Wearable Device Dataset from Induced Stress and Structured Exercise Sessions
Bosch, Facundo Prada, Lara Bonomini, Paula
Bosch, Facundo
Prada, Lara
Bonomini, Paula
Type
Artículo de publicación periódica
Abstract
This dataset comprises physiological signals recorded via a wearable device (Empatica E4) during structured acute stress induction and both aerobic and anaerobic exercise sessions. Collected metrics include blood volume pulse (BVP), accelerometer-based activity, skin temperature, and electrodermal activity, along with self-reported stress levels. The stress protocol combines math and emotional tasks with rest intervals, while the exercise sessions involve defined cycling routines for aerobic and anaerobic conditions. The dataset includes recordings from 36 healthy volunteers during stress sessions, 30 participants for aerobic exercise, and 31 for anaerobic protocols. The data are organized into stress, aerobic, and anaerobic categories, with raw signal files (e.g., TEMP, EDA, BVP, ACC, IBI, HR) and tags for segmentation. Demographic information (age, weight, height) and self-reported stress scores are also included. Some limitations such as incomplete sessions or signal artifacts are documented. This resource is intended for research in stress and exercise detection, classification, and physiological signal processing, facilitating the development of machine learning models to distinguish among stress, aerobic activity, and anaerobic activity from noninvasive wearable sensor data.
Description
Date
2025-06-24
Journal Title
Journal ISSN
Volume Title
Publisher
PhysioNet
Collections
Research Projects
Organizational Units
Journal Issue
Keywords
WEARABLE SIGNALS, ACUTE STRESS, AEROBIC EXERCISE, ANAEROBIC EXERCISE, PHYSIOLOGICAL MONITORING, EMPATICA E4, SIGNAL PROCESSING, DATASET, NON-INVASIVE SENSING, MACHINE LEARNING