Preclinical models of autism spectrum disorder

Autism Spectrum Disorder (ASD) is characterized by a high degree of etiological and phenotypic heterogeneity. Preclinical studies play an important role in parsing this heterogeneity, as they allow for rigorous experimental control using a variety of invasive and non-invasive experimental methods. Numerous preclinical animal models have been developed for this purpose. While no single model is able to capture all of the complexity of ASD, together they provide a way of approximating the disorder. The majority of animal models for ASD fall under the category of genetic models and have been generated using the mouse as the model animal. Assessment methods for these models can be classified into behavioural, molecular-anatomical and anatomical-functional, with behavioural assessment considered to be the gold standard. Most neurobiological findings obtained using these methods come from genetic models and are heterogeneous, but some common patterns have been identified. For instance, the diverse genetic modifications used to create these models give rise to a variety of cellular phenotypes. Unsurprisingly, this variability in phenotype in also seen in the macroscopic neuroanatomical characteristics of the models. However, some models share similar patterns of morphological change across the brain, which allows for clustering into subgroups. While animal models play a crucial role in treatment development, the aforementioned variability amongst models makes it unlikely that a single treatment will be discovered for all instances of ASD. Patient classification into sub-populations defined on the basis of common mechanistic processes, arising from diverse genetic causes, seems the most feasible strategy for targeted treatment development moving forward.