Dynamic Causal Modeling of Insular, Striatal, and Prefrontal Cortex Activities During a Food-Specific Go/NoGo Task

BACKGROUND: This study aimed to investigate the dynamic interactions among three neural systems that are implicated in substance and behavioral addictions in response to food cues in young adults. These include an impulsive system involving the striatum, a reflective system involving the prefrontal cortex, and a homeostasis sensing system involving the insular cortex. METHODS: College students (N = 45) with various levels of body mass index were recruited. Functional magnetic resonance imaging data were acquired while participants performed food-related Go/NoGo tasks, with low-calorie and high-calorie food cues. Participants were scanned under both food satiety and deprivation conditions. Dynamic causal modeling was applied to the data to examine the causal architecture of coupled or distributed dynamics among the aforementioned systems. RESULTS: Participants showed difficulty inhibiting responses to high-calorie foods as suggested by higher false alarm rate and decision bias for low-calorie food Go task. This difficulty was enhanced during the food deprivation condition. Deprivation increased neural activity of both the insula and the striatum bilaterally in response to high-calorie foods during Go trials and anterior cingulate cortex and dorsolateral prefrontal cortex activity during NoGo trials. Dynamic causal modeling analysis revealed that food deprivation modulated the communications between the insula, striatum, and dorsolateral prefrontal cortex, and the modulations were positively associated with body mass index. CONCLUSIONS: The results support tripartite views of decision making. Deprivation states, such as hunger, trigger insular activity, which modulates the balance between impulsive and reflective systems when facing tempting food cues.