The tragedy of autism research has been the continuing divide between theory and therapy, and between professionals who have ideas about what should work with autism and families who have experiences of what does work with autism. A currently dominant theory of autism, proposed by us and others in 2004, posits that autism begins locally at the synapse as an abnormality of neural connectivity which produces a pathological decrease in entropy, or information capacity, within local neural networks. This local perturbation then interacts with normal programmes and gradients of activity-dependent development to perturb long-range information transfer between brain regions and between cognitive subsystems as brain and cognitive development unfold. A crucial corollary is that the social communicative deficits that pose the most obvious, the most diagnostic, and the most debilitating aspects of autism aren't necessarily the most ætiologically primary, or even the most therapeutically accessible or relevant. By trying to confront social deficits head-on, too many therapies work against — rather than with — an autistic cognitive style characterised not only by differences in social skills or motivations per se but also by differences in skills of cognitive and motor control and perceptual organisation that are prerequisite to normal social cognitive function.

Our work using both fMRI and event-related potentials, across several computer game-embedded experimental paradigms exercising these domain-general, non-social prerequisite skills, demonstrates that a prolonged, ‘sticky’ style of attention-related prefrontal cortical activation runs in families affected by autism spectrum conditions, is evoked especially by task-irrelevant distractor stimuli, and correlates with psychometric measures of autistic social communicative traits across individuals both within and beyond the categorically defined autism phenotype. We further demonstrate that fully one third of autistic children who lack communicative speech are characterised by an uneven developmental profile in which deficits in prerequisite gross, fine, and especially oral motor skills correlate with large discrepancies between intact receptive and impaired expressive language skills.

With clinical testing and feedback from users and their teachers and therapists we have designed — and will shortly be bringing to clinical trial — software for the iPad to develop manual motor and oral motor skills prerequisite to verbal communication by pointing and/or speaking, bootstrapping users from the iconic manual task of pointing and dragging to assemble sequences of puzzle pieces to the symbolic manual task of pointing at sequences of alphabetic letters, and the symbolic oral task of vocalising sequences of syllables. All of our computer games for measuring visual attention, and our iPad software for developing motor skills for communication, log all stimuli and responses so as to track progress and to enable opt-in access for researchers, and are freely available to families and to scientists.