Table 2.
EMA performance using reduced channel montages.
| 4 channel montage | MSE | r | p-value | 2 channel montage | MSE | r | p-value |
|---|---|---|---|---|---|---|---|
| Fp-T, T-O | 104.2 | 0.934 | 0.815 | Fp-T | 111.5 | 0.928 | 0.427 |
| Fp-C, C-O | 112.3 | 0.898 | 0.079 | T-O | 119.7 | 0.908 | 0.575 |
| Fp-C | 125.6 | 0.896 | 0.025 | ||||
| C-O | 162.6 | 0.829 | 0.016 | ||||
| C-T | 172.8 | 0.789 | 0.013 |
The EMA used all features and applied to recordings that passed strict artefact detection. MSE is the means square error, r is the adjusted correlation coefficient (linear mixed model), electrode locations are Fp – frontal polar, C – central, T – temporal, O – occipital, and montages are symmetric across hemispheres. The p-value is estimated using a Wilcoxon signed rank test for paired data, n = 101; the null hypothesis tested is that the MSE was not different from that estimated using an 8-channel EEG montage.