Predicting Cognitive Disorder Progression Using EEG Signals and Convolutional Neural Networks

Millions of people all over the world suffering from neurodegenerative medical condition known as cognitive disorder (CD). The cognitive disorder may disrupt the human life in great manner. Therefore, the prediction of CD at an early stage is crucial for precise and timely treatment. In the similar context in order to predict CD, this research presents a method that combines the advantages and characteristics of electroencephalography (EEG) signals with convolutional neural networks (CNN) and the Short-Time Fourier Transform (STFT). EEG signals are a rich source of data that make it easy to identify any changes in the patterns of brain activity linked to the advancement of CD. Using the intricate 2D-CNN architecture, which is well-known for its ability to identify hierarchical patterns, a predictive model for analyzing EEG data can be used to predict. Our model gives a remarkable accuracy rate of nearly 96%, in case of Alzheimer disease (AD) and 74% in case of Schizophrenia. The low accuracy in case of Schizophrenia due to its wave nature which are very similar to other cognitive functions. This can be very helpful in early detection and in developing personalized treatment strategies and timely advancements developed for individuals at risk of succumbing to CD's debilitating effects. Thus, it can be concluded that the given fusion of EEG signals and CNNs can be very useful in the early and accurate detection of CD, making it easier for doctors and specialists to give proper and timely treatment to the persons suffering from this disease.