What question did this study set out to answer?

The research aims to develop a novel framework for efficiently detecting Alzheimer’s Disease using EEG signals with advanced deep learning techniques.

May 15, 2026Open Access

Designing an efficient adaptive Res-Dense-LSTM-based Alzheimer disease detection framework with feature extraction process from EEG signal in the healthcare industry

Key Points

The research aims to develop a novel framework for efficiently detecting Alzheimer’s Disease using EEG signals with advanced deep learning techniques.
Developed an innovative deep learning framework incorporating feature extraction and deep learning techniques.
Employed Spatio Temporal Attention-based Deformable Convolution Transformer for feature extraction.
Utilized Adaptive Res-Dense-LSTM optimized by the Refined Position-aided Hippopotamus Optimization Algorithm for detection.
The proposed framework demonstrates improved accuracy in Alzheimer Disease detection compared to classical schemes.
The optimization algorithm significantly enhances the performance of the A-RS-LSTM in processing EEG signals.
Numerous experiments validate the framework's efficiency in detecting AD under various conditions.

Abstract

Abstract Alzheimer’s Disease (AD) is a regressive nervous condition, which leads to dementia and shrinking of brain cells. However, AD is incurable; timely detection and appropriate treatment policies can mitigate its development. Currently, the detection of AD is frequently executed by employing Electroencephalogram (EEG), an effective tool that has excelled in the field of medical science due to its ability to identify abnormal activities within the nervous system, which leads to AD. Yet, the existing applications struggle to detect AD through the EEG signal due to its varying nature. Moreover, traditional AD detection applications are costly and are vulnerable toward human errors. Additionally, medical professional are less satisfied with the existing machine learning approaches because of their poor interpretability throughout the detection tasks. Although numerous detection approaches have been designed, the appropriate and timely detection of AD is still a major research topic to be solved. However, classical detection approaches are ineffective in processing large quantities of EEG signals and possess greater computational complexity. Therefore, it is crucial to resolve the obstacles in the conventional detection techniques. Thus, an innovative deep learning-aided AD detection framework is implemented. Initially, essential EEG signals are sourced from the online resources. Next, the accumulated EEG signals are fed into the feature extraction phase, where the Spatio Temporal Attention-based Deformable Convolution Transformer (STA-DCT) is employed to execute the process. Later, Adaptive Res-Dense-LSTM (A-RS-LSTM) is employed for the AD detection process. Furthermore, the parameters in A-RS-LSTM are optimized by the Refined Position-aided Hippopotamus Optimization Algorithm (RP-HOA), which assists in obtaining more accurate AD detection outcomes. At last, numerous experiments are executed to validate the developed framework’s efficiency with the classical schemes.

Mark Helpful

Bookmark

Relay

View Full Paper