Personalizing cognitive rehabilitation through a web-based Task Generator: an evaluation study with stroke patients

[Zlynt]

Personalizing cognitive rehabilitation through a web-based Task Generator: an evaluation study with stroke patients

Authors: Faria, A. L., Pinho, M., & Bermúdez i Badia, S. (2016).
In Presented at the International Neuropsychological Society 2016 Mid-Year Meeting, London, UK.

Abstract

1. Objective

Combining the accessibility provided by computerized technologies with an objective modeling of clinical experts’ input, we developed the Task Generator (TG), a cognitive rehabilitation web-tool that generates a set of paper-and-pencil tasks, according to patients’ cognitive profiles. The main objective of this study is to assess the validity of this tool for personalizing cognitive rehabilitation.

2. Participants and Methods

Participants were 20 middle-aged (Mdn=61.50) stroke patients (9 right and 11 left hemisphere), with a median of 2.50 months post-stroke and 4 years of schooling. Their cognitive profile, assessed with the Montreal Cognitive Assessment (MoCA) was used to parameterize a TG training session of 45/60 minutes. The TG (http://neurorehabilitation.m-iti.org/TaskGenerator/) includes 11 paper-and-pencil tasks (Cancellation, Numeric Sequencing, Problems Resolution, Association, Comprehension of Contexts, Image Pairs, Scrambled Words, Labyrinths, Categorization, Actions Sequencing and Memory Recall) for training attention, memory, executive functions, and language.

3. Results

TG revealed an acceptable internal consistency (α=.786). Considering TG median performance, the correlation with MoCA is moderate (rs=.520,p=.019), which suggests that performance of the personalized tasks is not completely determined by cognitive functioning. For instance, performance in more difficult tasks has a very strong correlation (rs=.872,p<.001) with MoCA, indicating that more difficulty was assigned to participants with higher MoCA score. Strong correlations were found for tasks performance weighed by their demand concerning attention (rs=.686,p=.001), executive functions (rs=.742,p<.001) and memory (rs=.730,p<.001), with MoCA corresponding domains scores. For language, this correlation was moderate (rs=.475,p=.034). 4.Conclusions The TG is a useful reliable tool that easily generates highly personalized and adapted paper-and- pencil tasks.

Find the full article is available here