Deep Learning with ConvNet Predicts Imagery Tasks Through EEG

Deep learning with convolutional neural networks (ConvNets) has dramatically improved the capabilities of computer vision applications just through considering raw data without any prior feature extraction. Nowadays, there is a rising curiosity in interpreting and analyzing electroencephalography (EEG) dynamics ConvNets. Our study focused on ConvNets different structures, efficiency multiple machine algorithms optimization ConvNets, constructing for predicting imagined left right movements subject-independent basis EEG data. We adapted novel lower-upper triangularization based extreme machines (LuELM) to ConvNet architecture. Results showed that recently advanced methods field, i.e. adaptive moments batch normalization together dropout strategy, ability, outperforming conventional fully-connected widely-used spectral features. The proposed prediction model achieved improvements classification performances rates 90.33%, 91.00%, 89.67% accuracy, recall, specificity, respectively.