Epileptic EEG Signal Detection Performance Metrics

Experimental results comparing the performance of neural network models (EEGNet, AlexNet, Shallow ConvNet) combined with time-frequency transforms (CWT, STFT) for epileptic EEG signal detection. The data originates from a study by Canhui Wang, employing subject-independent validation and techniques like Focal Loss and dynamic data augmentation. The file is 5.5 KB in size.

Use Cases

• Compare the performance metrics of CWT versus STFT feature extraction methods for EEG signal classification.
• Analyze the impact of the Squeeze-and-Excitation (SE) attention module and depthwise separable convolution optimizations on EEGNet model results.
• Evaluate the effectiveness of layered convolution for extracting 'time-frequency' features in the Shallow ConvNet model.
• Assess model robustness improvements attributed to the use of Focal Loss and dynamic data augmentation during training.

Strengths

• Data is derived from a structured experiment comparing three established neural network architectures (EEGNet, AlexNet, Shallow ConvNet).
• The study employs subject-independent validation, a method designed to test model generalizability across individuals.
• The dataset is openly available under a CC BY 4.0 license, facilitating reuse and sharing.

Limitations

• The dataset is very small at 5.5 KB, indicating it likely contains only summary results, not raw EEG signals or extensive model outputs.
• Specific details such as the number of subjects, EEG recording parameters, and raw performance metrics are not provided in the input.
• The data is focused on a specific methodological comparison, limiting its direct applicability to other EEG analysis tasks.

Provenance

Source

Figshare, authored by Canhui Wang.

Collection Method

Generated from experiments comparing CWT and STFT feature extraction with optimized neural network models for EEG classification.

Freshness

Last updated on 2026-03-20.