{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:I6U63OQ4J6PLBF3X4BB6KSWTGF","short_pith_number":"pith:I6U63OQ4","schema_version":"1.0","canonical_sha256":"47a9edba1c4f9eb09777e043e54ad33164bc851738a0bd82f69a52881901078e","source":{"kind":"arxiv","id":"2605.16739","version":1},"attestation_state":"computed","paper":{"title":"EmoMind: Decoding Affective Captions from Human Brain fMRI","license":"http://creativecommons.org/licenses/by/4.0/","headline":"EmoMind decodes continuous 34-dimensional affect from fMRI to rewrite neutral scene descriptions into subject-specific affective captions.","cross_cats":["cs.AI","cs.CL","q-bio.NC"],"primary_cat":"cs.LG","authors_text":"Bilal A. Mohammed, Lin Gu, Ruogo Fang","submitted_at":"2026-05-16T01:32:45Z","abstract_excerpt":"Decoding visual experience from brain activity has advanced substantially, but cur- rent brain-to-text systems largely recover semantic content while discarding affect. Additionally, language models can generate emotional text when prompted with categorical labels, but such labels collapse rich inter-subject variability into coarse discrete bins. We present EmoMind, the first end-to-end pipeline for decoding affective captions directly from fMRI signals. EmoMind first retrieves a semanti- cally grounded neutral scene description from brain-decoded visual features, then rewrites it using a cont"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.16739","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-16T01:32:45Z","cross_cats_sorted":["cs.AI","cs.CL","q-bio.NC"],"title_canon_sha256":"46921164230298f3f5de006dea355fd3bd2c3ecdd70953d9ce67eb2f337af9d5","abstract_canon_sha256":"9ba21d3b72a17b93265e9867fbea60ff02e92d4669667d300d99af90a6d49151"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:02:39.184335Z","signature_b64":"Qs0K3SEtf6/mIUjjYWzfqsp04qZFwno18E/9q5Hf1UoBYQCXAE4hQowGv2iQUNqc8GfeNBZFcqSrt9bn6FaVCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"47a9edba1c4f9eb09777e043e54ad33164bc851738a0bd82f69a52881901078e","last_reissued_at":"2026-05-20T00:02:39.183476Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:02:39.183476Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EmoMind: Decoding Affective Captions from Human Brain fMRI","license":"http://creativecommons.org/licenses/by/4.0/","headline":"EmoMind decodes continuous 34-dimensional affect from fMRI to rewrite neutral scene descriptions into subject-specific affective captions.","cross_cats":["cs.AI","cs.CL","q-bio.NC"],"primary_cat":"cs.LG","authors_text":"Bilal A. Mohammed, Lin Gu, Ruogo Fang","submitted_at":"2026-05-16T01:32:45Z","abstract_excerpt":"Decoding visual experience from brain activity has advanced substantially, but cur- rent brain-to-text systems largely recover semantic content while discarding affect. Additionally, language models can generate emotional text when prompted with categorical labels, but such labels collapse rich inter-subject variability into coarse discrete bins. We present EmoMind, the first end-to-end pipeline for decoding affective captions directly from fMRI signals. EmoMind first retrieves a semanti- cally grounded neutral scene description from brain-decoded visual features, then rewrites it using a cont"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"EmoMind significantly outperforms label-prompted GPT-4 on all three axes (subject-specificity, structural geometry, causal control) across two independent emotion fMRI datasets, with largest gains on metrics requiring person-specific affective structure.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That a continuous 34-dimensional emotion vector decoded from fMRI accurately captures rich inter-subject affective variability and that classifier-free guidance against an identity-preserving null branch enables controllable interpolation between semantic fidelity and affective expressivity without introducing artifacts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"EmoMind is the first end-to-end pipeline that decodes continuous affective captions from fMRI by combining brain-decoded visual features with a 34D emotion vector and classifier-free guidance to balance semantic fidelity and affective expressivity.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"EmoMind decodes continuous 34-dimensional affect from fMRI to rewrite neutral scene descriptions into subject-specific affective captions.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6c14daa33544d026954e25ac3394b110560549ba4f27b7c3f4cea554f55fcb1e"},"source":{"id":"2605.16739","kind":"arxiv","version":1},"verdict":{"id":"d7267ac4-cec5-459d-8a26-c9b83ca976af","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T20:19:59.205050Z","strongest_claim":"EmoMind significantly outperforms label-prompted GPT-4 on all three axes (subject-specificity, structural geometry, causal control) across two independent emotion fMRI datasets, with largest gains on metrics requiring person-specific affective structure.","one_line_summary":"EmoMind is the first end-to-end pipeline that decodes continuous affective captions from fMRI by combining brain-decoded visual features with a 34D emotion vector and classifier-free guidance to balance semantic fidelity and affective expressivity.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That a continuous 34-dimensional emotion vector decoded from fMRI accurately captures rich inter-subject affective variability and that classifier-free guidance against an identity-preserving null branch enables controllable interpolation between semantic fidelity and affective expressivity without introducing artifacts.","pith_extraction_headline":"EmoMind decodes continuous 34-dimensional affect from fMRI to rewrite neutral scene descriptions into subject-specific affective captions."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16739/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_compliance","ran_at":"2026-05-19T20:32:04.369953Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T20:31:19.166046Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T19:01:56.335849Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.464689Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"ff9f713620c4c9912b3f89a19716c4153dbea62d7573492927ceda93a320fb8a"},"references":{"count":35,"sample":[{"doi":"","year":2024,"title":"Mind captioning: Evolving descriptive text of mental content from human brain activity.Science Advances, 2024","work_id":"68b3d936-f9df-4ae9-9e82-c886df6e8eaa","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Semantic reconstruction of continuous language from non-invasive brain recordings.Nature Neuroscience, 26(5):858–866, 2023","work_id":"57ce806d-bf09-46c5-8449-044b65fb20ce","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Reconstructing the mind’s eye: fMRI-to-image with contrastive learning and diffusion priors.Advances in Neural Information Processing Systems (NeurIPS), 36, 2024","work_id":"6c034156-669f-47dc-8dd7-135008f503ed","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1909,"title":"CTRL: A Conditional Transformer Language Model for Controllable Generation","work_id":"c90b18cd-644d-4503-8eca-953249853cff","ref_index":4,"cited_arxiv_id":"1909.05858","is_internal_anchor":true},{"doi":"","year":2020,"title":"Plug and play language models: A simple approach to controlled text generation","work_id":"8101356d-7ad3-420c-b273-dc666d382ad0","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":35,"snapshot_sha256":"00d35730472895fbd3d3ac6d5d0fd91e5d5d266e9f03723f6217ecee3ac23660","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"a3bb418f69026f8ec16347bcf308de5d64ba847ccff7eb85f1ff7736d9017b22"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.16739","created_at":"2026-05-20T00:02:39.183608+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.16739v1","created_at":"2026-05-20T00:02:39.183608+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16739","created_at":"2026-05-20T00:02:39.183608+00:00"},{"alias_kind":"pith_short_12","alias_value":"I6U63OQ4J6PL","created_at":"2026-05-20T00:02:39.183608+00:00"},{"alias_kind":"pith_short_16","alias_value":"I6U63OQ4J6PLBF3X","created_at":"2026-05-20T00:02:39.183608+00:00"},{"alias_kind":"pith_short_8","alias_value":"I6U63OQ4","created_at":"2026-05-20T00:02:39.183608+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF","json":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF.json","graph_json":"https://pith.science/api/pith-number/I6U63OQ4J6PLBF3X4BB6KSWTGF/graph.json","events_json":"https://pith.science/api/pith-number/I6U63OQ4J6PLBF3X4BB6KSWTGF/events.json","paper":"https://pith.science/paper/I6U63OQ4"},"agent_actions":{"view_html":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF","download_json":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF.json","view_paper":"https://pith.science/paper/I6U63OQ4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.16739&json=true","fetch_graph":"https://pith.science/api/pith-number/I6U63OQ4J6PLBF3X4BB6KSWTGF/graph.json","fetch_events":"https://pith.science/api/pith-number/I6U63OQ4J6PLBF3X4BB6KSWTGF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF/action/storage_attestation","attest_author":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF/action/author_attestation","sign_citation":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF/action/citation_signature","submit_replication":"https://pith.science/pith/I6U63OQ4J6PLBF3X4BB6KSWTGF/action/replication_record"}},"created_at":"2026-05-20T00:02:39.183608+00:00","updated_at":"2026-05-20T00:02:39.183608+00:00"}