Shared input in three fly maps had correlations of 0.08 to 0.11 with calcium coactivity of 11 wing muscles in selected female flight bouts
We tested whether shared input to motor neurons was related to wing muscle coactivity in selected female flight bouts. We used all connection rows in MANC v1.2.1, Male CNS v1.0, and the BANC v888 compiled v3 table. For 11 matched muscles, we calculated input cosine separately on each side and then averaged sides. We compared the 55 pair scores with within-movie calcium correlations from 377 movies, 74 session groups, and 33 recording days. Each day had equal weight. The correlations were 0.0973 for MANC, 0.1079 for Male CNS, and 0.0833 for BANC. Full muscle-label tails were 0.2355, 0.2169, and 0.2698. Label tests within fixed attachment groups also gave no clear match. The anatomical pair scores agreed strongly across maps: r = 0.9978 for MANC and Male CNS, and 0.9786 for MANC and BANC. We scanned all 170,782,903 released connection rows and tested complete two-edge input walks. The tested unsigned walk models gave correlations from 0.0114 to 0.1104. Source quality rules reproduced the reported total of 72,219 wingbeats and left the direct scores nearly unchanged. The result limits a simple count-based input model. It does not test fast spike coordination or exclude correlated drive, signed dynamics, or measurement effects. BANC and the upstream and quality checks were added after the primary result. The same selected recordings are used for all three maps. The package preserves exact analysis inputs, source hashes, code, figures, results, and an offline repeat with identical numerical tables.
connectome · Drosophila · wing motor neurons · muscle calcium · structure and function · shared input · flight control · reproducible benchmark
References
8- usesdoi:10.1038/s41586-024-07600-z
Its shared-input analysis and motor-module hypotheses motivate the fixed cosine test.
- usesdoi:10.1038/s41586-024-07293-4
We analyze its released steering muscle calcium data and check its quality rules.
- usesdoi:10.22002/aypcy-ck464
We use version 1.0 of main_muscle_and_wing_data.h5 and preserve exact trace extracts.
- usesdoi:10.7554/eLife.97769
We use the MANC v1.2.1 compiled connectivity and neuron metadata.
- usesdoi:10.1101/2025.10.09.680999
We use the Male CNS v1.0 flat-connectome data with confidence threshold 0.5.
- usesdoi:10.1038/s41586-026-10735-w
We add a female BANC check after the primary result.
- usesdoi:10.7910/DVN/7WTH1N
We use Dataverse dataset version 3.0, BANC v888 metadata and the v3 simple edge table.
- supportsdoi:10.1016/j.cub.2025.12.024
It describes rapid mechanosensory flight pathways. Our calcium coactivity test does not test that timing mechanism.
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesA prior from shared motor input did not reduce mean error in a fixed linear model of wing motion from calcium on 31 fly recording days
We use its verified motor target matches, full input extracts, and source quality rules. We add wing motion outcomes and a new prediction task.
- usesAt fixed target scales and gain 0.8, source variances bounded by one million fitted 11 wing muscle correlations in two signed fly connectome models
Supplies matched wing muscle identities, trace extracts and quality rules.
- Wiring similarity of fly wing steering motor neurons in four connectomes matched flight co-activation of basalar and first-axillary muscles (Spearman 0.65) but not of the other six muscles (-0.12)
[Legacy relation: reproduces] We repeat its test (input cosine vs co-activation, 55 pairs) with 16.9 h of continuous, non-selected flight traces instead of triggered movies, and add FANC; we get Spearman 0.12 (FANC 0.13) vs its 0.08-0.11.
- usesAt gain 0.8, predicted transmitter signs gave higher wing coactivity scores than all 199 shuffles within cell class and side in two linear fly models
We reuse its verified muscle trace extracts, quality masks, target matches, and extraction rules. We extend its direct and unsigned path tests with signed paths of all lengths and conditional sign references.
- usesA fixed ridge model gave correlations near -0.15 between input and motion similarity for 11 wing muscles in three fly maps
We reuse the attributed direct input extracts, muscle matches, movie ledger, quality masks and source provenance. We test a new motion-profile target.