Paperonium

Brain imaging · Cereb Cortex · 2026

Two brain systems work together more in adults, the paper says. Its table: could be chance.

1 in 100every time1 in 20the usual barrarer than1 in 20The text: passes, even after correctingabout1 in 8Its own Table 3,before any correcting
1 in 100every time1 in 20the usual barrarer than1 in 20The text: passes,even after correctingabout1 in 8Its own Table 3,before any correcting

How often chance alone would give this adult–child gap with no real difference

The paper says

In scans of 15 children and 16 adults chatting with ChatGPT, the summary says two brain systems, one for focus and self-control, one that picks out what matters, work together more in adults. The results add: this survives the standard correction for running many tests.

Its own data show

The text gives no number for this gap. Table 3 does: p = 0.130, so chance alone would produce a gap this big about 1 time in 8. That fails the usual bar before any correcting; the table doesn’t mark it as passing. Correcting for its 21 pairs of systems pushes it to 0.228.

A p-value is roughly how often chance alone would produce a gap this big if there were no real difference: smaller is stronger; the usual bar: 1 in 20. More tests give chance more tries, so correcting can only raise it. The summary highlights two links stronger in adults, building its closing idea on this one. Only the other survives.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

Lower engagement of cognitive control in children on ChatGPT · Cereb Cortex (2026) · doi:10.1093/cercor/bhag167