6pm Wednesday 4 March 2009
MIT Department of Brain and Cognitive Sciences
Room 46-3002
In our zeal to discover what science might tell us about autism, we often don't stop to consider what autism can tell us about other scientific questions. As a spectrum condition with high incidence and a broad phenotype that shades into the continuum of normal variation, autism is well positioned to serve as a basis for studying normal human variation and diversity -- and in fact some of the same correlations amongst levels and domains of perceptual and cognitive processing as are found in autism are also found, subtly, in the normal population.
Key to understanding this link between autism and normality is the recognition of autism as a developmental disorder -- developmental not just nosologically but aetiologically. Autism's unfolding depends on the interaction of abnormal genetic and environmental antecedents with normal programmes and gradients of brain and cognitive development. Studying what goes wrong in autism can, therefore, tell us a great deal about what goes right in normal development. As well, comparing and contrasting autistic probands with their clinically unaffected sibs can reveal how familial factors permissive of autism might be developmentally translated into conditions determinative of autism.
Behavioural and physiological observations of autism reveal slowed shifting of attention between perceptual channels -- a deficit that, although not the sole or central issue in autism, may be part of a more general pattern of abnormal timing of brain activation and perturbed information transfer between brain regions and amongst cognitive subsystems. As a group, unaffected sibs resemble autism-spectrum sibs behaviourally in terms of task accuracy and physiologically in terms of the timing of activations within brain regions, but are distinguished by their normal degrees of functional connectivity between brain regions. These results suggest that abnormally delayed and prolonged frontal activation may be a familial marker permissive of autism, whereas abnormally low functional connecticity may be determinative of -- or at least specific to -- autism within these families.
Twentieth-century neuroscience was marked by advances in understanding mutations in single genes or lesions to single brain regions. Such disorders fulfilled scientists' need for univariate relationships, in which a single independent variable is manipulated and assayed for its dependent effect whilst all others are somehow held constant, and in which perturbed functions can be localised to a circumscribed tissue or allele. Twenty-first-century science is a multivariate science, undertaken more and more at the level of networks rather than single variables. As a disorder whose causation fans in through networks of interacting genetic and environmental causes and fans out through network perturbations of neural activity, autism is quintessentially an object of twenty-first-century scientific enquiry.
Scientific discoveries are determined at least as much by the questions that scientists ask as by the data that scientists observe. Historically, the basic and clinical science of autism has been marked by leading questions and hidden assumptions. Cognitive test procedures validated in people without autism are assumed to be equally sound measures for people with autism. Hypotheses are framed, and experiments designed, within separate theoretical apertures which all too often remain unintegrated with each other. Large variances within each tested domain of perceptual and cognitive function suggest that the covariance structure between these domains is likely to be very rich indeed: What relationships exist between measures of perception, attention, executive function, and social cognition, and what might such relations mean for developmental dependencies amongst these capacities? To explore this wide array of domains within a single set of subjects, and to test them in a format more engaging and ecologically valid than standard psychology exeriments, we embed experimental tasks in an entertaining video game that transparently logs behavioural data and synchronises with physiological recordings. Aimed at minimising anxiety, tasks are event-driven not time-driven, player-centred not computer-centred, and accommodative not anxiety-laden.
All our experience with autism science and with autistic people suggests that autism is most straightforwardly and productively viewed as the developmental response of a normal human mind to an abnormal perceptual and cognitive environment. In this sense, people with autism can be described as "human, but more so." Autism's intense perceptual style is marked by a difficulty sacrificing veridical detail to make way for abstractional power; in this regard, autism shines a spotlight on what Lacan has described as the violence of representation, a conflict fundamental to human cognition. As we seek to prevent and to cure autism, we must be careful to augment autistic social and communicative skills without deleting autistic perceptual skills, so as to give people with autism the ability to share their unique insights with the broader social world.