Exploring factors associated with severe cognitive fatigue symptoms in multiple sclerosis: A decision-tree approach

Cognitive fatigue is one of the most disabling symptoms in multiple sclerosis (MS). Despite the abundance of identified risk factors, causes of severe fatigue symptoms remain unclear. The aim of this project was to explore whether different variables are associated with cognitive fatigue severity using decision tree analysis. Data from a total of 18 variables were collected in patients with MS (N1=19) as well as in healthy subjects (N2=20). A decision tree function, implemented in MATLAB, was used to identify the predictors of fatigue severity, assessed with the Fatigue Scale for Motor and Cognitive Functions (FSMC), in four variable clusters (demographic, psychological, cognitive, and cerebral). Interactions between variables in each cluster were also investigated. Results showed that level of education, presence of depressive affects and degree of sleepiness were associated with severe symptoms in MS. In addition, the interaction between depression and anxiety symptoms can be linked to severe cognitive fatigue in healthy subjects. All these variables showed significant correlation (p < .05) with FSMC. Unfortunately, the model seems unable to predict severe symptoms for new individuals (perror = 0.57), probably due to the small sample size used during training. To conclude, decision trees models seem to be a promising method to explore relationships between different variables likely to play a role in the occurrence of severe symptoms of cognitive fatigue.