Table 1.
Classification accuracy as a function of channel number—VGG16 with symlet6 wavelet.
| Channel Number (Name) | 1 (Fp1) | 2 (F7) | 5 (Fp2) | 6 (F8) | 9 (F3) | 13 (F4) |
|---|---|---|---|---|---|---|
| Model precision—resting state | 94.62% (528/558) | 93.15% (517/555) | 92.02% (531/577) | 94.33% (516/547) | 98.62% (571/579) | 99.65 (576/578) |
| Model precision—picking (Level-K = 0) | 59% (70/118) | 51.63% (63/122) | 57.25% (75/131) | 53.90% (83/154) | 65.89% (85/129) | 64.06% (82/128) |
| Model precision—coordination (Level-K > 0) | 56.10% (92/164) | 52.15% (85/163) | 61.36% (81/132) | 56.11% (78/139) | 68.18% (90/132) | 64.93% (87/134) |
| Total model accuracy | 82.14% (690/840) | 79.16% (665/840) | 82.14% (690/840) | 80.59% (677/840) | 88.81% (746/840) | 88.69% (745/840) |