Detecting cognitive decline using speech only: The ADReSSo Challenge
This work addresses the societal and medical need for non-invasive, automated tools to assist in diagnosing and monitoring Alzheimer's disease, but it is incremental as it builds on a previous challenge.
The paper tackled the problem of detecting Alzheimer's dementia, predicting cognitive test scores, and forecasting cognitive decline using only speech data, achieving baseline accuracies of 78.87% for dementia detection and 68.75% for decline prediction, with an RMSE of 5.28 for score inference.
Building on the success of the ADReSS Challenge at Interspeech 2020, which attracted the participation of 34 teams from across the world, the ADReSSo Challenge targets three difficult automatic prediction problems of societal and medical relevance, namely: detection of Alzheimer's Dementia, inference of cognitive testing scores, and prediction of cognitive decline. This paper presents these prediction tasks in detail, describes the datasets used, and reports the results of the baseline classification and regression models we developed for each task. A combination of acoustic and linguistic features extracted directly from audio recordings, without human intervention, yielded a baseline accuracy of 78.87% for the AD classification task, an MMSE prediction root mean squared (RMSE) error of 5.28, and 68.75% accuracy for the cognitive decline prediction task.