Task Transfer Learning for EEG Classification in Motor Imagery-Based BCI System

Table 1

The average classification accuracy and accuracy range of three algorithms on the target dataset.

	Average accuracy (max-min)
	Transfer CSP	CSP	PSD
Sub 1	70.13% (79.58%-66.25%)	56.47% (70.83%-48.33%)	49.95% (56.25%-40.42%)
Sub 2	85.18% (91.43%-69.05%)	78.85% (92.38%-63.81%)	56.35% (70.00%-39.05%)
Sub 3	95.93% (98.79%-81.52%)	91.57% (99.39%-74.24%)	74.32% (88.79%-46.06%)
Sub 4	62.16% (78.61%-42.50%)	54.60% (69.72%-42.50%)	49.31% (56.39%-42.22%)
Sub 5	96.29% (99.39%-84.85%)	98.86% (100.00%-96.36%)	74.83% (91.82%-51.52%)