Multisensory temporal integration in autism spectrum disorders

Abstract

The new DSM-5 diagnostic criteria for autism spectrum disorders (ASDs) include sensory disturbances in addition to the well-established language, communication, and social deficits. One sensory disturbance seen in ASD is an impaired ability to integrate multisensory information into a unified percept. This may arise from an underlying impairment in which individuals with ASD have difficulty perceiving the temporal relationship between cross-modal inputs, an important cue for multisensory integration. Such impairments in multisensory processing may cascade into higher-level deficits, impairing day-to-day functioning on tasks, such as speech perception. To investigate multisensory temporal processing deficits in ASD and their links to speech processing, the current study mapped performance on a number of multisensory temporal tasks (with both simple and complex stimuli) onto the ability of individuals with ASD to perceptually bind audiovisual speech signals. High-functioning children with ASD were compared with a group of typically developing children. Performance on the multisensory temporal tasks varied with stimulus complexity for both groups; less precise temporal processing was observed with increasing stimulus complexity. Notably, individuals with ASD showed a speech-specific deficit in multisensory temporal processing. Most importantly, the strength of perceptual binding of audiovisual speech observed in individuals with ASD was strongly related to their low-level multisensory temporal processing abilities. Collectively, the results represent the first to illustrate links between multisensory temporal function and speech processing in ASD, strongly suggesting that deficits in low-level sensory processing may cascade into higher-order domains, such as language and communication.

Keywords: audiovisual; autism spectrum disorders; multisensory integration; sensory processing; speech perception; temporal processing.

Figure 1.

Stimulus types and trial format. A, Trial format. B, Tool stimuli. C, D, Speech stimuli. B–D, top to bottom, Individual frames from dynamic visual stimuli, the auditory waveform, and the auditory spectrogram. Trials began with a 500–1500 ms intertrial interval, followed by a stimulus presentation. After the stimulus presentation, a response screen appeared, and the next trail began after participants responded.

Figure 2.

Mean individual multisensory temporal binding windows were calculated as depicted in A and averaged together for ASD (red) and TD (black) groups (B). Significant differences across diagnostic groups were seen only with speech stimuli. No significant differences were seen between ASD and TD groups with either unisensory-auditory (C) or unisensory-visual (D) presentations. Error bars indicate SE. *p < 0.05.

Figure 3.

Rates of perceived syllables with audiovisual, incongruent McGurk trials are depicted for the ASD (red) and TD (black) groups in A. Individuals with ASD perceived the bound “da” percept less frequently and the auditory “ba” percept more frequently than TD controls. No significant differences were found in congruent audiovisual (B), unisensory-auditory (C), or unisensory-visual (D) presentations. Error bars indicate SE. *p < 0.05.

Figure 4.

Individuals with ASD (red, A) showed significant negative correlations between the width of their TBW and their rate of binding audiovisual speech. Individuals with TD (black, B) showed no such relationship.