Similarities in Autistic and Neurotypical visual-haptic perception when making judgements about conflicting sensory stimuli

A number of studies have shown that multisensory performance is well predicted by a statistically optimal maximum likelihood estimation (MLE) model. Under this model unisensory estimates are combined additively and weighted according to relative reliability. Recent theories have proposed that atypical sensation and perception commonly reported in autism spectrum condition (ASC) may result from differences in the use of reliability information. Furthermore, experimental studies have indicated that multisensory processing is less effective in those with the condition in comparison to neurotypical (NT) controls. In the present study, adults with ASC (n=13) and a matched NT group (n=13) completed a visual–haptic size judgement task (cf. Gori et al., 2008) in which participants compared the height of wooden blocks using either vision or haptics, and in a dual modality condition in which visual–haptic stimuli were presented in size conflict. Participants with ASC tended to produce more reliable estimates than the NT group. However, dual modality performance was not well predicted by the MLE model for either group. Performance was subsequently compared to alternative models in which the participant either switched between modalities trial to trial (rather than integrating) and a model of non-optimal integration. Performance of both groups was statistically comparable to the cue-switching model. These findings suggest that adults with ASC adopted a similar strategy to NTs when processing conflicting visual–haptic information. Findings are discussed in relation to multisensory perception in ASC and methodological considerations associated with multisensory conflict paradigms.