It has been established that children and adolescents with autism spectrum disorder (ASD) show a wide range of abilities to use spoken words and establish interactive conversations. Automatic measurement of such abilities in naturalistic environments would greatly facilitate assessment and monitoring of such individuals. If sufficient accuracy could be achieved with such automated tools, large enough studies could be performed to draw statistically meaningful conclusions.

In the current project, Mark Clements is testing the use of 16-hour home audio recordings using the LENA (Language ENvironment Analysis) device in older children and adolescents with ASD. Specific enhancements to the existing LENA analysis platform include the ability to diarize recordings for subjects aged 5 through 13, to detect nonverbal vocalizations such as laughter and whining, to identify child-directed speech and to determine when questions are posed. Other higher-level descriptors involve extraction of affect, computation of conversational interaction measures, detection of crosstalk and interruption events, and identifying emotional outbursts. Clements has demonstrated that subject-specific diarization based on a small amount of hand-labeled data yields acceptable accuracy. However, he is now assessing a newly developed system based on i-vectors, specifically designed for the environment at hand, which requires no such labeling at the onset.