Self-Regulation of Anterior Insula with Real-Time fMRI and Its Behavioral Effects in Obsessive-Compulsive Disorder: A Feasibility Study

Abstract

Introduction: Obsessive-compulsive disorder (OCD) is a common and chronic condition that can have disabling effects throughout the patient's lifespan. Frequent symptoms among OCD patients include fear of contamination and washing compulsions. Several studies have shown a link between contamination fears, disgust over-reactivity, and insula activation in OCD. In concordance with the role of insula in disgust processing, new neural models based on neuroimaging studies suggest that abnormally high activations of insula could be implicated in OCD psychopathology, at least in the subgroup of patients with contamination fears and washing compulsions.

Methods: In the current study, we used a Brain Computer Interface (BCI) based on real-time functional magnetic resonance imaging (rtfMRI) to aid OCD patients to achieve down-regulation of the Blood Oxygenation Level Dependent (BOLD) signal in anterior insula. Our first aim was to investigate whether patients with contamination obsessions and washing compulsions can learn to volitionally decrease (down-regulate) activity in the insula in the presence of disgust/anxiety provoking stimuli. Our second aim was to evaluate the effect of down-regulation on clinical, behavioural and physiological changes pertaining to OCD symptoms. Hence, several pre- and post-training measures were performed, i.e., confronting the patient with a disgust/anxiety inducing real-world object (Ecological Disgust Test), and subjective rating and physiological responses (heart rate, skin conductance level) of disgust towards provoking pictures.

Results: Results of this pilot study, performed in 3 patients (2 females), show that OCD patients can gain self-control of the BOLD activity of insula, albeit to different degrees. In two patients positive changes in behaviour in the EDT were observed following the rtfMRI trainings. Behavioural changes were also confirmed by reductions in the negative valence and in the subjective perception of disgust towards symptom provoking images.

Conclusion: Although preliminary, results of this study confirmed that insula down-regulation is possible in patients suffering from OCD, and that volitional decreases of insula activation could be used for symptom alleviation in this disorder.

Fig 1. Flow of the rtfMRI-neurofeedback training run.

Each run consisted of 6 baseline (duration: 30 seconds each) and 6 down-regulation blocks (duration: 27 seconds each). Immediately following each block of down-regulation, a monetary feedback was presented for 3 seconds. The same design was used for the transfer runs (5^th run of the each session) but no feedback was presented.

Fig 2. The Ecological Disgust Test and the flow of the experiment.

The experimenter approached the patient with a real-life, disgust-inducing object from a distance of 5 meters either with a ‘↓’ or ‘+’ cue (10 times each). The patient focused on the object (two newly chewed gums in this picture) as the experimenter approached him/her, and said “stop” whenever he/she felt that the object should not come any closer. In the pre-test, the meanings of the cues are not explained to the patients. In the post-test, during down-arrowed runs, patients used the cognitive strategies that they had learned in the rtfMRI-neurofeedback training sessions.

Fig 3. Picture-rating test.

Patients rate each picture at the dimensions of valence, arousal and OCD symptom provocation. During the ratings, skin conductance level and heart rate data were collected.

Fig 4. Differential brain activations during the down-regulation condition in Patient 1 on the first and second days of rtfMRI-neurofeedback training.

Colors red-yellow: increased activity on the second day as compared to the first day. Blue-white: decreased activation the second day as compared to the first day. The colored functional maps were overlaid on T1-weighted structural images of four representative axial brain sections covering insula, which is delineated by the green rectangle. Statistical significance was based on z-statistic threshold of -2.3 and 2.3 followed by multiple comparisons correction at the cluster level using Family-Wise Error (FWE) at p < 0.05.

Fig 5. Patient 1, the percentage of the hits from the right and the left anterior insula through rtfMRI-neurofeedback trainings respectively.

Blue columns show the percentage of the hits of the first, red columns show the second session of rtfMRI-neurofeedback training run by run. Due to technical difficulties, the first patient could complete only two runs of rtfMRI-neurofeedback training on the first session of rtfMRI-neurofeedback training.

Fig 6. Differential brain activations during the down-regulation condition in Patient 2 on the first and second days of rtfMRI-neurofeedback training.

Colors red-yellow: increased activity on the second day as compared to the first day. Blue-white: decreased activation the second day as compared to the first day. The colored functional maps were overlaid on T1-weighted structural images of four representative axial brain sections covering insula, which is delineated by the green rectangle. Statistical significance was based on z-statistic threshold of -2.3 and 2.3 followed by multiple comparisons correction at the cluster level using Family-Wise Error (FWE) at p < 0.05.

Fig 7. Patient 2, the percentage of the hits from the right and the left anterior insula through rtfMRI-neurofeedback trainings respectively.

Blue columns show the percentage of the hits of the first, red columns show the second session of rtfMRI-neurofeedback training run by run.

Fig 8. Differential brain activations during the down-regulation condition in Patient 3 on the first and second days of rtfMRI-neurofeedback training.

Colors red-yellow: increased activity on the second day as compared to the first day. Blue-white: decreased activation the second day as compared to the first day. The colored functional maps were overlaid on T1-weighted structural images of four representative axial brain sections covering insula, which is delineated by the green rectangle. Statistical significance was based on z-statistic threshold of -2.3 and 2.3 followed by multiple comparisons correction at the cluster level using Family-Wise Error (FWE) at p < 0.05.

Fig 9. Patient 3, the percentage of the hits from the right and the left anterior insula through rtfMRI- neurofeedback trainings respectively.

Blue columns show the percentage of the hits of the first, red columns show the second, green columns show the third and purple columns show the fourth session of rtfMRI-neurofeedback training run by run.