Alzheimer's disease (AD) is a progressive neurodegenerative disorder that affects memory and cognitive function. Early diagnosis of AD is important for timely intervention and treatment, which may improve the prognosis and quality of life for affected individuals. However, current methods such as clinical assessments and neuropsychological testing, can be time-consuming and costly, and may not be practical or accessible for all individuals. Automatic and early detection methods using deep learning techniques have the potential to be more efficient and scalable, and could be used to screen a larger population for the disease. To serve this purpose, we proposed hybrid EEG and fused CT-MRI based RPCA integrated deep transfer learning (HEMRDTL) model, combines data from electroencephalogram (EEG) and fused CT-MRI scans to detect AD at an early stage. The proposed HEMRDTL model use the concept of transfer learning, deep VGG-19 techniques, and robust principal component analysis (RPCA). Our approach involves extracting features from both fused CT-MRI and electroencephalogram (EEG) signals and combining them using VGG-19 for classification. For feature extraction from fused CT-MRI, we use a pre-trained VGG-19 model that was originally trained on the ImageNet dataset for image classification tasks. We fine-tune the model to extract features from the fused CT-MRI image, which are used to represent the structural properties of the brain. For feature extraction from EEG signals, we proposed modifies M-RPCA, a robust dimensionality reduction technique that is resistant to outliers and noise in the data. The features extracted from the EEG signals are used to represent the functional properties of the brain. We evaluate the performance of our approach using a large dataset of fused CT-MRI image and EEG signals and demonstrate that it outperforms several state-of-the-art methods for detecting Alzheimer's disease. Our results suggest that the combination of transfer learning, deep VGG-19 techniques, and RPCA is a promising approach for accurately detecting categories of Alzheimer's disease from both fused CT-MRI and EEG signals.
Measurement: Sensors, Volume 27, June 2023,