Spatial and Time Domain Feature of ERP Speller System Extracted via Convolutional Neural Network

Figure 7

Feature map of L₂ of H subject. Format is the same as Figure 5. High positive weight around P300 and P500 range were found in (a), (b), (c), and (d). (e)–(i) Moderate positive weight around P700 were also found.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)