Alzheimer’s Disease Detection Using Speech Dataset

Abstract

Alzheimer’s disease (AD) is an irrevocable and gradually increasing brain disease which in turn causes the brain cells to degenerate and die. This progressive disease can be slowed down with medical treatment if diagnosed at early stages. But the prevailing medical evaluations are quite expensive and takes a lot of effort and time. Also in the detection of dementia the changes in speech and language pattern plays a significant role as the speech production process starts in the left hemisphere of the brain and hence any decline in speech capabilities might indicate the presence of Alzheimer’s disease. Such features based on speech will provide a non-invasive and affordable tool for the detection of dementia that does not require any extensive infrastructure. This project describes the simple automated low-cost solution for detecting Alzheimer’s disease based on linguistic and acoustic features of the demented patients. In the proposed system, four neural models such as Convolution Neural Network (CNN), Long Short Term Memory (LSTM), CNN-LSTM and CNN features to LSTM are used where CNN deals with the acoustic features and LSTM deals with the linguistic features.