Abstract
Crow (1995a) has argued that schizophrenia is caused by a gene associated with the evolution of human language and cerebral specialisation. This paper suggests a mechanism for Crow's theory which requires only one new assumption for the right shift genetic model of handedness and cerebral dominance (Annett, 1978). The proposal is that the RS+ allele, whose normal function is to induce the left hemisphere to serve speech by impairing speech-related cortex in the right hemisphere, tends to lose its directional coding. It becomes agnosic (RS+ a) for right versus left and impairs the left or right hemisphere at random. Schizophrenia is likely to develop when the RS+ a gene is paired with a normal RS+ gene but only in the 50% of cases where both hemispheres are affected. In the 50% where RS+ a affects the right hemisphere, development is normal as in the RS+ RS+ genotype. The risks for schizophrenia in monozygotic and dizygotic twins and other relatives are as expected for 50% expression of a Mendelian gene which is paired with a particular allele, but not alternative alleles at the same locus. The frequency of homozygotes for the agnosic gene is about 4 in 10,000, the rate observed for autism. A random pattern of double hemisphere deficits would give scope for a range of developmental strengths and weaknesses as observed within the spectrum of autistic disorders. Tests of the model require brain-imaging studies sensitive to individual differences in hemisphere lateralisation and a search for a genetic locus with human and nonhuman primate alleles, together with a mutant of the human form with a frequency of about 2%.