Table 2 Subject-specific classification accuracy (in percentage %) and Kappa of state-of-the-art algorithms on the BCI IV-2a dataset.
From: CTNet: a convolutional transformer network for EEG-based motor imagery classification
|  | A01 | A02 | A03 | A04 | A05 | A06 | A07 | A08 | A09 | Average ± Std | p-value | Kappa |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
ShallowConvNet+ 27 | 82.64 | 55.21 | 92.01 | 74.31 | 72.92 | 59.72 | 81.60 | 83.33 | 79.51 | 75.69 ± 11.76 | 0.0020** | 0.6759 |
DeepConvNet+ 27 | 82.29 | 44.79 | 90.63 | 76.04 | 77.43 | 68.06 | 92.01 | 83.33 | 85.42 | 77.78 ± 14.42 | 0.1934 | 0.7037 |
EEGNet+ 28 | 88.19 | 56.94 | 93.06 | 71.18 | 70.49 | 62.85 | 87.15 | 82.64 | 84.03 | 77.39 ± 12.47 | 0.0059** | 0.6986 |
TSF-STAN37 | 88.3 | 81.7 | 92.2 | 77.6 | 63.3 | 67.5 | 90.0 | 95.0 | 91.7 | 83.0 ± 11.4 | 0.7695 | 0.765 |
Conformer+ 53 | 85.07 | 48.96 | 91.32 | 78.47 | 75.00 | 65.28 | 87.85 | 87.15 | 79.86 | 77.66 ± 13.35 | 0.0742 | 0.7022 |
MI-CAT24 | 90.62 | 54.51 | 91.32 | 72.57 | 63.19 | 62.85 | 87.15 | 85.07 | 84.03 | 76.81 ± 13.80 | 0.0273* | 0.692 |
CTNet (Proposed) | 90.97 | 73.61 | 96.53 | 84.72 | 77.08 | 64.24 | 86.11 | 84.38 | 85.07 | 82.52 ± 9.61 | - | 0.7670 |
