The neural substrates of subliminal attentional bias and reduced inhibition in individuals with a higher BMI: A VBM and resting state connectivity study.

Previous studies have shown that individuals with overweight and obesity may experience attentional biases and reduced inhibition toward food stimuli. However, evidence is scarce as to whether the attentional bias is present even before stimuli are consciously recognized. Moreover, it is not known whether or not differences in the underlying brain morphometry and connectivity may co-occur with attentional bias and impulsivity towards food in individuals with different BMIs. To address these questions, we asked fifty-three participants (age M = 23.2, SD = 2.9, 13 males) to perform a breaking Continuous Flash Suppression (bCFS) task to measure the speed of subliminal processing, and a Go/No-Go task to measure inhibition, using food and nonfood stimuli. We collected whole-brain structural magnetic resonance images and functional resting-state activity. A higher BMI predicted slower subliminal processing of images independently of the type of stimulus (food or nonfood, p = 0.001, εp2 = 0.17). This higher threshold of awareness is linked to lower grey matter (GM) density of key areas involved in awareness, high-level sensory integration, and reward, such as the orbitofrontal cortex [t = 4.55, p = 0.003], the right temporal areas [t = 4.18, p = 0.002], the operculum and insula [t = 4.14, p = 0.005] only in individuals with a higher BMI. In addition, individuals with a higher BMI exhibit a specific reduced inhibition to food in the Go/No-Go task [p = 0.02, εp2 = 0.02], which is associated with lower GM density in reward brain regions [orbital gyrus, t = 4.97, p = 0.005, and parietal operculum, t = 5.14, p < 0.001] and lower resting-state connectivity of the orbital gyrus to visual areas [fusiform gyrus, t = -4.64, p < 0.001 and bilateral occipital cortex, t = -4.51, p < 0.001 and t = -4.34, p < 0.001]. Therefore, a higher BMI is predictive of non food-specific slower visual subliminal processing, which is linked to morphological alterations of key areas involved in awareness, high-level sensory integration, and reward. At a late, conscious stage of visual processing a higher BMI is associated with a specific bias towards food and with lower GM density in reward brain regions. Finally, independently of BMI, volumetric variations and connectivity patterns in different brain regions are associated with variability in bCFS and Go/No-Go performances.

A hierarchical-drift diffusion model of the roles of hunger, caloric density and valence in food selection.

Decisions based on affectively relevant stimuli, such as food items, hardly follow strictly rational rules. Being hungry, the food's caloric density, and the subjective valence attributed to various foods are known factors that modulate food choices. Yet, how these factors relatively and altogether contribute to the food choice process is still unknown. In this study, we showed 16 healthy young adults low- and high-calorie food when hungry or fed, and we asked them to evaluate the valence of each visually-presented food. To compute the relative influence of hunger, caloric density and valence on food choice, we applied a hierarchical drift diffusion model (HDDM). Results indicated that hunger, caloric density and valence affected how fast participants accumulated information in favor of the chosen item over the other. When fed, participants were faster in choosing low-calorie foods and foods with a higher valence. Conversely, when hungry, participants were faster in choosing high-calorie foods, including food items with lower subjective valence. All in all, these findings confirm the complex nature of food choices and the usefulness of nuanced computational models to address the multifaceted nature of decision-making and value assessment processes affecting food selection.