognitive impaired (sMCI) vs. progressive MCI (pMCI) to evaluate their performance to predict future risk of dementia onset. A cortical GCN model handles the cortical surface mesh as graphicbased input, where cortical morphological data (thickness and curvature) are defined on each vertex (graph node) and used as predictive features. During training, batch normalization and dropout were applied to each graph convolutional layer to mitigate overfitting and enhance training stability.  \nResult: The cortical GCN classification model achieved a balanced accuracy of 0.736 in differentiating dementia (AD) and cognitively normal (CN) subjects and an average mean balanced accuracy of 0.644 for predicting sMCIfrom pMCI.  \nConclusion: This project demonstrated the effectiveness of using cortical GCN to achieve early prediction for future dementia onset for subjects with risk of AD.  \nThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.  \n© 2025 The Alzheimer’s Association. Alzheimer’s & Dementia published by Wiley Periodicals LLC on behalf of Alzheimer’s Association.  \nAlzheimer’s Dement. 2025;21(Suppl. 2):e107666.  \n[https://doi.org/10.1002/alz70856_107666](https://doi.org/10.1002/alz70856_107666)  \nwi[leyonlinelibrary.com/journal/alz](leyonlinelibrary.com/journal/alz)  \n1of2  \nFuture work will focus on independent validation of NACC/ADRC data to evaluate the model’s generalizability, as well as enhance the model’s explainability through techniques such as Grad-CAM and integrated gradient. These efforts will provide deeper insights into AD’s neuroanatomical aspects and contribute to developing more accurate, transparent, and effective diagnostic tools.","cbCaitMTB3CCLTco","https://ap.wps.com/l/cbCaitMTB3CCLTco","pdf",372874,"English","# Abstract\n## Background\n## Method\n## Result\n## Conclusion\n## Future work","[{\"question\":\"What is the main goal of this poster presentation?\",\"answer\":\"To develop an explainable cortical graph convolutional network that predicts elevated dementia risk by learning early cortical atrophy patterns from MRI-derived cortical surface morphonology.\"},{\"question\":\"How were the MRI data processed and models evaluated?\",\"answer\":\"T1 MRI baseline data from ADNI were processed using FreeSurfer (v7.4). Five-fold crossvalidated models used 90% for training and 10% for independent testing, with stratification by gender and clinical diagnosis.\"},{\"question\":\"What performance was reported for dementia and progression risk prediction?\",\"answer\":\"The cortical GCN achieved a balanced accuracy of 0.736 for differentiating dementia (AD) versus cognitively normal (CN), and an average mean balanced accuracy of 0.644 for predicting stable MCI (sMCI) versus progressive MCI (pMCI).\"},{\"question\":\"How will the study be extended in future work?\",\"answer\":\"Future work will validate the model on independent NACC/ADRC data to assess generalizability and improve explainability using methods such as Grad-CAM and integrated gradients.\"}]","Cortical Graph Neural Networks to Predict Dementia Risk Based on MRI-Derived Cortical Surface Morphonology - Poster Presentation | PDF"]