Paperonium

Brain science · PLoS Biol · 2026

The summary says 27% of brain cells responded. The paper counts 27 of 357: about 8%.

8%Its own count27 of 35727%The summarytoo high0%102030
8%Its own count27 of 35727%The summarytoo high0%102030

Of the 357 brain cells recorded, the share that responded

The paper says

The study recorded 357 brain cells in 19 brain-surgery patients. The paper’s summary, its short overview up front, says 27% of these cells responded: they sent more or fewer signals depending on how fast the brain’s electrical activity was swinging up and down nearby.

Its own data show

The paper’s results section counts 27 such cells out of 357 and puts the share at 8%, which checks out: 27 ÷ 357 is about 8%. A true 27% would need about 96 cells. No other count or total in the paper gives 27%. The summary’s 27 matches the number of cells, not their share.

A percentage is how many out of every 100; 27 of 357 is about 8. A reader of the summary would conclude about 1 cell in 4 responds; the count says about 1 in 13. The cells are no fluke: the paper’s test would pick out about 1 cell in 100, about 4, by luck alone; it found 27.

“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 →

Human neuronal firing varies with the frequency of local field potential oscillations · PLoS Biol (2026) · doi:10.1371/journal.pbio.3003818