A Multitasking EEG Dataset for Cognitive Workload Assessment with Controlled Task Difficulty

Cognitive workload impacts attention, memory, and decision-making. We introduce iNCog-EEG (Ideal vs. Noisy Cognitive EEG for Workload Assessment), an EEG dataset from 40 participants performing multitasking activities with varying difficulty levels.

Participants completed four concurrent tasks: Math Problem, Memory Match (N-Back), Object (Visual) Tracking, and Inhibition Task (Click task), alongside a resting baseline. The dataset includes 30 clean recordings and 10 with natural artifacts, allowing the study of both artifact-free and noisy signals. It supports binary cognitive workload classification (No Workload vs.

Workload) and hierarchical classification (Low, Moderate, High workload). Validation through correlation, time-frequency, and topographic analyses shows consistent neural patterns, such as increased frontal theta and decreased posterior alpha activity. Statistical validation using repeated-measures ANOVA confirmed significant neural differences across workload levels, while baseline MLP (Multi-Layer Perceptron) and XGBoost (Extreme Gradient Boosting) models achieved up to 87.48% binary and 82.80% ternary classification accuracy, demonstrating reliable workload separability, even with noisy EEG.

This dataset, freely available, supports cognitive workload classification, signal quality assessment, and adaptive BCI (Brain Computer Interface) development. This research was funded and supported by the Qatar Research, Development and Innovation Council (QRDI Council) through the Undergraduate Research Experience Program (UREP), Grant No. UREP 32-0221-250164; by the 1st Place Award in the Launchpad Track of the Health Tech Hackathon 2025, organized by World Summit AI – Qatar 2025 in partnership with Qatar University; and by the Qatar University Internal Grant (Grant ID: QUCP-CAS-2026-1051).

The authors used AI-based writing assistance tools such as ChatGPT, QuillBot, and Grammarly to improve manuscript clarity and readability. We would like to express our sincere gratitude to all the participants who volunteered their time and effort for this study. Their contributions were invaluable in making this dataset possible.

We also wish to thank the Ethical Review Committee of Apollo Clinic - JMI Specialized Hospital, Dhaka, Bangladesh, for granting ethical approval for the data collection. Department of Electrical & Computer Engineering, Rajshahi University of Engineering & Technology, Rajshahi, 6204, Bangladesh Fariya Bintay Shafi, Md. Fahim Hossen Department of Electrical Engineering, Qatar University, Doha, Qatar Department of Social Sciences, Qatar University, Doha, Qatar Chief Executive Officer & Consultant Surgeon, Apollo Clinic, Dhanmondi (Licensee: JMI Specialized Hospital Ltd.), Dhaka, Bangladesh Department of Civil and Environmental Engineering, Qatar University, Doha, Qatar The authors declare no competing interests.