AI Summary & Key Takeaways
AI InsightThe results indicate that the proposed approach outperformed other techniques in AD diagnosis.
📌 3 Key Takeaways
Abstract
Alzheimer's disease (AD) represents a significant neurological disorder with a wide prevalence worldwide, characterized by cognitive and behavioral deficits resulting from brain degeneration. Despite extensive research efforts, a cure for AD has yet to be found. However, detecting the disease at its early stages can aid in slowing down its progression. However, accurate diagnosis of AD involves costly and arduous testing procedures that necessitate the evaluation of an experienced specialist. To address this, a new computer-aided diagnosis (CAD) system with high performance has been proposed to automatically diagnose AD using EEG signals. The MSPCA method was used for preprocessing to eliminate existing noise, followed by the application of the TQWT signal decomposition technique to EEG data. The Hjorth parameters, derived from the recorded data, were obtained as distinctive attributes for subsequent analysis, and the resulting features were tested using various classification algorithms. The obtained features were evaluated using different statistical techniques to determine their classification performance in distinguishing AD patients from healthy individuals. The results revealed that the k-nearest neighbor (KNN) classifier yielded the highest classification performance of 99.78%±0.004. The methodological framework under investigation, which draws on the Tunable Q-factor Wavelet Transform (TQWT), was evaluated through the application of diverse signal separation methodologies. The results indicate that the proposed approach outperformed other techniques in AD diagnosis. As such, the present study puts forth a Computer-Aided Diagnosis (CAD) framework that holds promise in augmenting the proficiency of experts in the diagnosis of Alzheimer's disease.