Despite several decades of research effort, autism remains a behaviorally-defined disorder without any biological markers. New imaging methods aim to bring us closer to the goal of a biologically-based screening test for ASD.
A study1 from Declan Murphy, Ph.D. and colleagues used magnetic resonance (MR) structural brain imaging in combination with sophisticated machine learning methods designed to classify based on small differences in five different local measures of brain size, such as thickness of the cerebral cortex. In their study of adults with ASD and individuals with ADHD or normal development, this method was found to classify the individuals with ASD in their study with as much as 90 percent accuracy, which is impressive given that the scan only takes 15 minutes.
Although the literature is full of reports of specific brain areas being larger, smaller, more or less active in individuals with autism compared with typically developing peers, this report offers a different message. The researchers did not find, nor do they believe one should expect to find, one thing that distinguishes the brains of those with autism. The reason is that we are learning that there are many different paths that lead to what we behaviorally define as ASD and each of those paths may correspond to distinct differences in the brain.
This report, as well as those of other groups performing similar image-based diagnostics later in the year, garnered considerable media attention. The authors caution that the technology is still far from substituting for traditional diagnostic measures. It is possible that early changes in brain anatomy might serve as subtle markers of risk for ASD, however. Scientists are currently using brain imaging to assess young infants who are at risk for ASD to see if these measures can provide a new method for early detection. Given the speed and availability of scanners, perhaps methods such as these (and others, see voice analysis story) can assist in screening for ASD.
There is also the hope for imaging as a concrete method of measuring therapeutic success. Will behavioral improvements also lead to measureable changes in MR images? Time will tell, but 2010 should be remembered as a year when imaging biomarkers for ASD made their mark.