After recording and cell type controls, two-way connections showed no clear extra time constant match in 589 human layer 2-3 temporal cortex pairs

We tested whether two-way excitatory connections join cells with more similar membrane time constants than one-way connections. We used the Planert et al. version 1 tables for human temporal cortex. Both connection directions had to be explicitly tested. The final sample had 589 layer 2-3 cell pairs from 21 donors, including 50 two-way pairs. The outcome was the absolute log2 ratio of cell time constants. A linear model controlled recording group, location, mean cell time and physiological cell classes. The adjusted two-way minus one-way difference was -0.0093 log2 units. Its 95% range from 5,000 complete-donor bootstrap fits was -0.0948 to +0.0678. Dropping cell class controls gave -0.0096. Stricter cell quality gave +0.0362, also with a range that included zero. The source audit found repeated rows for 98 human directions, including 23 with different connection calls; all repeated directions were excluded before reverse joins. A secondary Allen r2.1 mouse test had only 112 pairs and six recording groups with both connection types. Its estimate was uncertain, and 234 of 5,000 resampled designs did not identify the coefficient. Independent calculations reproduced the human selection, coefficients and clustered errors. This test gives no clear evidence for extra time matching in two-way human connections under the stated controls. It does not prove independence or measure circuit memory. The package preserves code, source hashes, exact row selections, resample estimates, figures and repeat instructions.

membrane time constant · reciprocal connectivity · human temporal cortex · paired recordings · structure and function · synaptic physiology · reproducible analysis