Back

Spatio Temporal Attentional EEGNet: An Enhanced Deep Learning Model for Cognitive Workload Detection

Ghosh, S.

2025-09-30 bioengineering
10.1101/2025.09.26.678835 bioRxiv
Show abstract

Deep learning has emerged as a powerful tool for extracting meaningful patterns from electroencephalography (EEG) signals, particularly for mental workload recognition in cognitive neuroscience. This study proposes STA-EEGNet (Spatio Temporal Attentional EEGNet), an enhanced variant of the EEGNet architecture that integrates spatio temporal attention mechanisms to improve EEG signal decoding. The model is trained and evaluated on the STEW: Simultaneous Task EEG Workload Dataset, demonstrating superior classification accuracy compared to existing approaches. Heatmap analyses of key convolutional filters in TSA-EEGNet network during testing reveal distinct spatiotemporal activation patterns associated with varying workload levels. High workload is characterized by increased activation in discriminative filters and reduced or inverted responses in suppressive filters, primarily over frontocentral and parietal regions. These findings highlight TSA-EEGNets ability not only to achieve state-of-the-art performance but also to provide interpretable insights into the neural dynamics underlying cognitive workload.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

1
IEEE Transactions on Neural Systems and Rehabilitation Engineering
● 49 papers in training set
Top 0.1%
18.2%
Similar papers in this journal
2
Biomedical Signal Processing and Control
○ 22 papers in training set
Top 0.1%
12.4%
Similar papers in this journal
3
Journal of Neural Engineering
○ 221 papers in training set
Top 0.5%
7.1%
Similar papers in this journal
4
PLOS ONE
● 5266 papers in training set
Top 22%
7.1%
Similar papers in this journal
5
IEEE Transactions on Biomedical Engineering
● 40 papers in training set
Top 0.1%
6.6%
Similar papers in this journal
50% of probability mass above
6
Sensors
○ 43 papers in training set
Top 0.2%
5.4%
7
IEEE Access
● 35 papers in training set
Top 0.2%
4.4%
8
Scientific Reports
○ 3612 papers in training set
Top 26%
4.0%
9
Imaging Neuroscience
● 282 papers in training set
Top 2%
3.2%
10
NeuroImage
○ 903 papers in training set
Top 4%
3.2%
11
Frontiers in Neuroscience
○ 256 papers in training set
Top 2%
2.3%
12
IEEE Journal of Biomedical and Health Informatics
● 37 papers in training set
Top 0.6%
1.9%
13
Neuroinformatics
○ 46 papers in training set
Top 0.4%
1.7%
14
Frontiers in Human Neuroscience
○ 77 papers in training set
Top 1%
1.7%
15
PLOS Computational Biology
● 1863 papers in training set
Top 15%
1.7%
16
Computers in Biology and Medicine
○ 128 papers in training set
Top 3%
1.4%
17
Computer Methods and Programs in Biomedicine
○ 28 papers in training set
Top 0.8%
1.1%
18
Neurocomputing
○ 13 papers in training set
Top 0.2%
1.1%
19
Human Brain Mapping
○ 329 papers in training set
Top 3%
1.1%
20
Physiological Measurement
○ 14 papers in training set
Top 0.4%
1.0%
21
Heliyon
○ 152 papers in training set
Top 7%
1.0%
22
Scientific Data
○ 209 papers in training set
Top 3%
0.8%
23
eneuro
● 439 papers in training set
Top 8%
0.8%
24
Brain Sciences
○ 55 papers in training set
Top 2%
0.6%
25
Journal of Neuroscience Methods
○ 122 papers in training set
Top 2%
0.6%
26
Bioengineering
○ 29 papers in training set
Top 1%
0.6%
27
Frontiers in Neuroinformatics
○ 41 papers in training set
Top 0.8%
0.6%
28
Neural Computation
● 39 papers in training set
Top 0.9%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.