On the importance of normative data in speech-based assessment

TL;DR

This paper demonstrates that augmenting sparse Alzheimer's datasets with normative data and using oversampling techniques significantly improves binary classification accuracy for AD detection.

Contribution

It introduces a novel approach of combining patient data with normative datasets and applies oversampling to enhance AD classification performance.

Findings

Outperforms state-of-the-art in AD classification

Combining normative and patient data is effective

Oversampling improves model accuracy

Abstract

Data sets for identifying Alzheimer's disease (AD) are often relatively sparse, which limits their ability to train generalizable models. Here, we augment such a data set, DementiaBank, with each of two normative data sets, the Wisconsin Longitudinal Study and Talk2Me, each of which employs a speech-based picture-description assessment. Through minority class oversampling with ADASYN, we outperform state-of-the-art results in binary classification of people with and without AD in DementiaBank. This work highlights the effectiveness of combining sparse and difficult-to-acquire patient data with relatively large and easily accessible normative datasets.