Dataset on Emotional Dysregulation in Depression

This dataset contains electroencephalography (EEG) recordings collected from 15 adults diagnosed with severe Major Depressive Disorder (MDD) (HDRS-17 > 20). Participants viewed standardised emotional images from the International Affective Picture System (IAPS) across three affective categories: Low Valence High Arousal (LVHA) (Class 0), Neutral (Class 1), and High Valence Low Arousal (HVHA) (Class 2). Each stimulus block consisted of 12 images (10 seconds each), followed by a one-minute resting period, with continuous EEG recorded throughout. Baseline data were also collected for two minutes at the start of each session.EEG was recorded using a 19-channel RMS EEG-32 SuperSpec system (256 Hz), following the international 10–20 electrode placement system. Signals were preprocessed using smoothing filters, Chebyshev band-pass filtering, and wavelet-based denoising to remove ocular and muscular artefacts. The dataset includes preprocessed EEG data as well as features derived across the five major frequency bands (delta, theta, alpha, beta, gamma).From the processed EEG, 966 features were extracted, including: Relative spectral power (570 features), Frontal asymmetry indices (54 features), and Band ratios such as theta/alpha and gamma/alpha (342 features).These features were organised into three machine-learning pipelines: Stimulus classification (LVHA, Neutral, HVHA), Rest-state classification (rest 1, rest 2, rest 3), and Paired rest vs stimulus classification.Models including Decision Trees, Random Forests, XGBoost, and Logistic Regression were trained using these features, with SHAP-based interpretability used to identify the most important neural markers. Findings highlight distinctive roles of frontal asymmetry, parietal alpha/theta power, temporal gamma activity, and functional connectivity patterns in depression.This dataset supports research in affective neuroscience, EEG biomarkers, mental-health diagnostics, and machine-learning–based classification of emotional states in psychiatric populations.