I’m taking part in this meeting with the heads of CHU and innovation. I’ll be presenting our advances in the “Pelargos” project. In a nutshell, autism is “born” in utero, resulting in changes to the way the brain is constructed and later manifesting itself as the syndrome and its characteristics. As a result, we set out to identify from birth those babies who would later be diagnosed with an ASD syndrome. Using a Machine Learning program, we analyzed all the parameters normally collected in a maternity hospital (here in Limoges) from children with or without ASDs, and compared them. We can identify almost 100% of neurotypical children (NTs) and around half of those with ASDs. This program needs to be validated by another study of the same type carried out in another maternity hospital and then revalidated in a larger trial carried out in several French maternity hospitals. Our aim is to make France the 1st country in the world to have such early identification, in line with the wishes of the authorities for decades, as recently confirmed by a solemn vote in the French National Assembly (1). The advantage is that we can intervene earlier with psycho-educational approaches, which we know to be all the more effective the earlier they are introduced during the highly plastic period of the first 2-3 years.
Ba-Biomedical and its 2 experts in Machine Learning and big data analysis – Hamed Rabiei and Marilyn Begnis – have been working on these projects for the past 4 years. The resulting Pelargos project has also identified interesting parameters in the pathogenesis of autism. For example, one of the criteria used by the program to identify babies with autism is that their Cranial Perimeter (CP) is larger than that of neurotypicals, as early as the 2nd trimester in some cases, and virtually all of them shortly before birth! It has been known for years that the brains of children and adolescents with ASDs are larger than those of NTs. Clearly, brain volume is impacted in utero by events that will result in ASD. We have been able to validate this result, in experimental work carried out on mice showing notable differences between “autistic” and control brains. Better still, brain and neuronal growth continue to increase in “autistic” mice and not in control mice, suggesting differences in parturition and preparation for birth.
Our task is to convince maternity units and hospitals of the importance of this approach, which is not easy because they are understaffed and saturated… But we take charge of funding the Clinical Research Associates who will collect the data, and of course we take care of the rest – i.e. the analysis and validation of our program.
Incidentally, an interesting recent study shows a connection between socio-economic background and the size of brain regions such as the cingulate cortex – a region that plays a major role in behavior – during the prenatal period. A great deal of research is demonstrating the crucial role of in utero development in neurological or psychiatric syndromes. Our work falls well within these lines. It should also be emphasized that our measurements do not include any genetic parameters, once again underlining the limits of a purely genetic approach to the emergence of ASDs.
