Leveraging Multimodal Information in Speech Data for the Non-Invasive Detection of Alzheimer’s Disease

Alzheimer’s disease (AD) is the most common form of dementia, affecting millions of individuals worldwide. Early identification and treatment of AD are critical for advanced planning and preventing the gradual emergence of harsher symptoms. Speech data analysis is a promising non-invasive tool for the early identification of AD. This study investigates the possibilities of employing machine learning (ML) and deep learning (DL) approaches for speech data processing describing methods for detecting Alzheimer’s disease using unimodal and multimodal data analysis of a patient’s speech. Thus, in this study, we analyzed speech recordings from participants and classified them into AD and normal control (NC) categories. According to experimental findings, using multimodal models might increase the accuracy and robustness of speech-related diagnoses, leading to the early identification of AD. With its insights into the possibilities of speech data analysis and multimodal models for this purpose, this work adds to the expanding body of knowledge on non-invasive methods for AD identification.