Brain-Heart Connection for Psychological Health

Abstract

With growing stress in our lives, there has been a major impact on our hearts and brain, leading to several mental and heart-related problems that sometimes are fatal and so severe that they leave a spot on us for the rest of our lives. To see the effect of the stress level on our brain and heart, we have designed a model to analyze the EEG (Electroencephalography) and the ECG (Electrocardiography) signals. Our study aimed to correlate two biosignals, EEG and ECG. We have focused on two important emotions, namely valence and arousal. We used the DREAMER dataset, which consists of both ECG and EEG signals. We evaluated nine different classifiers, including nearest neighbors, linear SVM, RBF SVM, Gaussian process, Decision tree, Random-forest, Neural net, AdaBoost, and Naive Bayes. We found AdaBoost had the best mean accuracy (97%) but with the longest processing time of around 5-10 milliseconds, whereas other classifiers had a mean runtime of around 1 millisecond. We discuss three things: preprocessing and feature extraction of the dataset, evaluation of classifiers for arousal and valence, and data visualization for correlation of arousal and valence values for all the extracted features.

Keywords

• electroencephalography (eeg)
• electrocardiography (ecg)
• dreamer dataset
• classifiers