Binding of sights and sounds: age-related changes in multisensory temporal processing

Abstract

We live in a multisensory world and one of the challenges the brain is faced with is deciding what information belongs together. Our ability to make assumptions about the relatedness of multisensory stimuli is partly based on their temporal and spatial relationships. Stimuli that are proximal in time and space are likely to be bound together by the brain and ascribed to a common external event. Using this framework we can describe multisensory processes in the context of spatial and temporal filters or windows that compute the probability of the relatedness of stimuli. Whereas numerous studies have examined the characteristics of these multisensory filters in adults and discrepancies in window size have been reported between infants and adults, virtually nothing is known about multisensory temporal processing in childhood. To examine this, we compared the ability of 10 and 11 year olds and adults to detect audiovisual temporal asynchrony. Findings revealed striking and asymmetric age-related differences. Whereas children were able to identify asynchrony as readily as adults when visual stimuli preceded auditory cues, significant group differences were identified at moderately long stimulus onset asynchronies (150-350 ms) where the auditory stimulus was first. Results suggest that changes in audiovisual temporal perception extend beyond the first decade of life. In addition to furthering our understanding of basic multisensory developmental processes, these findings have implications on disorders (e.g., autism, dyslexia) in which emerging evidence suggests alterations in multisensory temporal function.

Fig. 1

Simultaneity Judgment Paradigm. Profile of the temporal relationship between stimuli used in assessments.

Fig. 2

Left: Bar graph displays group difference in multisensory temporal window size (i.e., distribution width at 3/4 maximum) for children (C) and adults (A). Error bars represent ±one standard error of the mean (SEM). Middle/Right: Scatterplot of individual window sizes for each child (circles) and adult (squares). Solid lines represent mean values and dotted lines denote the area encompassing ±1 SEM.

Fig. 3

Graph displaying windows for representative child and adult subjects. Note that the discrepancy in window size between subjects is primarily comprised of differences in the left side of the distribution.

Fig. 4

Grand averaged group data distributions. Children were significantly more likely to report trials as simultaneous at moderate and long SOAs in which the auditory stimulus preceded the ring flash. Error bars represent ±1 SEM, **sig Bonferroni p < 0.0033.

Fig. 5

Smoothed pseudocolor plots depicting mean probability of simultaneity judgment for each group on consecutive trials (top left) where warmer colors indicate a higher probability of simultaneity report and cooler colors represent a higher likelihood of asynchrony report. Trials are aligned from first (top) to last (bottom). Contrast plots reveal differences in behavioral report between groups (bottom) and across assessments (top right). For these contrast plots warmer and cooler colors represent positive and negative remainders, respectively; green is neutral or no change. Note the consistency of responses across assessments for both groups (top right) and the group difference in the simultaneity report at negative SOAs (bottom).