EPIStress: A multimodal dataset of Physiological signals to measure cognitive stress in epilepsy patients

The EPIStress dataset is curated from twenty epilepsy patients undergoing cognitive stress elicitation experiments in laboratory settings and in situ settings. The patients wore two wearable devices during the experiments: Muse S EEG headband and Empatica E4 wrist band. Consequently, electroencephalography (EEG), photoplethysmography (PPG), acceleration (ACC), electrodermal activity (EDA), and temperature (TEMP) sensor data were recorded to assess the physiological response during cognitively stressful and relaxed tasks. The laboratory tasks include mental arithmetic, Stroop, N-Back, and Sudoku at predefined difficulty levels, and the in-situ experiments include independently choosing tasks with at least two difficulty levels. Each task in both environments is self-rated using two 5-point Likert scales of mental workload and stress, and a pairwise NASA-TLX questionnaire. The dataset also provides a patient-reported list of seizure-provoking and -inhibiting factors. The dataset has the potential to explore the objective relationship between epileptic seizure and stress, as indicated in the subjective reports from patients found in existing literature. A dataset descriptor paper is under review and will be made available once published for a detailed description of the dataset.