Applications of classical epidemiological methods have led to general acceptance of a multifactorial threshold model of schizophrenia, whereby multiple genetic influences and environmental factors aggregate together, additively and/or interactively, to form a continuum of disease liability. Studies that make use of clinical diagnostic categories are limited in what they can tell us about the precise nature of the aetiological factors involved in schizophrenia because of the inherent limitations of categorical measures in reflecting a continuous liability scale. We propose that new methodologies that can represent this liability continuum on a quantitative scale are needed to aid in the search for schizophrenia susceptibility genes and in elucidating the pathophysiological mechanisms by which these genes and environmental factors act. Here we review evidence concerning structural brain abnormalities as markers of aetiological influences in schizophrenia, concluding that use of quantitative neuroanatomical measures will facilitate the detection of predisposing gene loci and help to clarify whether nongenetic influences contribute independently of or interactively with genetic factors in influences in schizophrenia, concluding that use of quantitative neuroanatomical measures will facilitate the detection of predisposing geneloci and help to clarify whether nongenetic influences contribute independently of or interactively with genetic factors in influencing disease liability. The eventual discovery of such genes will spawn a new era of epidemiological research, in which the effects of specific predisposing genotypes and specific environmental factors on schizophrenia phenomenology and pathophysiology can be studied.

Investigating gene-environment interaction in schizophrenia

Family, twin, and adoption studies have demonstrated that schizophrenia is a substantially heritable disorder. These studies are reviewed in Chapter 10. Still, attempts at isolating specific genes that confer vulnerability to schizophrenia have so far been only moderately successful (Moldin, 1997).