{"id":"W4389829528","doi":"10.31234/osf.io/srmx7","title":"Challenges in Multi-Task Learning for fMRI-Based Diagnosis: Benefits for Psychiatric Conditions and CNVs Would Likely Require Thousands of Patients","year":2023,"lang":"en","type":"preprint","venue":"","topic":"Functional Brain Connectivity Studies","field":"Neuroscience","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; Institut universitaire en santé mentale de Montréal; Institut Universitaire en Santé Mentale de Québec; Université de Montréal; Institut Universitaire de Gériatrie de Montréal; Centre Hospitalier Universitaire Sainte-Justine","funders":"National Institute of Mental Health; Compute Canada; Canadian Institutes of Health Research; National Institutes of Health; Canada First Research Excellence Fund; Simons Foundation Autism Research Initiative; Health and Care Research Wales; Courtois Foundation; Consortium canadien en neurodégénérescence associée au vieillissement; Wellcome Trust; Institut de Valorisation des Données; Fondation Brain Canada","keywords":"Machine learning; Medical diagnosis; Artificial intelligence; Computer science; Context (archaeology); Biobank; Robustness (evolution); Task (project management); Sample size determination; Benchmark (surveying); Copy-number variation; Psychology; Data science; Cognitive psychology; Medicine; Statistics; Bioinformatics; Cartography; Biology; Mathematics","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01960032,0.001976008,0.002832138,0.0009554187,0.001639281,0.002067765,0.003040374,0.003491642,0.004478476],"category_scores_gemma":[0.04501303,0.0008955623,0.001737754,0.00149602,0.001856481,0.003450261,0.003802852,0.005118608,0.004118553],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001386062,"about_ca_system_score_gemma":0.0022088,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00684102,"about_ca_topic_score_gemma":0.008925238,"domain_scores_codex":[0.9919606,0.005322852,0.0003515353,0.001459755,0.0006429819,0.0002622853],"domain_scores_gemma":[0.9633887,0.02605953,0.001036232,0.005380712,0.002578217,0.001556562],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.006715909,0.002714969,0.06681383,0.001514574,0.001964637,0.001066429,0.0009369308,0.213183,0.01660607,0.007267171,0.04685392,0.6343626],"study_design_scores_gemma":[0.0005264744,0.001270805,0.01965982,0.0001769118,0.0002114068,0.000818717,0.0006156042,0.9011031,0.00567323,0.05948148,0.01032787,0.0001344987],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3897673,0.01557095,0.5385526,0.03211285,0.001569347,0.00102741,0.005208467,0.007980351,0.00821072],"genre_scores_gemma":[0.7576676,0.001571119,0.2267892,0.003564015,0.0007617741,0.0008388875,0.005093787,0.0004047683,0.003308833],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01960032,"threshold_uncertainty_score":0.1036576,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2216063752147581,"score_gpt":0.3504629288367373,"score_spread":0.1288565536219792,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}