Positive EPSP weights gave no clear loop product increase over shuffled weights in 151 human epilepsy resection panels

We tested whether measured synaptic strengths add short feedback to fixed human cortical connection graphs. We used positive mean excitatory postsynaptic potentials (EPSPs) as relative edge weights. A loop score multiplies the weights around each closed path. The main sample had 151 epilepsy resection recording panels with 916 connections, 70 reciprocal pairs, and 30 three cell loops. We compared the recorded scores with weight shuffles that kept each graph and its strength values fixed. The reciprocal score was 67.605, against an exact shuffle mean of 64.736. The ratio was 1.044, and the two tail rank from 19,999 draws was 0.677. The three cell ratio was 1.065, with a rank of 0.713. Shuffles within source or target neurons also gave no clear reciprocal excess. A smaller tumor resection sample gave a higher reciprocal score, but its patient resample range included one. Fixing its reciprocal strength set gave a pairing ratio of 1.006. The scores are short return terms under a scalar linear rule with common leak. Voltage means are not calibrated coupling gains. These fixed slice graphs do not establish live cortical feedback or equal effects across patients. The package preserves source tables, exact means, all draws, code, figures, source checks, and repeat instructions.

human cortex · synaptic strength · EPSP · feedback · directed cycles · weight placement · patch clamp