Since biological criteria are not available to define objective phenotypes in psychiatric disorders, clinical phenomenology has been used to classify psychiatric diagnoses. In recent years, traditional diagnostic constructs have been increasingly challenged, especially in the cases where prominent symptom dimensions cut across diagnostic boundaries. This is particularly evident in schizophrenia (SZ) and bipolar disorder (BD) where an extensive overlap exists in many clinical characteristics, including psychosis, cognition, mood and behavioral symptoms. Moreover, recent genetic (Wellcome Trust Case Control Consortium, 2007) and neurophysiological (Thaker, 2008) studies have provided evidence for putative molecular and biological markers that overlap the two diagnoses. These observations challenge the traditional dichotomous model of SZ and BD and advocate for dimensional distinctions to enable studies of biological disease criteria.
One of the approaches that seeks to define biological markers which support valid diagnosis is the identification of endophenotypes: the specific heritable characteristics of brain anatomy, biochemistry or function that provide a direct measure of brain behavior, and, perhaps, more direct molecular underpinnings for disease definition than does the clinical syndrome (Gottesman and Shields, 1973). One of the broadly studied putative endophenotypes in psychiatric syndromes is cognitive dysfunction, as suggested by Burdick et al. (2006) and Bora et al. (2009). The established heritability of cognitive function and the described neural networks mediating specific aspects of cognition, as well as the availability of objective measurement tools to assess cognitive performance provide support for examining cognitive function as a promising endophenotype for SZ and BD (Ivleva et al., 2010). Several cognitive endophenotypes have been described in probands with SZ (SZP) and BD (BDP), with deficits in executive function, declarative memory and processing speed consistently found in both diagnoses (Schretlen et al., 2007; Hill et al., 2008; Arts et al., 2008; Smith et al., 2009; Stefanopoulou et al., 2009). Deficits in working memory and sustained attention are common in SZP (Braff, 1993; Keefe et al., 1995; Cornblatt and Malhotra, 2001), whereas are less consistently reported in BDP (Clark et al., 2002; Glahn et al., 2010), with some describing them as state-dependent (Kumar et al., 2010). Cognition deficits in SZP are thought to be more severe and stable “trait” measures, while these deficits in BDP are described as less debilitating and partially state-dependent, although attenuated impairments remain during periods of euthymia (Ferrier et al., 1999; Rubinsztein et al., 2000; Thompson et al., 2005). In addition, deficits in executive function, working and declarative memory, and processing speed have been reported in biological relatives of SZP (SZR) (Faraone et al., 1995; Toomey et al., 1998; Cannon et al., 2000; Asarnow et al., 2002; Sitskoorn et al., 2004), as well as in relatives of BDP (BDR) (Glahn et al., 2004; Kieseppa et al., 2005; Hill et al., 2008; Glahn et al., 2010). Deficits in attention are well documented in SZR (Snitz et al., 2006; Sponheim et al., 2006), whereas reports in BDR are less consistent (Gourovitch et al., 1999; Bora et al., 2009; Kumar et al., 2010). Growing evidence suggests that BDP may have distinctive cognitive profiles depending on presence of lifetime psychosis; in that, probands with psychotic BD show deficits in domains of executive function, verbal and non-verbal declarative memory, working memory, processing speed and attention, similar to those found in SZP (Glahn et al., 2006; Glahn et al., 2007; Martinez-Aran et al., 2008; Reichenberg et al., 2009). Mild psychosis phenotypes (e.g., schizotypy) have been linked to poorer cognitive performance in SZR (Cannon et al., 1994; Niendam et al., 2007). Interestingly, a single study of BDR with lifetime subclinical psychotic-like symptoms showed better performance in motor speed, verbal declarative memory and attention compared to non-psychotic BDR (Jabben et al., 2009).
Overall, evidence suggests that psychosis may be a key clinical dimension associated with overlapping cognitive characteristics in SZ and BD. Therefore, in this study we examined putative cognitive endophenotypes within the SZ - psychotic BD boundary contrasting the two clinical paradigms: traditional dichotomous DSM-IV diagnoses (SZ, BD) and the psychosis dimension defined as manifestation of lifetime psychotic symptoms independent of the DSM-IV diagnoses. In accordance with the classic endophenotype criteria (Gottesman and Shields, 1973), we conducted this examination in a family sample of probands with SZ and psychotic BD, and in their first-degree biological relatives. Establishing similarities and differences in the endophenotypic signatures of SZ and BD families may provide important insights for future genetic studies, and may improve conceptualizations about the common and distinct aspects of pathophysiology, about clinical heterogeneity, and about clinical boundaries of the two psychotic disorders. We asked the question, whether there are aspects of cognitive dysfunction that distinguish the traditional diagnostic groups (SZ and BD), or whether cognitive dysfunction is common to probands and relatives across the SZ/BD psychosis dimension. We conducted two sets of analyses with the cognitive measures, one contrasting individuals with the two categorical diagnoses, and the other contrasting individuals with and without lifetime psychoses. The traditional DSM-IV diagnosis analyses examined cognitive performance in probands with SZ and BD, and their relatives in the four study groups: SZP, BDP, SZR and BDR. The psychosis dimension analyses examined cognitive performance in individuals with and without lifetime psychoses and contrasted the three study groups: 1) psychosis probands (SZP and BDP, combined), 2) relatives with psychosis spectrum disorders (defined in Methods), and 3) relatives without psychosis spectrum disorders. We hypothesized that: 1) within traditional diagnoses, SZP and BDP would show similar cognitive performance on measures of working memory, declarative memory, executive function and attention; SZR and BDR would show similar performance on these measures; whereas relatives would show higher cognitive performance than probands; and, 2) within the psychosis dimension, probands would show lower cognitive performance on the measures of working memory, declarative memory, executive function and attention compared to relatives without psychosis spectrum disorders; whereas relatives with psychosis spectrum disorders would show cognitive performance on these measures intermediate between probands and non-psychotic relatives. This is the first study to directly compare putative cognitive endophenotypes in probands and relatives from both diagnostic groups (SZ and psychotic BD) using a psychosis “diagnosis vs. dimension” contrast. The proband groups included stable on medications out-patients, similar in age, education, IQ and socio-economic status, providing a common baseline of relevant characteristics for cognitive performance comparisons.