Obesity and regional brain activations during saliency processing
Despite indistinguishable performance on the SST, lean women showed greater brain activations to stop as compared to go trials than obese women. Saliency processing in regions including the left inferior parietal cortex, supplementary motor area, bilateral insula, and bilateral cuneus inversely correlated with BMI across lean, overweight, and obese women, with the right anterior insula showing the most significant association.
As part of the ventral attention system, the left inferior parietal cortex responds to infrequent but behaviorally relevant stimuli (22). The cuneus is a visual area which shows greater activation with attention to visual stimuli (7). Medial frontal cortex including the supplementary motor area responds to conflict, such as the stop trials, as compared to nonconflict go trials (15). Similarly, the right anterior insula increased activation during stop trials (23,24). Taken together, a network of brain regions responded to saliency processing and differentiated between obese and lean individuals.
There is growing evidence that dopamine is a major effector of saliency processing. Dopamine neurons spike at the appearance of salient, nonreward stimuli (25,26) and mediate decision making based on the salience of visual cues (27). Patients with Parkinson's disease, a disorder of decreased nigrostriatal dopamine, have less memory for angry faces, which are thought to be more salient (28).
Indeed, studies have linked obesity with altered dopaminergic neurotransmission. Severely obese individuals (BMI >40) were shown to have lower dopamine D2 receptor levels and these levels correlated to their BMI (29). Furthermore, striatal dopamine D2 receptor levels positively correlated with brain metabolism in the prefrontal cortices, and were lower in obese subjects (30). Similar to rats showing drug addictions, obese rats that have been exposed to a “cafeteria diet” of cakes, bacon, frosting, etc. have been found to have decreased dopamine D2 receptor and increased reward thresholds (31).
Deficits in saliency processing have also been reported in individuals with drug addictions or other mental conditions such as attention deficit hyperactivity disorder, in which altered dopaminergic neurotransmissions are implicated. For instance, cocaine dependent individuals show hypoactivity in presupplementary motor area, insula, and cingulate cortex during nogo as compared to go trials (32). Individuals with attention deficit hyperactivity disorder showed decreased prefrontal cortical responses to nogo trials in go/nogo tasks (33,34), paralleling our findings with obese subjects in the SST. Taken together, evidence is accumulating to suggest shared neural processes underlying obesity, drug addictions, and attention deficit hyperactivity disorder, which all appear to implicate the dopaminergic systems (35).
Implications and limitations of the current findings
There is some evidence that altered cortical functions as observed in obese individuals may be reversible after weight loss. For instance, after gastric bypass surgery, there was a reduction in brain activity to high-calorie foods as subjects lost weight (36). Gunstad and colleagues showed improvement cognitive performance in obese people after bariatric surgery and weight loss (37). In an earlier study, Kretsch et al. (38) found that after obese women dieted and lost weight, they showed marked improvement in word recall performance.
These findings also indicate a need to measure and control for BMI in brain imaging studies of mental illnesses that frequently coexist with obesity. Obesity has been associated with a major lifetime risk of both axis I and axis II disorders (39). The need for consideration of BMI was reflected by a recent study examining brain metabolite levels in alcohol-dependent patients, which discovered that many of these brain metabolites are altered as a function of BMI, independent of alcohol dependence (40).
There are some limitations to this study. First, we determined whether an individual is obese, overweight, or lean by BMI, which does not take into account muscle composition. Therefore, the current study focused on females, who are less likely to have muscle mass as a confound. Future studies should use either a skinfold test, where a piece of skin is measured by calipers at a standardized location to determine subcutaneous fat, or waist circumference in relation to height measurements to allow for the estimation of obesity in both men and women. Second, obesity is known to be associated with medical conditions. Although our participants reported no conditions worthy of medical attention, we did not perform a thorough medical examination on our subjects. On the other hand, in a study of 68 patients seeking surgical treatment for obesity, Boeka and Lokken (3) showed that cognitive dysfunction in obese individuals was independent of other common comorbidities including diabetes, sleep apnea, and hypertension. Third, the current results were not statistically significant when evaluated at a threshold corrected for multiple comparisons. These results thus must be deemed preliminary and replicated in the future. Finally, we also acknowledge that a contrast between stop and go trials in the SST could potentially involve processes related to attentional monitoring and response inhibition. For instance, the presupplementary motor area is implicated in response inhibition (7). Thus, future studies of a larger number of participants are needed to disambiguate how these different psychological processes are compromised in obesity (7,15,16).
To summarize, we demonstrated that obesity is associated with altered saliency processing, broadly consistent with other findings of cognitive impairment and altered catecholaminergic function in obese individuals.