A Machine Learning Approach to Personalize Computerized Cognitive Training Interventions

MELINA VLADISAUSKAS; LAOUEN MAYAL LOUAN BELLOLI; DIEGO FERNANDEZ SLEZAK; ANDREA P. GOLDIN

Frontiers in Artificial Intelligence

Frontiers

Año: 2022 vol. 5

Executive functions are a class of cognitive processes critical for purposeful goal-directedbehavior. Cognitive training is the adequate stimulation of executive functions andhas been extensively studied and applied for more than 20 years. However, there isstill a lack of solid consensus in the scientific community about its potential to elicitconsistent improvements in untrained domains. Individual differences are consideredone of the most important factors of inconsistent reports on cognitive training benefits,as differences in cognitive functioning are both genetic and context-dependent, andmight be affected by age and socioeconomic status. We here present a proof ofconcept based on the hypothesis that baseline individual differences among subjectswould provide valuable information to predict the individual effectiveness of a cognitivetraining intervention. With a dataset from an investigation in which 73 6-year-olds trainedtheir executive functions using an online software with a fixed protocol, freely availableat www.matemarote.org.ar, we trained a support vector classifier that successfullypredicted (average accuracy = 0.67, AUC = 0.707) whether a child would improve, ornot, after the cognitive stimulation, using baseline individual differences as features. Wealso performed a permutation feature importance analysis that suggested that all featurescontribute equally to the model?s performance. In the long term, this results might allowus to design better training strategies for those players who are less likely to benefit fromthe current training protocols in order to maximize the stimulation for each child.