Scans that reveal flaws in the brain’s circuitry may one day be used to diagnose autism, researchers from McLean Hospital say in a report published last week.

Nicholas Lange and colleagues write in the journal Autism Research that their neuroimaging studies were able to distinguish people with autism from control subjects 94 percent of the time using an MRI that took about 10 minutes. The scans focused on three parts of the brain that are known to be impaired in autism: areas involving language, social and emotional function. currently autism is diagnosed after hours of interviews with parents and observation of a child, usually 2 to 3 years old, who does not appear to be developing normally.

“we all want to know what we can and cannot change about the disorder and how to improve the lives of people with autism,’’ Lange said in an interview. “this is a major step toward that goal and more research needs to be done before it is ready for clinical use.’’

For the study, the researchers examined the brains of 30 people from 7 to 28 years old with high-functioning autism and compared them to the brains of 30 similar people with normal development. The MRIs showed differences between the two groups in how tiny fibers between nerve cells were organized.

Out of the 30 people with autism, 28 were correctly identified based on scans that suggested information could not flow as easily through faulty neural connections. Out of the 30 controls without autism, three were incorrectly labeled.

In an interview, Lange made an analogy to the electrical wiring between a lamp and a wall socket. If the insulation on the wire is faulty, electricity cannot flow properly. In the same way, if the fibers between nerve cells in the brain are not well insulated with myelin, or if the fibers are disorganized, signals cannot travel as well.

They did not see differences between the brains of older and younger people with autism, hinting that problems in brain development may occur before age 7. Lange said further research will look at younger people and at people with more severe forms of the disorder.

An autism specialist not involved in the study also said more research is needed before the technique can be considered for use as a diagnostic tool. Dr. Stewart H. Mostofsky, director of the Laboratory for Neurocognitive and Imaging Research and medical director of the Center for Autism and Related Disorders at the Kennedy Krieger Institute in Baltimore, noted that the biggest differences in the brain scans showed up in areas central to language.

“The critical issue is not distinguishing children with autism from typically developing children but distinguishing children with autism from children with developmental language disorders or language impairment,’’ he said. “it is certainly possible that the primary [difference] is not necessarily autism presenting but rather the language impairments that are associated with autism.’’ ELIZABETH COONEY