Single-edge exclusions restore bounded cell-gain fits without a unique edge diagnosis in seven selected Allen mouse groups

We tested whether failures of a current-to-voltage cell-gain model identify specific connections in Allen cortical slices. We used the paired amplitudes and selection from a prior signed study. At an assumed independent factor 1.10 error bound on each amplitude, five mouse inhibitory groups and two mouse excitatory groups failed the model. Exhaustive search found that one edge exclusion restored fit in every failing group. Each group had two to six alternative minimum repairs. No edge was present in every repair. In six groups every minimum repair removed all cycle constraints; in one inhibitory group a constraint remained. At zero tolerance, the nine testable mouse groups needed eleven exclusions in total. No human group was testable. These are conditional sensitivity and diagnosis results for released fits. They do not identify faulty measurements or biological mechanisms. Sparse slice graphs, missing paired measurements and assumed errors limit general use. Code, aggregate results, a figure, source hashes and external retrieval steps support reproduction.

Allen Institute · synaptic amplitudes · cell gains · fault localization · measurement cycles · minimum exclusions