A Mu Rhythm View of Neurofeedback Training Effects of Face Processing Impairments in Autism

Background: Face processing plays an important role in understanding nonverbal cues in everyday social interactions. Research has shown deficits for those with autism spectrum disorder (ASD) in the perception of faces and reduced activation in the social brain, including the mirror neuron system (MNS). Previous studies have indicated that MNS activity could be assessed through power suppression of mu rhythms recorded over the sensorimotor cortex.

Objectives: The current study monitored MNS activity during face processing through mu suppression before and after neurofeedback training (NFT).  NFT is a learning methodology that involves operant conditioning of EEG frequency bands, including the mu rhythm.  It has been used for modifying cortical resonances and behavior through activity-dependent brain reorganization.

Methods: High functioning ASD children, along with matched, typically developing (TD) children were exposed to 20 weeks of NFT before and after exposure to a 1-back memory paradigm using static faces (angry, disgusted and happy faces).  We hypothesized that modifying the dynamics associated with mu rhythm leads to activity-dependent brain reorganization and therefore normalization of behavioral responses, including responses to emotional faces.

Results: Both the ASD and TD groups learned to modulate mu power across training sessions. Furthermore, the ASD group showed significant improvement in behavior as assessed by the Autism Treatment Evaluation Checklist (ATEC) after 20 weeks of training. Additionally, this group showed significant mu suppression while observing hand movement following training.  This was not shown prior to training.
The TD group showed similar absolute mu power over both hemispheres during building observation (baseline condition).  In contrast, the ASD group showed higher absolute mu power on the right compared to the left hemisphere before training. Following training, this lateralization in the ASD group disappeared.  Moreover, the TD group showed significant mu suppression over the right hemisphere to the presentation of disgusted and happy faces. However, there was no significant mu suppression in the ASD group during any conditions before or after training. Analysis of absolute mu power showed a highly negative correlation with age in the TD group. In the ASD group, no such relationship was observed (except for over the right hemisphere in the disgusted face condition) before training.  However, after NFT, mu power over the right hemisphere displayed a negative correlation with age during face observation, similar to that seen in TD children.

Conclusions: The behavioral results and the mu suppression findings confirm that NFT contributes positively in children with ASD as has been shown in previous reports. However, during face processing, mu suppression showed no improvement following training. One possible explanation is that the brain area to which NFT mainly applied was the hand area in the right hemisphere. This region may not be sufficient to influence the whole neuronal circuit responsible for face processing. According to the findings from absolute mu power comparison, the right hemisphere became more TD-like in the ASD children following training.