A couple of studies have come out recently that seem to support my theory of how neurofeedback benefits some individuals on the autistic spectrum.

In a study done at the Massachusetts General Hospital, magnetoencephalography (MEG) was used to study the brains of young men on the autistic spectrum and compare them with normal controls.  The results showed that both local and long-range connectivity was significantly compromised in the autistic individuals.  The results were so striking that the experimenters could predict with 90% accuracy whether a MEG scan was that of someone on the autistic spectrum. 

Another study out of the University of Montreal in Canada used magnetic resonance imaging (MRI) to study brain structure before and after neurofeedback treatment.  They demonstrated changes in both white and gray matter in the brain.  Some of these differences where clearly in the tracts that are associated with long-range connectivity in addition to localized cerebral structural changes.

I find it plausible to believe that neurofeedback would foster both more organized local activity and long-range connectivity in the brain of those with symptoms of autism.  This would account for those who show calmer behavior, better focus, and improved skills.

-Local and long-range functional connectivity is reduced in concert in autism spectrum disorders.  Sheraz Khan, et al.  Proceedings of the National Academy of Sciences, January 14, 2013.

-Neurofeedback Training Induces Changes in White and Gray Matter.  J. Ghaziri, et al.  Clinical EEG and Neuroscience, March 26, 2013.