Initially intact neural responses to pain in autism are diminished during sustained pain.

The final sample was small: 15 adults with autism and 16 typical controls were analysed. Attrition was substantial: 24 adults with autism and 21 typical adults were recruited, but only 16 in each group completed the scan, and one autism participant was excluded for head movement. All participants were adults aged 18 to 50 years with IQ 80 or higher, so the results may not generalise to children, adults with lower IQ, or the wider autism spectrum. The pain protocol was difficult to tolerate, which may have selected participants who could remain in the scanner and complete the task. Pain ratings were collected after each run on a 1 to 10 scale, not during the early, intermediate and late phases, so they may not reflect the neural response at each time window. The self-injury subgroup analysis was exploratory and very small: 8 autism participants with self-injury and 4 without self-injury, and the authors state a definitive conclusion was not possible. The autism group tended to have lower performance IQ (p = 0.08), although other demographic variables did not differ. The paper does not show that any treatment, therapy or assessment method changes pain outcomes.

The supplied text does not include a conflict-of-interest declaration. The listed publication types include research support from NIH extramural and non-U.S. government sources.

Do not conclude that people with autism do not feel pain or that this explains self-injury. The autism group had normal early pain-network responses and similar pain ratings to controls, and the sustained neural difference was seen in a small fMRI study without treatment outcomes.