To BERT or Not To BERT: Comparing Speech and Language-based Approaches for Alzheimer's Disease Detection

TL;DR

This paper compares speech and language-based approaches for Alzheimer's detection, finding that fine-tuned BERT models outperform feature-based methods on a recent dataset, highlighting the importance of linguistic features.

Contribution

It provides a comparative analysis of traditional feature-based methods and BERT-based models for Alzheimer's detection using speech data.

Findings

BERT models outperform feature-based approaches in AD detection

Linguistic features are highly important for cognitive impairment detection

Fine-tuning BERT improves classification performance

Abstract

Research related to automatically detecting Alzheimer's disease (AD) is important, given the high prevalence of AD and the high cost of traditional methods. Since AD significantly affects the content and acoustics of spontaneous speech, natural language processing and machine learning provide promising techniques for reliably detecting AD. We compare and contrast the performance of two such approaches for AD detection on the recent ADReSS challenge dataset: 1) using domain knowledge-based hand-crafted features that capture linguistic and acoustic phenomena, and 2) fine-tuning Bidirectional Encoder Representations from Transformer (BERT)-based sequence classification models. We also compare multiple feature-based regression models for a neuropsychological score task in the challenge. We observe that fine-tuned BERT models, given the relative importance of linguistics in cognitive impairment detection, outperform feature-based approaches on the AD detection task.