Reassignment within recording and cell type raised expected second-response sums by 5.9% on average in 23 human cortical donors

We tested separate assignment of response parameters in human temporal cortex layer 2-3. The sample had 780 detected connections from 23 donor labels. We kept every connection and first response peak fixed. We permuted 50 ms pulse ratios within each recording and ordered cell-type pair. Exact calculation gave a mean donor increase of 5.86% in the expected second-response sum. Its descriptive 95% donor bootstrap range was 4.22% to 7.64%. Expected sums increased in 21 donors. Four aggregate second-to-first ratios crossed one. The stricter target-cell and source-type rule gave a 1.70% increase. A signal-to-noise filter of at least 10 reduced the main increase to 0.93%, with a range of 0.13% to 1.76%. Filtering changed both the measurements and the exchange groups. A shared ratio denominator can also cause inverse association from noise. The study therefore does not identify biological covariance or a change in neuronal firing. We provide an exact assignment audit, standard formula derivations, source hashes, code, results, figures, and repeat instructions. Independent checks passed. A full repeat from fresh source bytes matched all 33 numeric files. Original source data remain external under their own terms.

human cortex · paired-pulse ratio · joint parameter assignment · functional connectivity · synaptic dynamics · reproducible analysis