Copy number variations (CNVs; i.e. deletion or duplication) of the human chromosome 16p11.2 locus are among the strongest genetic risk factors for autism spectrum disorder (ASD), and predispose for a range of other neurodevelopmental conditions including schizophrenia, epilepsy and intellectual disability. Numerous case reports and populations studies have characterized the clinical profiles of 16p11.2 deletions and duplications; however, the underlying molecular mechanisms remain poorly understood due to insufficient basic science investigations. In Part I of my thesis, we characterized the behavioral, synaptic, and cellular pathologies associated with 16p11.2 deletions and duplications in mice and identified aberrant synaptic transmission in both models in the prefrontal cortex (PFC), a brain region critical for sociability and higher-level cognitive functions. Specifically, 16p11.2 deletion mice (16p11.2+/-) displayed impaired NMDA receptor-dependent glutamatergic transmission and regional hypoactivity in PFC, while 16p11.2 duplication mice (16p11.2dp/+) exhibited GABAergic synaptic deficits accompanied by neuronal hyperactivity in PFC. Notably, the divergent excitability profiles in PFC of 16p11.2+/- and 16p11.2dp/+ mice are the first opposing molecular phenotypes identified in 16p11.2 CNVs, and offer an intriguing bidirectional explanation for the associated behavioral pathologies. Moreover, restoring homeostatic excitability to PFC neurons in either model was sufficient to ameliorate the observed social and cognitive deficits, implicating excitation-imbalance in PFC as a core mechanism. These findings have increased the potential for therapeutic interventions by identifying a single molecular endpoint in both CNVs which may be pharmacologically targetable for the alleviation of social and cognitive symptoms. Mouse models are commonly used in ASD research, as mice are very social creatures. However, the structure and significance of these social interactions are complex, making sociability a difficult variable to assess experimentally. Several assays exist for evaluating sociability in mice, though these tests fail to address certain facets of social behavior, and also suffer from inconsistent usage across laboratories. A number of studies performed by different groups with identical ASD mouse models have reported inconsistent findings about social behaviors, due to the use of different behavioral testing protocols. In Part II of my thesis, we sought to improve the existing methodology for assessing social behavior in mice, both through the modification and optimization of existing protocols and the development of entirely novel assays. We first describe a standardized protocol of the three-chamber social preference test - a common method for evaluating sociability in ASD mouse models - which offers more robust sensitivity to social deficits than a commonly used protocol in the literature. Secondly, we propose a novel behavioral test for evaluating social motivation, which integrates components of the social approach test and the elevated plus maze. Through these efforts, we hope to enable more accurate, consistent and detailed behavioral investigations of mouse models across the ASD research field.