Normal behavior requires that brain regions interact with one another, and these interactions depend on the task; they change as attention shifts and as an action is planned and executed. Jonathan Victor and his colleagues at Weill Cornell Medical College in New York have developed a new technology to delineate these moment‐to‐moment changes in brain activity patterns so that they can be compared between typically developing children and those with autism.

Magnetic resonance brain imaging (MRI) and functional MRI have been used to show that the connections between brain regions and the way brain regions interact are abnormal in autism. Neither method, however, can be used to observe brain activity patterns that last less than a second. Since normal behavior requires rapid sequences of brain activity patterns, another approach is needed to delineate them.

In a pilot study, Victor’s team examined the feasibility of an approach that is based on computational analysis of the electroencephalogram, which is recorded by electrodes placed on the scalp. The recordings are analyzed to extract signatures of correlated patterns of activity between brain regions, and these patterns are further analyzed to determine their relation to a behavior (such as viewing a familiar face vs. an unfamiliar face).

The study demonstrated that this approach may be feasible, even in infants. That is, the analysis identifies patterns of interaction between brain regions that depend on the task. The crucial next step is to determine whether these patterns of interaction can be used to distinguish between typically developing children and children who ultimately are diagnosed with autism. If so, the approach may provide a route to early diagnosis and intervention.