Brain Mapping in Adolescents With Very Early Onset Schizophrenia

March 1, 2003

Schizophrenia is a devastating psychiatric disorder that affects 1% of the population worldwide. Patients often suffer their first psychotic outbreak in their late teens or early 20s. Despite advances in neuroleptic drugs, many patients' symptoms remain refractory to treatment, with recurrent episodes of auditory and visual hallucinations, bizarre delusions, depression, and social withdrawal that can last an entire lifetime.

Schizophrenia is a devastating psychiatric disorder that affects 1% of the population worldwide. Patients often suffer their first psychotic outbreak in their late teens or early 20s. Despite advances in neuroleptic drugs, many patients' symptoms remain refractory to treatment, with recurrent episodes of auditory and visual hallucinations, bizarre delusions, depression, and social withdrawal that can last an entire lifetime. Neuroimaging studies now suggest that schizophrenia is a disorder of brain development, with anatomic abnormalities present at disease onset. Teen-agers with a severe, early-onset form of schizophrenia also exhibit a dynamically spreading wave of cortical gray matter loss, detectable in sequential magnetic resonance imaging scans. The tissue loss begins in a small region of the parietal cortex and moves forward to engulf frontal and temporal systems. These deficits correlate with psychotic symptom severity and may link with cortical dopamine or serotonin dysfunction. The shifting pattern of deficits is distinct from the neurodegeneration observed in the dementias and may be an exaggeration or derailment of the neuronal remodeling that normally occurs in late teen-age brain development. Computerized tracking of these cortical deficits will help understand how neuroleptic drugs decelerate or block the disease process. Cortical deficits are also detectable in patients' first-degree relatives, who are at greatly increased genetic risk for schizophrenia (10% lifetime risk). In the future, these dynamic and genetic brain maps may predict imminent onset of the disease, identifying pre-symptomatic brain changes in family members who are candidates for early interventions.

One of the greatest enigmas in contemporary psychiatry is why schizophrenia strikes in the late teen-age years or young adulthood, often without warning. With an average age of onset around 20 to 25 in men and 25 to 30 in women, psychotic outbreaks in schizophrenia may include delusions, hallucinations and bizarre thoughts (i.e., positive symptoms). Negative symptoms include chronic depression, flattened affect, poverty of speech, loss of motivation and social decline. If these symptoms are untreated, the active phase of florid psychotic symptoms may last forever; however, they may be controlled to a degree by antipsychotics. Even when medications are effective, psychotic outbreaks are often replaced by a residual phase of poverty of thought or blunted affect. Around 20% of patients have a single psychotic outbreak, and 35% have multiple episodes without severe functional or personality impairments (Green, 1999). The remainder of patients has relatively static (10%) or progressive (35%) functional impairments between psychotic episodes.

Unlike Alzheimer's disease, where amyloid plaques, neurofibrillary tangles and neuronal loss are pervasive in the brain at autopsy, there are no widely accepted pathologic hallmarks of schizophrenia. This is frustrating, as a biochemical marker could provide the basis for a diagnostic test and a target for drug action or disease prevention. Nonetheless, multiple lines of evidence suggest that cortical neurotransmitter function is disturbed in schizophrenia. Abnormalities in cortical dopamine, serotonin, glutamate,

-aminobutyric acid (GABA) and norepinephrine have been intensively investigated. Classical antipsychotics (e.g., haloperidol [Haldol]) alleviate positive symptoms by blocking dopamine (D2-type) receptors in the limbic and prefrontal cortices of the brain, systems that regulate emotion and executive function. Newer atypical antipsychotics, including clozapine (Clozaril) and olanzapine (Zyprexa), powerfully block the 5-HT2 serotonin and D4 dopamine receptors and tend to outperform haloperidol in reducing negative symptoms (Bilder et al., 2002). Intriguingly, genetic studies suggest that 2% of patients with schizophrenia exhibit a chromosomal deletion in region 22q11, which harbors the gene encoding catechol-O-methyltransferase (COMT), a powerful inactivator of dopamine. Mice with targeted deletion of this gene have excess dopamine in the prefrontal cortex. This is consistent with the notion that patients with this deletion (which confers a 25% to 30% lifetime risk of schizophrenia) may suffer from a functional excess of cortical dopamine, and this may be responsible for their positive symptoms.

Advances in neuroimaging, and structural MRI in particular, have empowered the search for biological markers of schizophrenia. Reduced cortical and hippocampal volume are found consistently in patients with schizophrenia, and the ventricular and sulcal cerebral spinal fluid spaces are often enlarged. Diffuse gray matter deficits are observed on MRI even in first-episode patients, where the confounding effects of medication on brain structure are ruled out. There is great interest in identifying when these anatomical deficits first appear. If their origins were identified, it may be possible to pinpoint precisely where, and when, an active pathological process begins. This could allow earlier disease detection and targeted interventions.

Abnormal Brain Development

Over 20 years of studies suggest that the origins of schizophrenia, as well as multiple risk factors, may lie in childhood or embryonic brain development. Obstetric risk factors, which confer a later risk for schizophrenia, include fetal malnutrition, extreme prematurity, hypoxia and ischemia (Cannon et al., 2002b). People born in winter months (Kirch, 1993) or exposed to the influenza virus in the second trimester (Mednick et al., 1988) may also have an increased incidence of schizophrenia. Some studies have contested this association, but others suggest that early viral exposure may increase risk for other psychiatric disorders as well (Akil and Weinberger, 2000). Given the array of proposed risk factors, disrupted brain development may play a causative role in schizophrenia. If this is the case, a key puzzle is why there is a long gap between an early cerebral insult and the emergence of symptoms 20 or more years later. To explain this, some theorists favor a two-hit (or diathesis-stress) model, in which an early developmental or genetic anomaly must be compounded by psychological trauma, viral infection or some currently unknown trigger later in life for the disease to be expressed.

Renewed interest in the developmental hypothesis comes from recent brain imaging studies. These identify a drastic remodeling of brain structure in the teen-age years and beyond. Well into adolescence, there are growth spurts in myelination (Thompson et al., 2000), and dramatic waves of gray matter loss (Giedd et al., 1999a; Sowell et al., 1999). In a landmark paper based on MRIs of healthy subjects, Giedd et al. (1999b) built quadratic growth curves between the ages of 4 and 20 for gray matter volumes in each lobe of the brain. Perhaps surprisingly, the overall volume of gray matter declined sharply after the age of 12, especially in frontal and parietal cortices. This process continued through adolescence and beyond, with the latest decrements occurring in the frontal cortex (Sowell et al., 1999). Since schizophrenia typically strikes at a time when these developmental changes are still occurring, an intriguing hypothesis is that a normal teen-age process of dendritic remodeling and synapse elimination (sometimes called pruning) may be accelerated or otherwise derailed in schizophrenia (Feinberg, 1982). This excessive pruning may reach a threshold level where cortical information processing is disrupted. It may also account for the increased neuronal packing density seen in some cortical layers in postmortem studies of patients with schizophrenia (Selemon et al., 1995).

Early-Onset Schizophrenia

In a large-scale effort to map the trajectory of brain changes during development, Judith Rapoport, M.D., and her colleagues at the National Institute of Mental Health have scanned over 1,000 children and teen-agers with high-resolution MRIs (Rapoport et al., 1999). Most of these children have been scanned every two years since 1992, producing a remarkable time-lapse movie showing their brain development. Among those patients scanned at NIMH were 50 adolescents (30 boys, 20 girls) with early-onset schizophrenia (EOS). These patients satisfied DSM-III-R/DSM-IV criteria for diagnosis of schizophrenia before age 13. Rigorous clinical and cognitive evaluations revealed their symptoms were continuous with the adult disorder; many patients resemble poor-outcome adult cases (Rapoport and Inoff-Germain, 2000).

Brain Mapping

My colleagues and I studied these brain scans in collaboration with the NIMH group. We developed computerized methods to pinpoint rates of brain growth and gray matter loss in individual children and teen-agers and visualized these patterns as color-coded three-dimensional maps. Combining data from multiple subjects, we compared the amount of gray matter in multiple cortical regions across subjects and across sequential scans. This analysis produced color-coded three-dimensional maps of the cortex showing loss rates and group differences, and highlighting regions where these changes link with outcome measures or symptom severity. The brain maps for patients with EOS, published in the Proceedings of the National Academy of Sciences (Thompson et al., 2001), revealed a surprisingly dynamic pattern of disease progression. At their first scan (mean age=13), patients showed a 10% gray matter deficit, which was confined to a small region of the parietal cortex involved in spatial association, compared to healthy controls. Over the five succeeding years, this brain tissue loss swept forward, like a forest fire, into frontal and temporal brain regions.

Male and female patients showed a similar, dynamically spreading pattern of deficits. The frontal eye fields lost tissue fastest (5% per year); frontal and temporal regions were spared initially, but were subsequently engulfed. By age 18, gray matter was reduced by up to 20% to 25% in some brain regions.

Symptoms and Medication

The spreading deficits correlated, in some respects, with functional decline as well. Total frontal loss rates correlated with negative symptoms (total scores on the Scale for Assessing Negative Symptoms [SANS]) at final scan (p<0.038). This makes sense, as negative symptoms are thought to derive in part from reduced dopaminergic activity in frontal cortices. At an individual level, rates of temporal loss correlated strongly with Scale for Assessing Positive Symptoms (SAPS) total scores at final scan (p<0.015, left hemisphere; p<0.004, right hemisphere). Faster losses in both the superior temporal gyri and the entire temporal cortices were significantly associated with a more severe clinical profile of positive symptoms (e.g., hallucinations or delusions). While tissue loss rates were not significantly linked with the rate of change in SAPS scores from baseline (p>0.05), and SAPS scores were not linked with the amount of tissue at baseline (p>0.05), loss rates were a good predictor of positive symptoms at follow-up (i.e., the remaining symptoms that were refractory to medication).

To rule out confounding medication effects, a second medication- and IQ-matched control group was also studied, consisting of patients diagnosed with psychosis not otherwise specified. These patients exhibited only mild tissue loss, essentially confined to superior frontal cortices in a highly circumscribed pattern. Importantly, the pervasive, unrelenting spread of tissue loss was specific to schizophrenia and was not medication induced (although its rate could well be modulated by medication).

An exciting open question is how strongly different antipsychotic drugs combat this wave of loss. Clinical studies in adult patients, using MRI to assess cortical integrity, suggest that atypical drugs decelerate overall gray matter loss, while gray matter progressively declines in patients taking haloperidol alone (J. Lieberman, M.D., personal communication, 2001). In both teen-age- and adult-onset patients, these progressive losses appear to continue for a long period after diagnosis (here seven years), offering a window of opportunity for interventions. Dynamic brain maps, such as the ones shown here, may help evaluate the spatial selectivity of these medication responses when comparing different neuroleptics. Computational brain maps also make it easy to stratify cohorts into subgroups with different symptom profiles. Patterns of brain change in responders may, in the future, be compared with groups who remain refractory to treatment.

Pathologic Mechanism and Specificity

A shifting pattern of deficits raises perplexing questions. Is the wave of gray matter loss an exaggeration of a "normal wave" of gray matter pruning? Or is it a separate process entirely that begins in the teen-age years? Could it be a neurodegenerative process, similar to the progressive wave of gray matter loss seen in Alzheimer's disease (Thompson et al., 2003)?

The strongest evidence against extensive neuronal loss in schizophrenia is the lack of reactive gliosis. Glial cell swelling and proliferation is the brain's natural response to neuronal cell death, and it is strikingly absent in postmortem studies. Nonetheless, in vivo proton magnetic resonance spectroscopy studies show that frontal N-acetylaspartate is reduced, and this is a good marker of neuronal integrity. Positron emission tomography also shows reduced frontal lobe glucose metabolism in both early- and adult-onset patients. Impaired cortical activation is also seen in functional MRI studies of working memory and executive function. These deficits are clearest in tasks (e.g., Wisconsin Card Sorting) that place heavy demands on frontal systems.

Rather than widespread cell loss (as in the dementias), it is more likely that neuronal shrinkage, reductions in dendritic complexity and synaptic loss, as well as vascular changes, may underlie the gray matter changes observed here in schizophrenia. An intriguing hypothesis is that schizophrenia patients may suffer an abnormal intensification of whatever regional gray matter loss occurs in healthy subjects at their specific age of onset. This idea may reconcile why deficits in EOS and adult-onset cases differ, somewhat, in their scope and severity. Predominantly frontal and temporal gray matter losses tend to be reported in adult-onset patients, while more pervasive losses are seen in early-onset studies. Frontal and temporal lobes are the last to show gray matter loss in healthy controls. These systems may be especially vulnerable if the disease begins in young adulthood, while they are being actively remodeled. Additional cortical systems may be vulnerable if the disease hits earlier, in the early teen-age years.

Genetics, the Prodrome and the Future

Two final studies illustrate the usefulness of neuroimaging as a biomarker in schizophrenia. In a recent twin study (Cannon et al., 2002a), we detected frontal and temporal cortical deficits (around 5% to 8%) in healthy relatives of patients who are at increased genetic risk for developing the disease. These deficits were correlated with the degree of genetic affinity to a patient (i.e., worse deficits in identical twins of patients than fraternal twins, for example, as the latter share fewer genes with a patient). Patients' siblings and children have a 10% lifetime risk of developing schizophrenia, much greater than the 1% risk in the general population.

Isolation of a brain deficit that is an index of liability for schizophrenia is important for two reasons. First, a heritable biomarker, found in at-risk relatives, can be used in genetic association and linkage studies. Differences between MRIs of relatives can be covaried with the number of alleles they share with a patient at a candidate marker locus, to see if particular genetic loci play a role in increasing liability. The notion of genetic linkage (which links trait variation with allelic variation at a marker locus) can be generalized to the notion of a brain map of linkage, showing brain deficits that are linked with allelic variations (e.g., Thompson et al., 2003). As neuroimaging and genetic databases increase in size and content, the merger of genetics and neuroimaging will empower the search for (and characterization of) susceptibility genes. This is likely to clarify their effects on brain phenotype and disease progression in human populations.

Secondly, analysis of brain changes in those at genetic risk may help identify relatives who are in the prodromal (i.e., pre-symptomatic) phase of the disease. Since individual outcomes depend heavily on how early the disease process is detected, relatives with both genetic and neuroimaging risk markers may be able to use this information proactively. Armed with this knowledge, they may in the future opt for early interventions and drug treatment before the ravages of the disease have set in.


Special thanks go to the members of the NIMH Child Psychiatry Branch, the University of California, Los Angeles, Laboratory of Neuroimaging, and the Queensland Center for Magnetic Resonance, for their key role in the studies summarized here.




Akil M, Weinberger DR (2000), Neuropathology and the neurodevelopmental model. In: The Neuropathology of Schizophrenia: Progress and Interpretation, Harrison PJ, Roberts GW, eds. New York: Oxford University Press, pp189-212.


Bilder RM, Goldman RS, Volavka J et al. (2002), Neurocognitive effects of clozapine, olanzapine, risperidone, and haloperidol in patients with chronic schizophrenia or schizoaffective disorder. Am J Psychiatry 159(6):1018-1028 [see comment].


Cannon TD, Thompson PM, van Erp TG et al. (2002a), Cortex mapping reveals regionally specific patterns of genetic and disease-specific gray-matter deficits in twins discordant for schizophrenia. Proc Natl Acad Sci U S A 99(5):3228-3233.


Cannon TD, van Erp TG, Rosso IM et al. (2002b), Fetal hypoxia and structural brain abnormalities in schizophrenic patients, their siblings, and controls. Arch Gen Psychiatry 59(1):35-41.


Feinberg I (1982), Schizophrenia: caused by a fault in programmed synaptic elimination during adolescence? J Psychiatr Res 17(4):319-334.


Giedd JN, Blumenthal J, Jeffries NO et al. (1999a), Brain development during childhood and adolescence: a longitudinal MRI study. Nat Neurosci 2(10):861-863 [letter].


Giedd JN, Jeffries NO, Blumenthal J et al. (1999b), Childhood-onset schizophrenia: progressive brain changes during adolescence. Biol Psychiatry 46(7):892-898 [see comment].


Green B (1999), A review of schizophrenia. Psychiatry Online. Available at: Accessed February 18, 2003.


Kirch DG (1993), Infection and autoimmunity as etiologic factors in schizophrenia: a review and reappraisal. Schizophr Bull 19(2):355-370.


Mednick SA, Machon RA, Huttunen MO, Bonett D (1988), Adult schizophrenia following prenatal exposure to an influenza epidemic. Arch Gen Psychiatry 45(2):189-192.


Rapoport JL, Giedd JN, Blumenthal J et al. (1999), Progressive cortical change during adolescence in childhood-onset schizophrenia. A longitudinal magnetic resonance imaging study. Arch Gen Psychiatry 56(7):649-654.


Rapoport JL, Inoff-Germain G (2000), Update on childhood-onset schizophrenia. Curr Psychiatry Rep 2(5):410-415.


Selemon LD, Rajkowska G, Goldman-Rakic PS (1995), Abnormally high neuronal density in the schizophrenic cortex. A morphometric analysis of prefrontal area 9 and occipital area 17. Arch Gen Psychiatry 52(10):805-818.


Sowell ER, Thompson PM, Holmes CJ et al. (1999), In vivo evidence for post-adolescent brain maturation in frontal and striatal regions. Nat Neurosci 2(10):859-861 [letter].


Thompson PM, Giedd JN, Woods RP et al. (2000), Growth patterns in the developing human brain detected using continuum mechanical tensor mapping. Nature 404(6774):190-193.


Thompson PM, Hayashi KM, de Zubicaray G et al. (2003), Dynamics of gray matter loss in Alzheimer's disease. J Neurosci 23(3):994-1005.


Thompson PM, Vidal C, Giedd JN et al. (2001), Mapping adolescent brain change reveals dynamic wave of accelerated gray matter loss in very early-onset schizophrenia. Proc Natl Acad Sciences U S A 98(20):11650-11655.