Valuable videos

Accurately tracking repetitive movements in people with autism is a challenge. But characterizing these issues is incredibly important for more effectively diagnosing the disorder, for better understanding the neurological basis of the motor issues often seen in autism, and potentially for assessing new treatments.

For example, more sensitive measures for measuring repetitive movement could help doctors distinguish between Rett syndrome, a genetic disorder that is often linked to autism, and classic autism — especially when genetic testing for Rett is not an option.

To fill this gap, my collaborator Paul Greene and I have developed a fine-grained video coding system to identify and characterize the kinematic features — which include velocity, acceleration and trajectory — of these repetitive movements. In a new study, published 3 January in Frontiers in Integrative Neuroscience, we showed that this approach can provide a more nuanced description of repetitive behavior than questionnaires or clinical rating scales do.

We first described our scoring system in 2009, using it to track the prevalence and characteristics of repetitive movements in developmentally impaired preschool children with and without autism. We found that children with autism are more likely to have repetitive movements than those with other developmental disabilities.

A second study, published in July, showed that our video coding system can identify striking differences between hand stereotypies in children with autism and those in children with Rett syndrome.

Our approach is unique because it is entirely based on a detailed characterization of each movement. We purposefully kept away from interpreting what causes these abnormal movements or what functions they might provide. As such, we provide clinicians and researchers with a systematic, descriptive approach for classifying these distinct movements.

The goal of our study is to emphasize the value of video analysis compared with widely used questionnaires given to parents and caregivers, or clinical rating scales. Our method provides real numbers, such as frequency, velocity and duration of different movements during a fixed period of time, which can be used in quantitative statistical analysis. Rating scales, in contrast, are qualitative measures over longer period of time and are less precise.

Using selected video clips, we are able to draw attention to the variability within a category of repetitive movements, such as hand movements. We can also point out differences and similarities of particular characteristics, such as the duration, among different categories of repetitive movements.

Another advantage of this approach is to be able to return to the videos and reanalyze the data with a different research question in mind, such as do the hand stereotypies involve objects, for how long and with what type of object?

So far, most studies of repetitive movements in autism have used parent or caregiver questionnaires, such as the Repetitive Behavior Scale-Revised. These are less time-consuming and allow for larger data collection and broader contexts than does our video analysis.

However, these instruments do not provide the precise information necessary to study these debilitating movements in depth. This type of fine-grained detail could be especially useful for assessing how age or different treatments affect repetitive movements.

Ultimately, these types of video assessments, combined with automated computerized recording of the movements, such as with lightweight accelerometers, will allow for much-needed investigations that may unravel the physiological basis of these stigmatizing and disrupting motor behaviors.

Sylvie Goldman isco-director of the Human Clinical Phenotyping Core of the Rose F. Kennedy Intellectual and Developmental Disabilities Research Center at the Albert Einstein College of Medicine in New York.