Biotypes : Promise and P Commentary

This issue of Biological Psychiatry contains a thoughtprovoking article by Ivleva et al. (1) reporting on data generated as part of the Bipolar-Schizophrenia Network for Intermediate Phenotypes (B-SNIP) project. The B-SNIP project is an important effort that was started to help identify the similarities and differences across individuals with a spectrum of psychotic disorders, including schizophrenia, schizoaffective disorder, and bipolar disorder with psychosis. In addition, the B-SNIP project also recruited the first-degree relatives of these individuals to provide further information about what characteristics might be “endophenotypic” markers of risk for psychosis compared with characteristics present only among individuals with manifest illness. The B-SNIP project assessed a variety of measures in their sample that were motivated by previous research in psychosis, including in-depth clinical assessments, measures of a range of cognitive functions (including cognitive control), electrophysiological measures of sensorimotor reactivity, and both structural and functional neuroimaging (resting-state functional connectivity). One of the key goals of the B-SNIP project was to directly assess the neurobiological validity of the traditional categorical distinctions between schizophrenia, schizoaffective disorder, and psychotic bipolar disorder. In addition, they wished to determine whether they might be able to identify other ways of understanding the structure or organization of psychopathology among such individuals with psychosis. Accordingly, in previous work (2), B-SNIP used cognitive and electrophysiological data along with clustering techniques to identify three subgroups of individuals, or Biotypes. These Biotypes showed evidence of within-group homogeneity in terms of cognitive control and electrophysiological function, but differences between groups. The three Biotypes showed graded impairment in cognitive control compared to healthy controls, with individuals in Biotype1 showing the most impairment, individuals in Biotype2 showing less impairment but still significantly worse than control subjects, and individuals in Biotype3 not differing from control subjects. In contrast, in terms of sensorimotor reactivity, individuals in Biotype1 and Biotype3 showed significant reductions compared to controls (with a larger difference in Biotype1 than Biotype3), while individuals in Biotype2 showed significantly enhanced reactivity. Most importantly, diagnoses were not evenly distributed across Biotypes, because all three diagnoses were clearly represented across all three Biotypes, although there were more individuals with schizophrenia in Biotype1 and more individuals with bipolar disorder in Biotype3. In addition, Ivleva et al. also found that the level of impairment in the relatives of the probands also sorted more strongly as a function of Biotype than it did by diagnosis. In the original paper reporting on the creation and validation of the Biotypes (2), Clementz et al. used analyses of gray