Temporal processing as a source of altered visual perception in high autistic tendency
Introduction
Visual perception in autism spectrum disorders (ASD) is characterised by a local attentional style, with a bias towards more local elements at the expense of the global components of an image. Atypical sensory perception is now included in DSM-5 (American Psychiatric Association, 2013), and this divergent style of visual processing has been linked to the poor processing of eye gaze and facial expressions observed in Autism Spectrum Disorder, as well as less efficient global form and motion processing (Dakin and Frith, 2005, Simmons et al., 2009). The prevalence, in early studies, of raised thresholds for motion coherence in neurodevelopmental profiles including ASD is suggestive of an early motion processing vulnerability, encapsulated in the dorsal stream hypothesis (Braddick et al., 2003). Due to the strong afference of magnocellular inputs to the motion areas MT/V5+, a common conclusion was that there is a processing abnormality in the afferent Magnocellular (M) visual system. The current view, however, is somewhat mixed (as set out in Section 1.2).
The Magnocellular (M) and Parvocellular (P) pathways work in parallel, processing low-level visual features. M cells recorded from the lateral geniculate nucleus (LGN) show high luminance contrast gain with responses saturating at high contrast while P cells show lower contrast gain, without saturation, as well as high red/green sensitivity (Kaplan et al., 1990). While the koniocellular pathway contributes to the blue–yellow chromatic system, its contributions to other aspects of perception are less clear (Robson and Kulikowski, 2012, Sun et al., 2012). The influence of visual afferents on cortical processing and behaviour has been most clearly determined through LGN lesion studies (Merigan and Maunsell, 1990, Schiller et al., 1991). Parvocellular LGN lesions strongly affect acuity and high spatial frequency processing, as well as chromatic sensitivity. Magnocellular LGN lesions reduce neural activity in primate area MT (considered as a starting point for the dorsal cortical stream) and also strongly reduce motion sensitivity at low stimulus contrast (Merigan et al., 1991). However, opposite direction discriminations could be made at contrast threshold, suggesting that the lesions did not directly act on motion direction perception (Merigan et al., 1991). There is a similarity in human psychophysics where the common neural substrate of different motion perception tasks has been questioned (Goodbourn et al., 2012). M and P lesions produce approximately the same degree of effect on neurons in V4 – considered to be a part of the ventral cortical stream (Ferrera et al., 1994). The M pathway's superior conduction velocity (a result of larger axon diameter), gives a latency advantage in primary visual cortex (V1) – often referred to as the'magnocellular advantage (Laycock et al., 2007). Such advantage allows for foreground/background segmentation information to be reported back to V1 (Hupe et al., 1998) prior to the arrival of P pathway information, possibly playing a role in object perception (Bullier, 2001).
Direct physiological evidence of M or P abnormalities in ASD is limited. We have extended sampling to the Broader Autism Phenotype (BAP) of individuals with normal IQ exhibiting high and low autistic tendencies, as measured by Baron-Cohen's (Baron-Cohen et al., 2001) Autism Spectrum Quotient (AQ). This scale was established to measure autistic tendency for both typically developing and ASD individuals with IQ in the normal range. VEP recordings using a flickering pseudo-random binary m-sequence stimulus show that at low contrast, the magnocellular system in High vs Low AQ individuals is less responsive for short latency (N50) first order kernel responses (Sutherland and Crewther, 2010). During rapid visual stimulation, groups with high autistic traits exhibit larger second order magnocellular responses indicative of poorer neural recovery after stimulation (Jackson et al., 2013, Sutherland and Crewther, 2010) for conditions of stimulation at high temporal frequency.
Pseudorandom contrast fluctuations of checkerboard patterns (Frey et al., 2013) showed no VEP difference between ASD and control children when presented centrally, though peripheral stimulation differences in amplitude and latency were noted.
The most common protocols for assessing motion perception in autism are coherent motion thresholds, biological motion sensitivity and direction selectivity sensitivity. While early psychophysical studies found raised motion coherence thresholds in those with ASD (Milne et al., 2002; Spencer et al., 2000), the literature is now more divided. Of the 18 articles investigating motion coherence thresholds in ASD and the broader phenotype, published from 2000 to 2014, 9 found significantly raised thresholds (Annaz et al., 2010, Atkinson, 2009, Milne et al., 2002, Pellicano et al., 2005, Robertson et al., 2012, Robertson et al., 2014; Spencer et al., 2000; Spencer and O’Brien, 2006, Tsermentseli et al., 2008), 4 found significant interactions between autism and motion task parameters (such as dot lifetime) (Jackson et al., 2013, Manning et al., 2013, Robertson et al., 2012, Sutherland and Crewther, 2010) while 5 found no significance (de Jonge et al., 2007, Del Viva et al., 2006, Greimel et al., 2013, Jones et al., 2011, Milne et al., 2006). The direct interpretation of motion coherence perception in terms of magnocellular function has been complicated because of a lack of attention to psychophysical methodology and parametric targeting (reviewed, Kaiser and Shiffrar, 2009; Simmons et al., 2009). Dot size and dot velocity differences across studies were tabulated by Simmons, but perhaps the parametric variation most crucial for segregating magnocellular from parvocellular function – dot contrast, has yet to be fully accounted for. Most studies have employed high contrast dots. Certainly, the current motion coherence literature has not attempted to probe magnocellular contributions by using fast, low luminance contrast, large, limited lifetime dots.
For biological motion sensitivity as a function of ASD spectrum, the results are also mixed. 14 Studies (Annaz et al., 2012, Annaz et al., 2010, Blake et al., 2003, Cook et al., 2009, Freitag et al., 2008, Koldewyn et al., 2010, Koldewyn et al., 2011, Kroger et al., 2014, Nackaerts et al., 2012; O'Brien et al., 2014; Pelphrey et al., 2007; Rutherford and Troje, 2012; Swettenham et al., 2013; van Boxtel and Lu, 2013) found impaired biological motion sensitivity in those with ASD while 4 had null findings (Jones et al., 2011, Murphy et al., 2009, Parron et al., 2008, Saygin et al., 2010).
Despite the majority of findings in favour of motion perceptual abnormality in ASD and the BAP, there are null findings for direction of motion contrast sensitivity (de Jonge et al., 2007, Koh et al., 2010) and motion perception suppression (Foss-Feig et al., 2013), as well as null findings for flicker contrast sensitivity (Bertone et al., 2005, Pellicano et al., 2005) at low temporal frequencies (10 Hz and 6 Hz). Also, in addition to the parametric caveats raised by Simmons et al. (2009), other between group differences in performance ability (e.g. in error rates, critical in PEST methodologies) (Crewther and Sutherland, 2009), eye movement patterns, and other attentional functions, need further investigation.
When flicker of a luminant stimulus reaches frequencies above the nervous system's ability to respond, this intermittent light stimulus appears as constant and the point at which this percept occurs is termed the flicker fusion threshold (McKendrick and Johnson, 2003). As the neuronal class with highest temporal cut-off frequencies in primate LGN, it is reasonable to assert that flicker fusion threshold in human represents the temporal capacity of the magnocellular system.
The separation of M and P responses via temporal thresholds (flicker fusion) is theoretically well founded, on the basis that primate studies show higher temporal frequency cut-offs for M-cells cf P-cells (Benardete and Kaplan, 1999, Kaplan et al., 1990, Usrey and Reid, 2000, Xu et al., 2002).
In monkey, temporal cut-off frequencies in LGN neurons are in the range 31–55 Hz for the M-pathway (Levitt et al., 2001, Merigan, 1980, Movshon et al., 2005) whereas the P-pathway's temporal capacity, as adjudged by chromatic fusion at the level of LGN neurons, is lower – 15–28 Hz (Levitt et al., 2001, Movshon et al., 2005, Schiller et al., 1991). In human, the critical flicker frequency measured psychophysically depends considerably on luminance and modulation depth (temporal contrast), but plateaus at around 60 Hz (Hecht and Schlaer, 1936), whereas estimates of colour fusion range from 11 to 18 Hz (Truss, 1957) depending on luminance (Lee et al., 1990).
The increased amplitude of magnocellularly generated VEP nonlinearities in those with High cf Low AQ, led to a prediction of lower flicker fusion frequencies in ASD (Jackson et al., 2013, Sutherland and Crewther, 2010), however a clinical research study has yet to eventuate. The current study aimed to delineate the temporal properties of magnocellular and parvocellular visual pathways in individuals scoring high and low on self-reported autistic traits. The psychophysical performance of the magnocellular system across groups was assessed using a luminance flicker fusion task focussing on low temporal contrasts, while the temporal performance of the parvocellular system was assessed using a red–green colour fusion task. The two experiments were designed so that separate statements on M and P processing in high versus low autistic tendency groups could be made. Given the reliability from previous experiments (Jackson et al., 2013, Sutherland and Crewther, 2010) of the relation between abnormalities in nonlinear VEP and in visual perception, resulting in strong discriminants for AQ class, we predicted that the High AQ group would show diminished fusion thresholds compared to Low AQ, particularly at low contrast.
Section snippets
Participants
Upon approval from the Swinburne University Human Research Ethics Committee, participants were recruited through an online version of the Baron-Cohen et al. (2001) Autism Spectrum Quotient (AQ), with possible scores ranging from 0 to 50 (Opinio, Object Planet Inc., Oslo, Norway). 108 Participants completed an online version of the AQ with those scoring 13 and below selected for a Low AQ group (n=22), and those scoring 18 and above selected for the High AQ group (n=17). Informed consent was
Results
Flicker fusion thresholds for each of the luminance contrast levels showed a pattern of increasing frequency with contrast. This effect was most marked for the 5% contrast value, while at higher contrast response appeared to saturate. This behaviour was exhibited for both the High AQ and Low AQ groups (see Fig. 1) and it is evident that overall, the Low AQ group showed higher flicker fusion values than the High AQ group, although there tended to be a convergence at high contrast levels.
Testing
Discussion
When responding to M-targeted stimuli (achromatic, low contrast flicker), individuals with High AQ show lower flicker fusion frequencies than those with Low AQ. In addition, although P-targeted isoluminant colour fusion did not generate a significant difference between High and Low AQ groups, a significant interaction between AQ group and performance on the two tasks was found when low contrast flicker fusion and red/green colour fusion were analysed together.
The flicker fusion frequencies for
Acknowledgments
The authors acknowledge the support of the National Health and Medical Research Council of Australia through Project Grant (APP1004740).
References (74)
Impaired recognition of emotions from body movements is associated with elevated motion coherence thresholds in autism spectrum disorders
Neuropsychologia
(2009)- et al.
Normal and anomalous development of visual motion processing: motion coherence and dorsal-stream vulnerability
Neuropsychologia
(2003) Integrated model of visual processing
Brain Res. Brain Res. Rev.
(2001)- et al.
Reduced sensitivity to minimum-jerk biological motion in autism spectrum conditions
Neuropsychologia
(2009) - et al.
The more he looked inside, the more piglet wasn't there: is autism really blessed with visual hyperacuity?
Biol. Psychiatry
(2009) - et al.
Vagaries of visual perception in autism
Neuron
(2005) - et al.
Spatial and motion integration in children with autism
Vis. Res.
(2006) - et al.
Perception of biological motion in autism spectrum disorders
Neuropsychologia
(2008) - et al.
The temporal dynamics of coherent motion processing in autism spectrum disorder: evidence for a deficit in the dorsal pathway
Behav. Brain Res.
(2013) - et al.
The dynamics of primate retinal ganglion cells
Prog. Brain Res.
(2001)