Autism screening by YouTube?

A new paper by Vincent Fusaro and colleagues offers a potentially intriguing solution to the question of how to screen the many thousands of infants born every year for some of the signs and symptoms of autism spectrum disorder (ASD) – screening YouTube videos by non-expert assessment.

Authors conducted a preliminary study on the effectiveness of non-clinical raters to apply scoring measures from one of the gold standard assessment tools for autism – the autism diagnostic observation schedule (ADOS) – to a selection of public videos of children uploaded to the popular video sharing website; some videos self-reported to be indicative of an ASD and others not.

Traditionally administered by trained raters using a variety of scenarios and prompts to elicit certain types of behaviours with scoring based on a standardised set of criteria, the ADOS has dominated autism assessment for many years. Recent revisions to the algorithm used by ADOS to determine whether autism or an ASD may be present in line with the revised DSM-5 criteria for the condition, demonstrate how adaptive this instrument has proved to be.

Fusaro and colleagues reported some accuracy in the ratings provided by their four non-clinical volunteers who analysed the videos in terms of determining possible autism/ASD or not. Scored results also showed a relatively high level of overlap with a blinded expert analysis of the videos.  Fusaro says

Our results also demonstrate the potential for video-based detection of autism in short, unstructured home videos and further suggests that at least a percentage of the effort associated with detection and monitoring of autism may be mobilized and moved outside of traditional clinical environments.

This was a preliminary study which requires further replication. Because some of the non-clinical raters used in the trial were also used to locate the videos from YouTube, there may have been some bias in the final reports provided by raters. The reliance on questions from only one module of the ADOS normally reserved for children who have little or no phrase speech is also another shortcoming of the Fusaro trial.

Nevertheless, the potential implications from these results, if replicated, could be profound. Not least the implication for non-expert screening of home videos uploaded to video sharing sites directing healthcare providers to those children who may require a more thorough examination for the presence of autism or ASD.


* Fusaro VA. et al. The Potential of Accelerating Early Detection of Autism through Content Analysis of YouTube Videos. PLoS ONE. 2014; 9: e93533.

Further commentary on this study can be found at: