Paperonium

Brain imaging · Hum Brain Mapp · 2009

Paper: Alzheimer’s brain thinning detected within a year. Its own numbers say it could be chance.

1 in 1001 in 1,000every time1 in 20the usual bar1 in 1,000The paper’s textabout1 in 4Its own score
1 in 1001 in 1,000every time1 in 20the usual bar1 in 1,000The paper’s textabout1 in 4Its own score

How often chance alone would produce this much thinning in 9 Alzheimer’s patients

The paper says

The paper’s new way to measure, from brain scans, how thick the brain’s outer layer is found it thinned on average in 9 Alzheimer’s patients scanned twice, a year apart. Chance alone, it says, would produce that only 1 time in 1,000 (p = 0.001).

Its own data show

Beside 0.001 the paper prints the score that figure should come from: the average change weighed against how much patients varied. For 9 patients that score, 1.2, gives 0.26, about 1 time in 4; 1 in 1,000 needs a score near 5. Rebuilding the score from the paper’s table of averages gives 0.26 too.

A p-value is how often chance alone would give a change this big; under 0.05, 1 in 20, counts as convincing. At 0.26, the paper’s only test on real patients can’t tell the thinning from chance. It also suggests rival program FreeSurfer errs because, by its measure, healthy brains thickened on average (p = 0.001), which they shouldn’t. Redone: 0.86.

“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 →

Measurement of Cortical Thickness from MRI by Minimum Line Integrals on Soft-Classified Tissue · Hum Brain Mapp (2009) · doi:10.1002/hbm.20740