A Novel Deep Neural Network for Robust Detection of Seizures Using EEG Signals

Table 2

Configurations of 8 models using 10-fold cross-validation for the A vs. B vs. C vs. D vs. E cases.

Block	Parameter	M1	M2	M3	M4	M5	M6	M7	M8
Conv 1	Number of kernels	20	20	20	20	20	20	20	20
	Size of receptive field	5	5	5	5	40	40	40	40
	Dropout rate	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2
Conv 2	Number of kernels	40	40	40	40	40	40	40	40
	Size of receptive field	3	3	3	3	20	20	20	20
	Dropout rate	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2
Conv 3	Number of kernels	80	80	80	80	80	80	80	80
	Size of receptive field	3	3	3	3	10	10	10	10
	Dropout rate	0.2	0.2	0.2	0.2	0.2	0.2	0.2	0.2
FC 1	Number of neurons	32	32	64	64	32	32	64	64
	Dropout rate	0.2	0.5	0.2	0.5	0.2	0.5	0.2	0.5
FC 2	Number of neurons	16	16	32	32	16	16	32	32
	Dropout rate	0.2	0.5	0.2	0.5	0.2	0.5	0.2	0.5
A vs. B vs. C vs. D vs. E	Acc	90.47	88.61	91.89	91.20	93.37	91.62	93.55	92.92
	Sen	75.98	70.86	79.66	77.79	83.33	78.85	83.73	82.30
	Spe	94.00	92.72	94.92	94.45	95.83	94.71	95.93	95.57