Table 6.

Mean performances of the predictive models – N170 amplitude.

Face	Dataset	Accuracy %	Sensitivity	Specificity	AUC
Fear	Albeniz vs. Mozart	32.75	0.28	0.38	0.31
	Albeniz vs. Chopin	62.00	0.52	0.73	0.72
Happy	Albeniz vs. Mozart	66.88	0.60	0.74	0.67
	Albeniz vs. Chopin	37.00	0.33	0.41	0.37
Neuter	Albeniz vs. Mozart	63.50	0.68	0.60	0.64
	Albeniz vs. Chopin	49.88	0.54	0.46	0.50

These models highlight that in happy faces ERP amplitude is modulated as in neuter faces (in Albeniz vs. Mozart predictive model).