Evaluating Multimodal Driver Displays of Varying Urgency for Drivers on the Autistic Spectrum

Friday, May 12, 2017: 5:00 PM-6:30 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
L. S. Shim1, P. Liu2, I. Politis3, P. Regener4, S. Brewster5 and F. Pollick6, (1)School of psychology, University of Glasgow, Glasgow, United Kingdom, (2)School of psychology, University of Glasgwo, Glasgow, United Kingdom, (3)University of Cambridge, Cambridge, United Kingdom, (4)Glasgow University, Glasgow, UNITED KINGDOM, (5)School of Computing Science, University of Glasgow, Glasgow, United Kingdom, (6)School of Psychology, University of Glasgow, Glasgow, United Kingdom
Background: Individuals with ASD demonstrate alterations in sensory processing within different modalities and uncommon responses to sensory stimuli across multiple sensory domains. It is important to note that abnormal sensory processing does not necessarily mean worse performance, as there is evidence of enhanced perceptual functioning in autism. Recent studies show differences of multisensory perception between typically developed and ASD individuals that potentially can lead to difficulties in the integration of multiple sensory signals in everyday situations, such as driving. This is relevant since diverse modes of sensory information can be used to warn a driver and multisensory warnings incorporating two or more modalities show promise in their effectiveness.

Objectives: To evaluate the effectiveness of multisensory warning signals, designed to indicate different levels of urgency, in individuals on the autism spectrum.

Methods: The study involved 20 male (10 typically developed and 10 ASD) participants. The ASD group was defined by having an AQ score over 26 and had an average AQ score of 40.3. Two experiments were conducted: (1) Experiment 1 examined perceive urgency and annoyance with the warnings and (2) Experiment 2 measured recognition accuracy of the level of urgency and the reaction time during a simulated driving task.

Warning design: The warnings used in this study were similar to those of a previous study (Politis, I. et al (2013)). Totally, twenty-one different signals were obtained from the combination of: (1) three levels of designed urgency (LDU) and (2) seven sensory conditions which included unimodal, bimodal and trimodal combinations of audio, visual and tactile.

Results: The results of Experiment 1 revealed that there was no difference between TD and ASD groups in the perceived urgency of the warning signals, though the autism spectrum group reported lower annoyance with the signals than TD group ((F(1,58) = 13.22, r = 0.43, p < 0.01) ). Experiment 2 showed that while both groups exhibited high accuracy in correctly reporting urgency level, the autism spectrum group performed better (Q(1) = 11.80, p < 0.01). Moreover, the fastest overall reaction times obtained were by the autism spectrum group when the warning included a visual component, with vision alone (F(1,57) = 5.62, r = 0.30, p < 0.001) producing the quickest response.

Conclusions: This study compared how typically developed people and individuals with ASD responded to a set of multimodal combinations to alert drivers to events of varying urgency. Two group difference were found in that the ASD group reported the warning signals to be less annoying than the typically developed participants and the ASD group showed an advantage in response time when the warning included a visual modality, in particular for the vision-only condition. These results highlight that while there are similarities, substantial performance differences exist between typical and autism spectrum individuals, suggesting that consideration of these differences can contribute to the design of effective warning signals.