Animal models have long been employed to replicate some of the behavioral and biochemical characteristics of autism (see Converging Signaling Pathways). The models are chosen for study either because they have behaviors reminiscent of autism, or because they have received genetic or environmental manipulations believed to be linked, directly or indirectly, to autism.

Yet, only with the recent progress of detailed genetic studies in developmental disorders have these models been based on the actual genetic differences found in humans with autism. Some of these newer models for autism include mouse models of medical genetic syndromes that show overlap with autism, e.g., Fragile X syndrome, Rett syndrome and Tuberous Sclerosis (see Potential Reversal of Neurodevelopmental Disorders). However, no model existed that contained the precise genetic defect found in anyone whose autism is not caused by one of these other genetic syndromes. This changed in October 2007, when researchers in Texas reported they had succeeded in replacing the mouse neuroligin-3 gene with a human version containing the exact mutation discovered in 2004 to be the cause of autism in a Swedish family with two affected brothers. Excitingly, the initial exploratory studies have found the "humanized neuroligin-3" mouse has several unusual behaviors, including deficits in some social behaviors and an increased ability for spatial learning in a swimming test.

This mouse provided the research community with a strong new tool to directly assess the neurobiology, behavioral deficits and, conceivably soon enough, treatment approaches for autism. Such models are a vital part of the drug discovery process because measurement of changes in their behaviors can be used as surrogate markers for preclinical evaluation of new therapeutics.