We compiled data included in the Primal Health Research Database (www.primalhealthresearch.com) to test the hypothesis that when two pathological conditions or personality traits share the same critical period for gene-environment interaction, we should expect further similarities, particularly from clinical and pathophysiological perspectives. The keywords 'autism' and 'anorexia nervosa' (but not bulimia nervosa) lead to studies suggesting that for both conditions the perinatal period is critical. We take this example to look at other possible links between these pathological entities. From a clinical perspective, several teams have independently emphasized the importance of autistic traits in anorexia nervosa. Deficits in the processing of oxytocin have been demonstrated in both cases. Autistic groups have significantly lower blood oxytocin levels than normal groups, and oxytocin levels increase with age in the normal group only. In autistic groups there is a high ratio of intermediates of oxytocin synthesis (OX-T) to the nonapeptide oxytocin (OT). On the other hand, it has been reported that the level of oxytocin in the cerebrospinal fluid of anorexic women is significantly lower than the level of oxytocin in bulimic and control subjects. Scanning data reveal similar asymmetric functions with left hemisphere preponderance in autistic spectrum disorders and anorexia. A comparative study of the mirror neurons systems is another promising avenue for research. Such an accumulation of similarities from a great diversity of perspectives suggests that anorexia nervosa might be considered a female variant of the autistic spectrum. A plausible interpretation is that prenatal exposure to male hormones might protect against the expression of this disease: girls who have a twin brother are at low risk for anorexia nervosa, compared with girls who have a twin sister, and with controls; furthermore genetic linkage analyses do not detect change on the X chromosome. From an overview of the database, the perinatal period appears to be critical for all disorders related to the capacity to love (including love of oneself), to the potential for aggression (including self-destructive behaviours), or to sociability. Is the perinatal period critical for the organisation of the oxytocin system? This is an important question at a time when we learn that the widely used synthetic oxytocin can probably diffuse across the placenta. On the other hand, where the genesis of metabolic types is concerned, it is prenatal life that appears to be critical.