The purpose of this study is to determine whether upregulating the left amygdala during positive autobiographical memory recall via real time functional magnetic resonance imaging neurofeedback will lead to an improvement in clinician administered ratings of depressive symptoms. The investigators predict that patients with major depressive disorder receiving left amygdala neurofeedback will increase their amygdala response during positive autobiographical memory recall compared to those receiving control feedback from a region not involved in emotional processing and that this ability will be associated with clinically significant improvement.

Major depressive disorder (MDD) is the leading cause of years lived with disability worldwide. Traditional pharmacological and/or psychological interventions are ineffective in up to one-half of patients, and treatments (such as electroconvulsive therapy, vagus nerve stimulation, and deep brain stimulation) available for severely ill patients who do not respond to standard interventions are invasive, and associated with potentially significant side effects. Therefore, there is a need to explore and develop novel non-invasive treatments.

One such non-invasive method is real-time functional magnetic resonance imaging neurofeedback (rtfMRI-nf), which allows a person to see and regulate the fMRI signal from his or her own brain. Emerging evidence suggests rtfMRI-nf has clinical utility in reducing symptoms of chronic pain, tinnitus, and Parkinson's disease. The goal of the current study is to leverage recent advances in rtfMRI-nf to determine whether this procedure can be adapted as treatment for MDD. While amygdala activity is exaggerated in response to negative stimuli in MDD, evidence further suggests that the amygdala response to positive stimuli is attenuated in MDD and normalizes with remission. Therefore, the target for our rtfMRI-nf procedure is the left amygdala. Participants will be randomly assigned to receive rtfMRI-nf from either the left amygdala or the left horizontal segment of the intraparietal sulcus (HIPS; a region not involved in emotional processing) and to increase the activity within that region to a target level by thinking of positive autobiographical memories. This neurofeedback condition will alternate with periods of rest and counting backwards in order to allow participants to disengage from memory contemplation. A final run without neurofeedback information will be included to determine whether participants can maintain the learned amygdala elevation during positive memory recall in the absence of neurofeedback. Participants will complete two sessions within a one-week period. Clinical ratings will be taken at the time of each scan to determine whether the amygdala rtfMRI-nf procedure results in improvement of depression symptoms, and changes within the emotional regulation network that occur with successful amygdala regulation will be examined. Furthermore, the investigators aim to determine whether the rtfMRI-nf procedure will alter assessments of emotional processing conducted within three days prior to, and following completion of, the rtfMRI-nf procedure.

Specific Aim 1: In individuals with MDD, determine the degree to which rtfMRI-nf enhances voluntary control over neural activity in the amygdala, co-modulates other brain regions within the emotion regulation circuitry, and alters depressive symptom severity ratings.

• Hypothesis 1.1: MDD participants receiving rtfMRI-nf regarding blood oxygen-level dependent (BOLD) activity within their amygdala can learn to voluntarily regulate this activity in response to positive stimuli. MDD participants receiving rtfMRI-nf regarding left amygdala activity will demonstrate greater activity in this region while contemplating positive autobiographical memories (AMs) than MDD participants receiving rtfMRI-nf regarding BOLD activity in the left HIPS, a region not involved in emotion.
• Hypothesis 1.2: The investigators hypothesize enhancing control over the amygdala via rtfMRI-nf will increase connectivity strengths between the amygdala and prefrontal regions involved in modulating emotional behavior including the pregenual anterior cingulate cortex and ventromedial prefrontal cortex.
• Hypothesis 1.3: The investigators hypothesize participants showing the greatest enhancement of amygdala activity in response to rtfMRI-nf also will show the greatest improvement in depressive symptom severity ratings at the end of the study.

Specific Aim 2: In MDD patients, determine the degree to which rtfMRI-nf from the amygdala restores a normative mood-congruent processing bias during the processing of emotionally valenced stimuli.

• Hypothesis 2.1: During the performance of a backward masking task in which emotional faces are presented below conscious awareness, the investigators hypothesize MDD participants will initially show a processing bias toward negative stimuli in the amygdala that will reverse to a processing bias toward positive stimuli in participants receiving active vs HIPS rtfMRI-nf
• Hypothesis 2.2: The investigators hypothesize that MDD patients will initially show a mood-congruent processing bias toward negative stimuli on the P1Vital Emotional Test Battery that will reverse to a bias toward positive stimuli following amygdala (vs HIPS) rtfMRI-nf.

Results from this project will lead to new insights into the plastic neurobiological mechanisms that govern recovery from MDD and promote novel, non-invasive approaches to MDD treatment.

Inclusion Criteria:

• clinical diagnosis of major depressive disorder
• right handed
• adult aged 18-55
• currently depressed

Exclusion Criteria:

• clinically significant or unstable cardiovascular, pulmonary, endocrine, neurological, gastrointestinal illness or unstable medical disorder
• met Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria for alcohol and/or substance abuse or substance dependence (other than nicotine) within 12 months prior to screening
• endorse suicidal intent or have made a suicide attempt within the preceding three months
• history of traumatic brain injury
• inability to complete MRI scan due to claustrophobia or general MRI exclusions (e.g., shrapnel inside body)
• current pregnancy or breast feeding
• a primary language other than English
• received psychotropic drugs for at least 3 weeks (8 weeks for fluoxetine) prior to scanning (Effective medications will not be discontinued for the purposes of the study)

2 weeks

large proportion of patients excluded primarily because of medication status or not meeting DSM-IV-TR criteria for major depressive disorder, limiting the generalizability Patients were only followed for 1 week after the final rtfMRI-nf session