{"id":"W3132764260","doi":"10.1016/j.jneumeth.2021.109098","title":"Grad-CAM helps interpret the deep learning models trained to classify multiple sclerosis types using clinical brain magnetic resonance imaging","year":2021,"lang":"en","type":"article","venue":"Journal of Neuroscience Methods","topic":"Multiple Sclerosis Research Studies","field":"Medicine","cited_by":214,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Calgary","funders":"Canadian Institutes of Health Research; Natural Sciences and Engineering Research Council of Canada; Multiple Sclerosis Society of Canada","keywords":"Interpretability; Artificial intelligence; Convolutional neural network; Discriminative model; Pattern recognition (psychology); Computer science; Voxel; Deep learning; Neuroimaging; Functional magnetic resonance imaging; Magnetic resonance imaging; Generalizability theory; Machine learning; Neuroscience; Medicine; Psychology","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":[0.0009664863,0.001664513,0.0006014062,0.0009789468,0.0004416434,0.001516432,0.001537647,0.00171807,0.01831731],"category_scores_gemma":[0.005282025,0.0004756163,0.0009590974,0.0004370142,0.0002804649,0.001182246,0.001099636,0.001906892,0.005353293],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000610462,"about_ca_system_score_gemma":0.001133588,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009484208,"about_ca_topic_score_gemma":0.01932414,"domain_scores_codex":[0.9998323,0.00003174505,0.00001280864,0.00006417882,0.00002817889,0.0000308523],"domain_scores_gemma":[0.9991885,0.0003505312,0.00005101671,0.0001175811,0.0002444047,0.00004802578],"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":[0.0008501893,0.0004307594,0.02263306,0.0005037447,0.0003259534,0.0005274959,0.0002624903,0.09148648,0.01657963,0.007134656,0.1356442,0.7236213],"study_design_scores_gemma":[0.00007458389,0.0001102861,0.003649469,0.0001671201,0.00006952539,0.0001688915,0.0000968708,0.9526826,0.01380607,0.01345695,0.01568607,0.00003159002],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1376348,0.002305465,0.7032382,0.003466159,0.001795735,0.0003962819,0.01382925,0.1217962,0.01553775],"genre_scores_gemma":[0.6067653,0.0007540772,0.3582393,0.002109126,0.0004342161,0.0004158705,0.01165477,0.004730793,0.01489666],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01831731,"threshold_uncertainty_score":0.06127751,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.343061916807897,"score_gpt":0.4768399233011524,"score_spread":0.1337780064932554,"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."}}