Direct evidence of neurodevelopmental differences is available (Weinberger, 1995). One source is a remarkable piece of opportunistic research by Walker and colleagues (Walker, Grimes, Davis, & Smith, 1993; Walker & Lewine, 1993) who studied home movies of families in which one child later developed schizophrenia. Facial expression of emotion and general motor functions were rated blind to that child's identity among siblings. The preschizophrenic children were distinguished on both accounts, some with fairly gross but transitory motor differences. These may point to the basal ganglia of the brain as being involved in the underlying mechanism, reminding us that subtle motor disturbances are apparent at the beginning of schizophrenia, before any treatment (Gervin et al., 1998).

Such developmental differences have now been demonstrated in large, population-based or epidemiological samples. In the British 1946 birth cohort, a group of several thousand people born in 1 week in March 1946 have been studied regularly throughout their lives. Their mothers were asked about development when the children were age 2 years, before anyone knew what would happen later on. All the milestones of sitting, standing, walking, and talking were slightly though clearly delayed in those who developed schizophrenia as adults, but there was nothing that would have alarmed parents at the time. There were other indications that language acquisition was different before onset of schizophrenia. Nurses were more likely to notice a lack of speech by 2 years in the children who developed schizophrenia as adults, and school doctors noted speech delays and problems in them throughout childhood.

Developmental differences have been replicated in similar cohort studies in other domains, such as bladder control, fine motor skill, and coordination during late childhood and adolescence (Cannon et al., 1999; Crow, Done, & Sacker, 1995). The motor and language delays were replicated and extended in a birth cohort study from Dunedin, New Zealand (Cannon et al., 2002), where over a thousand children have been followed during childhood. Those who indicated in their mid-20s that they had experienced symptoms suggestive of schizophrenia, mania, and other disorders were compared with those who said that they had never had such phenomena.

Figure 7.2 shows how a summary motor performance score was lower throughout most of childhood for those who experienced a schizophreniform disorder than that of the other groups. Figure 7.3 indicates that there was also a receptive language problem in those who later had hallucination, delusion, and thought disorder.

Developmental differences before onset of schizophrenia were observed during the first year of life in the North Finland 1966 birth cohort. This comprises about 12,000 babies due to be born in this geographical area during 1966 (Rantakallio, 1969). Their early development was charted in the first year of life and later linked to information about who had developed schizophrenia through adolescence and into the early 30s (Isohanni et al., 2001). Figure 7.4 shows the incidence of schizophrenia in male subjects according to how quickly the little boys learned to stand without support or “toddle” during the first year. The figure for girls was similar.

It is clear that not only was there an effect whereby the later a boy learned to toddle, the greater his chance of developing schizophrenia in later life, but also that this effect seemed to hold true throughout the range of variation in reaching this milestone, all of which might be considered normal. If one were looking only for very late developers, then one might be more likely to find them within the preschizophrenia group than in those who did not develop the illness. However, this approach would completely obscure the widespread nature of this association, the meaning of which is considered later on in this chapter.

There is another finding apparent from Figure 7.4. For the boys who passed the milestone early, in the 9-month and 10-month categories, the relatively few individuals who developed schizophrenia all did so in their mid-teens to mid-20s; their period of risk seems fairly short. For those who were later developers, the period of risk is longer; these groups are still accruing cases of schizophrenia into their early 30s and beyond. It may be that the overall risk period for schizophrenia is shorter when neurodevelopment is more efficient, and longer when it is less efficient.

Bleuler's description of schizophrenia most obviously implies differences in behavior and temperament. Studies in this area have also moved on through retrospective research methodologies to cohort designs. Sophisticated rating scales for the retrospective assessment of behavior and personality demonstrate differences prior to psychosis, with the most common being characteristics of a shy, “schizoid” habit (Ambelas, 1992; Cannon-Spoor, Potkin, & Wyatt, 1982; Foerster, Lewis, Owen, & Murray, 1991; Gittleman-Klein & Klein, 1969).

Robins (1966) undertook a pioneering, historical cohort study in which she followed a group of boys who had been referred to a child guidance clinic in St. Louis, Missouri. Here antisocial behavior was associated with later schizophrenia. Watt and Lubensky (1976; Watt, 1978) traced the school records of people with schizophrenia who came from a geographically defined neighborhood in Massachusetts. Girls who were to develop schizophrenia were introverted throughout kindergarten into adolescence. Boys who were to become ill were more likely to be rated as “disagreeable,” but only in the later school grades (7 to 12). This pattern has been identified (Done, Crow, Johnson, & Sacker, 1994) in a British cohort using a similar set of behavioral ratings and in the Dunedin cohort mentioned above (Cannon et al., 2002). The 1946 British birth cohort contained children's own ratings of their behavior at age 13 years and teachers' ratings 2 years later. These data showed no evidence of antisocial traits in the preschizophrenia group, but a strong association with shy, “schizoid” behaviors at both ages. The two views gave a very similar picture; the shyer someone seemed as a child, the greater the risk. Other studies, however, remind us of the varied childhood psychiatric conditions that predate schizophrenia (Kim et al., in press).

The behavioral differences seem to persist toward the prodrome, though are independent from it. Malmberg, Lewis, David, and Auerbach (1998) studied a sample of some 50,000 men conscripted into the Swedish army at age 18 to 20 years when they underwent a range of tests and assessments. Four behavioral variables at age 18 were particularly associated with later schizophrenia: having only one or no friends, preferring to socialize in small groups, feeling more