9/18/25 Noah Fram: Rhythm as a prism for predictive coding in autism
CCN brown bag
Noah Fram, PhD
Postdoctoral Research Fellow, Otolaryngology
Date: Thursday, September 18, 2025
Time: 12:10-1:00pm
Location: 316 Wilson Hall
Rhythm as a prism for predictive coding in autism
Over the last decade, predictive coding theories of autism have become more prominent. The proliferation of these theories has, in turn, exposed a need for novel approaches to assess predictive processes in real time, rather than relying solely on surprisal-based outcome measures.
Musical rhythm is an ideal vehicle for such approaches, since it is ecologically valid, sensitive to experimental manipulation, and substantively implicated in several domains impacted in autism such as social interaction, communication skills, and motor function. Using eye-tracking in toddlers, we investigated the synchronization of visual attention to the beat of child-directed song. While present across diagnostic categories, this synchronization is stronger in nonautistic toddlers. When the beat is disrupted, diagnostic gaps in visual prediction emerge: nonautistic toddlers show reduced reliance on temporal patterns when their predictability is reduced, while autistic toddlers show a preference for a pattern-based predictive model than a duration-based one. We interpret these findings to support predictive coding theories involving priors less sensitive to stimuli in autism.
We also performed a simulation study, modeling predictions of an auditory rhythm’s tempo under different sets of cognitive parameters. These simulations revealed observable variation in the inferred tempo under nuanced variation in model parameters, implying that by fitting models to multiple data modalities, we can assess fine-grained heterogeneity in temporal prediction mechanisms in autism. Our ongoing work is further testing this finding using a combination of EEG, eye-tracking, and behavioral synchronization.