Comparing Real-time fMRI Neurofeedback Versus Sham for Altering Limbic and Eating Disturbances in Anorexia Nervosa

Aim 1: Establish that rt-fMRI neurofeedback of limbic activity can correct neural disturbances in AN. Hypothesis 1: Compared to the sham group, the amygdala neurofeedback group will show reduced amygdala activation to aversive images, which will increase with repeated training. This effect will generalize to non-neurofeedback test runs. Hypothesis 2: Compared to the sham group, the amygdala neurofeedback group will exhibit enhanced task and resting amygdala-prefrontal cortex (PFC) connectivity, which will increase with repeated training. Enhanced amygdala-PFC connectivity will be associated with less amygdala reactivity to aversive images during the emotion regulation task.

Aim 2: Identify the impact of rt-fMRI neurofeedback targeting limbic functioning on symptoms of AN. Hypothesis 1: Compared to the sham group, the amygdala neurofeedback group will exhibit improvements in self-reported emotion regulation and eating disorder symptoms over the study visits. Hypothesis 2: Compared to the sham group, the amygdala neurofeedback group will engage in less restrictive eating (i.e., will consume more calories) at a post-training test meal. Hypothesis 3: Across groups, decreased aversive amygdala reactivity and enhanced amygdala-PFC connectivity will predict reduced emotion dysregulation and eating disorder symptoms, and less restriction.

Inclusion criteria

• DSM-5 diagnosis of AN, with the exception of body image disturbance and intense fear of weight gain
• Ability to read and speak in English
• Right-handed

Exclusion criteria

• Medical instability or current pregnancy (self-reported)
• Acute suicidality, current substance use disorder, psychosis, or mania
• Contraindication for fMRI as determined by CMRR safety screening standards
• History of neurological disorder/injury (e.g., stroke; head injury with > 10 minutes loss of consciousness)