Cracking the brain’s genetic code

In PNAS, Fjell et al. (1) reveal a remarkable connection between genetics and neuroscience—a new discovery that reveals how genes shape our path through life. Their article draws thought-provoking links between modern genetics and how our brains develop, and even how we age: They find unsuspected patterns in how our genetic code drives childhood development; these patterns may even explain why our brains decline in characteristic ways late in life.

In their prior work (2–4), the authors discovered “modules” of genetic organization in the human brain: clusters of key brain systems, each shaped by a common genetic blueprint. With this new genetic map of the brain as a guide, Fjell et al. (1) now show how these genetic modules change in unique ways, as we progress through life, using brain scans collected from over 1,000 people aged 4–88 y.

The ebb and flow of brain development and aging follow their new genetic blueprint. Each of their genetic modules has a distinct profile of development and aging, pointing to organized clusters or “communities” of brain regions with common genetic determination and common trajectories of aging. The results are remarkable, and shed light on some unresolved questions in neuroscience, which are outlined in this commentary. If their genetic findings hold up in additional samples, they point to more efficient ways to identify genes and pathways that influence brain development and aging, and causal mechanisms that drive our brains toward mental illness or protect us.

The new findings are encouraging for large-scale “biobanking” initiatives, many of which are starting up worldwide. Several biobanking efforts seek to discover factors that help or harm the brain, and our risk for psychiatric illness. As explained below, the new findings point to a roadmap that may speed up discoveries in psychiatric genetics. To see why, we need to consider classical neuroscientific theories of brain development and aging, and how large-scale imaging projects have led us to question and revise them.

During the past 20 y of research, a highly dynamic picture of brain development has emerged, perhaps more dynamic than was suspected before modern brain scanning with MRI. Based on classical neuroanatomic data, it had long been known that the brain develops in a characteristic sequence. However, neuroimaging revealed, perhaps surprisingly to many researchers, that the sequence of brain remodeling exhibits a “shifting” pattern of tissue growth and loss. These changes can be seen and measured in brain scans, well into the teenage years and young adulthood. One “evolutionary-developmental” hypothesis suggests

Ultimately, the new paper by Fjell et al. offers us a roadmap to understand why the brain ebbs and flows in its unique pattern of development and decline.

that in humans and primates, basic brain functions develop first. Primary visual and sensorimotor areas, critical for survival, are quick to mature in infancy. Higher order language areas then undergo a burst of growth, followed by limbic systems involved in motivation and emotion. Much later to develop are our frontal brain regions, crucial for planning and self-control, and these frontal brain regions continue to mature well into our early 20s.

Pioneering work in the 1990s by Judith Rapoport and her colleagues at the National Institute of Mental Health created “time-lapse” movies tracking these shifts in the brain’s development. This landmark project scanned children with MRI every 2 y, from 4–21 y of age (5). The most remarkable finding was a “wave” of development, sweeping forward into the frontal lobes. This dynamic pattern, captured using MRI scans, offered clues about teenage risk for addictive behaviors and why mental illnesses, such as schizophrenia and major depression, sometimes strike without warning in late adolescence.

On the other hand, MRI scans of the aging brain show shifts in the pattern of decay that unravel in the reverse sequence (6). People differ in how their brain loses tissue, but memory systems tend to decline fastest; time-lapse movies of Alzheimer’s progression show sensory and visual areas largely intact until late in the disease, which are cortical brain areas that develop first in infancy.

Why should the brain develop in this sequence? Why would it unravel in the opposite way? Some suspect an evolutionary reason: If phylogenetically older parts of the brain develop early, perhaps newer genetic programs come online later to cope with the brain’s growing demands for higher cognition; others point to a neuroprotective effect of myelination, a heavy coating of lipids that speeds up neural communication in visual and sensory circuits, perhaps protecting them from late-life pathologies, such as amyloid, tau, and vascular damage (7). Hallmarks of Alzheimer’s disease build up fastest in “plastic” brain regions, such as the hippocampi. These regions are memory areas that are relentlessly remodeled throughout life, yet are so highly vulnerable to the dementing diseases of old age.

The paper in PNAS by Fjell et al. (1) points to a different, genetic explanation for these patterns—that the brain has distinct units, each with a common genetic program. What is fascinating is that these genetic units develop and age in unique ways, now detectable with a population analysis of brain MRI. A distinctive rise and fall for each genetic unit may explain the shifting patterns of growth and decay; each contributes, in turn, to the palindromic waves of change that we all undergo as our brains age.

Fjell et al.’s report (1) is the culmination of several years of work using twin and family studies to detect these genetic modules in the brain: sets of brain regions influenced by the same or overlapping sets of genes. Well over 100 papers have compared brain measures in identical and fraternal twins; around one-third to one-half of the individual variation in brain measures is due to genetic variations among us, and much of the rest is shaped by environmental factors ranging from education and diet to trauma and stress or by unmodeled and unknown sources of variance. When genetic models are fitted to model signals in brain scans, almost all brain measures are strongly genetically influenced: cortical gray matter thickness (8), the size of subcortical structures (9), the hemodynamic response in functional MRI (10), and even the pattern of brain synchrony seen with EEG (11). The brain’s anatomical network, the so-called structural “connectome,” is under remarkably tight genetic control: Powerful algorithms are now screening these networks for markers in the genome that affect them (12).

Finding Genetic Clusters

Fjell et al.’s recent discovery of genetic modules in the brain (1) relies on a classical method in quantitative genetics that allows one to say what fraction of the correlation between two measures is due to common genetic factors. We can find out when common genetic variants cause two measures to be correlated, such as gray matter and intelligence quotient. Genetic analysis of correlated brain traits quickly refueled the acrimonious debate between hereditarian and environmental points of view on cognition, and educational attainment. New methods have emerged to assess genetic correlations from genome-wide association studies (13), which test over 1 million genetic variants for associations with a trait. Large-scale, genome-wide screening of hundreds of thousands of people reveals that genetic risk factors for schizophrenia, bipolar illness, and major depression overlap (14). The discovery of these overlapping genetic pathways has sped up the search for treatable targets in these brain disorders. Furthermore, by computing genetic correlations, not just for diseases but for measures of the brain’s structural variability, genetic modules have emerged, leading to a new way to partition the brain along genetic lines (2–4).

In their new report, Fjell et al. (1) take what is known about these genetic modules further. Plotting the trajectories of cortical thinning for each genetic cluster, an established marker of brain maturation and aging that has often been linked with cognitive performance, each cluster follows the now-accepted model of growth and decline. We see fastest decline in the teenage years and a plateau in old age. This finding was once highly controversial. When the correlations are tested between the lifespan trajectories of these brain regions, however, each genetic cluster ages in a different way, supporting the idea of a genetic mosaic of sectors that grow and decline in a sequence.

The impact of this discovery for brain mapping initiatives is tantalizing. The Enhancing Neuro Imaging Genetics through Meta-Analysis (ENIGMA) Consortium and other large scientific consortia, including Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE), recently performed a worldwide genetic screen of brain data from over 30,000 people, and found eight common variants in the genome that affect the size of key structures within the brain; some of these genetic changes are common in the population and appear to age the brain by about 3 y (9). Worldwide genomics efforts are now so highly powered that studies can unearth genetic markers that account for as little as 0.25% of the variance in brain measures, with strikingly consistent effects across 30 countries of the world. Encouraged by this finding, ongoing genetic screens are underway in tens of thousands of brain datasets to isolate common and rare variants, and epigenetic changes, that affect brain integrity, and even synchrony and activity in functional MRI (15).

More Efficient Screening

The results reported by Fjell et al. (1) suggest that a complementary and more biologically informed approach may help us to discover genetic markers that affect the brain and disease risk: Rather than screen the brain’s regions, one by one, for common variants that affect each, we should first group brain systems into genetic modules with more coherent genetic determination. Genetic clustering of diffusion imaging data suggests that this approach may discover key genes more efficiently (16). Better models of gene action in the brain, spatially and over time, should expedite global efforts to find causal variants for mental illnesses, factors that affect childhood development, and dementing conditions in old age. Partnerships between genomics consortia, such as the Psychiatric Genomics Consortium (PGC), and imaging consortia, such as ENIGMA, will add data and insight in this quest. Large-scale neuroscience efforts are now compiling brain growth and aging charts for over 10,000 people, shedding light on long-debated left- and right-hemisphere differences (16). These discoveries will help to refine the genetic mosaic hypothesis, where the brain’s genetic modules show distinct paths through life, rising and falling according to a script, like voices in a musical fugue.

Ultimately, the new paper by Fjell et al. (1) offers us a roadmap to understand why the brain ebbs and flows in its unique pattern of development and decline. A full knowledge of all of the factors that shape our brains throughout life remains a “mystery inside an enigma.” These new genetic mosaics offer a valuable key to crack the brain’s genetic code, however.