ABSTRACT
Subtle volumetric deviations of several brain structures have been consistently demonstrated when comparing patients with schizophrenia to controls. Some of these are global, such as a prominent enlargement of lateral ventricles and a small reduction in cortical grey matter, whereas others are more regional, such as reduced volume of medial temporal lobe structures, prefrontal cortex and thalamus (Shenton et al., 2001; Wright et al., 2000). Most studies have utilized hypothesis-based region-of-interest techniques to investigate these deviations, by which a small number of structures are chosen for measurement and delineated manually. These region-of-interest studies are laborious and con®ne the areas explored for deviations to a small fraction of the brain. Other studies have been performed using automated computational neuroanatomy techniques to explore the entire brain for tissue deviation. These computational neuroanatomy studies have identi®ed volume de®cit in portions of the prefrontal, medial temporal, lateral temporal and thalamic regions, as well as in other areas less commonly chosen for region-of-interest measurement, such as the insula and anterior cingulate gyrus (Ananth et al., 2002; Hulshoff Pol et al., 2001; Kuperberg et al., 2003; Sigmundsson et al., 2001; Wright et al., 1999a). It remains unclear whether brain deviations identi®ed in case±control studies are linked to the illness process itself, genetic risk for schizophrenia or environmental risk factors such as obstetric complications. It is also
uncertain whether such deviations are additionally associated with bipolar disorder or genetic risk for bipolar disorder.
