Mapping differential responses to cognitive training using machine learning.

Area of Science:

• Cognitive Science
• Machine Learning
• Developmental Psychology

Background:

• Working memory (WM) training aims to improve cognitive functions.
• Analyzing individual differences in response to cognitive training is crucial.
• Existing methods often focus on task-specific changes, potentially missing broader patterns.

Purpose of the Study:

• To apply unsupervised machine learning techniques to identify differential change trajectories in children undergoing WM training.
• To investigate how task representations evolve post-training.
• To distinguish between task-specific and domain-general changes and identify distinct response profiles.

Main Methods:

• Utilized self-organizing maps (SOMs) to represent multivariate cognitive training data.
• Employed K-means clustering to identify subgroups of children based on training response.
• Allocated children (N=179) to pre-identified subgroups (from N=616 sample) before and after WM training.

Main Results:

• SOM analysis indicated changes in cognitive processes underlying WM task performance post-training.
• K-means clustering identified four distinct subgroups of children responding differently to training.
• Fluid intelligence scores predicted individual children's improvement trajectories.

Conclusions:

• This study presents a novel machine learning approach for analyzing cognitive training data.
• The findings demonstrate a method to differentiate task-specific from domain-general cognitive changes.
• The approach allows for the identification of varied response profiles to cognitive interventions in children.