Creating and sharing knowledge for telecommunications

Prediction of Alzheimer disease from handwriting tasks with explainability techniques

Moreira, A. ; Ferreira, A. ; Leite, N. Leite

Prediction of Alzheimer disease from handwriting tasks with explainability techniques, Proc Inforum - Simpósio de Informática, Lisbon, Portugal, Vol. , pp. - , September, 2024.

Digital Object Identifier:

 

Abstract
Neurodegenerative diseases are caused by the progressive de
generation of nerve cells. Alzheimer and Parkinson are two well-known neurodegenerative diseases such that they progressively decrease the cog nitive abilities and the motor skills of an individual. Without a cure, we aim to slow down their impact by resorting to rehabilitative therapies
and medicines. The early diagnosis is still a key factor to delay their progression. In this context, the analysis of handwriting dynamics for specific tasks is found to be an effective tool to provide early diagnosis. Recently, the Diagnosis AlzheimeR WIth haNdwriting (DARWIN)
dataset was introduced. It contains handwriting samples from people affected by Alzheimer’s as well as a control group with 174 participants executing 25 specific handwriting tasks, including dictation, graphic, and copy categories. In this paper, we investigate the use of the DARWIN dataset, with classification and explainability techniques. We identify the
most relevant and decisive handwriting features for predicting Alzheimer. From the original set of 450 features, composed of (x,y)-coordinates, pressure, and timestamp metrics, we found that the time spent to perform the in-air movements plays a decisive role at predicting Alzheimer.