Biomarkers for autism

Perhaps one of the primary stumbling blocks for the various strands of autism research is the lack of objective 'markers' to denote cases of autism from not-autism. Sure, we have the various diagnostic descriptions of autism based on observable behaviour and analysis of developmental history, and these do a pretty good job of providing diagnostic information. The problem however is that the 'label' of autism, whilst good for assessing strengths and weaknesses, accessing the relevant services and opportunities and in some cases providing an 'explanation' for the presentation of specific behaviours, is not actually a very objective and definitive concept when it comes to studying autism. Don't believe me? How about the relative stability / instability of autism presentation?

I must add that not for one minute I am suggesting that autism is not autism or there is anything wrong with the 'observe and checklist' method currently used; there is not - its the only thing we have at the moment and the clinicians who carry out such assessments are indeed skilled. My point is that the lack of markers outside of just presented behaviour is a significant hindrance to autism research and in particular the various searches for genetic and environmental factors potentially linked to the condition.

A new paper published in Nature Reviews by Walsh and colleagues* discusses some of the issues around biomarker research in autism. The paper is open-access and can be viewed here.

It is quite a good read and provides a good overview of the various challenges facing biomarker research in autism (or should that be autisms).  Issues such as heterogeneity, co-morbidity and the changing ideas of what autism is both diagnostically and culturally, are all covered in the paper. Gone perhaps are the days when autism was going to be diagnosed on the basis of one or perhaps two genes, compounds or anatomical structures to be replaced by something altogether more complex and multi-disciplinary. I am taken back to my post a few days ago on the presentation by Prof David Amaral and the sterling work at the MIND Institute on more than on 'type' of autism.

It strikes me that one of the ways forward in trying to determine biomarkers for autism is to concentrate less on a 'universal' biomarker for autism and instead similar to the MIND Institute model, looking at subgroups. Move away from the autism diagnostic label as your starting point and focus on other things, such as response to a particular intervention or strategy and compare your 'responders' and 'non-responders' to ascertain any biological/genetic/anatomical differences. Not only would this inform about which people on the autism spectrum might benefit from which interventions, but eventually one might assume that such markers would delineate those all important subgroups which might lead into some interesting areas.

Just a thought.

* Walsh P. et al. In search of biomarkers for autism: scientific, social and ethical challenges. Nature Reviews: Neuroscience. Vol 12. October 2011.