A Neuroimaging Study of the Effort-Reward Imbalance Framework for Cognitive Fatigue in Individuals with Multiple Sclerosis

Background: Cognitive fatigue is one of the most pervasive yet least understood symptoms in persons with multiple sclerosis (PwMS). The current study examined whether the effort-reward imbalance, a cognitive neuroscience-based framework, can explain cognitive fatigue. The effort-reward framework posits that cognitive fatigue results from a mismatch between effort and reward processing. We hypothesized that 1) high-demand and low-reward conditions would be associated with cognitive fatigue, and 2) structural connectivity between and cerebral activation within fronto-striatal brain regions would be associated with the effort-reward tradeoffs and fatigue. Methods: Twenty PwMS and 20 cognitively healthy controls (HC) underwent fMRI during a computerized switching task with independent high- and low-demand (effort) and reward conditions. We assessed fatigue with the Visual Analog Scale of Fatigue (VAS-F) before the start of the task and after each condition. Mixed effects models were used to estimate the association between effort, reward, VAS-F, FA values between and BOLD activity within frontal (dorsolateral prefrontal cortex, DLPFC; orbitofrontal cortex, OFC; ventromedial prefrontal cortex, vmPFC; and dorsal anterior cingulate cortex, dACC) and striatal (dACC and nucleus accumbens, NAcc) regions previously implicated in cognitive fatigue. Results: We found that subjects with MS reported higher VAS-F scores than their HC counterparts (p=.01). We did not find a relationship between effort or reward conditions on VAS-F, although the conditions were associated with task performance and cerebral activation within ROIs. During the high-demand conditions, PwMS showed significantly greater right-DLPFC (p<.01) and lower bilateral dACC (p<.01) BOLD activity than the HC group. During the low-reward presentation, participants showed significantly lower bilateral vmPFC BOLD activity (p<.05), and across conditions, PwMS showed significantly greater bilateral-NAcc activity (p<.01) than the HC group. We did not observe a relationship between VAS-F and DLPFC, OFC, vmPFC, or NAcc activation. However, we found that in PwMS, right-dACC activation during the high-demand condition was associated with lower VAS-F scores ([beta]=-.33, 95% CI: -.65- -.01, p<.05). Conclusions: The current study implicates the dACC as a key region underlying fatigue in MS. The observed cerebral activation patterns during a switching task suggest potential feedback loop differences in cognitive control (DLPFC and dACC) and reward (vmPFC and NAcc) circuits between PwMS and HC. The loop pattern differences, and the link between fatigue and dACC, support a view that fatigue stems from or leads to a lack of top-down command based on an evaluation function that assesses the utility of effortful action, mediated by the dACC (i.e., "Is it worth it?"). If computational resources are more limited in MS due to the disease, fatigue signals would be more widely used to protect these resources. As such, we consider fatigue an adaptive outcome that influences how and for what valuable cognitive resources are spent to preserve these resources. Our results provide more specificity to the effort-reward imbalance framework by suggesting that a value assessment mechanism is driving the imbalance. More studies are necessary to test these observations. However, we offer the dACC as one neural target to further examine as a step toward developing a neurophysiologically-based fatigue assessment and treatment system.