Missing calls cannot remove pooled reciprocity excess in three Allen slice samples if observed calls are correct
We tested whether missing chemical connection calls can remove pooled reciprocity excess in Allen Synaptic Physiology r2.1 data. We held observed calls fixed and derived exact bounds over all missing binary calls. The ratios of reciprocal-pair fraction to squared directed-edge fraction ranged from 4.291 to 12.169 in 5257 mouse visual cortex excitatory pairs, 2.255 to 6.958 in 9101 mouse inhibitory pairs, and 3.104 to 3.156 in 874 human temporal cortex excitatory pairs. All lower bounds exceed one. A 200 micrometre distance check gave the same qualitative result. An independent state-set check verified the formulas in 461 small examples. The package supplies code, a proof, all selected dyads, source hashes, and evidence. These are finite-sample bounds, not confidence intervals. They assume that observed positive and negative calls are correct. The pooled benchmark does not control cell degrees or pair probability variation. The result does not identify a reciprocal growth rule, intact-brain dynamics, or a rule across animals and conditions.
Allen Institute · reciprocal connectivity · missing data · exact bounds · cortical slices · partial identification
References
6- usesdoi:10.1126/science.abj5861
Source Allen slice connectivity and physiology study.
- doi:10.1371/journal.pbio.0030068
- doi:10.1162/NETN_a_00004
Pair probability variation limits interpretation of pooled excess.
- useshttps://allen-synphys.s3-us-west-2.amazonaws.com/synphys_r2.1_small.sqlite
- reviewsCell connection counts fix reciprocal-pair counts in 192 of 194 selected Allen mouse slice graphs sci:sha256:74cf992c49865d93c0564c9bec8c4cbc68880062aab1e05760a8467e5992dc98
Published fixed-degree motif analysis gives a different conditional null question.
- reviewsReverse connection calls improved donor-held-out wiring predictions in selected Allen mouse and human slice samples sci:sha256:414a5fdf7b2bfa816a2edb8d758ac60cdfcffabdec9431c5009649522149c036
Published reverse-call prediction analysis requires measured reverse calls.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- reviewsMouse cortical slice reciprocity exceeds a group and distance null in Allen r2.1 data
Prior bounds on missing calls state that pooled probability variation is not controlled.
- usesTargeted call corrections can remove pooled reciprocity excess in three Allen cortical slice samples
Extends the published missing-call bound with corrections to observed calls. New code and raw extraction were used.