Skip to main content

Transforming the understanding
and treatment of mental illnesses.

Celebrating 75 Years! Learn More >>

 Archived Content

The National Institute of Mental Health archives materials that are over 4 years old and no longer being updated. The content on this page is provided for historical reference purposes only and may not reflect current knowledge or information.

“Covert” Neurofeedback Tunes-up the Social Brain in ASD

Playing game boosts brain circuit connectivity – and maybe sociability

Science Update

Young people with autism spectrum disorder (ASD) thought they were just playing a picture puzzle game while undergoing functional magnetic resonance imaging (fMRI) scanning. In fact, the game was rigged – by their own brain activity.

The more participants spontaneously activated social brain circuitry known to be under-connected in ASD, the more pieces of a puzzle filled-in to reveal the picture. Since the game was controlled by circuit activity, the participants were unknowingly tuning-up their own brains. Resting state scans following the sessions revealed increased communications between two key networks of the social brain that typically don’t talk with each other enough in ASD. What’s more, participants’ parents noted improvements in their children’s social behavior linked to this boost in circuit connectivity.

“Rather than specifying explicit tasks, the game provided positive reinforcement for spontaneously achieved desired states of circuit activity,” explained Alex Martin, Ph.D., chief of NIMH’s Section on Cognitive Neuropsychology.  “Such implicit training resulted in significant long-term connectivity increases in the suspect circuits that were correlated with improvements in behavior. This proof-of-principle suggests that covert neurofeedback may have potential as an intervention in ASD – and perhaps in other disorders of circuit under-connectivity.”

However promising, the researchers note that results of this small exploratory study would have to be confirmed in larger samples, and the approach would need to be adapted to a more accessible and cost-effective form of neuroimaging before it might become a practical treatment alternative.

The NIMH researchers and others had earlier linked such resting state under-connectivity in social brain circuits with symptom severity and a worsening course of symptoms in ASD. So the researchers wondered whether training to boost such connectivity might improve social behavior.

Conducting fMRI scans while the brain is at rest predict how well brain networks communicate with each other when performing behaviors. In this case, the networks in question are normally activated during social behaviors. However, training on explicit social behavior tasks only serves to re-activate dysfunctional circuit patterns underlying the social deficits seen in the disorder.

Enter post-doctoral fellow Michal Ramot, Ph.D. She devised a training scheme based on open-ended, implicit, non-social tasks. The researchers told participants in the scanner – 17 young adult males with ASD and 10 without ASD – to do whatever might fill-in pieces of a picture puzzle on a computer screen. Participants were promised monetary rewards and bonuses for completing the puzzles.

Unbeknownst to the participants, a computer running the game was programmed to fill-in pieces of the picture whenever the scanner detected increased chatter between the two social brain networks that are under-communicative in ASD. The more participants spontaneously activated the targeted social brain circuitry, the more pieces of the picture appeared.

“Showing faces or other social stimuli would only activate networks wired in a sub-optimal way, so we found a way to bypass this,” explained Ramot. “We show pictures that are completely unrelated to social behavior that we thought participants would like, while still being relatively abstract – such as mechanical things or board games. Participants never really reach the ‘true’ strategy, because there is no explicit strategy for getting these networks to co-activate. That’s why the scanner is useful. When we do neurofeedback, we go directly to the networks.”

Resting state scans taken immediately following each of the four training sessions revealed increased functional connectivity between the two targeted brain regions in most participants with ASD. Participants without ASD did not show a significant change in functional connectivity between the targeted networks. Parents of participants with ASD noted improvements in their children’s social behavior that correlated with the increased connectivity. While falling short of an overall statistically significant change in sociability, the researchers demonstrated that change in the behaviorally-relevant networks correlated with a change in behavior.

These findings led the researchers to become interested in understanding how long the observed changes in connectivity would last. To find out, they had participants come back between 5 and 56 weeks following the main training session for a single shorter training session. Brain scans taken at this session indicated that that the changes in circuit connectivity seen following the main training were mostly preserved even over the longest time period. 

In the future, instead of fMRI, which is expensive, clinicians might be able to use more readily-available and cost-effective EEG technology to develop a more practical application of covert neurofeedback, said Martin. But more research is needed. “Researchers would first have to determine what EEG signals reflect the same activation of the social brain circuit networks corresponding to the fMRI signals,” he added.


Direct modulation of aberrant brain network connectivity through real-time NeuroFeedback.  Ramot M, Kimmich S, Gonzalez-Castillo J, Roopchansingh V, Popal H, White E, Gotts SJ, Martin A. Elife. 2017 Sep 16;6. pii: e28974. doi: 10.7554/eLife.28974. PMID: 28917059