Using CNN Saliency Maps and EEG Modulation Spectra for Improved and More Interpretable Machine Learning-Based Alzheimer’s Disease Diagnosis

Table 8

Age prediction accuracy comparison against a random regression model for experiment 1 and experiment 4 settings.

Tasks	Exp. 1		Exp. 4
	Nonrandom	Random	Nonrandom	Random
T1	10.33  2.02	9.94  1.55	9.82  1.65	9.71  1.75
T2	9.83  1.85	9.74  1.45	9.40  1.69	9.48  1.65
T3	8.05  1.20	8.55  1.55	7.81  1.18	8.19  1.36
T4	9.74  2.77	9.30  2.42	9.70  2.64	9.26  2.63
T5	12.19  2.63	11.97  2.14	11.28  1.82	11.63  2.46