Castellanos, FX, PP Lee, W Sharp, NO Jeffries, DK Greenstein, LS Clasen, JD Blumenthal, RS James, CL Ebens, JM Walter, A Zijdenbos, AC Evans, JN Giedd and JL Rapoport. 2002. Developmental trajectories of brain volume abnormalities in children and adolescents with attention-deficit/hyperactivity disorder. JAMA 288:1740-1748.

Press coverage of this study:
New York Times
Los Angeles Times

Lingering doubts have persisted among skeptics that attention-deficit/hyperactivity disorder (ADHD) is not a real phenomenon but instead the creation of overly protective or reluctantly disciplinarian parents, or of a pharmaceutical industry wanting to justify sales of drugs to the last remaining new demographic in the pharmaceutical marketplace, children.

This paper should dispel those doubts resoundingly. Castellanos et al. show that the brains of children with attention-deficit/hyperactivity disorder (ADHD) average smaller than children without ADHD. Their study provides compelling evidence that ADHD is a real, biologically-based phenomenon not imagined by parents nor teachers nor caused by medical treatment for the condition.

From the earliest age studied, the trajectory of brain growth in ADHD children paralleled that of normal children, but at each age brain size overall, and that of most brain subregions, were smaller than normal.

Children diagnosed with ADHD but not receiving medical treatment also had reduced brain size.

Because differences between ADHD children and controls were apparent even at the youngest age (ages ranged from 5 to 18) Castellano's research does reveal when the differences in brain growth initially arose. They could arise from genetic differences, they could involve prenatal interference with brain development, or they could appear in early childhood.

What did they do? Castellanos et al. used magnetic resonance imaging equipment (MRI) to scan the brains of 152 children (89 male; 63 female) diagnosed with ADHD and compared measured brain volumes with 139 unrelated healthy children (83 male; 56 female). Some of the subjects were scanned up to 4 times over the course of the 10-year study. Their MRI scanner allowed them to examine not only overall brain volume, but also the sizes of different brain subregions, and to calculate the volumes of gray and white matter compartments in frontal, temporal, parietal and occiptal lobes, as well as basal ganglia and cerebellum.

The research team then used statistical procedures that allowed them to test for differences between ADHD and normal children, and within ADHD for the effects of medical treatment. These procedures adjusted for variables such as differences in overall body size, gender, age, etc.

What did they find? Brain volume was slightly smaller in ADHD patients. On average, after controlling for size and other variables, cerebral and cerebellar volumes were 3.2% and 3.5% smaller in ADHD patients. These differences were highly significant statistically. The largest difference observed was a 9.2% decline in temporal white matter.

Using an additional statistical step Castellanos et al. determined that the cerebellum decreased disproportionately more than other brain parts.

And looking at the effect of medication, Castellanos' team observed that unmedicated ADHD total cerebral and cerebellar volumes were smaller than controls, but indistinguishable from medicated ADHD patients. Hence the brain volume differences observed between ADHD and controls are not caused by medication.

While medicated ADHD patients did not differ from controls on any white matter measures, "robust differences from controls remained for all gray matter measures."

Examining changes in brain volume over time, the scientists found that for all but one region studied, differences observed in the youngest patients continued through teenage years. In other words, there was no acceleration in brain volume growth that might have allowed the ADHD volumes to "catch up" to normal brain size. Equally important, these data on developmental trends establish that the volumetric differences are evident at the youngest ages measured. This indicates that ADHD begins very early in life, possibly even in the womb.

More detailed analysis revealed that measured brain subregions were smaller in ADHD patients than controls at all ages, with one exception in later teenage years.

The exception involved the caudate region, where volume normally peaks around the age of ten and then declines. This is shown in the left graph, below. ADHD caudate volume tracked beneath that of controls at the peak and then on through early-teenage years. In mid-teenage years and beyond, however, caudate volume in ADHD patients did not fall proportionately as far as normal caudate volume. The greater decline in normal caudate volume then led to convergence of normal brains with ADHD brains in this one parameter.

By comparison (right), cerebellar volume continues to grow throughout the ages studied. At each stage in development, ADHD volume is lower than control volume. It would appear that ADHD cerebellar growth may approach a plateau during teenage years more rapidly than that of the controls.

What does it mean? The volume measurements reported by Castellanos et al. clearly indicate that that brains of control and ADHD patients differ. This is important, because it demonstrates that ADHD has a biological basis with measurable physical manifestations. Their work also shows ADHD is not caused by caused by medication. Instead it is related to changes in patterns brain growth initiated in early childhood, or in the womb. These changes persist throughough childhood and teenage years.

The results reveal little about the detailed structural changes that almost certainly underlie the volume differences. It therefore remains impossible to say what stimulates the changes in growth or what mechanisms link the overt changes in brain volume reported here with the behavioral manifestations of ADHD.

Interviewed by the Los Angeles Times, the scientists said that they believe the data indicate ADHD is caused prenatally.