Explainable Nested Dilated Adaptive Deep Learning System-based Epilepsy Detection with High Frequency Oscillations in EEG Signal

The irregular as well as synchronized electrical discharges within the neural circuits of the brain are known as Epilepsy. With the aim of mitigate the seizure complexities on patients, enhancing the diagnostic accuracy and therapeutic efficiency is essential. The seizure-specific characteristics are visible in ElectroEncephaloGraphy (EEG) recordings. The sharp and spike wave patterns in the EEG signals are used to differentiate the abnormal and normal activities of the brain. Numerous signal processing techniques have been introduced in existing works to recognize the abnormal patterns. Neurologists examine EEG traces to identify epileptiform events. This research introduces a novel epilepsy detection method. Initially, the needed EEG signals are taken from the standard data resources and given to the proposed Nested Dilated Adaptive Graph Attention Network with Gated Recurrent Unit layer (NDAGAT-GRU) for detecting epilepsy seizures. Here, the Explainable deep learning method, such as Gradient-weighted Class Activation Mapping (Grad-CAM), is used to generate better scalability regarding the detection result. HFOs-based epilepsy detection helps to recognize the brain tissue which starts the seizure activity. To improve the detection network performance, the hybrid algorithm named Fused Northern Goshawk with Clouded Leopard Optimization (FNG-CLO) is included. The efficiency of the designed approach is demonstrated via the experimental results. In the estimation phase, the designed approach has attained higher accuracy values of 93.23%, 94.92%, 94.60%, 94.16% an 94.32% in terms of Linear, ReLu, Tanh, Softmax and Sigmoid activation function than other conventional methods that shows its superior detection performance at an earlier stage, Thus, it helps to improve proper decision making and better clinical enhancement through mitigating unnecessary hospital admissions.

Authors

Publication Details

Journal
International Journal of Artificial Intelligence Tools
Published
2026-11-13
DOI
https://doi.org/10.1142/s0218213026500338
Primary Topic
EEG and Brain-Computer Interfaces
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Explainable Nested Dilated Adaptive Deep Learning System-based Epilepsy Detection with High Frequency Oscillations in EEG Signal

Jessy John, Aswin Raj Valsarajan
4 links
International Journal of Artificial Intelligence Tools
EEG and Brain-Computer Interfaces
article

Explainable Nested Dilated Adaptive Deep Learning System-based Epilepsy Detection with High Frequency Oscillations in EEG Signal

Jessy John, Aswin Raj Valsarajan
article en

Abstract

The irregular as well as synchronized electrical discharges within the neural circuits of the brain are known as Epilepsy. With the aim of mitigate the seizure complexities on patients, enhancing the diagnostic accuracy and therapeutic efficiency is essential. The seizure-specific characteristics are visible in ElectroEncephaloGraphy (EEG) recordings. The sharp and spike wave patterns in the EEG signals are used to differentiate the abnormal and normal activities of the brain. Numerous signal processing techniques have been introduced in existing works to recognize the abnormal patterns. Neurologists examine EEG traces to identify epileptiform events. This research introduces a novel epilepsy detection method. Initially, the needed EEG signals are taken from the standard data resources and given to the proposed Nested Dilated Adaptive Graph Attention Network with Gated Recurrent Unit layer (NDAGAT-GRU) for detecting epilepsy seizures. Here, the Explainable deep learning method, such as Gradient-weighted Class Activation Mapping (Grad-CAM), is used to generate better scalability regarding the detection result. HFOs-based epilepsy detection helps to recognize the brain tissue which starts the seizure activity. To improve the detection network performance, the hybrid algorithm named Fused Northern Goshawk with Clouded Leopard Optimization (FNG-CLO) is included. The efficiency of the designed approach is demonstrated via the experimental results. In the estimation phase, the designed approach has attained higher accuracy values of 93.23%, 94.92%, 94.60%, 94.16% an 94.32% in terms of Linear, ReLu, Tanh, Softmax and Sigmoid activation function than other conventional methods that shows its superior detection performance at an earlier stage, Thus, it helps to improve proper decision making and better clinical enhancement through mitigating unnecessary hospital admissions.

International Journal of Artificial Intelligence Tools
Good health and well-being
EEG and Brain-Computer Interfaces
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon