Deep learning based treatment remission prediction to transcranial direct current stimulation in bipolar depression using EEG power spectral density

Bipolar disorder is characterized by marked changes in mood and activity levels and is a leading cause of disability worldwide. We sought to investigate the application of deep learning methods to electroencephalogram (EEG) signals to predict clinical remission after 6 weeks of home-based transcranial direct current stimulation (tDCS) treatment. Pre-treatment resting-state EEG acquired from 21 bipolar participants was used for this work. A hybrid 1DCNN and GRU model, with input from power spectral density values of theta, beta and gamma frequency bands of the AF7 and TP10 electrodes, achieved a treatment remission prediction accuracy of 78.5% (sensitivity 81.4%, specificity 74.64%).