Visual Spatial Channel Bandwidth Varies with Autistic Trait Level

Friday, May 12, 2017: 5:00 PM-6:30 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
M. H. Laurie1 and D. R. Simmons2, (1)College of Medicine & Veterinary Medicine, University of Edinburgh, Edinburgh, United Kingdom, (2)School of Psychology, University of Glasgow, Glasgow, United Kingdom of Great Britain and Northern Ireland

Visual processing in autism is the topic of ongoing debate (Simmons et al, 2009). One of the most basic measures of visual performance is contrast sensitivity - the detectability of small changes in light intensity. Previous studies have reported lower, higher and equal sensitivities in autistics, relative to controls, depending on the stimulus used (Adams et al, 2010; Koh et al, 2010). Stimulus complexity potentially also affects contrast sensitivity differentially (Bertone et al, 2005), and autistics have greater difficulties when perceiving “masked” stimuli (Greenaway, Davis & Plaisted-Grant, 2013). These difficulties are potentially linked to neural noise theories of autism (Simmons et al. 2009; Dinstein et al, 2015), given physiological research that relates increased noise to increased channel bandwidths (i.e. information taken in) in neural processors, which is in turn a suggested explanation for sensory filtering difficulties in autism (Plaisted et al. 2003). Furthermore, sensory difficulties vary with autistic trait level, as measured by the Autism Spectrum Quotient (AQ; Baron-Cohen et al. 2001; Robertson & Simmons, 2013).


To investigate the potential link between autistic trait level and the bandwidth of visual spatial channels.


38 participants completed the study, which involved two visual tasks and completing the AQ. Both tasks were presented in a standard two-interval forced choice paradigm. Task 1 was a contrast detection task, which measured the thresholds (i.e. minimum contrast) for detecting a sinusoidal variation in luminance (Gabor patch) centred at three spatial frequencies (SFs): 0.5, 1, and 2 cycles per degree(cpd). In task 2, participants were asked to detect a 1-cpd target stimulus in the presence of three different masks (0.5-, 1-, and 2-cpd) – whilst the mask contrast was held constant (10x detection), the target stimulus contrast was adjusted to measure thresholds. The size of the mask varied with SF to counteract stimulus spatial bandwidth differences. A ratio of the thresholds (masked vs. unmasked) for the target stimulus was calculated to estimate the SF processing bandwidth at 1-cpd.


Groups of individuals with high (N = 18; mean AQ = 24) and low (N = 20; mean AQ = 9) levels of autistic traits were formed by a median split. Bootstrapped t-tests (iterations = 1,000) revealed group differences in both tasks. For unmasked stimuli, participants with higher levels of autistic traits had significantly higher contrast sensitivities (lower thresholds) at 0.5-cpd (t=-47.65,p<.001) and 1-cpd (t=-8.28,p<.001) and lower sensitivity (higher thresholds) at 2-cpd (t=53.87p<.001). Additionally, the threshold ratio (estimation of SF bandwidth at 1-cpd) was significantly higher for those with higher levels of autistic traits at all SFs (t=-34.62;-11.43;-58.66,p<.001).


These data suggest that (1) contrast sensitivity varies with autistic trait level, and (2) those with higher levels of autistic traits have broader spatial frequency channels, in line with theoretical predictions (Plaisted et al. 2003; Simmons et al. 2009), thereby lending support to neural noise theories of altered perception in autism, and meriting further testing in the autistic population.