What VFB says DNp32 releases
The class descending neuron of the posterior brain DNp32 is typed Ms neuron and adult peptidergic neuron (a neurosecretory cell of the adult brain), after Namiki et al. 2018. Each EM reconstruction then carries its own classifier prediction, and the three connectomes do not agree.
| Reconstruction | Predicted transmitter | Confidence | Source |
|---|---|---|---|
| FlyWire L / R | dopamine | 51% / 56% | Eckstein et al. 2024 |
| Hemibrain DNp32_R | octopamine | 43% | Eckstein et al. 2024 |
| Male CNS L / R | serotonin | 58% / 57% | Berg et al. 2025 (preprint) |
| FAFB skid 430 | no prediction | ||
Three different monoamines at ~50% on a curated peptidergic cell reads as the classifier reacting to dense-core-vesicle ultrastructure, not as evidence for any one of them. Safer wording: an Ms-expressing descending neuron with an unresolved monoamine prediction.
Mushroom body output onto DNp32
Across hemibrain, FlyWire and male CNS, weight ≥ 1: 15 MBON → DNp32 pairs, 63 synapses, nothing above 8. Two types recur in every dataset.
| Pair | Male CNS | FlyWire | Hemibrain |
|---|---|---|---|
| MBON35 (γ2) → DNp32 | 8, 7 | 6 | 2 |
| MBON20 (γ1γ2) → DNp32 | 7, 4 | 6, 3 | 7 |
| MBON17 → DNp32 | – | 4 | 2 |
| MBON04 / 15 / 25 / 34 | 2 (MBON15) | – | 3 / – / 1 / 1 |
| DNp32 → MBON20 (reciprocal) | 5 | – | 3 |
So there is a weak MBON20 ⇄ DNp32 loop that survives in two datasets, and MBON35 is the strongest single input from the mushroom body. The MBONs are not drawn above; they live in the brain view of the earlier query.
Flight and halteres: what the data actually holds
DNp32 → IN18B011 14 / 14 → sternotrochanter MN 81 / 41 · tergotrochanteral (jump) MN 23 / 21 · tibia extensor MN 31 · STTMm 23 · MNhl59 24
DNp32 → IN12B031 14 → sternotrochanter MN 18 · MNhl02 10 · MNhl01 9
DNp32 → IN06B064 15 → sternal anterior rotator MN 3 · hg4 MN 2
DNp32 → DNd03 27 → PS100 36 · GNG650 18 · MNwm36 2 · DVMn 1a–c 2
This matches the curated anatomy: the class definition lists synaptic input/output in T1–T3 leg neuropils, tectulum and gnathal ganglion, not wing or haltere neuropil. The 18B hemilineage is the takeoff/jump circuit, and the tergotrochanteral MN is the jump muscle, so if DNp32 has a role around flight it is upstream of takeoff, not in wing or haltere steering. If the flight/haltere link comes from a different source (imaging, a different DN identity), that is where to look next.
Brain-side output
DNp32 has as many synapses in the brain as in the cord. The strongest brain targets are the GABAergic WED055 cluster in the wedge, the DN–DN partners DNge075 and DNd03, and the serotonergic CSD interneuron. Direct output to PAM dopaminergic neurons exists but is 2–3 synapses.
| Target | Transmitter (predicted) | DNp32 → | → DNp32 |
|---|---|---|---|
| WED055_b (several cells) | GABA | 43, 36, 34, 24… | 0 |
| DNge075 | ACh | 42, 28 | 6, 19 |
| CB3660 / CB3630 | Glu | 38, 29 | 0, 4 |
| CSD | 5-HT | 38 | 4 |
| AVLP075 | Glu | 33, 28 | 0 |
| DNd03 | Glu / OA | 27, 16 | 15, 6 |
| DNp32 (contralateral) | — | 16 | 8 |
Method and sources
Skeletons are the male CNS v1.0 reconstructions (Berg et al. 2025, CC-BY 4.0) as served by VFB in JRC2018U space, resampled to 2 µm. The JRC2018U brain and JRCVNC2018U cord outlines are decimated VFB template meshes; the cord outline was fitted into the brain-space layout with an affine transform (median residual 6.5 µm), so treat it as context, not registration.
Connectivity is from VFB's NeuronNeuronConnectivityQuery on the two male-CNS DNp32 cells and their eleven strongest targets, plus query_connectivity for MBON ↔ DNp32 across all datasets. Every neuron in the viewer links to its VFB page.
Links
DNp32_R (MaleCNS:524968) · DNp32_L (MaleCNS:556286) · DNp32 class · Namiki et al. 2018 · Berg et al. 2025 · Eckstein et al. 2024