It has been known since the groundbreaking work of Carlos Pardo (see his profile in my book) that the brain in autism is in an immune-activated state. In postmortem samples, there is evidence of microglial activation and cytokine increases. The latter increases were even seen in cerebral spinal fluid from living patients, so this immune-activated state appears to be ongoing and possibly life-long. Several papers in recent years, including a new one from my Caltech colleagues Nicole Tetreault and John Allman, have confirmed various aspects of the initial Pardo work. In her new publication, Tetreault demonstrates a very significant increase in the density of microglia in two different areas of postmortem autism cortex. In a talk at the recent IMFAR meeting in Toronto, a group led by Nakamura reported that they could visualize activated microglia in living brains using a PET (positron emission tomography) method, and found more such cells in autistic subjects than in controls. In another new paper, Morgan and colleagues at UCSD report that in autism brains, microglia are frequently much nearer neurons than in control brains. In many cases, the microglia appear to be encircling neurons, suggesting a very intimate interaction. While the meaning of these findings is not entirely clear at this point, microglia are known as the brain’s own immune cells, where they mop up debris. Microglia are also known to prune synapses – removing inactive connections. Thus, it is possible that such pruning is more extensive in autism. Could it be that more synaptic connections are inactive in autism than in control brains, or is it that the microglia are activated and inappropriately pruning synapses, or even trying to remove the neurons themselves? I suggested in my book that the immune activation seen in the autistic postmortem brain may begin with immune activation in the fetus, driven in some cases by maternal immune activation during infection or stress. More on this when/if our new paper on this topic is approved for publication.