The ADHD Brain Reaches a Developmental Peak Three Years Later

Timing one landmark

A magnetic resonance scan of a child’s head can be sorted by computer into white matter, grey matter and cerebrospinal fluid. Two surfaces are then fitted to the grey matter, one at its outer edge and one at the boundary with the white matter beneath it, and the distance between those surfaces is measured in millimetres at 40,960 points across the cortex. That distance is cortical thickness.

Across childhood the measurement rises, reaches a maximum, and then falls through adolescence. Fitting a curve to repeated scans of the same child, and to scans of many children of different ages, allows the age of that maximum to be estimated point by point. The age of peak cortical thickness serves as an index of how far a region has progressed, and it can be calculated only where the data actually describe a rise followed by a fall.

Eight hundred and twenty-four scans, all acquired on the same machine from 223 children with ADHD and 223 typically developing children matched for sex, age and measured intelligence, were analysed in this way by Shaw and colleagues at the National Institute of Mental Health, with collaborators at the Montreal Neurological Institute. The results appeared in the Proceedings of the National Academy of Sciences in 2007. Just over half the children were scanned more than once, at a mean interval of about three years.

The median age by which half of all cortical points had attained peak thickness was 10.5 years in the ADHD group and 7.5 years in the control group. The order in which regions matured was the same in both groups, with primary sensory and motor areas peaking first and the association areas that integrate their signals peaking last.

Regionally the difference was uneven. It was largest in the middle prefrontal cortex, where the ADHD group reached peak thickness approximately five years after the controls, and about two years in the superior and medial prefrontal cortex. Posterior regions were involved as well, the middle and superior temporal cortex extending into the middle occipital gyri peaking at 10.6 years against 6.8. The parietal comparison could not be made cleanly, because in the ADHD group the superior parietal lobules and postcentral gyri followed a straight line rather than a curve with a turning point.

One region ran the other way. In the primary motor cortex the ADHD group peaked slightly earlier, at 7.0 years against 7.4.

Reading a three-year interval

A delay, in this study, is a statement about when a curve turns. The scans record the timing of one anatomical landmark in two groups; they do not record what any child could do on the day of the scan. Clinical outcome data were unavailable for most of the ADHD participants, so the authors could not examine whether the timing of a child’s peak related to how that child fared.

The paper does not answer whether the two trajectories eventually meet. The age range covered was insufficient to identify when adolescent thinning levels off into stable adult dimensions, and the authors write that they predict this endpoint would also arrive later in ADHD. A prediction in a discussion section is a hypothesis awaiting a study. What the paper establishes is that one landmark arrived roughly three years late.

The regions with the greatest delay support the suppression of unwanted responses and the deliberate control of attention. That correspondence is why the finding drew notice, and it remains a correspondence. Timing was measured in anatomy, ADHD was established by interview and rating scale, and the two were connected at the level of groups. Nothing in these scans explains any individual child’s difficulties, and no scan can be read backwards to account for a particular afternoon.

Some of the children with ADHD had been treated with stimulants before entering the study, in numbers the paper reports inconsistently, and while the group’s earlier work found no effect of stimulants on grey matter volume, an observational design leaves any conclusion about medication tentative. Children with an IQ below 80 were excluded, and the average measured intelligence in both groups was above 100.

What survives all of that is the unit. A difference of this kind is counted in years, and years are longer than a term, a report cycle, or the interval between one parents’ evening and the next. A child whose middle prefrontal cortex reaches its maximum at ten rather than seven has spent the intervening years meeting the same demands as everyone else, from a point earlier in a sequence every cortex follows in the same order. Group anatomy supports that description of children in aggregate and cannot certify it for a named one.

The study measured its interval in years. Years remain the honest unit for watching your own child change.