The process of diagnosing autism is complex, subjective, and often limited to only a segment of the population in need. With the recent rise in incidence to 1 in 88 children, the need for accurate and widely deployable methods for screening and diagnosis is substantial.

I recently gave a talk at the Autism Consortium Symposium held at Harvard Medical School on October 25 and attended by more than 600. In the talk I described my recent work on using artificial intelligence & machine learning to abbreviate the best-of-breed tools used for behavioral recognition of autism spectrum disorders. These approaches have helped me to identify a substantially shorter instrument that takes only minutes to complete but that achieves the same accuracy as the gold-standard tools.

we have improved our short questionnaire for screening and diagnosis of autism, and through the help of over 2000 participants have measured it to be over 98% accurate. The questionnaire contains only 7 questions and takes minutes to complete, 93% shorter than the standard tool and over 2 hours faster. If you are a careprovider for someone with autism, you can contribute to our research by answering the questions in this questionnaire here.

We recently designed a short, but effective behavioral questionnaire to diagnose autism. We have wrapped that questionnaire into an online survey in hopes that the entire community of caregivers will help us to validate the accuracy. So far, over 250 people have participated! We hope to bring that number to 500 by the end of the week -- April 1, 2011. Please help us spread the word.

More information: