A growth chart for the brain's white matter
Processing 35,120 brain scans from 50 studies, a team has built reference charts for white-matter tracts across the whole human lifespan, from birth to one hundred years. The brain now has the kind of growth chart children have long had for height and weight.
A paediatrician can plot a child’s height on a centile chart and say at a glance whether the line looks right. For the brain, until recently no equivalent existed. Whole-brain volume and grey matter have begun to get reference curves, but white matter, the wiring that carries signals between regions, had no standard against which an individual scan could be judged. A paper in Nature by Michael E Kim, Chenyu Gao, Karthik Ramadass and colleagues, published online on 13 May 2026, supplies that missing chart, mapping how the brain’s cables grow, mature and fade across a century of life.
The scale of the data is what makes it reference-grade. The team processed and standardized 35,120 diffusion scans drawn from 50 imaging studies around the world, covering typically developing and ageing participants from birth to one hundred years old. Diffusion imaging does not photograph structure directly; it tracks the diffusion of water molecules, which runs fast and straight along myelinated nerve fibres and slowly across them. From that signal the group derived, for named tracts, both macrostructure (volume) and microstructure (fractional anisotropy and related measures). The trajectories were modelled with generalized additive models for location, scale and shape, so that median, spread and the tails of each curve can all be read off.
The headline finding is that different parts of the wiring mature and decay on different clocks. White-matter volume rises through development, peaks in adolescence or early adulthood and then declines, but the peak age is tract-specific. Microstructure, measured by fractional anisotropy, increases through childhood, plateaus in adulthood and dips in later life, again with the timing varying pathway by pathway. The arcuate fasciculus that supports language and the corticospinal tract that drives movement do not follow the same schedule, and neither does the genu of the corpus callosum that bridges the two hemispheres. A single average, the paper makes clear, would have hidden all of it.
There is also a message about variability. Across the lifespan the spread of values is widest early in life, narrows to a minimum in middle age and widens again as people grow old. Middle age, on these charts, is the period when brains look most like one another; the extremes, at both ends of life, are where an individual deviates most from the crowd. The charts are released open access, so any new scan can be compared against the normative centiles and a z-score read off.
That is the practical point. White-matter disruption is implicated in dementia, in developmental disorders and in the aftermath of brain injury, but without a reference range a radiologist looking at one scan has few anchors. A chart turns a snapshot into a trajectory: a tract that sits at the third centile for a seventy-year-old means something quite different from the same measurement at age ten. The hope is that early deviations, rather than outright damage, become detectable.
The limits are worth naming. The cohort, drawn from 50 studies, is weighted toward convenience samples and high-income settings, so the charts encode one world’s typical development rather than humanity’s. Diffusion imaging also has well-known artefacts at older ages and across different scanners, which standardized processing only partly removes. And a normative band is not a diagnosis; it flags where to look, not what is wrong.
Yet the contribution is foundational rather than flashy. For decades brain imaging has compared patient groups against controls and published averages. It has lacked the boring, essential instrument of a growth chart. With 35,120 scans behind it, white matter finally has one.