Brain scans to investigate autism causes may be misleading

By Deborah Kotz, Globe Staff

Six months ago, a leading autism researcher declared to USA Today that autism brain science “has moved stunningly fast” over the past year or two with scientists racing to prove through functional magnetic resonancing imaging or fMRI that autism is caused by a miswiring in the brain.

“We’ve fundamentally moved around a corner where we will move much faster now,” Kevin Pelphrey, an associate professor of child psychiatry at the Yale School of Medicine’s Child Study Center, said in the upbeat article, with the tantalizing promise of new treatments to fix brain wiring.

Pelphrey, who has an autistic son, doesn’t seem so optimistic now. In an opinion piece he co-authored this week in the journal Nature, he called into question the reliability of the fMRI images on which the theory is based.

Those images in previous research studies suggested that distant regions of the autistic brain don’t communicate effectively with each other -- as they do in the brains of those who don’t have autism. This could explain why autistic individuals often don’t know how to read emotional cues or express deep feelings for friends or loved ones.

“This ‘connectivity hypothesis’ claims that the social and cognitive abnormalities in people with autism can be explained by a dearth of connections between distant regions of the brain,” wrote Pelphrey and MIT graduate student Ben Deen in Nature.

But recent studies have found that when people who don’t have autism move their head while undergoing fMRI, the connectivity patterns that result resemble patterns seen in the brains of those with autism.

“That’s a sobering discovery,” Pelphrey and Deen wrote. “It means that a major source of evidence for a leading hypothesis on autism, and one that several research teams have pursued for years, may arise from an artefact.”

The two proposed several solutions for minimizing this bias such as “scrubbing” the images to remove periods of high motion during the scans or to determine, on average, how much the head moves in each study participant over the course of a scan and to take that into account when computing measurements.

Software tools are available that would allow for this motion to be taken into account and a reanalysis of the data. “Revisiting fMRI studies with these approaches,” the authors concluded, “would help establish whether there really is a connectivity deficit in the brains of people with autism.”