Machine learning and EEG can classify passive viewing of discrete categories of visual stimuli but not the observation of pain

Table 4 Mean sample performance metrics for neutral—pain scenes classification

Metric	Cross validation		Cross-subject validation		Within-subject validation
Metric	Mean	SD	Mean	SD	Mean	SD
Accuracy	0.8038	0.0208	0.2837	0.0358	0.5065	0.0504
AUC	0.8348	0.0234	0.2747	0.0361	0.5123	0.0518
Brier Score	0.1480	0.0093	0.3966	0.0232	0.3044	0.0257
F1 Score	0.8344	0.0151	0.3866	0.0423	0.4798	0.0554
Precision	0.7277	0.0231	0.3379	0.0340	0.4960	0.0473
Recall	0.9788	0.0204	0.4553	0.0635	0.4682	0.0758

Optimal hyperparameters: Number of estimators = 735, Maximum depth = 46, Minimum samples to split = 28, Minimum samples at leaf = 17, Maximum features = sqrt, Bootstrap = False

ISSN: 1471-2202