The neural network basis of altered decision-making in patients with amyotrophic lateral sclerosis

Abstract

Objective: Amyotrophic lateral sclerosis (ALS) is a multisystem disorder associated with motor impairment and behavioral/cognitive involvement. We examined decision-making features and changes in the neural hub network in patients with ALS using a probabilistic reversal learning task and resting-state network analysis, respectively. Methods: Ninety ALS patients and 127 cognitively normal participants performed this task. Data from 62 ALS patients and 63 control participants were fitted to a Q-learning model. Results: ALS patients had anomalous decision-making features with little shift in choice until they thought the value of the two alternatives had become equal. The quantified parameters (Pαβ) calculated by logistic regression analysis with learning rate and inverse temperature well represented the unique choice pattern of ALS patients. Resting-state network analysis demonstrated a strong correlation between Pαβ and decreased degree centrality in the anterior cingulate gyrus and frontal pole. Interpretation: Altered decision-making in ALS patients may be related to the decreased hub function of medial prefrontal areas.