An EEG-ECG Dataset of Stress and Mental Workload during Cognitive Tasks

A modified version of the Montreal Imaging Stress Task (MIST) was employed to investigate stress-induced neural and physiological changes. The task was adapted to suit the experimental protocol and consisted of computer-based arithmetic problems presented under two conditions: a neutral baseline and a stress-inducing condition. A within-subject design was used, with each participant completing both conditions in separate sessions spaced two weeks apart to minimize carryover effects. Each session consisted of seven stages: task familiarization, an eyes-open resting baseline, four arithmetic task levels (AC1–AC4) of increasing difficulty, and a post-task recovery period.

In the neutral condition, the task was designed to assess cognitive workload without inducing stress. After a non-recorded training phase, EEG and ECG signals were recorded during a 5-minute eyes-open baseline while participants fixated on a dot. Participants then completed four arithmetic levels: single-digit addition and subtraction (AC1), small-scale multiplication (AC2), multi-operand expressions (AC3), and expressions including division (AC4). Immediate correctness feedback was provided after each response, with no time constraints or performance pressure. A 5-minute recovery period followed the final task.

The stress condition followed the same structure but incorporated psychosocial stressors, including response time limits, visible countdown indicators, accuracy manipulation to maintain performance below 50%, and evaluative feedback such as “time out” messages. Performance scores were displayed after each task block to increase performance pressure and induce stress.

ECG signals were recorded using bipolar limb leads (Lead I and Lead II) and amplified using an ECG acquisition system. Data were sampled at 1,000 Hz and bandpass filtered between 0.3 and 200 Hz. EEG data were recorded from eight scalp electrodes (Fp1, Fp2, F3, F4, P3, P4, T3, and T4) following the international 10–20 system, referenced to bilateral mastoids. Signals were acquired using an EEG amplifier at a sampling rate of 2,000 Hz, with electrode impedances maintained below 10 kΩ.

ECG signals were further high-pass filtered at 1 Hz and inspected for cardiac abnormalities. EEG preprocessing was conducted using dedicated analysis software, applying a 1 Hz high-pass filter and a 50 Hz notch filter. Artifact rejection excluded segments exceeding ±150 µV, eye-blink artifacts were manually removed based on frontal channels, and clean 2-second EEG segments were extracted for subsequent analysis.