Deep Learning Approach for Automatic Classification of Ocular and Cardiac Artifacts in MEG Data

Figure 5

(a) Topoplots showing spatial ICA components extracted from one sample recording in the validation dataset. (b) The corresponding temporal ICA components extracted from the same data. Components identified as artifacts are highlighted in red. ECG (blue) and EOG (green) signals are shown for comparison. (c) The corresponding classification results provided by the reference and the model. Reference values include CTPS for ECG in blue and Pearson’s correlation coefficients for EOG in green. The model probability (scores) is shown in orange. Thresholds used are indicated by dashed lines.