Integrating Pause Information with Word Embeddings in Language Models for Alzheimer's Disease Detection from Spontaneous Speech

Abstract

Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and memory loss. Early detection of AD is crucial for effective intervention and treatment. In this paper, we propose a novel approach to AD detection from spontaneous speech, which incorporates pause information into language models. Our method involves encoding pause information into embeddings and integrating them into the typical transformer-based language model, enabling it to capture both semantic and temporal features of speech data. We conduct experiments on the Alzheimer's Dementia Recognition through Spontaneous Speech (ADReSS) dataset and its extension, the ADReSSo dataset, comparing our method with existing approaches. Our method achieves an accuracy of 83.1% in the ADReSSo test set. The results demonstrate the effectiveness of our approach in discriminating between AD patients and healthy individuals, highlighting the potential of pauses as a valuable indicator for AD detection. By leveraging speech analysis as a non-invasive and cost-effective tool for AD detection, our research contributes to early diagnosis and improved management of this debilitating disease.