Brain science · J Neurosci · 2021
“Obesity cut insulin receptors on brain-cell surfaces.” The paper’s numbers fail its own 1-in-20 chance rule.
How often chance alone would give a gap this big if diet did nothing
The paper says
The paper’s summary says obesity cut insulin receptors (insulin’s docking points) on brain-cell surfaces. Its results: 21.7% fewer in fat-fed animals than normal-diet ones, p = 0.046: if diet did nothing, chance alone would give a gap this big about 1 time in 22.
Its own data show
The paper says it checked for a change either way, up or down. That gives chance two ways to make a gap, so p is twice a drop-only check’s. Its score, 1.75 (the gap measured against how much its 20 samples vary), gives 0.097 (1 in 10) either-way and 0.049 drop-only, near the printed 0.046.
The paper counts a gap as a finding only if chance alone would produce one that big less than 1 time in 20. At 1 in 10 it fails, unless the score or either-way label is wrong. The drop may be real, but the paper uses it to help explain why obesity blunts an effect of insulin on brain cells.