There are a series of visual characteristics and visual processing impairments and abnormalities reported for children with autistic symptoms.

There is evidence for abnormal visual processing of colour informationin autistic children. This manifests as poor chromatic discrimination (Heaton, Ludlow, & Roberson, 2008)(Franklin, Sowden, Burley, Notman, & Alder, 2008)(Franklin et al., 2010) but good colour memory (Heaton et al., 2008). Although originally outlined as test of executive functioning, autistic children, children with autism also show less interference effects for colour naming compared with typically developing children on the classic Stroop test e.g. (Simmons et al., 2009), which is probably related to verbal information processing. Some theories suggest that there is an association between verbal ability and perceptual discrimination (Heaton et al., 2008) such that normally developing children apply verbal labels to colours, and are therefore distracted when presented with colours carrying non-typical labels. Children with language difficulties, as typically expressed in autism spectrum disorders (ASD), however are thought to remember colours perceptionally, in other words they do not encode colour verbally, and may therefore have less interference effects compared to typically developing children (Simmons et al., 2009). A considerable body of research reports that children with autism have also impairments in facial recognition and in the understanding of facial emotional expressions (Simmons et al., 2009). This might be related to the finding that ASD children appear to process and/or attend the eye region of faces less effectively than typically developing children e.g. (Rutherford, Clements, & Sekuler, 2007)(Spezio, Adolphs, Hurley, & Joseph Piven, 2007)(Simmons et al., 2009) and more likely fail to form view-invariant face representations (Wolf et al., 2008). In addition, the ability to perform face-eye region special discriminations might be coupled to verbal intelligence (Rutherford et al., 2007). Furthermore, several studies have observed an enhanced ability to perform visual search tasks in autistic patients such as the ability to disembed a target figures from a complex background in the "Embedded Figures Test". Likewise, autistic children show an superior performance in perceptional grouping as well as mental rotation tasks (e.g. Block-Design subtest of the WISC-III; (Shah & Frith, 1993)(Venter, Lord, & Schopler, 1992)(Goldstein, Beers, Siegel, & Minshew, 2001). Latter association was also observed for autistic-like traits (Stewart, Watson, Allcock, & Yaqoob, 2009).

Aim: Assuming the dimensionality of autism, the proposed fellowship aims to understand the mechanisms linkingvisual functioning to autistic symptoms using a general population sample in order to provide guidance for strategies aiming at intervention and treatment of autistic symptoms.

1) We aim to study chromatic information processing, interference processes between colour perception and colour naming, the understanding of facial expressions as well perceptional grouping and mental rotation tasks with respect to autistic symptoms. Findings will be used to create a profile of visual characteristics and impairments which are associated with specific autistic symptoms in the general population. We will use Growth mixture modelling for longitudinal measures of autistic traits and mixture modelling for cross-sectional autistic data to accomplish this task.

Required variables in ALSPAC:

Outcomes:

4 SCDC measures (8-17 years), CCC

Predictors:

Visual processing of chromatic information: L'Anthony desaturated D15, Stroop (8,11y)

Facial expression based tasks:DANVA (8y Happy, Sad, Angry, Fearful faces), CVI related "face recognition" items (13 y) (Williams, 2010)

Perceptional grouping tasks:CVI related "crowded scenes" items (13y) (Williams et.al, 2010)

Mental rotation tasks:WISC-III Block Design (8y)

Moderators: Verbal IQ (8 y); Colour Ishihara test (12y)

Sensitivity analysis:This will performed by adding less autism-specific visual characteristics like refractive error (7,10,11,12,15 y), strabismus (7y) and visual acuity (7, 11y) to the derived mixture models.

2) Given the genetic basis of autism and autism-specific visual tasks (e.g. h2 Stroop = 0.5, Stins et. al, 2009) we also aim to identify genetic variants which are related to both visual and autistic traits.

a. We propose genome-wide analyses using the Stroop test, the DANVA and the perceptional grouping items of the CVI. Replication will be sought internally within ALCPAC by dividing the sample into a discovery and a replication cohort and within independent samples such a QIMR (Faces task). Genome-wide analysis of the Block-Design test is performed within CHIC.

b. Selection of SNPs for functional network work construction will be enhanced through a second step of GWAS analyses where we combine each autistic symptoms score (i.e. CCC and SCDC) with each studied visual trait (Stroop test, the DANVA, the perceptional grouping items of the CVI, in addition to the WISC-block design) into a bivariate outcome to identify SNPs which preferentially affect both visual and autistic outcomes. For this analysis we will select preferentially trait combinations which have been supported through GMM and cross-sectional mixture modelling (see step 1).

c. Functional networks will be constructed with Ingenuity pathway analysis (http://www.ingenuity.com/products/answers.html).