fMRI and autism: Does imaging research miss the mark?

Recent studies have hinted at a potential role for functional MRI (fMRI) in the diagnosis of children and adolescents with autism. However, researchers may be putting the cart before the horse by focusing on imaging rather than biology, according to a Health Imaging interview with Nicholas Lange, PhD, a biostatistician at Harvard Medical School in Boston.

“Until its solid biological basis is found, any attempts to use brain imaging to diagnose autism will be futile,” Lange wrote in an editorial published in the November issue of Nature, in which he outlined several issues with the fMRI-based approach.

The first stumbling block is the lack of clinical knowledge about the biology of autism. Historically, in medicine, great advances are made when there is a breakthrough on the biological basis of the disorder, Lange said. Currently, there is no blood test, genetic test or other physical measurement tool that provides information about the biological basis of autism. fMRI observations can not be linked to a biological reference point.

Another issue relates to the difference between fMRI and other imaging modalities, such as PET, CT and MR spectroscopy. Other modalities provide results in units of physical measurement, such as diffusion rates of water, uptake of tracers or concentrations of metabolites.

In contrast, fMRI results are based on the percentage difference between an active state and an empirically determined baseline characterized by a lack of activity. The difference between those percentages is relatively small, or on the order of 1 to 3 percent, says Lange.

Another complicating factor relates to the multitude of possible interpretations of fMRI hot spots. One part of the brain can have several functions. The posterior temporal gyrus, for example, might be interpreted as the seat of music or the seat of language, depending on a researcher’s scientific bias, says Lange.

On a related note, multiple disorders—autism, dyslexia, learning disabilities—share common symptoms. “It isn’t surprising that any child with a language problem will show lower activity in the superior temporal gyrus because that is where language reception and expression are oriented. But scanning a child and showing low activity does not indicate the child is autistic. A child with dyslexia or other language problems would show similar phenomenology.”

Despite the shortcomings of fMRI, Lange sees a role for brain imaging in understanding autism. He cited several examples, including:

• Volumetric MRI has shown early abnormal enlargement of the brain in one in five children with autism;
• fMRI can show where people with autism focus during social interaction; and
• PET has identified differences in the distribution of serotonin and dopamine receptors in the brains of people with autism compared with typical controls.