Crespi & Badcock (C&B) have elegantly summarized the evidence for psychoses and autism to be considered as “diametrical disorders of the social brain.” Importantly, their work blends the concept of differential evolutionary selection pressures on male and female brains (Lindenfors et al. Reference Lindenfors, Nunn and Barton2007) with contemporary neurobiological observations. Although the authors have described autism and psychoses as “diametrically opposite phenotypes” (target article, Abstract), evidence from other studies (reviewed in this commentary) suggests that the relationship between them might be more complex.

This intriguing complexity is illustrated by the striking comorbidity of “seemingly incompatible” syndromes (e.g., 34.8% of adults with atypical autism had schizophrenia spectrum disorders; Mouridsen et al. Reference Mouridsen, Rich and Isager2008). Also, in the recently postulated “network disease hypothesis,” it was estimated that 20–60% of autism-predisposing genetic polymorphisms also predispose the bearer to bipolar disorder, and 20–75% of these polymorphisms also predispose the bearer to schizophrenia (Rzhetsky et al. Reference Rzhetsky, Wajngurt, Park and Zheng2007). Moreover, DISC1 gene, which was reported to co-segregate with schizophrenia (Mackie et al. Reference Mackie, Millar and Porteous2007), is also associated with autism and Asperger's syndrome (Kilpinen et al. 2007).

In addition to these shared genetic diatheses, evidence for potentially similar pathogenesis for both autism and schizophrenia does exist (although the studies summarized in Table 1 of the target article suggest contrasting pathogenetic mechanisms). Both autism and schizophrenia are associated with (1) deficient insulin-like growth factor-1 (Riikonen et al. Reference Riikonen, Makkonen, Vanhala, Turpeinen, Kuikka and Kokki2006; Venkatasubramanian et al. Reference Venkatasubramanian, Chittiprol, Neelakantachar, Naveen, Thirthall, Gangadhar and Shetty2007), (2) masculinized digit ratio (Manning et al. Reference Manning, Baron-Cohen, Wheelwright and Sanders2001; Procopio et al. Reference Procopio, Davies and Marriott2006), (3) smaller corpus callosum (see Waiter et al. Reference Waiter, Williams, Murray, Gilchrist, Perrett and Whiten[2005] in relation to autism; see Venkatasubramanian et al. Reference Venkatasubramanian, Jayakumar, Gangadhar, Janakiramaiah, Subbakrishna and Keshavan[2002] and Woodruff et al. Reference Woodruff, McManus and David[1995] in relation to schizophrenia), and (4) increasing paternal age (Reichenberg et al. Reference Reichenberg, Gross, Weiser, Bresnahan, Silverman, Harlap, Rabinowitz, Shulman, Malaspina, Lubin, Knobler, Davidson and Susser2006; Sipos et al. Reference Sipos, Rasmussen, Harrison, Tynelius, Lewis, Leon and Gunnell2004).

These shared genetic and pathogenetic abnormalities, as well as clinical comorbidity between psychoses and autism, suggest an extended hypothesis that, although genomic imprinting is likely to have an influence on the pathogenesis of psychoses and autism, it might contribute to phenomenological antithesis (opposite symptomatic manifestations) between as well as within these disorders. This hypothesis derives indirect support from the recently emerging view that no unitary cause (genetic or cognitive) can explain the diverse symptoms of autism (Happé et al. Reference Happé, Ronald and Plomin2006). Extending the thesis by Happé et al. Reference Happé, Ronald and Plomin(2006), I propose that at least three evolutionary significant dimensions – namely, systemizing ability, schizotypy, and affect regulation (with each dimension being independent and possibly having multiple components) – need careful consideration with relevance to the impact of genomic imprinting on these disorders. Interestingly, the prototypical disorders related to these three dimensions (systemizing ability – autism; schizotypy – schizophrenia; affect regulation – bipolar disorder) are all influenced by genomic imprinting (Luedi et al. Reference Luedi, Dietrich, Weidman, Bosko, Jirtle and Hartemink2007).

Recently, Luedi et al. Reference Luedi, Dietrich, Weidman, Bosko, Jirtle and Hartemink(2007) have listed genes that have been proved or predicted with high confidence to be imprinted in loci linked to these three disorders. The ratio of paternal to maternal (P:M) effects (parent-of-origin as well as paternal/maternal expression) of the proposed imprinted genes in these three disorders are as follows: autism (P:M=5:4), schizophrenia (P:M=3:2), and bipolar disorder (P:M=3:6) (Luedi et al. Reference Luedi, Dietrich, Weidman, Bosko, Jirtle and Hartemink2007; supplemental research data online). These P:M ratios suggest that the influence of genomic imprinting on these disorders may not be exclusively paternal or maternal. The P:M profiles of autism and schizophrenia suggest more of paternal than maternal effects, whereas the P:M profile of bipolar disorder suggests the reverse.

Based on this, one might speculate that autism and bipolar disorder are “phenomenologically antithetical” disorders (at either end of spectrum) with schizophrenia being “in the middle.” Interestingly, the multidimensional model of schizotypy (Kwapil et al., in press) suggests that “positive schizotypy” is related to psychotic-like experiences and mood disorders (tending more towards bipolarity), whereas “negative schizotypy” is related to negative and schizoid symptoms (tending more towards autism). Similar descriptions for affective instability (Lara et al. Reference Lara, Pinto, Akiskal and Akiskal2006), as well as systemizing ability (Baron-Cohen Reference Baron-Cohen2006), can explain the spectrum of symptoms within a specific dimension, as well as the paradoxical coexistence of apparently “polar” or phenomenologically antithetical conditions.

Genetic studies of these three hypothesized dimensions might offer valuable insights on the pathogenesis of these neuropsychiatric disorders. Because autism, schizophrenia, and bipolar disorder are affected by genomic imprinting (Luedi et al. Reference Luedi, Dietrich, Weidman, Bosko, Jirtle and Hartemink2007), it is plausible that the three hypothesized dimensions that are adversely affected in these disorders might also be influenced by imprinting. Indeed, the schizotypal dimension in schizophrenia and bipolar disorder has been reported to be associated with homozygous high activity catechol-O-methyltransferase (COMT). Moreover, hypomethylation of membrane-bound COMT promoter gene has been reported to be a potential molecular basis for the shared symptoms of schizophrenia and bipolar disorder (Abdolmaleky et al. Reference Abdolmaleky, Cheng, Faraone, Wilcox, Glatt, Gao, Smith, Shafa, Aeali, Carnevale, Pan, Papageorgis, Ponte, Sivaraman, Tsuang and Thiagalingam2006). Significantly, methylation-mediated processes constitute one of the important mechanisms for genomic imprinting (Reik & Walter Reference Reik and Walter2001). This suggests that schizotypy might be influenced by imprinting.

Imprinting is postulated to offer evolutionary adaptations (Wilkins & Haig Reference Wilkins and Haig2003). Interestingly, these hypothesized dimensions are evolutionarily significant because a balanced mix of these dimensions might offer survival advantages (James Reference James2003; Nettle & Clegg Reference Nettle and Clegg2006), whereas an imbalance might lead to psychiatric disorders. An imbalanced imprinting polarized to one of these three dimensions might result in one disorder (e.g., extreme schizotypy might result in schizophrenia), whereas an imbalanced imprinting adversely influencing more than one dimension might result in comorbidity of phenomenologically antithetical conditions. Also, the imprinting mechanisms influencing these hypothetical dimensions might be more intricate because of additional factors like “leaky imprinting” (which can result in “gain-of-function” as well as disorder; Delong Reference DeLong2007), environmental influences (Jirtle & Skinner Reference Jirtle and Skinner2007), and interaction with sex-related biological differences.

To summarize, Dobzhansky's statement that, “Nothing in biology makes sense except in the light of evolution” (Dobzhansky Reference Dobzhansky1973), emphasizes the need for evolutionarily informed approaches to understand diseases and disorders. However, contemporary medicine focuses predominantly on “proximal-etiology” whereas the “distal-etiology” based evolutionary approach has mostly been neglected (Nesse et al. Reference Nesse, Stearns and Omenn2006). C&B's target article is an example for promising applications of evolutionary approaches in understanding complex human diseases. Although the studies summarized by C&B strongly emphasize the influence of genomic imprinting in psychoses and autism, additional literature reviewed in this commentary illustrates the complexity of this influence; and I hypothesize that genomic imprinting contributes to phenomenological antithesis between as well as within these disorders.