Stanford fMRI study: group-averaged brain data hides how individuals actually think

Stanford Medicine researchers analyzed fMRI scans of more than 4,000 nine- and ten-year-olds performing a stop-signal task and found that conclusions drawn from group averages frequently invert what is happening inside any single child’s brain. Slower reaction times correlated with increased default mode network activity at the group level, but within individuals the same slowdown corresponded to decreased activity in that network — the opposite signal. Published in Nature Communications, the work argues that conventional averaging methods systematically mischaracterize the neural dynamics behind cognitive control.

The team, led by Percy Mistry, Nicholas Branigan, and Vinod Menon, also identified subgroups with distinct adaptive or maladaptive regulation patterns and showed that inhibitory control is not a single capacity but a set of separable proactive and reactive pathways that different children recruit differently. Some apparent group effects turned out to be driven entirely by one subgroup, masking opposite behavior in others.

The implication is methodological: neuroscience and psychiatry research that relies on cohort averages — including much of the existing literature on ADHD, bipolar disorder, and addiction — may be describing a statistical artifact rather than any real individual brain. The authors push for person-level temporal modeling as the baseline approach, with downstream consequences for how behavioral therapies and classroom interventions for inhibitory control are designed.