Courchesne et al counted the number of cells in the prefrontal cortex of 7 boys with autism and 6 non-autistic control boys, aged 2-16 years old. The analysis was performed by a neuropathologist who was blind to the theory behind the study and to which brains were from which group. That's good.
They found that the total brain weight of the brain was increased in autistic boys, by about 17% on average. But the number of neurons in the prefrontal cortex was increased by an even higher margin - about 60%. The difference was specific to neurons - glial cell counts were normal. Of the 7 autistic boys, 4 also had intellectual disability - an IQ less than 70. However, the 3 without showed broadly similar results.
As well as having more prefrontal neurons, there were also some other issues in some but not all of the autism brains. Two had prefrontal cortical abnormalities - dysplasia in one case and abnormal cell orientation in another. And no fewer than 4 had flocculonodular lobe dysplasia in the cerebellum.
None of the nonautistic brains had any abnormalities reported but they don't seem to have looked very closely in the controls because that was based on "coroner's report only", rather than a detailed neuropathological exam...
It's a nice piece of work, but very small. These postmortem neuropathology studies always are because postmortem brain samples are in short supply, especially for disorders like autism.
In fact, it's so small, that doing statistics on these data is not really meaningful. The authors do some stats and get some impressive p values but we should take those with a pinch of salt and just look at the individual data (see the scatterplots above).
Now, prefrontal cortical neurons are generated while you're still in the womb. New ones can't be created after you're born - numbers can only decrease. So the increased neuron count in autism must have a very early origin, either genetic or caused by pre-natal environmental factors. Unless the timeline for cell genesis is totally different in autism.
Still, it casts doubt on the idea that, in the brain, bigger is always "better". Assuming that we consider autism to be "bad" - which I'm not saying is necessarily right, but it's fair to say most people do assume that - then the common practice of equating volume increases with all kinds of good things seems rather silly.
Courchesne E, Mouton PR, Calhoun ME, Semendeferi K, Ahrens-Barbeau C, Hallet MJ, Barnes CC, & Pierce K (2011). Neuron number and size in prefrontal cortex of children with autism. JAMA : the journal of the American Medical Association, 306 (18), 2001-10 PMID: 22068992