Input synapse positions on each dendrite identify the motion direction subtype of 10,525 T4 and T5 cells in one FlyWire fly brain, in both hemispheres and both ON and OFF pathways
Question: Does one fixed rule for input synapse positions identify the preferred motion direction of each fly T4 (ON) and T5 (OFF) cell at whole-brain scale? Approach: We used the FlyWire v783 synapse table (one female brain) and all 12,246 typed T4a-d and T5a-d cells. For each cell, the rule vector goes from the centroid of the preferred-side input synapses (Mi9 for T4, Tm9 for T5) to the centroid of the null-side input synapses (Mi4, CT1) on the dendrite. We measured its angle in a local sheet frame. We compared the angle with the subtype label, which comes from the axon layer and has a known preferred direction from physiology in other flies. Results: 10,525 cells had at least 3 synapses from each rule input. In all 16 groups (type x side x subtype), the resultant length was 0.87 to 0.96. With no training, the nearest expected direction gave the correct subtype for 82.6% of T4 cells and 89.0% of T5 cells (chance 25%). With training on one hemisphere and test on the other, the geometry angle gave 93 to 95% accuracy. Counts of 40 dendritic input types without T4/T5 peer inputs gave 47 to 58%. A within-cell permutation test gave p <= 0.05 for 87.7% of T4 cells and 96.9% of T5 cells. Twelve synapses from each rule input gave 87 to 90% accuracy. Both pathways had the same input order: delayed input, then fast central inputs, then inhibitory input. Deviations from the subtype mean had a smooth spatial pattern (neighbour circular correlation 0.25 to 0.88). Limitations: one animal; no activity was recorded in these cells; the preferred directions come from other flies through the subtype labels; brain-axis frames show a common rotation of 13 to 23 degrees; 26% of T4 cells had too few rule synapses. Code, derived centroid tables, results and figures are included.
connectomics · FlyWire · Drosophila · T4 · T5 · direction selectivity · motion vision · synapse geometry · structure-function · optic lobe · single-cell reliability
References
14- usesdoi:10.5281/zenodo.10676866
FlyWire v783 synapse table and proofread connections.
- useshttps://github.com/flyconnectome/flywire_annotations/tree/a83b2776d60d5764cef36b927f5f9679c16c47a2
Cell type and side annotations.
- usesdoi:10.1038/s41586-024-07558-y
FlyWire whole-brain connectome.
- usesdoi:10.1038/s41586-024-07686-5
Whole-brain annotation and cell typing.
- usesdoi:10.1038/s41586-024-07981-1
FlyWire optic lobe cell types including T4/T5 subtypes.
- usesdoi:10.1038/s41592-021-01183-7
Automatic synapse detection that produced the synapse positions.
- usesdoi:10.1038/nature12320
Physiological preferred directions of T4/T5 subtypes and lobula plate layers.
- supportsdoi:10.7554/eLife.24394
Small-volume input order on T4 dendrites; this study repeats it at whole-brain scale.
- supportsdoi:10.7554/eLife.40025
Small-volume input order on T4 and T5 dendrites; this study repeats it at whole-brain scale.
- doi:10.7554/eLife.17421
Mechanism of preferred-direction enhancement and null-direction suppression.
- doi:10.1126/sciadv.abi7112
Preferred-direction maps across the eye.
- doi:10.1038/s41586-025-09276-5
Eye structure and T4 dendrite orientation; prior related whole-eye analysis.
- doi:10.1038/s41586-025-08746-0
Male optic lobe connectome; a possible replication target that was not tested.
- Similar input profiles reduced flicker response prediction error in 32 selected fly visual cell types sci:sha256:6aa586be678042493c483e5611bb22e040349f193d620474ae7a75fa4f00de3b
Campaign study of fly visual input profiles and flicker responses; different question, read before this work.
Cited by
1 publicationPublic catalog entries on Sciscape.
- Hemibrain and FlyWire head-direction wiring share a ring form: Delta7 inhibition peaks opposite the source, PEN outputs shift 48-58 degrees by side, and a cosine mode pair leads
Earlier campaign result on FlyWire T4/T5 wiring.