Table 1.

Means and standard deviations of decoding accuracy across participants for each tested configuration.

	Comparison			Accuracy (%)
	Class A	Class B	(Classes A+B)/2	Class A	Class B	p-value
Intra-block leave-one-trial-out
Passive elicitation	Positive	Neutral	85.41 ± 14.43	84.08 ± 18.25	86.73 ± 13.13	<0.001
	Negative	Neutral	81.53 ± 16.05	78.37 ± 21.92	84.69 ± 13.86	<0.001
	Positive	Negative	58.37 ± 15.86	57.14 ± 20.00	59.59 ± 19.36	0.007
Active elicitation	Positive	Neutral	89.80 ± 13.84	88.16 ± 16.29	91.43 ± 13.23	<0.001
	Negative	Neutral	91.53 ± 13.04	89.80 ± 17.85	93.27 ± 10.49	<0.001
	Positive	Negative	45.10 ± 21.32	44.08 ± 23.09	46.12 ± 24.90	1.000
Inter-block cross-validation
Passive x Active elicitation	Positive	Neutral	56.94 ± 21.77	50.20 ± 26.18	63.67 ± 24.55	0.364
	Negative	Neutral	54.69 ± 22.72	44.90 ± 29.59	64.49 ± 25.34	1.000
	Positive	Negative	50.82 ± 13.82	53.47 ± 34.73	48.16 ± 32.89	1.000
Active x Passive elicitation	Positive	Neutral	50.10 ± 16.19	42.04 ± 25.25	58.16 ± 18.56	1.000
	Negative	Neutral	55.92 ± 18.33	50.61 ± 29.75	61.22 ± 19.43	0.341
	Positive	Negative	50.41 ± 14.28	44.08 ± 34.39	56.73 ± 33.75	1.000
Inter-block leave-one-trial-out
Neutral			83.27 ± 12.19	84.90 ± 13.09	81.63 ± 14.63	<0.001
Negative	Active elicitation	Passive elicitation	72.86 ± 17.91	70.61 ± 24.19	75.10 ± 18.50	<0.001
Positive			75.71 ± 17.68	77.55 ± 19.85	73.88 ± 20.90	<0.001

The first and third configurations were analyzed using the leave-one-trial-out (LOTO) method (intra-block and inter-block LOTO, respectively), while the second configuration (inter-block cross-validation) was trained using data from either the active or passive elicitation condition and tested with data from the remaining condition. Significant accuracies (p < 0.05) are highlighted in bold.