[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85706-en":3,"doc-seo-85706-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85706,137441390410,"Hazel","https://ap-avatar.wpscdn.com/avatar/2000252f4ab5702993?_k=1776741390130283984",8,"Research & Report","Cross-Subject Modeling for Widefield Calcium Imaging via Atlas-Aligned Spatiotemporal Tokenization","Large-scale, multi-subject widefield calcium imaging enables brainwide cortical dynamics study, yet high-dimensional spatiotemporal structure and task-irrelevant activity restrict modeling to single-session settings, limiting scalability and generalization. Prior multi-subject pretraining for other modalities exists, but multi-subject foundation modeling for widefield calcium imaging and subject-invariant zero-shot decoding remain largely unaddressed. WiCAT is proposed as a multisubject self-supervised model using atlas-grounded spatiotemporal tokenization to learn globally shared representations. It supports lightweight decoding and strong zero-shot behavior decoding across unseen subjects and reconstructs left-out brain regions across datasets, surpassing baselines.","Cross-Subject Modeling for Widefield Calcium Imaging via Atlas-Aligned Spatiotemporal Tokenization  \nMohammad Hosseini 1 Eray Erturk 1 Saba Hashemi 2 Maryam M. Shanechi 1 2 3  \narXiv :2607 .09754v 1 [ cs .CV] 5 Jul 2026  \nAbstract  \nLarge-scale, multi-subject widefield calcium imaging provides unprecedented access to brainwide cortical dynamics. However, the high dimensionality, complex spatiotemporal structure, and substantial task-irrelevant activity in widefield recordings have largely restricted modeling efforts to single-session analyses, limiting scalability and generalization. While multi-subject pretrained models have been explored for some neural modalities, multi-subject models for widefield calcium imaging have not yet been demonstrated; further, subject-invariant zero-shot behavior decoding remains elusive for multi-subject models across neural modalities more broadly. As a first step toward foundation modeling of widefield data, we introduce WiCAT, a multisubject model that leverages self-supervised pretraining to both outperform single-session models and enable zero-shot behavior decoding on unseen subjects. WiCAT introduces an atlasgrounded tokenization scheme without sessionspecific components and learns globally shared spatiotemporal representations. Across multiple widefield datasets, the pretrained model supports lightweight downstream decoding, transfers across subjects, tasks, and datasets, and outperforms baseline models. Notably, the model also achieves robust zero-shot continuous behavior decoding and left-out brain region reconstruction on unseen subjects. Code: [https://github](https://github) . com/ShanechiLab/WiCAT/  \n1Ming Hsieh Department of Electrical and Computer Engineering, USC Viterbi School of Engineering, University of Southern California, Los Angeles, CA, USA 2Thomas Lord Department of Computer Science, University of Southern California, Los Angeles, CA, USA 3Alfred E. Mann Department of Biomedical Engineering, University of Southern California, Los Angeles, CA, USA. Correspondence to: Maryam M. Shanechi \u003C[shanechi@usc.edu](shanechi@usc.edu) >.  \nProceedings of the 43 rd International Conference on Machine Learning, Seoul, South Korea. PMLR 306, 2026 . Copyright 2026 by the author(s) .  \n1. Introduction  \nWidefield calcium imaging is an optical neural imaging modality that enables the recording of neural activity across large portions ofthe cortex, and in some settings near wholecortex coverage, in behaving animals (Cardin et al., 2020 ; Saxena et al., 2020) . By measuring fluorescence signals reflecting intracellular calcium dynamics, widefield imaging provides access to mesoscale neural activity with relatively high temporal resolution compared to imaging techniques that infer neural activity indirectly via vascular responses, while offering substantially broader spatial coverage than electrophysiological recordings (Nietz et al., 2022) . Recent experimental efforts have produced unprecedented largescale widefield datasets spanning multiple animals, recording sessions, behavioral tasks, and laboratories (Musall et al., 2019 ; Couto et al., 2021 ; Raut et al., 2025 ; Findling et al., 2025 ; Kondo et al., 2025) . Thus, widefield imaging has become an important tool for studying distributed population dynamics, large-scale functional organization, and their relationship to complex behavior (Musall et al., 2019 ; Cardin et al., 2020 ; Ren & Komiyama, 2021 ; Nietz et al., 2022) . Despite these advantages, widefield calcium imaging data present significant challenges for learning generalizable representations. Recordings are high-dimensional image time series with rich spatiotemporal structure, containing both task-relevant signals and substantial task-irrelevant or spontaneous activity (Musall et al., 2019 ; de Vries et al., 2020 ; Nietz et al., 2022 ; MacDowell et al., 2024 ; Hosseini & Shanechi, 2025) . Given the above difficulties, existing approaches to modeling widefield imaging data are","cbCaimzxlB5PyStT","https://ap.wps.com/l/cbCaimzxlB5PyStT","pdf",9948364,1,30,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"What problem does WiCAT address in widefield calcium imaging modeling?\",\"answer\":\"Existing methods often model each session or subject separately, which limits scalability and cross-subject generalization. WiCAT targets the ability to learn shared representations and enable subject-invariant zero-shot behavior decoding for multi-subject widefield data.\"},{\"question\":\"How does WiCAT build multi-subject shared representations?\",\"answer\":\"WiCAT uses self-supervised pretraining with an atlas-grounded spatiotemporal tokenization scheme that avoids session-specific components. It learns globally shared spatiotemporal representations across subjects.\"},{\"question\":\"What capabilities does WiCAT demonstrate on unseen subjects?\",\"answer\":\"Across multiple widefield datasets, the pretrained model supports lightweight downstream decoding and transfers across subjects, tasks, and datasets. It also achieves robust zero-shot continuous behavior decoding and left-out brain region reconstruction on unseen subjects.\"}]",1784205712,76,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"cross-subject-modeling-for-widefield-calcium-imaging-via-atlas-aligned-spatiotemporal-tokenization","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/cross-subject-modeling-for-widefield-calcium-imaging-via-atlas-aligned-spatiotemporal-tokenization/85706/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does WiCAT address in widefield calcium imaging modeling?","Question",{"text":75,"@type":76},"Existing methods often model each session or subject separately, which limits scalability and cross-subject generalization. WiCAT targets the ability to learn shared representations and enable subject-invariant zero-shot behavior decoding for multi-subject widefield data.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does WiCAT build multi-subject shared representations?",{"text":80,"@type":76},"WiCAT uses self-supervised pretraining with an atlas-grounded spatiotemporal tokenization scheme that avoids session-specific components. It learns globally shared spatiotemporal representations across subjects.",{"name":82,"@type":73,"acceptedAnswer":83},"What capabilities does WiCAT demonstrate on unseen subjects?",{"text":84,"@type":76},"Across multiple widefield datasets, the pretrained model supports lightweight downstream decoding and transfers across subjects, tasks, and datasets. It also achieves robust zero-shot continuous behavior decoding and left-out brain region reconstruction on unseen subjects.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":21,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]