Schizophrenia Research Forum - A Catalyst for Creative Thinking

Lag in Verbal Growth Marks Vulnerability to Schizophrenia

7 February 2013. Falling behind in verbal development during the teenage years increases the risk for schizophrenia in adulthood, reports a study published online January 16 in JAMA Psychiatry. Led by James MacCabe of King’s College London, United Kingdom, in collaboration with colleagues at the Karolinska Institute in Stockholm, Sweden, the study tracks cognitive changes in young men between the ages of 13 and 18 using information from Swedish population databases. The study finds that a slower elaboration of cognitive ability relative to peers during adolescence is more tightly associated with future psychosis than cognitive performance at 18 years of age, and suggests that this reflects a neurodevelopmental program gone awry.

“This time during adolescence is when the finishing touches are put on the process of brain maturation,” MacCabe told SRF. “We found this is where it appears something goes wrong in people who develop schizophrenia later.”

Cognitive deficits are well-known precursors to psychosis, but their origins remain unclear. Children who do poorly in school or in certain cognitive tests are at an increased risk for developing schizophrenia as adults (MacCabe, 2008), and two years prior to psychosis onset, people exhibit various cognitive deficits (Seidman et al., 2010; see also SRF related news story). But what do such cognitive shortcomings signal? Do they stem from adolescent-stage brain development that does not unfold correctly? Or do they reflect the prodromal rumblings of psychotic illness onset? A third idea is that they are linked to social factors, such as urban living or low socioeconomic status, that are in turn associated with increased schizophrenia risk.

The new study tried to differentiate among these possibilities by looking for changes in cognition well before psychosis onset. Though previous studies have addressed this in younger children, this is the first time the crucial teenage window of 13-18 years of age has been assessed in a study with a prospective design, where cognitive abilities had been measured prior to the development of psychosis.

A repurposed cohort
First author MacCabe and colleagues drew their sample totaling 10,717 individuals from a cohort originally assembled for a program begun in the 1950s to study education and career development in Sweden. This program had tested cognitive abilities in the verbal, spatial, and inductive reasoning domains in a subset of 13-year olds in 1953, 1967, 1972, and 1977. The standardized tests were designed to be comparable to cognitive tests given at 18 years of age, when all Swedish men are evaluated for military training. The researchers then used the national patient register to identify those who were later diagnosed with schizophrenia, other nonaffective psychoses, bipolar disorder, or other affective psychoses. To differentiate between cognitive changes related to neurodevelopment versus the prodrome, the researchers excluded anyone who developed psychosis within one year of his 18-year-old assessment, as determined by hospitalization records.

In each of these psychosis groups, a relative decline was found in verbal abilities between ages 13 and 18. Though termed a “decline,” it didn’t necessarily reflect a loss of an ability possessed at age 13; rather, it meant they weren’t improving as quickly as the unaffected population did. For example, at age 13, the schizophrenia group’s verbal ability was similar to that of the unaffected group, but at age 18 they fell behind, as though unable to keep pace with the unaffected group’s progress.

This kind of verbal decline was significantly associated with schizophrenia, with greater declines associated with a greater risk. The association remained after adjusting for other social factors, such as urban living and parent educational level, suggesting the association is more than a byproduct of a person’s social milieu. A similar association arose between verbal ability decline and other nonaffective psychoses, but not for bipolar disorder or other affective psychoses. In fact, the bipolar group on average outperformed the unaffected population on these cognitive measures at both ages, but their improvement in verbal and inductive reasoning scores was not quite as marked.

Neurodevelopment versus prodrome
The researchers argue that this cognitive decline stems from a faulty neurodevelopmental program rather than the prodrome, because it preceded psychosis onset by a fair amount: by 10 years in the schizophrenia group, whose average age of onset was 28 as determined by age of first hospitalization. When the researchers more stringently tested their idea by excluding those who developed psychosis before age 25 (leaving two-thirds of their original affected group), the association endured for schizophrenia and other nonaffective psychoses.

The wide gap between the timing of the cognitive decline and the onset of psychosis suggests discrete stages along the path to psychosis. Declines tied to adolescence like the ones found here could reflect derailed brain development, whereas those linked to illness onset could signal a brain succumbing to illness. The lagging verbal abilities reported in this study may follow earlier cognitive declines in working memory and attention (Reichenberg et al., 2010), and MacCabe and colleagues suggest that this could reflect decreases in gray matter volume in speech areas, possibly related to overexuberant synaptic pruning (Ramsden et al., 2011). Though future work will have to discover the neural underpinnings of these changes, the findings highlight the teenage brain as an important source of clues to adult-onset psychotic illness.—Michele Solis.

MacCabe JH, Wicks S, Löfving S, David AS, Berndtsson A, Gustafsson JE, Allebeck P, Dalman C. Decline in Cognitive Performance Between Ages 13 and 18 Years and the Risk for Psychosis in Adulthood: A Swedish Longitudinal Cohort Study in Males. JAMA Psychiatry. 2013 Jan 16:1-10. Abstract

Comments on News and Primary Papers

Primary Papers: Decline in cognitive performance between ages 13 and 18 years and the risk for psychosis in adulthood: a Swedish longitudinal cohort study in males.

Comment by:  Terry Goldberg
Submitted 12 February 2013
Posted 12 February 2013

Intellectual decline in schizophrenia has been an area of active investigation for well over two decades. In a twin paradigm, we reported evidence for intellectual decline in the schizophrenia proband, using both reading level and current IQ (Goldberg et al., 1990). In a later study we determined that intellectual decline is present in about 50 percent of patients and appears to be the modal presentation (Weickert et al., 2000).

In the current paper, MacCabe and colleagues attempt to pinpoint the timing of cognitive decline in a Swedish population-based study. They found strong evidence for decline in verbal ability in the 13-18 years of age period. Nevertheless, there are several unexpected features to the report.

First, though it is “population based,” the schizophrenia/schizoaffective N = 50 and bipolar N = 18 are rather small compared to many other studies in the area.

Second, the schizophrenia group has a relatively late onset (28 years at first hospitalization). This has important implications for the authors’ claim that the decline was not related to a prodrome and invites some skepticism.

Third, the authors comment that decline does not reflect frank deterioration, but rather a slowing of developmental gains. I agree with this, but this could have been directly addressed by examining raw scores, not standardized scores. See work which suggests slowed maturation by taking this former approach (Bedwell et al., 1999).

Conceptually, it is unclear what is behind the decline. One possibility not commented upon is that compromises in episodic memory prevent the acquisition of new information, including vocabulary. Hence, the language decline would be secondary to a failure in memory. A study by Elvevåg et al. (Elvevåg et al., 2003) is in keeping with this view.


Bedwell JS, Keller B, Smith AK, Hamburger S, Kumra S, Rapoport JL Why does postpsychotic IQ decline in childhood-onset schizophrenia? Am J Psychiatry 156:1996-7,1999. Abstract

Elvevåg B, Kerbs KM, Malley JD, Seely E and Goldberg TE. Autobiographical memory in schizophrenia. An examination of the distribution of memories. Neuropsychology 17: 402-409, 2003. Abstract

Goldberg TE, Ragland DR, Gold J, Bigelow LB, Torrey EF and Weinberger DR: Neuropsychological assessment of monozygotic twins discordant for schizophrenia. Archives of General Psychiatry 47: 1066-1072, 1990. Abstract

Weickert T, Egan MF, Weinberger DR and Goldberg TE: Cognitive impairments in patients with schizophrenia displaying preserved and compromised intellect. Archives of General Psychiatry 57: 907-913, 2000. Abstract

View all comments by Terry Goldberg

Primary Papers: Decline in cognitive performance between ages 13 and 18 years and the risk for psychosis in adulthood: a Swedish longitudinal cohort study in males.

Comment by:  Timothea Toulopoulou
Submitted 22 February 2013
Posted 22 February 2013

Adolescence is a period of dramatic brain maturation and reorganization, and may also be a time of particular risk for pathophysiological processes underpinning neurodevelopmental models for psychiatric disorders such as schizophrenia. It, therefore, represents a particularly good period to examine for cognitive changes, as indices of that neurobiological risk, as they occur during that time of development. MacCabe’s excellent work complements numerous other streams of research that point in a similar direction, i.e., that a failure of normal verbal maturation during adolescence indicates later psychosis risk. However, without wanting to take anything away from the value of the study, we are left wondering whether we have been distracted by large samples, and association studies, for too long, perhaps confusing bigger with better? Epidemiological studies cannot tell us about the underlying neurobiology of the changes that occur during this period of development, and isn’t that where we should focus?

View all comments by Timothea Toulopoulou