Azmitia group originally developed a developmental hyperserotonemia model (DHS) in rat for ASD [37] and found several abnormal behaviors seen in ASD including fewer olfactory-based social interactions, reduced ultrasonic vocalizations induced by maternal separation, seizures, hyper-responsiveness to auditory and tactile stimuli, and decreased alteration in the spontaneous alteration task [37, 38]. The DHS rat was generated by treatment with a nonselective serotonergic agonist, 5-methoxytryptamine, during development (from gestational age 12 days to postnatal day 20 (PND20)). This procedure results in a significant loss of 5-HT terminals and mimics hyperserotonemia in ASD. The same group reported that DHS rats lost oxytocin-containing cells in the paraventricular nucleus of the hypothalamus [38].

Epidemiological studies have indicated that valproate (VPA) or thalidomide (THAL) exposure to pregnant women causes a higher incidence of ASD [39–41]. VPA is a recognized teratogen implicated in the increased risk for a low myelomeningocele lesion, after use in the first trimester for maternal epilepsy [42]. VPA-treated rats not only show abnormal behavior but also have many types of abnormalities related to 5-HT, such as hyperserotonemia, increased hippocampal 5-HT, and abnormal migration or differentiation of 5-HT neurons in the raphe nucleus [43–47]. These results were seen when a single dose of VPA (600–800 mg/kg) or THAL (500 mg/kg) was administered to pregnant dams at E9 by intraperitoneal injection or oral administration. The effect and history of THAL is reviewed by Miller et al. [48].

Acute tryptophan (Trp) depletion is a widely used dietary method to assess the function of 5-HT. Although Grockett et al. and Van Donkelaar et al. pointed out concerns on the effect of acute Trp depletion on 5-HT release or synthesis [49, 50], Trp depletion studies have provided evidence that 5-HT signaling is important in understanding the mechanism of autism.

Trp is a dietary precursor of 5-HT in vivo, so Trp depletion causes 5-HT reduction in the whole body including the brain via the blood brain barrier. Trp depletion has been shown to increase pain sensitivity [51], acoustic startle [52], and muricidal behavior [53] in animals. Importantly, people with ASD have a significantly lower ratio of Trp to other large neutral amino acids, and Trp depletion leads to worsened autistic symptoms including repetitive behaviors, exacerbation of anxiety, self-hitting, irritability, and perseveration [54–57]. Conversely, some papers have shown that acute administration of Trp is associated with improved emotion recognition [58] or reduced cortisol via increased hypothalamic serotonergic activity [59]. In rats, the method of acute Trp administration and its effect has been relatively well validated compared to mice, though the behavioral alterations have not always been reproducible. Acute administration of a Trp-free diet in rats induces a 40 % decrease in extracellular levels of 5-HT in the cortex, and subchronic administration (5 day depletion of Trp) causes almost undetectable levels of 5-HT in brain [60]. In mice, acute Trp depletion leads to the reduction of 5-HT in the brain [61–63]. Chronic Trp depletion or limitation in mice causes altered 5-HT turnover in the brain, impaired nesting behavior, impaired formation of contextual fear memory, defensive aggression, or enhanced social dominance [62, 64, 65].

ASD is believed to have a strong genetic basis because (1) genetic mutations or chromosomal duplications/deletions are found in ASD patients, (2) there is a relatively high risk of autism diagnosis for children with an affected sibling, (3) siblings or parents of ASD children tend to have similar behavioral features found in probands more frequently than controls, and (4) monozygotic twins showed higher concordance rates (70–90 %) than dizygotic twins [66]. Recently, genetic engineering technology has generated many mouse models for ASD with mutations in genes or chromosomal deletions/duplications found in patients with ASD (Table 2). There are three classes of mouse models for ASD: single-gene mutations (monogenic heritable ASD), copy number variations (CNV), and inbred strains. Studies on mouse models with a gain-of-function mutation of Nlgn3 (R451C) and a loss-of-function mutation in Nlgn4 (D396X) seen in human ASDs [67] revealed that these mutations can cause autistic-like behaviors and synaptic dysfunction [68–70]. In CNV models, 15q11–13 duplication, 16p11.2 duplication/deficient, and 17p11.2 duplication model mice are reported to have serotonergic abnormalities [71, 72] or dose-dependent micro/macrocephaly [73] in addition to autistic-like behaviors [71–75]. Finally, BTBR T+tf/J (BTBR) and BALB/c, as inbred strains, are reported to have autistic features or serotonergic disturbances. Comprehensive studies in BTBR revealed that BTBR mice not only have autistic-like behavioral abnormalities, but also the absence of the corpus callosum or reduced sulfate concentration, both of which are found in ASDs [76–79]. BALB/c is also an inbred line and shows low sociability and high levels of anxiety. Importantly, BALB/c also shows a low level of brain 5-HT, caused by a single nucleotide polymorphism at C1473G that produces an amino acid change from 477Pro to 477Arg in the Tph2 gene [80–85]. This amino acid change is not seen in C57BL/6 or 129/X1SvJ strains, and BALB/c mice have about a 50 % reduction of 5-HT in the brain compared to 129X1SvJ [84] mice. Recent association studies have revealed a genetic association between certain Tph2 alleles and ASD; however, how this mutation affects 5-HT synthesis remains unclear [86]. Moreover, the BALB/c strain has a large brain to body weight ratio and shows underdevelopment of the corpus callosum compared to other strains. These abnormalities might contribute to abnormal behaviors [87, 88].

Crawley et al. proposed behavioral procedures for evaluating model mice for ASD [89]. These behavioral tests are designed to identify the core elements found in ASD including deficits in social interactions, impaired communications, frequent stereotyped/repetitive behaviors, and narrow, restricted interests. In this section, we introduce representative behavioral tests for analyzing mouse behaviors related to ASD.