Drosophila melanogaster head bristle mechanosensory neuron (BMN) projections and their synapses.

(A,B) Bristles of the anterior (A) and posterior (B) head. Color-coded dots on the right indicate bristle classifications whose names are abbreviated (full names below). Dorsal and ventral views are shown in Figure 1 – figure supplement 1. Su bristles are classified, but no associated BMNs have been identified. (C,D) Reconstructed BMN projections in the ventral brain (left, previously described in (Eichler et al., 2024)) and their corresponding pre- and postsynaptic sites (right, this study), colored by type according to the bristles that they innervate. Shown are the anterior (C) and dorsal (D) views. Overlapping projection zones are evident where synapses of different BMN types spatially intermingle, whereas segregated zones show little or no color mixing. Arrows indicate the projection directions for each incoming BMN nerve bundle. Note: BMNs are from the anatomical left side of the head, but are displayed inverted on the right as previously described (Schlegel et al., 2024). Scale bars, 50 μm (full brain) and 20 μM (anterior and dorsal zoom views). Medial and lateral views of the projections and synapses are shown in Figure 1 – figure supplement 2 and Figure 1 – figure supplement 3, respectively. (E) Relative numbers of total synapses for each head BMN type. BMNs are named according to the bristle population that they innervate (e.g. BM-Taste neurons innervate Taste bristles). Rectangles correspond to individual BMNs whose relative areas indicate the number of total pre- and postsynaptic sites. Colors indicate individual BMNs of the same type. Underlying data is in Supplementary file 1. Abbreviations used to identify the bristles (A) and BMNs (E) are as follows: antennal (Ant), frontal (Fr), orbital (Or), frontoorbital (FrOr), ocellar (Oc), interocellar (InOc), vibrissae (Vib), vertical (Vt), dorsal occipital (dOcci), dorsal postorbital (dPoOr), interommatidial (InOm), ventral occipital (vOcci), ventral postorbital (vPoOr), Taste (Taste), haustellum (Hau), and maxillary palp (MaPa). Panels A and B were reproduced under the terms of the CCBY license from Figure 1A,B of Eichler et al. (Eichler et al., 2024).

Head BMN type synapse numbers plotted onto their corresponding head bristles.

(A) Table indicating for each BMN type: numbers of BMNs (n), percent of total head BMN synapses, and the average number of synapses. The percent of total number of synapses was calculated using the total number of input and output synapses for each type, divided by the total synapses for all BMN types. The average number of synapses was calculated using the sum of input and output synapses for each type, divided by the number of BMNs in the type. Box plots of BMN input and output synapses by type are shown in Figure 2 – figure supplement 1. (B-C) Dot shading on the anterior and posterior head indicates the percent total head BMN input and output synapses (B) or the average synapses per BMN type (C). *Calculations were done using only BMN synapses from connections to partners that were pre- or postsynaptic by at least five synapses. Therefore, fewer than the total number of BM-InOm neurons were included in the analysis because they did not meet this threshold. Underlying data is in Supplementary file 1.

Pre- and postsynaptic connectome of the BMNs.

(A) Sankey diagram of normalized synapse fractions between BMNs and different partner categories. Black bars in the center of the plot represent the normalized total of postsynaptic sites for each given BMN type, and colored bars along the periphery represent partner categories. All synapse fractions were normalized for visualization by making them proportional to the total number of output synapses for the given BMN type. Boxes to the left of BMNs are presynaptic and those to the right are postsynaptic. Each partner category (except motor neurons) has one subset that is purely pre- or postsynaptic, and another that is both pre- and postsynaptic (pre/post). These subsets are separated and displayed on both sides of BMNs to reveal what proportion of presynaptic input arises from postsynaptic partners. Colors for A match the partner categories in B. Raw and normalized underlying data are in Supplementary file 4 and Supplementary file 5, respectively. (B) Sunburst plots showing the composition of partners (top) or synapses (bottom) that are pre- and postsynaptic to BMNs at a 5 synapse connection threshold. Inner layer shows the proportion of partners or synapses that are presynaptic, postsynaptic, or pre- and postsynaptic (pre- and postsynaptic synapses (bottom plot) include pre/post neurons). Middle layer categorizes the partners or associated synapses as being interneurons, descending or ascending neurons, motor neurons, or BMNs. Outer layer categorizes the partners or synapses based on neurotransmitter prediction. White in the outer rings indicate neurotransmitters could not be predicted. Underlying data are in Supplementary file 6.

Motor neurons projecting to the proboscis or antennae are postsynaptic to the BMNs on the same appendages.

(A) Anterior view of reconstructed BMNs projecting into the brain from the anatomical left side of the head. Note: although the BMNs are from the left side of the head, they are displayed in the right brain hemisphere to be consistent with how they are displayed in FlyWire.ai (Schlegel et al., 2024). (B) Anterior, dorsal, and lateral views of reconstructed motor neurons that are postsynaptic partners of the BMNs with a 5 synapse connection threshold. Colors indicate the nerve that the motor neuron axons project through to the periphery, including the labial (LabNv, magenta), pharyngeal (PhyNv, green), and antennal (AntNv, blue) nerves. Nerves in both brain hemispheres are indicated with colored arrows. Note: to maintain consistency with FlyWire.ai, all neurons in this manuscript are displayed in the opposite brain hemisphere (e.g left hemisphere neurons are shown on the right; (Schlegel et al., 2024)). (C) Heatmap indicating the number of presynaptic BMNs of each type that are connected with different motor neurons. Grayscale shading scale maximum indicates 22 presynaptic BMNs. Colors indicate nerve class. FlyWire.ai neuron identification numbers are shown for each motor neuron. Underlying data are in Supplementary file 7. (D-F) Colored dots indicate bristles that are innervated by BMN types presynaptic to the indicated motor neuron nerve class, including the labial (D), pharyngeal (E), and antennal (F) nerve classes. (D,E) Posterior views. (E) The connections of BM-Ant neurons with one of the seven pharyngeal motor neurons (C) is not shown. (F) Anterior (left) and posterior (right) views.

BMN pre- and postsynaptic partners in the CNS.

(A) Anterior view of reconstructed BMNs projecting into the brain from the anatomical left side of the head. Note: to maintain consistency with FlyWire.ai, all neurons in this manuscript are displayed in the opposite brain hemisphere (e.g left hemisphere neurons are shown on the right; (Schlegel et al., 2024)). (B) Anterior, dorsal, and lateral views of reconstructed synaptic partners of the BMNs with a 5 synapse connection threshold (BMN and motor neuron partners not shown). Colors indicate whether the partners are presynaptic (green), postsynaptic (magenta), or pre/post (black) to the BMNs. Cervical connective (CvC) is indicated and contains descending and ascending neuron axons to and from the VNC, respectively. (C-J) Anterior view of all pre- (C,E,G,I) and postsynaptic (D,F,H,J) partners of BMNs. Different categories of synaptic partners are shown in columns, including all partners (C,D) ascending (E,F), descending (G,H), and interneurons (I,J). The numbers of the different types are indicated in each panel. Underlying data are in Supplementary file 6.

Most BMN partner input and output synapses are distributed in the ventral brain.

(A-D) Brain neuropil distributions of input or output synapses of the presynaptic (A), postsynaptic (B), and pre/post (C,D) partner neurons. (A) Input synapses onto the BMN presynaptic partners. (B) Output synapses of the postsynaptic partners. (C,D) Input (C) and output (D) synapses of the pre/post partners. The total numbers of synapses contributing to each panel are indicated in the connectivity diagram in red. For pre/post partners, the number associated with the arrow from the BMNs indicates the BMN input synapses onto pre/post partners, while the arrow to the BMNs indicates pre/post partner output synapses onto BMNs. The other arrows going to and from the pre/post neurons refer to input and output synapses that are not to or from BMNs. Presynaptic partner outputs and postsynaptic partner inputs are not shown, but their synapse numbers with BMNs or other neurons are indicated in black. The 6 neuropils with the most partner synapses are labeled in each panel, with the first and second highest being indicated with bold font. Darker red shades indicate areas with higher synapse counts. Shading is not linear because the gnathal ganglia (GNG) contains most BMN partner synapses. To highlight neuropils with fewer partner synapses, the 2 neuropils with the most synapses are shaded dark red (top shade indicator). All other neuropils are shaded based on the fraction of synapses relative to the neuropil with the second highest number of partner synapses (gradient indicator). Shading is calculated for each panel, and cannot be compared across panels. See Supplementary file 8 for a complete table of BMN partner pre- and post synapse numbers in all neuropils. Synapses are from ascending, descending, and interneurons, and do include BMNs or motor neurons. Synapses of left side BMN partners shown. Abbreviations: Gnathal ganglia (GNG), saddle (SAD), antennal mechanosensory and motor center (AMMC), vest (VES), flange (FLA), anterior ventrolateral protocerebrum (AVLP), posterior ventrolateral protocerebrum (PVLP), prow (PRW), lateral accessory lobe (LAL), inferior posterior slope (IPS), wedge (WED). Raw and normalized underlying data for partner synapses in the top 31 neuropils (total synapse count > 100) are in Supplementary file 9.

BMN presynaptic inputs are mostly GABAergic.

(A-B) BMN presynaptic (A) and pre/post (B) partner neurotransmitters and their numbers of synapses onto BMNs. Table (left) indicates synapses onto BMNs from each partner category. Partners are predicted to use acetylcholine (ACh), GABA, glutamate (Glut), or serotonin (5-HT). Unknown indicates synapses whose neurotransmitter identities could not be determined by the prediction algorithm (Eckstein et al., 2024). Grid grayscale shades indicate synapse numbers, with black indicating 1,598 (A) or 8,657 (B) synapses (scale indicator on right). Schematic (middle) shows BMN inputs from different presynaptic partners. Edge grayscale indicates the relative numbers of synapses from each category onto BMNs (scale indicator on right). Partner connections are indicated as excitatory ACh (arrow), inhibitory GABA (ball), and modulatory 5-HT (box). Color codes for each neurotransmitter are used in C-J. (C-J) Anterior views of reconstructed BMN presynaptic (C-F) or pre/post (G-J) partner categories, including BMNs (C,G), ascending (D,H), descending (E,I), and interneurons (F,J). Colors indicate the predicted neurotransmitters for each partner, including GABA (purple), ACh (teal), Glu (mustard), and 5-HT (red). n indicates the total number of neurons in each group. Colored numbers in parentheses represent the number of neurons with predicted neurotransmitter identities, indicated using the color code from panels (A,B). See Supplementary file 6 for a complete table of partners, neurotransmitter predictions, and synapse counts.

BMN synapses with postsynaptic partners.

(A-B) BMN synapse numbers onto postsynaptic (A) and pre/post (B) partners with different neurotransmitter profiles. Table (top or left) indicates the number of BMN synapses onto partners from each partner category. Partners are predicted to use acetylcholine (ACh), GABA, glutamate (Glut), or dopamine (DA). Unknown indicates synapses whose neurotransmitter identities could not be determined by the prediction algorithm (Eckstein et al., 2024). Grid grayscale shades indicate synapse numbers, with black indicating 16,939 (A) or 9,012 (B) synapses (see scale indicators). Neurotransmitter color codes are used in C-F. Schematic illustrates the outputs from BMNs onto postsynaptic partners. Edge grayscale indicates the relative numbers of synapses from BMNs onto each category (see scale indicators). Partners are indicated as excitatory ACh (arrow), inhibitory GABA (ball), and excitatory or /inhibitory Glu (box). DA and motor neurons not shown. (C-F) Anterior views of reconstructed BMN postsynaptic partner categories, including BMNs (C), ascending (D), descending (E), and interneurons (F). Colors indicate the predicted neurotransmitters for each partner, including GABA (purple), ACh (teal), Glu (mustard), and DA (red). n indicates the total number of neurons in each group. Colored numbers in parentheses represent the number of neurons with predicted neurotransmitter identities, indicated using the color code from panels (A,B). See Supplementary file 6 for a complete table of partners, neurotransmitter predictions, and synapse counts.

Clustering based on BMN postsynaptic connectivity similarity.

(A,B) Heatmap of cosine similarity among BMNs based on all output partners in the CNS (ie. ascending, descending, and interneurons; BMN-BMN connections and motor neurons excluded). Row and column order was determined by Ward’s hierarchical clustering method applied to pairwise cosine similarity test scores with a threshold of ≥ 0.3 applied. Clustering of the rows and columns is the same as the dendrogram. Clustering was performed on all BMNs (A) and excluding BM-InOm neurons (B). Colored bars bordering the headmap indicate the BMN type (C, colors) or BMN head region (D, grayscale). (C,D) Bristle populations on the anterior and posterior head, marked with labeled and color-coded dots indicating their classification by bristle/BMN type (C) or head region (D). Head regions grayscale code is shown on the top right, and correspond to the dorsal (black), posterior (dark gray), and ventral head (light gray), and eyes (white). (E) Simple network diagram of individual BMNs (nodes) based on cosine similarity scores (edges) with a threshold of ≥ 0.3 applied, colored by BMN type. Colored dots indicate BMN type as shown in C, with the exception of the BM-InOm neurons that are shown transparent. Node positions were computed with igraph’s default Fruchterman–Reingold force-directed layout: edges (weighted by cosine similarity) act as springs attracting similar neurons, while all nodes repel to avoid overlap. Consequently, neurons with higher similarity scores (i.e., similar postsynaptic connectivity profiles) are drawn close together, forming visually distinct clusters with varying overlap. Underlying data are in Supplementary file 6, Supplementary file 11, and Supplementary file 12.

BMN pre- and postsynaptic connectomes show somatotopic organization.

(A) Bristle populations on the anterior and posterior head, marked with labeled and color-coded dots indicating their classification. Bristle colors used in B-E refer to their corresponding BMN types. (B-E) Network diagrams of connections between BMN types (colored dots; synapses pooled, each node represents all BMNs of that type) and their presynaptic (B,C) or postsynaptic partners (D,E) shown in gray. Squares correspond to interneurons while down and up triangles indicate descending and ascending neurons, respectively. Each node represents an individual partner. BMN/BMN connections and motor neurons are excluded. Neurons that are both pre- and postsynaptic are included in all network diagrams in B-E. B and D show the entire network diagram, while C and E are zoomed cutouts with edge colors removed to better demonstrate position of BMN type nodes relative to each other. (B,D) Edges are in colors corresponding to the BMN type inputs (B) or outputs (D). (C,E) Edges are gray. Ellipses indicate BMN types in particular head regions, including the dorsal, posterior, and ventral head (see Figure 9D for graphical view of regions on the head). Node positions were computed using the visNetwork-igraph interface to employ a Fruchterman–Reingold force-directed layout: each directed edge acts like a spring whose strength is proportional to the edge weight (synapse count), drawing highly connected nodes together, while all nodes repel one another to prevent overlap. Consequently, neurons with many or strong synaptic connections cluster closely in the 2D embedding, whereas weakly or unconnected neurons are pushed to the periphery. Underlying data are in Supplementary file 6 and Supplementary file 10.

Pre/post neuron pairs form feedback loops with particular BMN types.

Connections between BMN types (colored dots) and neurons that are reciprocally pre- and postsynaptic to particular BMN types (gray). Squares correspond to individual interneurons while down and up triangles indicate descending and ascending neurons, respectively. Only edges contributing to reciprocal BMN and pre/post neuron connections are shown. Edges are in colors corresponding to the BMN type outputs, while gray edges correspond to BMN GABAergic inputs. All edges with non-reciprocal connections with pre/post neurons were removed. Bold text indicates the general head location of BMNs on the plot, revealing somatotopy-based connectivity with pre/post neurons (i.e. ventral, dorsal, posterior, and the ventral/dorsal transition). Node positions were computed using a Fruchterman–Reingold force-directed layout: each directed edge acts like a spring whose strength is proportional to the edge weight (synapse count), drawing highly connected nodes together, while all nodes repel one another to prevent overlap. Consequently, neurons with many or strong synaptic connections cluster closely in the 2D embedding, whereas weakly or unconnected neurons are pushed to the periphery. Ellipses indicate neurons shown in B-I. See Figure 11 – figure supplement 1 and Supplementary file 13 for a complete account of the pre/post neuron connections with the BMN types and information used for pre:post ratio and reciprocal filtering. (B-I) Some BMN pre/post neurons include pairs of neurons from the left and right (L,R) brain hemispheres. Shown are pre/post neuron pairs that are connected with specific BMN types, including ascending neurons Asc_1 (L,R) (B), interneurons CB0786 (L,R) (C), CB2513 (L,R) (D), CB0516 (L,R) (E), CB0496 (L,R) (F), descending neurons DNge122 (L,R) (G), DNg20 (L,R) (H), and ascending neurons Asc_2 and _3 (I). All neurons are of the same cell type (morphologically similar pairs), with the exception of the non morphologically similar ascending neurons shown in I. (B) Asc_1 (L,R) are not a pair like the others because they are not annotated as such in Codex due to difficulties annotating ascending neurons in the FAFB dataset. However, their axons in the brain are morphologically similar. Colored dots in each panel correspond to BMN types that are connected to both neurons in the pair. Scale bar, 50 μm.

BMN postsynaptic hemilineages.

(A) Each hemilineage containing neurons that are postsynaptic to left-side BMNs is listed in the first column. In the second column, the number of post and pre/post neurons per hemisphere that are postsynaptic to BMNs is indicated (magenta numerator), along with the total number of neurons in each hemilineage (denominator). Both ipsilateral (left) and contralateral (right) neurons are included. The third column represents the percentage of total BMN-hemilineage connectivity that is accounted for by each hemilineage. This percentage is calculated as the number of neurons in the hemilineage that are connected to BMNs, divided by the total number of BMN-connected hemilineage neurons (122). (B-T) Neurons within each hemilineage that are postsynaptic targets of BMNs (magenta) and remaining hemilineage neurons that are not connected with BMNs (grey). The percentages below each hemisphere indicate the proportion of BMN postsynaptic partners within the hemilineage, calculated relative to the total number of neurons in that hemilineage. Information about the BMN postsynaptic partner developmental origins, synapse counts, and predicted neurotransmitters are in Figure 12—figure supplement 1. For a list of connected hemilineages and their relative connectivity with all BMN partners, see Figure 12—figure supplement 2. Underlying data are in Supplementary file 14.

Hemilineage 23b (LB23) neurons elicit head grooming movements.

(A) Published driver lines expressing in neurons whose activation elicit head grooming (Hampel et al., 2015; Seeds et al., 2014). Shown are the identifiers for each line and their activated head grooming movements. (B) Anterior view of a light microscopy (LM)-imaged brain of the R40F04-GAL4 driver line expressing GFP in LB23 neurons (arrow) and other neurons. Image is a confocal Z-stack maximum intensity projection, immunostained for Bruchpilot (magenta) and GFP (green). Scale bar, 100 μm. (C) EM reconstructed LB23 neurons from both brain hemispheres (21 neurons on left, 22 on right). (D,E) MCFO labeling of aBN2 neurons targeted by aBN2-spGAL4-1. Shown are, maximum intensity projections of two aBN2 neurons stained using tag-specific antibodies (anterior views). The neuron shown in E has a midline projecting branch (arrow). Panel top right corner indicates the number of labeled aBN2 neurons without (D) or with (E) the midline projecting branch, versus the total MCFO-labeled aBN2 neurons. Dotted line indicates the brain midline. Scale bar, 20 μm. (F) Labeling of two aBN2 neurons in the same brain hemisphere that both lacked the midline-projecting branch. The green neuron is shown in D. (G-I) The three left hemisphere EM reconstructed aBN2 neurons (names indicated in top right corner). Dotted line indicates the brain midline. (H) Midline projecting branch of aBN2_2 is indicated with an arrow. Neurons rendered from the FAFB reconstructions in CATMAID (Hampel 2020). The following are the flywire IDs for the different aBN2 neurons: aBN2_1 (720575940634683237), aBN2_2 (720575940618961857), aBN2_3 (720575940621355759).

Developmentally related LB23 neurons form a putative somatotopically organized BMN architecture that underlies grooming.

(A) Bristles on the anterior and posterior head, marked with labeled and color-coded dots indicating their classification. (B) Connections between BMN types (colored dots; synapses pooled) and their postsynaptic LB23 neurons (gray). Squares correspond to individual LB23 interneurons while triangles indicate descending neurons. BMN edges are weighted by synapse count and shown in their corresponding bristle colors in A. Bold text indicates the general head location of BMNs on the plot, revealing somatotopy-based connectivity with LB23 neurons (i.e. ventral, dorsal, and posterior head). Node positions were computed using a Fruchterman–Reingold force-directed layout: each directed edge acts like a spring whose strength is proportional to the edge weight (synapse count), drawing highly connected nodes together, while all nodes repel one another to prevent overlap. Consequently, neurons with many or strong synaptic connections cluster closely in the 2D embedding, whereas weakly or unconnected neurons are pushed to the periphery. (C) LB23 neurons are derived from NB7-4 neuroblasts that divide to produce another neuroblast and a ganglion mother cell. The ganglion mother cell divides once to produce an LB23 neuron and another neuron undergoing programmed cell death. Multiple rounds of this process produce the LB23 hemilineage that includes aBN2 neurons. (D) Hypothesized somatotopic BMN to LB23 parallel grooming circuit architecture. Each BMN type shows preferential connectivity with specific subpopulations of LB23 neurons. BMNs innervating neighboring bristles show overlapping connections with LB23 subpopulations. LB23 subpopulations elicit grooming of their corresponding presynaptic BMN locations. For example, aBN2 neurons are preferentially connected with BM-Ant neurons and elicit grooming of the antennae. (E) Connections of BMN types with antennal grooming LB23 neurons (previously named aBN2_1, _2, and _3). Shown connections have at least two neurons of a BMN type connected to the same aBN2 neuron, or are connected by greater than five synapses. Number of connected BMNs versus the total for each type is indicated and represented as a partially filled circle. Synapse numbers for each connection are indicated and represented by the width of the edges. Supplementary file 16 provides a list of Flywire IDs for LB23 neurons which can be used to filter data from Supplementary file 10. (F) BMNs and JON types that are connected with neurons in the previously described antennal command circuit (black boxes). The BMN connections are reported in this work from Codex, whereas Johnston’s organ neuron (JON) connections were described previously (Hampel et al., 2020) (see Discussion).

Features of the BMN second-order connectome: without vs with spatial resolution.

(A) General features of mechanosensory processing were identified by analyzing the connections of nearly all BMNs innervating anterior and posterior head bristles (black dots), independent of spatial resolution of individual BMN populations. (B) Brain-level organization. BMNs project to the ventral brain neuropils, the gnathal ganglia (GNG; not shown). BMN partners rarely connect directly to dorsal higher-order regions because most partner neurites reside within the GNG and other ventral neuropils collectively referred to as the subesophageal zone (SEZ). Some partners have ascending or descending projections to/from the ventral nerve cord (VNC); however, the FAFB dataset does not include the VNC, so neurites within the VNC are not reconstructed. (C) Major SEZ-localized features (red box in A): local postsynaptic feedforward excitation and inhibition, descending postsynaptic feedforward excitation, ascending presynaptic inhibition, and local feedback inhibition. Partners are labeled “post,” “pre,” or “pre/post.” GABAergic neurons are shown as ball-and-stick symbols and cholinergic neurons as arrows. Features shown were prioritized by BMN pre- and postsynaptic synapse counts (largest fractions; Figures 7-8). (D-G) Spatially linking BMNs to their bristle populations reveals additional organizational features. (D) Head bristle/BMN populations include both neighboring and distant groups (Eichler et al., 2024). (E) BMNs project to somatotopically organized zones in the GNG, with neighboring populations overlapping more than distant populations (e.g., BM-Ant/BM-Fr neurons overlap; BM-MaPa neurons do not) (Eichler et al., 2024). (F) Using this spatial framework, the present study shows that BMN connectivity similarity is proximity-dependent, revealing somatotopically organized pre-and postsynaptic pathways. For example, neighboring BM-Ant and BM-Fr neurons share overlapping sets of connected postsynaptic partners, consistent with spatially organized excitatory and inhibitory pathways. Panel F was adapted and updated from Eichler et al. (2024). (G) The present study further identifies proximity-based inhibitory feedback hubs: neighboring BMNs show reciprocal connectivity with the same pre/post partners, indicating head-region-specific feedback inhibition (Figure 11). Panels D and E are reproduced from Eichler et al. (2024) (Figure 8D-F) under a CC BY license.

Proposed grooming circuit architecture suggests parallel mechanosensory pathways that are somatotopically organized.

(A,B) This architecture facilitates aimed grooming of specific head and body locations (A) and controls hierarchical suppression during sequential grooming (B). (A) When a mechanical stimulus like dust touches a specific area (e.g., antennae), local mechanosensory neurons detect it. These neurons connect to excitatory postsynaptic circuits that elicit aimed grooming of those locations. This study emphasizes the connectome of bristle mechanosensory neurons (BMNs) responsible for grooming different head locations, such as the eyes (red), proboscis (orange), and antennae (aqua). The grooming pathway for body locations (gray) extends this parallel structure to any body part, like the abdomen or wings. (B) Dust covering multiple areas simultaneously causes competition among pathways that elicit exclusive grooming actions that are prioritized through hierarchical suppression (unidirectional connections between circuits shown). For clarity, only the nearest-neighbor connections are depicted. In this hypothesized structure, each circuit inhibits all lower-ranking circuits in the hierarchy. For instance, eye grooming occurs first as it inhibits all subsequent actions, like grooming the proboscis, antennae, and body areas. An alternative to the depicted unilateral inhibition is a sensory gain-mediated suppression mechanism (e.g. mechanosensory presynaptic inhibition). (C) Mechanosensory neurons from different head locations project to distinct, somatotopically organized zones in the ventral brain and elicit aimed grooming of those locations, including the antennae (via JONs [Johnston’s organ neurons; not analyzed in this study] and BMNs), eyes (BMNs), and proboscis (BMNs). The projections of mechanosensory neurons in the brain connect with postsynaptic circuits (indicated by boxes) that elicit grooming via descending pathways that activate movement pattern generators in the ventral nerve cord. (D-G) Bristles on the anterior (D), posterior (E), ventral (F), and dorsal (G) locations on the male head. The bristles on the right side are marked with color-coded dots for classification. Bristle full names are listed in the Figure 1 legend. Panels A-C were modified under the terms of the CCBY license from Figure 1 – figure supplement 1 of Eichler et al. (Eichler et al., 2024). Panels D-G were reproduced from Figure 1A-D of Eichler et al. (Eichler et al., 2024).

BMN type ventral brain projections.

(A) Bristles of the anterior and posterior head. Bristle populations are marked with labeled and color-coded dots indicating their classification. Bristle population names are abbreviated (see Figure 1 legend for full names). (B-E) Reconstructed BMN projections colored by type according to the bristles that they innervate. Shown are the anterior (B), dorsal (C), medial (D), and lateral (E) views. Arrows indicate the projection directions for each incoming BMN nerve bundle. Scale bars, 20 μm. Panel A was modified under the terms of the CCBY license from Figure 6C,D of Eichler et al. (Eichler et al., 2024).

BMN type ventral brain synapses.

(A) Bristles of the anterior and posterior head. Bristle populations are marked with labeled and color-coded dots indicating their classification. Bristle population names are abbreviated (see Figure 1 legend for full names). (B-E) Reconstructed BMN projections (gray) and their pre- and postsynaptic sites, colored by type according to the bristles that their corresponding BMNs innervate. Shown are the anterior (B), dorsal (C), medial (D), and lateral (E) views. These views highlight both overlapping projection zones, visible as intermingled synapses of different colors from neighboring BMN types, and segregated zones, where synapses from distinct BMN types remain spatially separated with minimal color mixing. Arrows indicate the projection directions for each incoming BMN nerve bundle. Scale bars, 20 μm. Panel A was modified under the terms of the CCBY license from Figure 6C,D of Eichler et al. (Eichler et al., 2024).

Distribution of BMN synapses.

(A) Box plots of BMN pre- and post synapses by type. Each point (dot) indicates the number of synaptic inputs- (green) or outputs (magenta) for an individual BMN. Points are horizontally dispersed for easier visualization. Box indicates the interquartile range (IQR) for the particular BMN type, horizontal bar in box indicates the median, diamond indicates the mean, and vertical lines extend to furthest points within 1.5 times the IQR. “X” indicates outliers. (B) Table indicating cumulative number of inputs, outputs, and partners for each BMN type. n in gray indicates the number of left side BMNs that were reconstructed (Eichler et al., 2024). n in green and magenta indicate the number of BMNs that were connected with at least 5 synapses to any given pre- or postsynaptic neuron, respectively. The number of inputs or outputs refers to synapses, while the number of partners refers to pre- or postsynaptic neurons. *The total number of pre- or postsynaptic partners is lower than the total of connected partners because some partners are connected with multiple BMN types. Underlying postsynaptic and presynaptic data is in Supplementary file 2 and Supplementary file 3, respectively. (C-F) BMN synapse locations, colored by BMN inputs (green) and outputs (magenta); shown in anterior (C), dorsal (D), medial (E), and lateral (F) views. Reconstructed BMNs are shown in gray.

Percent of the total motor neuron synaptic inputs from each BMN type.

Heatmap indicating the percent of the total synaptic inputs from any neuron that are from each BMN type. Grayscale shading maximum indicates that BM-Ant neurons occupy 6.28% of the total presynaptic sites onto antennal nerve class neuron 720575940621351412. Underlying data are in Supplementary file 7. Note: to observe the extent of BMN synaptic inputs into the motor neurons, the analysis shown here includes reconstructed BMNs from the left and right sides of the head. Left side BMNs are shown in Figure 3C. Right side BMN types include BM-Taste, -Hau, -Ant, -Fr, -Or, -Oc, and -Vt/PoOc neurons, of which the BM-Fr neurons are the only type that are uniquely connected from the right side but not the left.

BMN connectome with pre/post neurons shows somatotopic organization.

(A) Network diagram of 34 pre/post neurons whose pre:post ratio is within 0.1-10 (gray) with different BMN types projecting into the left brain hemisphere (colored dots). Squares correspond to individual interneurons while down and up triangles indicate descending and ascending neurons, respectively. Edges are in colors corresponding to the BMN type outputs, while gray edges correspond to BMN GABAergic inputs. Node positions were computed using a Fruchterman–Reingold force-directed layout: each directed edge acts like a spring whose strength is proportional to the edge weight (synapse count), drawing highly connected nodes together, while all nodes repel one another to prevent overlap. Consequently, neurons with many or strong synaptic connections cluster closely in the 2D embedding, whereas weakly or unconnected neurons are pushed to the periphery. Underlying data are in Supplementary file 13 which includes a complete account of the pre/post neuron connections with the BMN types.

BMN postsynaptic partner developmental origins.

(A) BMN synapse numbers onto postsynaptic partners that use different neurotransmitters. Table indicates the number of BMN synapses onto postsynaptic descending neurons and interneurons. These neurons are subdivided based on developmental birth timing (Schlegel et al., 2024), including putative primary neurons of embryonic origin, part of a postembryonic hemilineage, or those with no hemilineage assigned. Partners are predicted to use acetylcholine (ACh), GABA, glutamate (Glut), or dopamine (DA). Unknown indicates synapses whose neurotransmitter identities could not be determined by the prediction algorithm (Eckstein et al., 2024). Grid grayscale shades indicate synapse numbers, with black indicating 15,226 synapses. Neurotransmitter color codes are used in B-G. (B-G) Anterior views of BMN postsynaptic descending neurons (B-D) and interneurons (E-G) belonging to putative primary neurons (B,E), hemilineage neurons (C,F), and neurons with no hemilineage assigned (D,G). Colors indicate the predicted neurotransmitters for each partner, including GABA (purple), ACh (teal), Glu (mustard), and DA (green). n indicates the total number of neurons in each subdivision. Colors in parentheses represent predicted neurotransmitter identities using the color code from panel (A). Underlying data are in Supplementary file 14.

BMN postsynaptic partners.

Neurons in each postsynaptic partner category and their connectivity to head BMNs. Left columns: different categories of partners including interneurons, descending neurons, BMNs, ascending neurons, and motor neurons. Interneurons and descending neurons are subcategorized as putative primary, assigned to a hemilineage, or not assigned to a hemilineage. Those assigned to a hemilineage are further subcategorized by the individual hemilineage they belong to. Centercolumns: n indicates the number of neurons in each category found to be postsynaptic to BMNs, expressed also as the percentage (%) of total BMN partners accounted for by each category (i.e., [n / 432]*100). Right columns: n indicates the total number of BMN output synapses corresponding to each category, expressed also as the percentage (%) of total BMN outputs accounted for by each category (i.e., [n / 65409]*100) . Underlying data are in Supplementary file 14.