Table 1 Dataset 1 long window (\(w=1\,\text{s}\)) and amplitude features grand average (across all participants) accuracy and F1 score of sLDA classification, shown by condition, types of features and types of TFRs calculated with Rényi entropy
From: Detection of motor imagery based on short-term entropy of time–frequency representations
cond | Accuracy (%) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
TFR | – | tfrsp | tfrrsp | tfrgabor | tfrrgab | tfrpwv | tfrrpwv | tfrspwv | tfrrspwv | |
Feature type | Ampl. | Entropy | Entropy | Entropy | Entropy | Entropy | Entropy | Entropy | Entropy | |
Right vs up | VtG | 64.07 | 52.87 | 54.83 | 53.62 | 53.96 | 53.61 | 53.67 | 52.74 | 52.56 |
noVtG | 60.44 | 52.72 | 53.33 | 52.56 | 54.21 | 52.38 | 51.10 | 51.49 | 52.64 | |
Right vs base | VtG | 86.91 | 64.68 | 60.78 | 69.97 | 66.95 | 75.04 | 88.29 | 66.92 | 64.52 |
noVtG | 85.68 | 64.85 | 61.48 | 67.42 | 68.47 | 76.25 | 87.93 | 67.84 | 63.43 | |
Up vs base | VtG | 85.81 | 60.48 | 59.90 | 67.20 | 67.29 | 71.94 | 88.19 | 63.55 | 60.93 |
noVtG | 84.59 | 65.05 | 61.91 | 67.72 | 69.57 | 77.77 | 88.19 | 68.30 | 64.40 | |
F1 score (%) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Right vs up | VtG | 62.83 | 51.08 | 54.41 | 53.35 | 52.95 | 52.15 | 54.20 | 51.28 | 51.57 |
noVtG | 59.04 | 51.96 | 52.69 | 52.28 | 53.62 | 50.92 | 49.78 | 50.88 | 51.94 | |
Right vs base | VtG | 84.19 | 64.44 | 60.35 | 69.87 | 66.54 | 72.15 | 87.30 | 66.50 | 64.23 |
noVtG | 82.52 | 63.60 | 61.31 | 67.34 | 67.55 | 73.35 | 86.69 | 67.21 | 63.86 | |
Up vs base | VtG | 82.30 | 59.76 | 59.49 | 67.50 | 66.47 | 68.59 | 86.93 | 63.17 | 60.47 |
noVtG | 80.55 | 64.14 | 61.84 | 67.87 | 68.50 | 75.03 | 87.06 | 67.40 | 63.84 | |