Autistic-spectrum conditions and psychotic-spectrum conditions (mainly schizophrenia, bipolar disorder, and major depression) represent two major suites of disorders of human cognition, affect, and behavior that involve altered development and function of the social brain. We describe evidence that a large set of phenotypic traits exhibit diametrically opposite phenotypes in autistic-spectrum versus psychotic-spectrum conditions, with a focus on schizophrenia. This suite of traits is inter-correlated, in that autism involves a general pattern of constrained overgrowth, whereas schizophrenia involves undergrowth. These disorders also exhibit diametric patterns for traits related to social brain development, including aspects of gaze, agency, social cognition, local versus global processing, language, and behavior. Social cognition is thus underdeveloped in autistic-spectrum conditions and hyper-developed on the psychotic spectrum.;>We propose and evaluate a novel hypothesis that may help to explain these diametric phenotypes: that the development of these two sets of conditions is mediated in part by alterations of genomic imprinting. Evidence regarding the genetic, physiological, neurological, and psychological underpinnings of psychotic-spectrum conditions supports the hypothesis that the etiologies of these conditions involve biases towards increased relative effects from imprinted genes with maternal expression, which engender a general pattern of undergrowth. By contrast, autistic-spectrum conditions appear to involve increased relative bias towards effects of paternally expressed genes, which mediate overgrowth. This hypothesis provides a simple yet comprehensive theory, grounded in evolutionary biology and genetics, for understanding the causes and phenotypes of autistic-spectrum and psychotic-spectrum conditions.

Autism and schizophrenia are presented as the extremes of disorders affecting the social brain. By viewing human cognition impairment in terms of competence and performance, a variety of social brain disorders can be identified along the autistic-psychotic continuum.

Co-morbidity of schizophrenia and autism is low because interpersonal concerns of schizophrenic patients presuppose developmental achievements that are absent in autism. Autism may arise if primary anxiety is not overcome at a key developmental stage by affective synchronisation between infant and caregiver. Schizophrenic patients will have learned to regulate primitive anxiety by affectively attuning to narrow social networks but remain highly vulnerable to exclusion from larger groups.

The theory of psychosis and autism as diametrical disorders offers a tractable and testable view of normal and abnormal human cognitive variation as a function of opposing traits grouped by their selection for maternal and paternal reproductive fitness. The theory could be usefully rooted and developed with reference to the lower-level perceptual and attentional phenomena from which social cognitive modules are developmentally refined.

Crespi & Badcock (C&B) present an appealing and parsimonious synthesis arguing that schizophrenia and autism are differentially regulated by maternal versus paternal genomic imprinting, respectively. We argue that animal models related to schizophrenia and autism provide a useful platform to explore the mechanisms outlined by C&B. We also note that schizophrenia and autism share certain risk factors such as advanced paternal age. Apart from genomic imprinting, copy number variants related to advanced paternal age may also contribute to the differential trajectory of brain development associated with autism and schizophrenia.

Crespi & Badcock (C&B) provide a novel hypothesis outlining a role for imprinted genes in mediating brain functions underlying social behaviours. The basic premise is that maternally expressed genes are predicted to promote hypermentalistic behaviours, and paternally expressed genes hypomentalistic behaviours. The authors provide a detailed overview of data supporting their ideas, but as we discuss, caution should be applied in interpreting these data.

The confounding effects of heterogeneity in biological psychiatry and psychiatric genetics have been widely discussed in the literature. We suggest an approach in which heterogeneity may be put to use in hypothesis testing, and may find application in evaluation of the Crespi & Badcock (C&B) imprinting hypothesis. Here we consider three potential sources of etiologic subtypes for analysis.

What is the status of the dichotomy proposed and the nosological validity of the contrasting pathologies described in the target article? How plausibly can dysregulated imprinting explain the array of features described, compared with other genetic models? We believe that considering alternative models is more likely to lead in the long term to the correct classification and explanation of the component behaviours.

In the target article, Crespi & Badcock (C&B) propose a novel hypothesis based on observations that a large set of phenotypic traits exhibit diametrically opposite phenotypes in autism-spectrum versus psychotic-spectrum conditions. They propose that development of these conditions is mediated in part by alterations in “genomic imprinting.” This hypothesis is based on the model of the Prader-Willi and Angelman syndromes. The authors have produced a masterful discussion of the differences between psychosis and autism. Of course, another article could be written on the similarities.

Autism and psychosis manifest private speech disruptions analogous to their diametrical opposition along the autism-psychosis continuum. Autism has naturally suppressed private speech with predictable structural deficits when it does surface; psychosis has overt but ineffectual private speech with similar structural deficits. These private speech oppositions are best understood in the context of the control processes of cognitive-computational architectures.

Studies of the role of imprinted genes in psychological phenomena are long overdue. The target article is comprehensive, presenting a wealth of important and convergent evidence, and provides an excellent point of departure for further research. However, the authors' evidentiary grasp exceeds the explicatory capacity of the proposed model. Greater genotypic and phenotypic precision would significantly enhance its predictive power.

Imprinted genes are predicted to affect interactions among relatives. Therefore, variant alleles at imprinted loci are promising candidates for playing a causal role in disorders of social behavior. The effects of imprinted genes evolved in the context of patterns of asymmetric relatedness that existed within social groups of our ancestors.

Combined with recent evolutionary psychological theories, Crespi & Badcock's (C&B's) intragenomic conflict theory of the social brain suggests that schizophrenics are more religious, and autistics are less religious, than the normal population. Combined with the generalized Trivers-Willard hypothesis (gTWH), it suggests that schizophrenics have more daughters, and autistics have more sons, than expected.

Crespi & Badcock (C&B) convincingly argue that autism and schizophrenia are diametric malfunctions of the social brain, but their core imprinting hypothesis is less persuasive. Much of the evidence they cite is unrelated to their hypothesis, is selective, or is overstated; their hypothesis lacks a clearly explained mechanism; and it is unclear how their explanation fits in with known aspects of the disorders.

By conceiving of autism and psychosis as diametrically opposite phenotypes of underactive and overactive mentalizing, respectively, Crespi & Badcock (C&B) commit themselves to a continuum view of intercorrelated mentalizing functions. This view fails to acknowledge dissociations between mentalizing functions and that psychotic people show a mixture of both hypo- and hyper-mentalizing.

A number of studies suggest that women find artistically creative men attractive, especially in the short-term mating context. Artistic creativity (but not mathematical or technical creativity) is linked to psychosis-proneness. I hypothesise that in preferring artistically creative men, women may be choosing paternal genotypes that make babies that are not excessively somatically demanding on them.

Crespi & Badcock (C&B) hypothesize that biases toward expression of paternally or maternally imprinted genes lead to the symptoms of autism spectrum disorders (ASD) and psychotic-spectrum disorders (PSD), respectively. We suggest that such genetic risk factors may act by inducing abnormalities in developmental and learning-related plasticity. We provide preliminary evidence of abnormal plasticity in ASD and suggest transcranial magnetic stimulation as a useful tool to investigate as well as influence cortical plasticity.

We argue that autism and psychosis spectrum disorders cannot be conceptualized as polar extremes of mentalizing ability. We raise two main objections: (1) the autistic-psychotic continuum, as conceptualized by the authors, excludes defining features of schizophrenia spectrum: negative symptoms, which correlate more strongly with mentalizing impairments; and (2) little evidence exists for a relationship between mentalizing ability and positive symptoms.

Autism and early-onset schizophrenia share common dimensions of social communication deficits. The possible role of common genetic factors has to be seriously considered, such as the serotonin transporter gene that influences the severity of social communication impairments (negative symptoms) and hallucinations (positive symptoms). Autism and the negative syndrome of schizophrenia might be at one extreme of a continuum, and paranoid schizophrenia (positive symptoms) at the other extreme.

Social cognition is a broad term, incorporating all aspects of social functioning from perceiving emotional stimuli to attributional style and theory of mind. Not distinguishing between these different capacities may confound the interpretation of the data deriving from studies of the relationship between psychiatric disorders and the social brain. The distinction between cognitive and affective components of social cognition is clearly exemplified by the abnormalities observed in psychopathy and Williams syndrome.