Analysis of the autistic brain in neuroscience
In neuroscience, we know that the brain is functional from birth: a vast neural network is already in place, including “algorithms” devoted to facial recognition, navigation, spatial and temporal perception, the discrimination of sounds and syllables in all languages, and the distinction between colours, for at least the first six months of the child's life.
After birth, experiences multiply the number of neural fibres, strengthen the wiring and increase connections between the body and different brain areas. The child then initiates primitive movements —slithering, crawling, and so on—which are the foundations of their motor development and the brain's internal coordination.
The cortical layers gradually thicken. A network of fine arteries supplies the brain with glucose, its main energy source, while the lymphatic system ensures nourishment, maintenance and toxin removal. Information processing becomes established in each brain sector, and new “brain algorithms” are created.
The balance between neurotransmitters and inhibitors develops slowly, and different brain “processors” organise data in increasingly elaborate ways. To strengthen the brain–body link, the thyroid gland plays a central role: it regulates energy production, the balance of glands and organs, and that of neurotransmitters and inhibitors.
What elements are likely to disrupt these developmental processes?
The brain is therefore ready before birth. In the autistic person described as a “savant”, this brain in a sense represents the pinnacle of genetic and physical evolution: it is extraordinarily refined and powerful, but also very fragile and exceptionally energy-hungry. Thanks to its neural connections and hyperdeveloped senses, it can access multiple levels of information: a way of functioning designed for a universe different from that of the allistic person. Its ultrafast “processors” handle data in microseconds, in a branching, associative mode: every piece of information triggers the mobilisation of a broad range of prestored and presynthesised data. This complexity demands considerable energy, and integrating this multilayered information requires rigorous learning.