Identifying subtypes for autism and narrowing down the heterogeneity of symptoms has been considered the holy grail of autism research. If one person with autism is not like another person with autism, can they at least be put into groups to speed up studies into causes, intervention and services? And how? This podcast explains two different studies that used the same statistical method but different children with autism to identify different groups. One of the things that helped define these groups was verbal ability and IQ. For the first time, comorbid symptoms like medical issues and psychiatric diagnoses are being taken into account. Already, this approach is helping scientists better understand why fever improves symptoms in some people with autism.