[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-328187-105":59,"doc-detail-328187-en":131},{"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":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","latent-graph-encoding-of-multimodal-neuroimaging-features-with-generative-ai-architectures","LATENT GRAPH ENCODING OF MULTIMODAL NEUROIMAGING FEATURES WITH GENERATIVE AI ARCHITECTURES","","Generative models are used to encode complex neuroimaging data for feature generation and reconstruction, but optimal architectures and latent-space processing are essential for revealing brain structure and function. A multimodal generative framework is developed for structural and functional MRI features, systematically evaluating encoding strategies, latent multimodal fusion, and generative model selection using GMV and sFNC. Graph-encoded modality-aware latent representations outperform vector and direct data-space encoders, with gMMVAE achieving superior fidelity, reconstruction quality, efficiency, and latent discriminability.",{"@graph":69,"@context":123},[70,84,106],{"@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":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/latent-graph-encoding-of-multimodal-neuroimaging-features-with-generative-ai-architectures/328187/",{"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/latent-graph-encoding-of-multimodal-neuroimaging-features-with-generative-ai-architectures/328187.png","ImageObject",300,407,{"name":92,"@type":93},"Aurelia","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-24","2026-09-21",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",11,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What problem does the paper address in multimodal neuroimaging generative modeling?","Question",{"text":113,"@type":114},"It addresses how to design optimal architectural frameworks for encoding neuroimaging features into latent spaces, including how to choose encoding strategies, latent fusion, and generative models for structural and functional MRI.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"Which MRI modalities and features are used in the framework?",{"text":118,"@type":114},"Structural MRI uses gray matter volume (GMV), and functional MRI uses static functional network connectivity (sFNC) features from a large neuroimaging dataset.",{"name":120,"@type":111,"acceptedAnswer":121},"What encoder and model design achieves the best results?",{"text":122,"@type":114},"Architectures using modality-aware graph encoding of functional connectivity into a lower-dimensional latent space outperform vectorized encoders or direct data space approaches, and the proposed multimodal graph VAE (gMMVAE) leads across multiple evaluation metrics.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},328187,1790217637,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":29,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145,"preview_page_count":4},1099514068365,"https://ap-avatar.wpscdn.com/avatar/10000253d8d9f28188e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1794309523707217728","LATENT GRAPH ENCODING OF MULTIMODAL NEUROIMAGING FEATURES WITH  \nGENERATIVE AI ARCHITECTURES  \nIshaan Batta∗+, Meenu Ajith∗, Vince Calhoun  \nCenter for Translational Research in Neuroimaging and Data Science (TReNDS): Georgia State University, Georgia Institute of Technology, and Emory University, Atlanta, USA  \narXiv :2607 .07027v 1 [ cs .LG] 8 Jul 2026  \nABSTRACT  \nWhile generative models enable encoding of complex neuroimaging data for feature generation and reconstruction, developing optimal architectural frameworks with appropriate encoding and latent space processes is crucial for studying structural and functional properties of the brain. We design a multimodal generative framework for structural and functional magnetic resonance imaging (MRI) features through systematic evaluation of encoding strategies, latent multimodal fusion, and generative model selection. Using structural gray matter volume (GMV) and static functional network connectivity (sFNC) features from a large neuroimaging dataset, we analyze generative frameworks involving variational autoencoders (VAEs), transformers, generative adversarial networks (GANs), and diffusion models. Architectures that employ modality-aware graph encoding of functional connectivity into a lower-dimensional latent space outperform vectorized encoders or direct data space approaches. The proposed multimodal graph VAE (gMMVAE) surpasses alternative generative variants across multiple metrics for generation fidelity, reconstruction quality, efficiency, and latent space discriminability, highlighting its potential for robust multimodal neuroimaging analysis.  \nIndex Terms— Generative AI, Multimodal learning, Neuroimaging, Graph Neural Networks, Data Reconstruction  \n1. INTRODUCTION  \nUnderstanding how brain structure and function relate to externally observed clinical and demographic variables has driven the development of neuroimaging studies leveraging magnetic resonance imaging (MRI) and modern machine learning methods [1, 2] . While one major line of studies focus on learning the associations for accurately predicting target variables [1], many unsupervised and semi-supervised learning frameworks have been developed to create lowdimensional representations for further associative analysis [3, 4] . Deep learning and generative learning models have  \n* these authors contributed equally to this work + Corresponding author: Ishaan Batta ([ibatta@gsu.edu](ibatta@gsu.edu));  \nThis work was supported by NIH grants 1R01AG090597, R01AG073949 awarded to V. Calhoun  \nenabled a more sophisticated way to map complex neuroimaging features into low-dimensional latent encodings for feature-to-target prediction [5] or feature-to-feature mappings for reconstruction and generation [6] . Information encoded in the latent space of these models can be utilized and studied for extracting meaningful associative patterns related to assessment variables and the features being encoded [7] . This can be achieved by studying the latent space itself, or also the prediction and reconstruction outputs yielded by the model.  \nFunctional MRI (fMRI) models brain dynamics and the connectivity between different brain regions [8], while structural MRI (sMRI) captures static features such as gray matter volume (GMV) [9] . Multimodal learning integrating both modalities provides more comprehensive views of brain dynamics [9] compared to unimodal analysis. Deep generative models encode these features into latent spaces either through direct voxel-level processing [5, 6], or using pre-computed features summarized at the level of brain regions or components, such as static functional network connectivity (sFNC) matrices and GMV [3] . The choice of encoders is crucial for downstream learning tasks as it determines the performance and explainability of the model. In case of sFNC features, graph encoders offer a relevant encoding method for functional connectivity [10], and these features enable tasks involving prediction","cbCaivlvT1Hu3vWa","https://ap.wps.com/l/cbCaivlvT1Hu3vWa","pdf",3208989,"English","# ABSTRACT\n# 1. INTRODUCTION\n# 2. METHODS\n## 2.1. Modality-aware encoders and multimodal fusion","[{\"question\":\"What problem does the paper address in multimodal neuroimaging generative modeling?\",\"answer\":\"It addresses how to design optimal architectural frameworks for encoding neuroimaging features into latent spaces, including how to choose encoding strategies, latent fusion, and generative models for structural and functional MRI.\"},{\"question\":\"Which MRI modalities and features are used in the framework?\",\"answer\":\"Structural MRI uses gray matter volume (GMV), and functional MRI uses static functional network connectivity (sFNC) features from a large neuroimaging dataset.\"},{\"question\":\"What encoder and model design achieves the best results?\",\"answer\":\"Architectures using modality-aware graph encoding of functional connectivity into a lower-dimensional latent space outperform vectorized encoders or direct data space approaches, and the proposed multimodal graph VAE (gMMVAE) leads across multiple evaluation metrics.\"}]","LATENT GRAPH ENCODING OF MULTIMODAL NEUROIMAGING FEATURES WITH GENERATIVE AI ARCHITECTURES | PDF",1789978801,15]