Decreased Habituation to Naturalistic Stimuli in Autism

Background:  Habituation plays a fundamental role in dynamically changing saliencies of environmental stimuli. Reduced habituation reduces stimulus suppression and immerses an individual in an unrelentingly salient, and hence potentially overwhelming, world. It would also compromise the ability to 'detach' attention from a given stimulus. Thus, an impairment in habituation can, in principle, help account for sensory hypersensitivities as well as sustained interest in a limited set of stimuli, two prominent correlates of autism.

Objectives:  In work that we presented last year, we showed that habituation to low-level sensory stimuli (metronomic auditory sequences) is reduced in autistic participants compared with neurotypical controls. This study aims to extend our previous work to more naturalistic stimuli.

Methods:  We investigated the habituation response to repeated presentations of naturalistic video stimuli, using a measure of galvanic skin response (GSR). Participants were shown 1 minute of neutral baseline stimuli, followed by five repetitions of an exciting 30 second video clip. The convenience and ease of collecting GSR signals allowed us to enlist a wide range of participants on the autism spectrum, including individuals with minimal language abilities.

Results:  We found consistent differences in the time-course of GSR signals from ASD participants compared with age-matched neurotypical controls. For the ASD group, the slopes of GSR signals over time showed a flat or increasing trend, indicating compromised habituation to the repeated video stimuli. Conversely, for neurotypical participants, GSR signals over time consistently showed a decreasing trend, indicating habituation to the stimuli. Further computational analysis of these results, using a machine learning classifier, showed differentiation between the neurotypical and ASD groups, based on the slope of the line of best-fit to the GSR signal over time. The classifier successfully differentiated between these two groups with ~75% accuracy.

Conclusions:  These results can help to explain several commonly observed features associated with individuals with ASD: restricted and repetitive interests, and associated behaviors such as repetitive intake of media such as videos and music. Furthermore, these results provide support for a broader hypothesis of reduced predictive abilities in autism. Given the simplicity and ease of administering the protocol, we believe that there are possible applications in the measurement of habituation profiles in infancy, which may help with early identification of autism. Furthermore, the results could inform the development of learning environments and interventions best suited to the unique sensory profiles of individuals with ASD. Importantly, we argue that these findings are not due to extraneous factors such as general arousal levels, agitation, movement, or inattention, although further study is required. Future investigations will assess the specificity of the reduced habituation profile in autism compared with other conditions, its occurrence in other sensory modalities, and with naturalistic stimuli in modalities other than vision. At this stage, however, the consistency of our results across a wide range of ages and cognitive abilities points towards reduced habituation as a possible endophenotype that may yield a better understanding of some key aspects of the autism phenotype.