fig6

Figure 6. (A) Schematic of the DNN; (B) Recall and accuracy of classification prediction using raw data with different numbers of channels; (C) Recall and accuracy of predictions using high-pass-filtered data (HP data), low-pass-filtered data (LP data), and raw data; (D) Recall and accuracy of classification predictions using low-pass-filtered (< 3 Hz) data with different numbers of channels; (E) Recall and (F) accuracy of predictions using low-pass-filtered (< 3 Hz) data with different numbers of channels; (G) Recall and accuracy of predictions for each hand movement; (H) Average response time for different hand-movement speeds; (I) Recall and accuracy of predictions using different hand-movement speeds; (J) Confusion matrix for predictions made with mixed data. The error bars represent the SD obtained from a five-fold cross-validation. DNN: Deep neural network; SD: standard deviation.