{"id":"W3119151570","doi":"10.1016/j.neuroimage.2021.117756","title":"Retrospective motion artifact correction of structural MRI images using deep learning improves the quality of cortical surface reconstructions","year":2021,"lang":"en","type":"article","venue":"NeuroImage","topic":"Advanced MRI Techniques and Applications","field":"Medicine","cited_by":97,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Institute of Biomedical Imaging and Bioengineering; National Institute on Aging; Canadian Institutes of Health Research; Janssen Research and Development; National Institutes of Health; Genentech; IXICO; H. Lundbeck A/S; Servier; GE Healthcare; Fujirebio US; Roche; University of Southern California; U.S. Department of Defense; Alzheimer's Disease Neuroimaging Initiative; BrightFocus Foundation; Pfizer; Michael J. Fox Foundation for Parkinson's Research; Northern California Institute for Research and Education; Johnson and Johnson; AbbVie; Merck","keywords":"Artificial intelligence; Computer science; Artifact (error); Image quality; Convolutional neural network; Neuroimaging; Computer vision; Motion (physics); Pattern recognition (psychology); Image (mathematics); Neuroscience; Psychology","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001494766,0.00009249248,0.0002386198,0.00002644958,0.0001618529,0.000009923281,0.00004289142,0.00004868201,0.00005682081],"category_scores_gemma":[0.0005138332,0.00007494924,0.0001061888,0.0002844291,0.0002325697,0.0001033502,0.00003945398,0.0003900194,8.031187e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005394403,"about_ca_system_score_gemma":0.0000416404,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006574992,"about_ca_topic_score_gemma":0.000009753369,"domain_scores_codex":[0.9989998,0.0001290265,0.0003571846,0.0002290494,0.0001645265,0.0001203529],"domain_scores_gemma":[0.9989536,0.0001202484,0.0002601602,0.0002856904,0.0003428224,0.0000375175],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","study_design_scores_codex":[0.0000348472,0.00003647686,0.02506224,0.00002232874,0.000009125333,0.000001881423,0.00007850958,0.001610561,0.966143,0.0004897407,0.000009735524,0.006501534],"study_design_scores_gemma":[0.0002154044,0.00009583303,0.5266711,0.00002412325,0.00006716164,0.0001193212,0.0005269552,0.0229568,0.4484828,0.0007384382,0.00003251981,0.00006952736],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.847718,0.00004549829,0.15103,0.0002037902,0.0001366958,0.0002113273,0.00001067279,0.00005908882,0.0005848666],"genre_scores_gemma":[0.9704728,0.00004596534,0.02924681,0.0000199918,0.00003791848,0.000003673321,0.00001237365,0.00001375349,0.0001467437],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5176603,"threshold_uncertainty_score":0.3056341,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03292777352009419,"score_gpt":0.3553017881469739,"score_spread":0.3223740146268797,"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."}}