Dimensional models A more recent strategy for addressing the heterogeneity of the phenotype in schizophrenia is the dimensional approach. While ‘categories’  traditionally arise from disease models,  ‘dimensions’ are often derived from the study of normal psychology; therefore, students of dimensional approaches have shown less concern about identifying brain - behaviour relationships. Dimensions define groups of symptoms that co-occur, but the cooccurrence is noted through statistical techniques such as factor analysis.  While categories classify individuals,  dimensions classify symptoms. Therefore, dimensions can overlap within a given individual and be additive. Bilder et al.  (1985)  identified three distinct clusters of symptoms using correlative analysis of symptom ratings and neuropsychological data, providing the first validating evidence for the distinction between symptom dimensions. 1  The disorganization cluster included alogia,  attentional impairment,  positive formal thought disorder and bizarre behaviour. 2  The blunted affect and volition cluster included affective flattening, avolition/apathy and anhedonia. 3  The florid psychotic cluster included delusions,  hallucinations and ‘breadth of psychosis’. Impaired neuropsychological performance was strongly association with the disorganized symptoms and, to a lesser extent, was associated with the blunted affect symptoms but was not associated with the psychotic symptoms. Andreasen and Grove (1986) replicated this finding in the second study, validating this distinction of three symptom dimensions,  and subsequently many others have had similar findings (Kulhara et al. 1986; Liddle 1987a,b; Arndt et al. 1991; Gur et al. 1991; Brown & White 1992; Minas et al. 1992; Miller et al. 1993). Arndt et al. (1995) found that the negative symptoms, unlike the psychotic or disorganized symptoms, are stable over time. Liddle (1987b) used factor analysis to examine the relationship between symptoms in a group of patients with schizophrenia and began using the often cited dimension names of ‘disorganization’,  ‘psychomotor poverty’ and ‘reality distortion’. He also provided further validating evidence for these dimensions with imaging data (Liddle et al. 1992;). Later studies have further validated the finding that the negative and disorganized dimensions are associated with impairments in cognitive functioning (Liddle 1987a,b; Liddle & Morris 1991). O’Leary et al. (2000) further specified the nature of the impairment by investigating a large sample (134) of patients with schizophrenia. Negative symptoms were related to generalized brain dysfunction, the disorganized symptoms were related to verbal processing abnormalities and the psychotic symptoms showed no relationship to cognitive impairment. One potential drawback of the dimensional approach is its disregard for the boundaries of the classical disease model which searches for the underlying construct of an illness. A comparable approach would be to search for the principal basis of a feature existing both in schizophrenia and other diseases, e.g. hallucinations, as they are found in schizophrenia, mania and epilepsy. The consequence of many investigations has been to treat dimensions as categories, resulting in attempts to localize specific symptoms.