The Scientist - Watching How the Brain Grows

Brain size is a lot like shoe size. It doesn't correlate with height, weight or even IQ, though boys tend to have larger brains (and feet) than girls. This lack of proportional comparison coupled with the fact that, like fingerprints, brains are unique, has created barriers to the better understanding of brain development. But recent imaging technology advances that factor out individual differences, as well as tools that automate data collection and quantitation, are allowing researchers to construct a detailed picture of the growing brain.

One such scientist is Judith Rapoport, director of the child psychiatry branch at the National Institute of Mental Health. Rapoport has marshaled her forces to track long-term changes in brain anatomy in the largest prospective study ever attempted of normal and abnormal children. One decade and thousands of scans later, she and her collaborators are reporting some unexpected findings that could have implications for treating and diagnosing children suffering from debilitating psychoses.

Rapoport says she feels unusually fortunate to be working in the National Institutes of Health's intramural program, where she has the luxury of time to establish the essential norms to understand brain development. In fact, Rapoport says that being at the NIH, where her projects have been afforded consistent funding, allows her to conduct research that can take years before there is a payoff.

Starting Young

Collaborating with researchers at the Montreal Neurological Institute, Rapoport has been studying brain anatomy and development in children 4 to 20 years old, ages that have been traditionally understudied and under-appreciated, she says. The Montreal group provided a sophisticated software package that combines tissue classification with anatomic information. This application, which refines earlier approaches that required manually measuring the volumes of different brain parts, is one of the keys to Rapoport's success. It provides highly automated, highly reliable measurements for over 100 brain structures, says Paul Thompson, assistant professor of neurology at the University of California, Los Angeles' Laboratory of Neuro Imaging. These measurements are essential when studying hundreds of brain features in thousands of individuals.

The results, says Rapoport, have been remarkable. In a lecture at the Annual Meeting of the Society for Neuroscience, she said, "One of the things we were able to find out, almost at once, was how unexpected the findings are just by doing something this simple—just by following a normative population." Rapoport's team has found, for example, that development is uneven across the brain—different parts peak with respect to volume or growth at different ages. In work published in 1999, they showed that the cerebellum is the last to mature, with its growth period extending into the mid-20s, long after other regions have ceased growing.¹ Doing what Rapoport calls the simplest sort of clinical genetics, comparing mono- and dizygotic twins, they have found that most areas of the brain, with the exception of the cerebellum, are highly heritable—more alike in monozygotic twins than in dizygotic siblings. The cerebellum, in contrast, is the least heritable—identical twins are no more alike than fraternal twins with respect to cerebellar size, suggesting that this particular part of the brain is more sensitive to environmental cues.

Another surprising and counter-intuitive finding, said Rapoport, is that development proceeds from the brain's anterior portions toward the posterior. Age determines where to find change: in an older person, it will be in the posterior. Though the analysis subdivides the brain into more than 100 structures, complexities still abound. Rapoport and others are trying to increase the resolution of their studies.

Morphing Brains

Thompson developed one technique that increases resolution by creating fine-structure maps of children's growing brains that have undergone significant changes in architecture. The program uses a tensor (color-coded) mapping strategy to track millions of physical landmarks in the growing brain, keeping it in focus as it morphs into new shapes. These landmarks enable alignment of scans taken at different times, warping them to make the two images fit over each other. Measuring the amount of warping reveals how much growth has occurred in each section. Thompson explains his technique's advantages this way, "Unlike earlier approaches, you see exactly which parts are growing fastest, as a color-coded picture. This can pinpoint exactly where it is that brain tissue is growing or shrinking, rather than just looking at the overall size of a part of the brain."

With this tool, Thompson and coworkers compared the brains of children from 3 to 15 years old, scanned at varying intervals—from 2 weeks to 4 years—and found a complex pattern of growth and loss in the developing brain.² In children aged 3 to 6, the greatest growth occurs in the frontal lobes, which are responsible for learning new skills and being able to think ahead. In older children, 7 to 15 years, the isthmus, which houses the language centers, showed the greatest activity. The growth rate in this region dropped off abruptly at puberty, coinciding with the end of a well known critical period for language learning. This age group also experienced an intense buildup of neural circuits in regions that handle mathematical thinking and the understanding of spatial relationships.

Martha Herbert, research neurologist at Massachusetts General Hospital's Center for Morphometric Imaging, found similar results to Thompson's in a study of cortical development in normal children. She finds the notion that regions of the brain grow at different rates fascinating. "It just pleases me every time biology is shown to be more complicated than you would think," she says. "The really cool thing about morphometry is that it can, on occasion, tell you where to look with your array of other research tools. Once you find regional differences between normal and affected people in a study of a disease, for example, tissue-based researchers can go in and study those regions looking for differences in many factors at the tissue level."

Brain Waves

With this information on normal brain development in hand, Rapoport has turned to childhood psychiatric disorders, such as attention deficit hyperactivity disorder (ADHD) and childhood onset schizophrenia (COS). She hopes that by having a large, normal population followed throughout childhood, studies of abnormal youngsters might reveal some differences that have clinical or diagnostic significance. There is additional impetus with COS, a rare disease. With a condition like schizophrenia, likely the result of the interplay among multiple genetic and environmental factors, identifying a target within the developing brain might shed some light on what triggers the disease, which has been elusive.

Rapoport's studies have revealed specific changes in the way the brain develops in each group of children. ADHD children, for example, have smaller brains overall, differences in frontal regions and unexpectedly in the basal ganglia and cerebellar vermis, the regions that control movement and planning. These changes are non-progressive, and remain fairly constant over time. By looking at monozygotic twins discordant for ADHD (one twin has it, the other doesn't), Rapoport found the same differences in brain structure, suggesting that the cause, at least in this cohort, is not genetic, but rather environmental.

COS poses a different picture. At first glance, this group resembles severe adult schizophrenics in that they have a smaller brain volume and larger ventricles. But when these children were rescanned, the researchers saw something quite different from what is seen in adult schizophrenics. In adults, the changes are largely static and non-progressive, while with COS patients, rescanning showed striking differences in comparison with control subjects.

A Schizophrenic Firestorm

Using Thompson's method for aligning anatomical landmarks, Rapoport further refined her analysis of COS brain development. Here the results were more striking still. Subtracting the COS scans from the healthy control group at the earliest time point, Rapoport's team found a gray matter loss of greater than 10% toward the back of the brain, in the parietal region, which with time, swept forward to encompass the rest of the brain. With these maps, Rapoport and Thompson pinpointed where the loss is taking place in schizophrenic children, and plotted its spread across the brain. Thompson says, "The idea of a wave of loss is intriguing. Despite some tentative ideas, we are puzzled about what causes it."

This finding raises some interesting questions about brain development. According to Thompson's earlier study, some tissue loss during adolescence is normal, but it is rather gentle, in the range of 1% per year, while with COS it can be 2% to 5% per year. Some children lost as much as 25% of their frontal lobe gray matter by the study's end.

Questions that remain unanswered include whether this wave of gray matter loss is characteristic of normal development and why it appears exaggerated in schizophrenic children. Alternatively, these disease-related changes in the pattern of brain growth might compound normal changes, leading to the cognitive deficits and other symptoms typical of COS.

This finding also raises some interesting possibilities in terms of treatment. Many drugs currently given to schizophrenics alter brain chemistry; clozapine, which is used extensively, alters dopamine activity. However, now that changes in brain structure, particularly losses in gray matter, have been detected, the emphasis might shift from chemistry-altering drugs to those that spare nerve cells.

Treasure Chest

Thompson says Rapoport, who followed her subjects individually by bringing them back every two years, has created a "treasure chest about brain development." Because of the variability in brain structure among individuals, only a study such as this could have revealed with any confidence the patterns of growth and loss. In addition, Rapoport's data on unique patient cohorts, such as the COS patients, required a prodigious nationwide recruiting effort, as well as careful screening of thousands of children to verify the diagnosis. The initial diagnosis can be difficult and prone to error, particularly with schizophrenic children.

And this is just the beginning. As researchers home in on the brain regions affected by childhood psychoses, they can begin to ask questions about the molecular events going on in the affected regions. As Rapoport puts it, "We hope that genetic studies on candidate molecules will be better informed as a consequence of these unexpected findings."

References
1. J.N. Giedd et al., "Brain development during childhood and adolescence: a longitudinal MRI study," Nature Neuroscience, 2:861-3, 1999.

2. P.M. Thompson et al., "Growth patterns in the developing brain detected by using continuum mechanical tensor maps," Nature, 404:190-3, 2000.

RESEARCH

Watching How the Brain Grows

MRI offers new insights into brain development

Starting Young

Morphing Brains

Brain Waves

A Schizophrenic Firestorm

Treasure Chest