family history and less than that of other nongenetic risk factors. The association is also often nonspecific (Kendler, 2005), with single gene variants being associated with multiple disorders. Moreover, genetic profiles vary greatly among affected individuals. Not everyone with the susceptibility variant in any one of the associated genes will develop the disorder, and not everyone with a particular disorder will have the susceptibility variant of any associated gene. Therefore, a single genetic variant will rarely be necessary or sufficient to produce a disorder, a point similar to findings on the association of environmental risk factors with MEB disorders (described later in this chapter and in Chapter 4). One strategy that has emerged to address the complexity of linking genes to disorders is to identify more narrowly defined behaviors, characteristics, or biological markers, termed “endophenotypes,” that correlate with specific disorders or that are common to more than one disorder. These endophenotypes can serve as a simpler, more readily identifiable focus of genetic studies (Caspi and Moffitt, 2006; Gottesman and Gould, 2003; van Belzen and Heutink, 2006).
Beyond finding associations between genetic variants and MEB disorders or endophenotypes, identifying the effects that specific genes have on molecular pathways, cellular organization, functioning of neural networks, and behavior is crucially important to developing effective intervention approaches based on the modifiable components of the pathways from genes to behavior. This level of genetic research requires experimental manipulations in animal models. Most commonly this involves modification of the genome of mice by inserting, deleting, or mutating specific genes and, in some cases, controlling where in the brain, in what cell types, and when during the course of development a gene is turned off or on. This extraordinary degree of spatial and temporal control over gene expression makes animal models invaluable in identifying the molecular processes of normal and pathological brain development. The disadvantage of animal models, however, is the difficulty of representing the complex cognitive, behavioral, and emotional symptoms experienced by humans. Although the effects of experimental manipulation on certain aspects of cognition and memory can be assessed through the ability of animals to learn and repeat standardized tasks, analogues of emotional experience and thought can be inferred only through behavior that must be correlated with subjective human experience (Cryan and Holmes, 2005; Joel, 2006; McKinney, 2001; Murcia, Gulden, and Herrup, 2005; Powell and Miyakawa, 2006; Sousa, Almeida, and Wotjak, 2006).
Animal models are proving to be of central importance in identifying the likely disturbances in molecular and cellular pathways caused by single gene mutations in some neurodevelopmental disorders, including the fragile X, Prader-Willi, Angelman, and Rett syndromes. Knowledge of those molecular pathways already has led to promising treatment approaches in animal