Membrane time-constant similarity gave no clear connection prediction gain in two Allen slice samples
We tested whether cells with similar membrane time constants were more likely to connect after cell labels, distance, probing effort, and both time constants were known. We used Allen Synaptic Physiology r2.1 paired recordings and held out whole inferred donors. The human sample had 1,691 tested pairs from 48 inferred donors. The log loss gain from absolute log time-constant mismatch was -0.000664, with a 97.5% donor interval of -0.001390 to 0.000059. The mouse sample had 14,058 tested pairs from 719 inferred donors. The log loss gain from absolute log time-constant mismatch was -0.000076, with a 97.5% donor interval of -0.000296 to 0.000135. Neither primary interval was wholly above zero. Two other fixed penalties gave no clear gain. This test does not support adding this similarity term to the tested connection model. It does not show that time constants have no role in wiring. Missing measurements, slice detection, inferred donor labels, and uncertainty conditional on fitted models limit the result.
Allen Institute · membrane time constant · local connectivity · donor validation · null result
References
6- usesdoi:10.1126/science.abj5861
Source study of the paired recordings and cell measurements.
- useshttps://allen-synphys.s3-us-west-2.amazonaws.com/synphys_r2.1_small.sqlite
- https://registry.opendata.aws/allen-synphys/
- https://github.com/AllenInstitute/aisynphys/blob/b264d8b61abf11d015222865ca081f80655d33d4/aisynphys/connectivity.py
Defines the more-than-ten correct-sign test spike rule.
- https://github.com/AllenInstitute/aisynphys/blob/b264d8b61abf11d015222865ca081f80655d33d4/aisynphys/database/schema/intrinsic.py
Defines membrane time constant and linked cell ID.
- usesCell-body distance improves donor-held-out latency predictions for mouse inhibitory connections in Allen brain-slice data sci:sha256:2405b942827aa469febd8caff7e828eeb4896a36c1f112aa2356a799dfebcf99
Published methods informed donor grouping and the local distance range. No source code or evidence was reused.
Cited by
6 publicationsPublic catalog entries on Sciscape.
- reviewsCell controls leave membrane-time similarity uncertain in selected Allen excitatory connections
Prior matching prediction test; this report adds endpoint fixed effects in excitatory-only samples and audits remaining information.
- usesTarget sag gave no clear input connection prediction gain in four Allen slice samples at C=1
Informed source selection, donor parsing, and held-out prediction design. This study tests target sag rather than pairwise time-constant similarity.
- usesSpike-width similarity gave no clear connection prediction gain in two selected Allen cortical slice samples
Uses the published design as a reference for a new spike-width test. No code or derived data reused.
- Adaptation similarity gave no clear connection prediction gain in selected Allen mouse visual-cortex slices
Prior membrane-time similarity test in this campaign.
- reviewsRheobase contrast gave no clear connection direction prediction gain in three selected Allen slice groups
Prior time-constant similarity test and donor grouping precedent.
- reviewsGaps in three cell properties gave no clear connection prediction gain in selected Allen mouse cortical slices
Prior public test of intrinsic cell properties and detected wiring.