Using Perceptual Signatures to Link Genotype with Neural Alterations in Autism Spectrum Disorder (ASD) and Fragile X Syndrome (FXS)

Background: The functional link between genetic alteration and behavioral end-state manifested in different neurodevelopmental conditions is rarely straightforward. As suggested by Belmonte & Bourgeron (2006), in order to study convergent and/or divergent neuro-cognitive phenotypes between genetically dissociable conditions, analysis at a neural network level must be considered. Objectives: To introduce a theoretical framework and paradigm that characterizes neural endophenotypes (Muller, 2007) underlying visual information processing in ASD and FXS, defined by characteristic information processing abilities referred to as perceptual signatures (reflect the integrity of neural networks mediating visual information processing; Bertone & Faubert, 2006). We argue that perceptual signatures may be useful for differentiating ASD and FXS from each other at a neural network level, and for providing a better understanding of plausible causal mechanisms underlying atypical information processing in each condition. Methods: Based in part on results from a series of studies assessing static and dynamic information processing in high-functioning autism and fragile x syndrome (FXS), perceptual signatures of the two conditions were compared. Given that similar perceptual paradigms were used to assess functioning for both conditions, a direct comparative with limited methodological constraints was possible. Results: The resulting perceptual signatures are consistent with a pattern of perceptual performance that is condition-specific only for luminance-defined information conditions. In general, enhanced sensitivity to luminance-defined static information is evidenced for the ASD group only whereas decreased sensitivity to simple dynamic information defines FXS performance. Decreased sensitivity to texture-defined information is manifested in both groups. These results suggest that neural networks mediating low-level information processing in ASD and FXS are divergent, or condition-specific, at the local neural network level only; the perceptual consequence of altered integrative neural networks mediating complex, texture-defined information are convergent, or non-specific (Bertone et al., 2003,2005; Kogan et al., 2004). Conclusions: We present a data-driven, causal model associating genetic perturbation with neuromodulatory consequences on local network connectivity that is specific to ASD and FXS. The model suggests altered lateral connectivity within primary visual areas in ASD, and dorsal-stream related dysfunction in FXS as the most biologically plausible type of atypical connectivity congruent with their respective perceptual signatures. We argue that this intermediate level of analysis is useful for suggesting condition-specific neural etiology, and for guiding genetic research by restricting the search for candidate genetic origins most consistent with the neuromodulatory effects on low-level networks underlying perceptual abilities in ASD and FXS. This suggestion is especially important within the context of evidence suggesting that atypical information processing may be one of autism's core deficits (Belmonte et al., 2004). Finally, such a theoretical framework is shared by others who support genetically-influenced, systemic neural models where causal genetic perturbations alter neural connectivity (and also tissue morphology), leading to atypical information processing capabilities which may at least in part define neurocognitive phenotypes manifested in ASD (Herbert, 2005) and FXS.