article · International Journal of Computational Intelligence Systems
Automated epileptic seizure detection from electroencephalogram signals often suffers from data leakage and high patient variability, which inflates reported performance. To overcome this, a leakage-aware deep learning framework analyses electroencephalogram windows through three distinct views: time-domain waveforms, short-time Fourier transform spectrograms, and continuous wavelet transform scalograms. Eight pretrained convolutional neural networks serve as independent modality experts. Rather than merging high-dimensional feature maps, the system applies a lightweight multilayer perceptron enhanced with a confidence-gated decision mechanism to fuse probability outputs from frozen experts. Evaluated on the CHB-MIT dataset under strict patient-specific, cross-validation, and leave-one-patient-out schemes, the method effectively mitigates temporal and subject-level leakage. A ResNet50 backbone provided the highest clinical efficacy, securing a mean balanced accuracy of 99.23 percent and a sensitivity of 99.50 percent, demonstrating dependable and realistic detection capabilities.
Automated seizure detection algorithms frequently fail in clinical practice because testing protocols inadvertently leak patient or temporal data, creating unrealistic performance claims. By enforcing rigorous, leakage-safe validation across diverse signal representations, this approach ensures that accuracy metrics reflect genuine real-world efficacy, helping clinicians and system developers build more trustworthy automated monitoring tools for individuals with epilepsy.
The method could enable automated seizure diagnostic and alert systems within clinical monitoring equipment or bedside decision-support software. Because it employs low-complexity probability fusion rather than heavy feature-map processing, it is computationally suitable for integration into monitoring devices. The technology currently represents applied research tested on a benchmark dataset, meaning prospective clinical trials and integration tests on medical hardware are necessary next steps.
AI-generated from the published abstract. Always read the original work before citing.
Automated epileptic seizure detection from Electroencephalogram (EEG) signals remains challenging due to the non-stationary nature of EEG activity, strong inter-patient variability, and the tendency of window-based evaluation protocols to produce overly optimistic results when temporal or subject-level data leakage is not properly controlled. This study proposes a leakage-aware multimodal deep learning framework for robust EEG seizure detection on the CHB-MIT scalp EEG dataset. Each EEG window is represented using three complementary views: time-domain waveform images, Short-Time Fourier Transform (STFT) spectrograms, and Continuous Wavelet Transform (CWT) scalograms. Eight pretrained convolutional neural network (CNN) backbones, including DenseNet201, EfficientNetB0, GoogleNet, MobileNetV2, ResNet18, ResNet50, VGG16, and VGG19, are independently trained as modality-specific experts. Instead of fusing high-dimensional feature maps, the proposed framework performs compact probability-level fusion by learning from the posterior outputs of frozen experts through a lightweight multilayer perceptron (MLP) enhanced with a confidence-gated decision mechanism. The main contribution of this work is a unified seizure detection framework that combines complementary EEG representations, low-complexity learnable probability-level fusion, confidence-adaptive decision refinement, and rigorous leakage-safe validation. Three evaluation scenarios were implemented: patient-specific evaluation with Group ID-based splitting and fallback handling, patient-wise 5-fold cross-validation, and Leave-One-Patient-Out (LOPO) validation. These protocols explicitly reduce temporal, recording-level, and patient-level leakage, thereby providing a more reliable assessment of both individualized and cross-patient generalization. Experimental results demonstrate that the proposed confidence-gated fusion consistently outperforms conventional MLP fusion across the evaluated backbones. Among the evaluated backbone models, ResNet50 provided the best overall clinical performance, achieving a mean balanced accuracy of 99.23%, a mean F1-score of 99.12%, the highest mean sensitivity of 99.50%, a mean specificity of 98.96%, and a mean precision of 98.85%. ResNet50 was selected as the best-performing model due to its superior seizure detection capability, as reflected by its highest sensitivity while maintaining comparable performance across all other evaluation metrics. These results demonstrate that adaptive probability-level fusion of waveform, STFT, and CWT representations can provide highly accurate and stable seizure detection while maintaining a leakage-aware evaluation framework. Overall, the proposed framework offers a robust and clinically relevant strategy for automated EEG seizure detection under realistic validation conditions.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1007/s44196-026-01544-0
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
