Background: Behavioral interventions for obesity have modest and variable effectiveness. Emotion dysregulation may hinder effective weight loss interventions, especially among those with comorbid depression. Whether neural mechanisms related to emotion dysregulation mediate the effect of behavioral intervention on improving health behaviors in obesity combined with depression is unknown. Hyperactivation of the amygdala and loss of functional connectivity with the anterior cingulate regions are neural characteristics of unhealthy responses to negative emotion and accompanying emotion dysregulation. Methods: Using data from a randomized clinical trial of 108 primary care adult patients with obesity and depression, we investigated whether changes in amygdala activity and connectivity following problem solving therapy for depression over 2 months mediate weight loss, physical activity, and/or dietary outcomes at 6 and 12 months. Results: Compared with controls (n=49), intervention participants (n=59) had a significant blunting of the otherwise unhealthy increase in activity of the bilateral amygdala at 2 months (right: -0.83, 95% CI -1.55 to -0.11; left: -0.86, -1.63 to -0.10). While increases in amygdala activity significantly correlated with decreases in leisure-time physical activity levels and total energy expenditure at 12 months in the control group, this relationship was tempered or reversed among intervention participants (for physical activity, right: usual care -692.9, -1365.3 to -20.6; interaction 652.2, -115.4 to 1419.9; left: usual care -686.0, -1127.9 to -244.1; interaction 664.7, 73.1 to 1256.4; for energy expenditure, right: usual care -1.82, -3.08 to -0.56; interaction 1.94, 0.50 to 3.38; left: usual care -1.72, -2.51 to -0.93; interaction 1.89, 0.84 to 2.94). Compared with controls, intervention participants had a significant increase of the otherwise unhealthy loss of amygdala-subgenual anterior cingulate connectivity at 2 months (0.78, 0.05 to 1.52). This increase in functional connectivity significantly correlated with a decrease in BMI at 6 months among intervention participants (usual care 0.04, -0.42 to 0.50; interaction -0.61, -1.20 to -0.02) and an increase in fruit and vegetable intake at 12 months across all participants (usual care 1.42, 0.27 to 2.57; interaction -0.95, -2.36 to 0.46). Conclusions: These findings highlight potential neural mechanism for optimizing the efficacy of behavioral intervention in comorbid obesity and depression.Trial registration: ClinicalTrials.gov#NCT02246413.URL: https://clinicaltrials.gov/ct2/show/NCT02246413.