Four pairs of missing delays need joint recovery to add constraints in selected Allen mouse cell-delay models

We tested whether missing delays must be planned together to add tests of an additive source-plus-target timing model. In Allen Synaptic Physiology r2.1 mouse visual cortex graphs, twelve missing delays can each add one constraint alone. Four other pairs each add one constraint only if both values are recovered. The pairs remain under stricter labels, with seven single recoveries. Under independent equal recovery chance p, attempts on all twenty primary candidates add an expected 12p + 4p squared constraints. This is seven at p = 0.5, compared with sixteen under certain success. Plans tied under certain success can differ by 0.25 expected constraints at p = 0.5. Human temporal cortex graphs add no constraint under full recovery. Exact subset calculations, numerical rank checks and exhaustive graph tests support the result. Recovery chances are assumptions. No new delays or model validity were measured. Code, aggregate evidence, source hashes and retrieval steps are supplied. External records are excluded.

Allen Institute · synaptic delay · measurement selection · graph cycles · recovery failure