Detecting Alzheimer's Disease Using Gated Convolutional Neural Network from Audio Data

TL;DR

This paper introduces a GCNN-based method for detecting Alzheimer's disease from speech data, achieving higher accuracy than traditional methods and applicable across languages due to its linguistic feature independence.

Contribution

It presents a novel GCNN approach that effectively captures temporal audio features for Alzheimer's detection with limited data and language independence.

Findings

Achieved 73.6% accuracy on Pitt Corpus

Outperformed conventional SMO by 7.6 percentage points

Effective with small datasets and language-independent features

Abstract

We propose an automatic detection method of Alzheimer's diseases using a gated convolutional neural network (GCNN) from speech data. This GCNN can be trained with a relatively small amount of data and can capture the temporal information in audio paralinguistic features. Since it does not utilize any linguistic features, it can be easily applied to any languages. We evaluated our method using Pitt Corpus. The proposed method achieved the accuracy of 73.6%, which is better than the conventional sequential minimal optimization (SMO) by 7.6 points.