Talk: No evidence for reduced precision-weighting of prediction errors in autistic adolescents' ERP…
Speaker: Emma Ward, she/her, Freie Universität Berlin (grid.14095.39)
Title: No evidence for reduced precision-weighting of prediction errors in autistic adolescents' ERPs or behaviour during adaptation
Emcee: Alberto Antonietti
Backend host: Ali Rigby
Details: https://neuromatch.io/abstract?submis...
Personal website: http://www.sites.google.com/view/emma...
Twitter: / emma_cogdev
Presented during Neuromatch Conference 3.0, Oct 26-30, 2020.
Summary: Predictive Processing accounts of autism posit that autistic individuals are less biased by expectations than those without autism when interpreting incoming sensory information, most recently formulated as stronger precision-weighting of prediction errors. Since precision-weighting is fundamental to all information processing, any differences between autistic and non-autistic individuals should be domain general and observable in both behaviour and brain responses.
The current study uses gaze-direction adaptation as a test case for the theoretical predictions. The current combined EEG and behavioural study with eye-tracking coregistration investigates whether increased precision-weighting of prediction errors is evident through smaller adaptation after-effects in the behaviour and ERP responses of autistic adolescents when compared to non-autistic peers.
Participants viewed stimuli used in previous studies (Jenkins et al., 2006; Pellicano, Ewing & Rhodes, 2013; Ward, Braukmann, Buitelaar & Hunnius, 2020) and responded with button presses whether they perceived the gaze direction as left, direct, or right. Ordinal multi-level modelling showed that both autistic and non-autistic adolescents responded consistent with behavioural adaptation (β = -1.49, SE = 0.16, z = -9.32, estimated p ＜ .001), with no effect of group (β = 0.41, SE = 0.30, z = 1.36, estimated p = 0.16) or interaction between group and experimental condition (β = -0.18, SE = 0.18, z = -1.01, estimated p = 0.31).
Cluster-based permutation testing of the ERP responses did not show the expected larger reduction in the N170 or P2 components in response to adapted than unadapted stimuli, and no group difference was detected.
The current study is one of the first to take advantage of neuroscientific methods to test specific predictions from theoretical accounts of Predictive Processing in autism. Combined with the few other available studies, the current findings raise challenges for the theory and suggest no fundamental difference in precision-weighting of prediction errors in autism.