2019 3rd International Conference on Trends in Electronics and Informatics (ICOEI) · 2019 · 89 citations · 7 references
ComputingBiomedical EngineeringNeuroscienceBiomedical Signal ProcessingMachine LearningTime Series AnalysisDeep LearningDepressionElectroencephalographyLong-term MemoryNeural NetworksTime Series PredictionDeep Learning For Eeg AnalysisDeep Learning For Time Series ForecastingDepression Symptom PredictionElectrophysiological RecordingsLong Short-term Memory NetworksMachine Learning For ElectroencephalogramNeural Time Series ForecastingRecurrent Neural Network Forecasting
Depression, a neurological disorder is the leading cause of disability worldwide. EEG recordings have found wide use in the diagnosis and analysis of various neurological disorders including depression. In this paper, LSTM (Long-short term memory) deep learning models are used in the prediction of trends of depression for the next time instants, based on the features extracted. The statistical time-domain feature encompassing the mean of amplitude of the data is extracted employing moving window segmentation from the acquired EEG signals. The model uses one LSTM layer with 10 hidden neurons for the prediction. Out of a total of 7000 mean values calculated from a sample of 30 patient records from each resting states, 5600 sample means were used to train the model. The proposed LSTM network could predict the next 1400 sample mean values accurately with root mean square error of 0.000064. The performance of the model is compared with CNN-LSTM and ConvLSTM. It is observed that LSTM predictor model works best for the prediction of trends of depression.
7
Sepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 98.1K citations
A Novel Depression Diagnosis Index Using Nonlinear Features in EEG Signals
U. Rajendra Acharya, Vidya K. Sudarshan, Hojjat Adeli et al. · European Neurology · 2015 · 275 citations
Computing, Neuroscience, Psychiatry +17