{"paper":{"title":"Towards a Virtual Neuroscientist: Autonomous Neuroimaging Analysis via Multi-Agent Collaboration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Multi-agent system NIAgent autonomously builds and refines neuroimaging analysis workflows to outperform fixed pipelines.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Carl Yang, Keqi Han, Lifang He, Songlin Zhao, Xiang Li, Yao Su, Yixuan Yuan","submitted_at":"2026-05-10T06:30:19Z","abstract_excerpt":"Transforming neuroimaging data into clinically actionable biomarkers is a knowledge-intensive and labor-intensive process. Standardized workflows such as fMRIPrep have improved robustness and efficiency, but they are statically configured and cannot reason about downstream objectives, deliberate over alternative strategies, or close the loop between intermediate evidence and subsequent decisions in the way a human researcher would. This lack of closed-loop adaptation often leaves domain experts trapped in a cycle of manual trial-and-error to tune parameters and remediate pipeline failures, sev"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on ADHD-200 and ADNI demonstrate that NIAgent outperforms standard workflow-based baselines in predictive performance while exhibiting sophisticated agentic behaviors, including strategy exploration and adaptive refinement.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The hierarchical verification framework integrating cohort-level metric screening with agentic visual inspection is sufficient to drive evidence-grounded workflow remediation without human intervention or additional safeguards.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"NIAgent is a multi-agent system using code-centric execution and hierarchical verification to autonomously build and adapt neuroimaging analysis workflows, showing better predictive performance than standard pipelines on ADHD-200 and ADNI data.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Multi-agent system NIAgent autonomously builds and refines neuroimaging analysis workflows to outperform fixed pipelines.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ea7c15d22a28dec2eda0715b67a33105553e9e33700fb4d0c4e655374802ea58"},"source":{"id":"2605.09366","kind":"arxiv","version":2},"verdict":{"id":"88a8c4ec-03d6-455f-ba3d-55f6810a7b28","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:07:13.156079Z","strongest_claim":"Experiments on ADHD-200 and ADNI demonstrate that NIAgent outperforms standard workflow-based baselines in predictive performance while exhibiting sophisticated agentic behaviors, including strategy exploration and adaptive refinement.","one_line_summary":"NIAgent is a multi-agent system using code-centric execution and hierarchical verification to autonomously build and adapt neuroimaging analysis workflows, showing better predictive performance than standard pipelines on ADHD-200 and ADNI data.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The hierarchical verification framework integrating cohort-level metric screening with agentic visual inspection is sufficient to drive evidence-grounded workflow remediation without human intervention or additional safeguards.","pith_extraction_headline":"Multi-agent system NIAgent autonomously builds and refines neuroimaging analysis workflows to outperform fixed pipelines."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.09366/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T19:36:40.893887Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T13:01:18.462392Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T10:19:43.387949Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"e9df7d7d7731bd60bf5bf7f263db0018343c8e1a13b80ac39f9dba312b7e7fdd"},"references":{"count":55,"sample":[{"doi":"","year":2016,"title":"The brain imaging data structure, a format for organizing and describing outputs of neuroimaging experiments.Scientific data, 3(1):1–9","work_id":"69555b26-5ec0-4992-89b8-aacab44fac96","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"fmriprep: a robust preprocessing pipeline for functional mri.Nature methods, 16(1):111–116","work_id":"9afc4b40-c942-4fb0-8444-781a9d37b185","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"A comprehensive overview of large language models.ACM Transactions on Intelligent Systems and Technology, 16(5):1–72","work_id":"20d85be0-14ba-4496-b49b-8ea9b5c4d811","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Agentic ai for scientific discovery: A survey of progress, challenges, and future directions","work_id":"972dac2c-8eda-4685-ab16-906c8d7431a3","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2012,"title":"The adhd-200 consortium: a model to advance the translational potential of neuroimaging in clinical neuroscience.Frontiers in systems neuroscience, 6:62","work_id":"22ebf784-a4b8-44d7-9e16-3273363648d9","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":55,"snapshot_sha256":"87887d949dc98d340e3094005834e5556b18e075626c6987471acda0523777f3","internal_anchors":4},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ab3ece7c498f0b519333d2dd7b40bc0ecc461f21a08380cdf70e89d6c00c2751"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}