Authors :
Presenting Author: Crystal Cooper, PhD – UT Arlington
Nichol Civitello, MS – University of Texas Arlington
F. Kathryn King, PhD – University of Texas Arlington
M. Scott Perry, MD – Cook Children’s Physician Network
Hunter Ball, PhD – University of Texas Arlington
Christos Papadelis, PhD – Cook Children's Health Care System
Rationale:
Depression is a commonly observed psychiatric illnesses in epilepsy, and risk of this comorbidity is higher with earlier epilepsy onset. Despite key clinical characterization of neuropsychiatric function in patients with epilepsy (EPI), there is a need for phenotyping affective and hedonic processing thought to underlie cardinal psychiatric symptoms, e.g., anhedonia. Reward processing deficits are well-documented in depression but remain underexplored in epilepsy. Here, we 1) characterized dysfunction in reward processing in adolescents EPI and typically developing controls (TD) using traditional and computational decision-making models, and 2) explore the relationship between these reward behavior and brain function using magnetoencephalography (MEG).
Methods:
A probabilistic reward task (PRT), was administered to adolescents aged 10–19, including TD controls (n=24) and adolescents diagnosed with epilepsy (EPI; n=24). Reaction time and response bias (i.e., reward learning) were assessed behaviorally using signal detection theory (SDT). Drift diffusion modeling (DDM) was applied to evaluate underlying cognitive processes related to decision-making and reward learning. MEG region-of-interest connectivity was used to measure active neural connectivity in sensory-motor and prefrontal regions during an affective decision-making task. Differences in behavioral measures were calculated between the groups using independent samples t-tests. Linear regressions were conducted to investigate the predictive relationship between reward-based decision-making metrics and brain function. See Fig. 1 for methods illustration.
Results:
EPI had slower RT overall than TD when performing the PRT (p< .05). Using SDT, reward learning was equivalent between the groups (p>.05) Using DDM, EPI exhibited lower boundary separation (p=.02), suggesting reduced cognitive control and less adaptive decision-making compared to TD. Additionally, non-decision time was longer in the EPI group (p=.04), indicating slower sensory or motor processing. There were lower values for drift rate in the rewarded condition for the EPI group (p=.03), suggesting a slower learning rate for rewarded stimuli as compared to the TD group. A negative association was found between non-decision time and connectivity strength between the left precentral and postcentral gyri in the TD group (p=.02), indicating that faster sensory encoding and motor response were associated with stronger connectivity in sensorimotor regions. See Fig. 2 for results.
Conclusions:
SDT modeling of reward processing did not reveal differences in response bias between TD and EPI groups. However, relative to TD, EPI participants showed differences in reward-based decision-making processes using DDM, with blunted sensory-motor processing being linked to connectivity in sensory-motor brain areas. This suggests that the separate aspects of this form of decision-making in EPI are more nuanced and need further investigation. Better understanding of reward-based behavior in epilepsy and their neural correlates may aid the development of early diagnostic tools and interventions to track and improve cognitive and mental health outcomes.
Funding: Jordan Elizabeth Harris Foundation