Figure 5
Classification results for the down-sampled raw time-series EEGs. (a) The inputs were classified with a fully connected neural network (FCNN), recurrent neural network (RNN) and convolutional neural network (CNN) for 1D input. (b) False positive (FP) numbers for the FCNN, RNN, and CNN. ***p  < 0.001 vs. FCNN, ###p  < 0.001 vs. RNN. (c) False negative (FN) numbers for the FCNN, RNN, and CNN. (d) The receiver operating characteristics (ROC) curve for the classification result of the FCNN. (e) The ROC curve for the classification result of the RNN. (f) The ROC curve for the classification result of the 1D CNN. The area under the curve (AUC) is presented for each ROC curve.
