Organization of the gravity-sensing system in zebrafish

Liu, Zhikai; Hildebrand, David G. C.; Morgan, Joshua L.; Jia, Yizhen; Slimmon, Nicholas; Bagnall, Martha W.

doi:10.1038/s41467-022-32824-w

Download PDF

Article
Open access
Published: 27 August 2022

Organization of the gravity-sensing system in zebrafish

Nature Communications volume 13, Article number: 5060 (2022) Cite this article

4642 Accesses
13 Citations
18 Altmetric
Metrics details

Subjects

Abstract

Motor circuits develop in sequence from those governing fast movements to those governing slow. Here we examine whether upstream sensory circuits are organized by similar principles. Using serial-section electron microscopy in larval zebrafish, we generated a complete map of the gravity-sensing (utricular) system spanning from the inner ear to the brainstem. We find that both sensory tuning and developmental sequence are organizing principles of vestibular topography. Patterned rostrocaudal innervation from hair cells to afferents creates an anatomically inferred directional tuning map in the utricular ganglion, forming segregated pathways for rostral and caudal tilt. Furthermore, the mediolateral axis of the ganglion is linked to both developmental sequence and neuronal temporal dynamics. Early-born pathways carrying phasic information preferentially excite fast escape circuits, whereas later-born pathways carrying tonic signals excite slower postural and oculomotor circuits. These results demonstrate that vestibular circuits are organized by tuning direction and dynamics, aligning them with downstream motor circuits and behaviors.

Neural dynamics and architecture of the heading direction circuit in zebrafish

Article Open access 24 April 2023

Locomotion-induced ocular motor behavior in larval Xenopus is developmentally tuned by visuo-vestibular reflexes

Article Open access 26 May 2022

Tiltable objective microscope visualizes selectivity for head motion direction and dynamics in zebrafish vestibular system

Article Open access 21 December 2022

Introduction

Central neuronal circuits mediate the transformation from sensory inputs to motor outputs. Motor output can be described by both the direction of movement and its temporal dynamics—fast and rapidly fatiguing versus slow and sustained contractions. Fast and slow motor behaviors emerge sequentially, set up by early-born and late-born motor circuits, respectively¹. In zebrafish, spinal motor neurons differentiate in sequence from those controlling fast movements to those controlling slow². Similarly, premotor neurons in spinal cord and brainstem governing fast escape movements develop earlier than premotor neurons driving slower locomotion^3,4,5. Developmental sequence also influences motor circuit connectivity in flies, linking the timeframe of motor circuit formation to temporal dynamics of movement^6,7. Both fast and slow behaviors can be elicited by a wide range of sensory inputs. It remains unclear whether sensory systems are also organized around principles set up by developmental sequence that align with the speed-dependent motor circuit architecture.

The vestibular system encodes information about head movement in space. Both translational acceleration and orientation with respect to gravity are encoded by hair cells arrayed underneath otoliths in the inner ear. Striolar hair cells carry high-pass signals which are predominantly relayed by irregular-firing vestibular afferents with phasic kinetics, whereas extrastriolar hair cells excite predominantly regular-firing vestibular afferents with tonic or phasic-tonic dynamics⁸. Both types of vestibular afferents relay head movement information into the brainstem, where they excite a variety of central targets. Central vestibular nucleus neurons innervated by vestibular afferents project directly to oculomotor and spinal neurons to drive behaviors⁹. This tight connection from sensory input to motor output suggests that vestibular circuits, like motor and premotor circuits, might be organized around speed-dependent principles. However, it has been difficult to link the patterning of central premotor circuits to hair cell organization because vestibular afferents are not known to be spatially organized by tuning direction or their temporal dynamics.

We examined the architecture of the gravity-sensing system in larval zebrafish (Danio rerio). Zebrafish begin to exhibit simple vestibular functions, such as postural control and vestibulo-ocular reflexes, as early as 3 days post fertilization^10,11,12 (dpf). Over the following days and weeks, vestibular behaviors improve and refine^13,14. The small size of the larval zebrafish brain makes it tractable for reconstruction with serial-section electron microscopy (ssEM) at synaptic resolution. In these animals, only the utricular otolith is responsible for gravity sensation^10,15, and accordingly we reconstructed all utricular hair cells, afferents, and four classes of central neurons receiving utricular inputs.

Directional tuning of utricular hair cells was topographically mapped onto the rostrocaudal axis of the utricular afferent ganglion. Afferents carrying inferred rostral or caudal tilt information excited compensatory central vestibulo-ocular reflex (VOR) circuits to stabilize eye position. Similarly, afferents carrying information about head movement in inferred ipsilateral or contralateral directions excited distinct elements of Mauthner cell escape circuits. Along with these directional maps, we identified cellular signatures of vestibular hair cells and afferents, including cilia length, synapse counts, and myelination, that indicated their developmental sequence. Inferred developmental sequence was mapped onto the mediolateral axis of the utricular ganglion and also correlated with central vestibular targets controlling distinct behaviors. Early-born sensory pathways are preferentially connected to drive early-born fast motor circuits, whereas later-born pathways govern movements mediated by later-born slower motor circuits. Collectively, these data revealed a sensorimotor transformation organized around movement speed, where phasic and tonic vestibular signals are preferentially used to regulate fast and slow movements respectively. Together, the directional and temporal tuning of vestibular circuits pattern the entire vestibulomotor transformation.

Results

Imaging the utricular system at synaptic resolution

Gravity and inertial forces are sensed by hair cells in the inner ear. The otolith, or in mammals the otoconial matrix, slides relative to the macula during head tilt or translation to deflect hair cell cilia. Utricular hair cells, which in larval zebrafish serve as the sole gravitational sensors^10,15, synapse onto the peripheral process of utricular afferents (schematic, Fig. 1a). These afferents, whose cell bodies reside within the vestibular ganglion, project axons that bifurcate and synapse in several brainstem nuclei that mediate behaviors like escape, posture, and the vestibulo-ocular reflex (VOR). We obtained an ultrathin section library of the larval zebrafish at 5.5 days post fertilization (dpf), which had originally been imaged at 18.8 × 18.8 × 60.0 nm³ per voxel or 56.4 × 56.4 × 60.0 nm³ per voxel depending on the region¹⁶. We re-imaged the peripheral and central areas of the right utricular circuit at 4.0 × 4.0 × 60.0 nm³ / voxel (Fig. 1b), sufficient resolution to visualize hair cell cilia (Fig. 1c) and vestibular afferent processes (Fig. 1d). The new images were aligned to the lower-resolution data and used to produce a reconstruction of the gravity-sensing system, including 91 utricular hair cells, 105 ganglion afferents, and ~45 target neurons in the ipsilateral vestibular brainstem (Fig. 1e–g; Supplementary Movie 1).

**Fig. 1: High-resolution serial-section electron microscopy of the gravity-sensing system.**

Characterization of direction tuning in utricular afferent circuits

We first established the tuning direction of every hair cell in the utricular macula. Hair cells are maximally depolarized by head tilts in the direction of their kinocilium relative to the cluster of stereocilia⁸. We measured directional tuning for each hair cell by drawing a vector from the center of mass of the stereocilia bases to the kinocilium base (schematic, Fig. 2a). Hair cell direction tuning vectors were displayed relative to soma position to yield a sensory map of the entire utricular macula (Fig. 2b), with vectors colorized by their inferred directional tuning. Consistent with prior observations^17,18, the line of polarity reversal (LPR) was found towards the lateral edge. On either side of the LPR, directional tuning varied smoothly from rostral tilt sensitive to caudal tilt sensitive. As a result, hair cells located near each other tended to have more similar tuning than those far apart (Fig. 2c).

**Fig. 2: The utricular afferent ganglion is organized in the rostrocaudal axis by directional tuning.**

Organization of the utricular ganglion by directional tuning

We next asked whether this organization of hair cells by directional tuning was reflected in the organization of the utricular afferent ganglion. Despite a rich literature of physiological responses in utricular afferents recorded in their axonal processes^19,20,21, the bone surrounding this ganglion in adult animals has historically prevented somatic recordings, rendering its sensory topography unknown. To create a map of utricular ganglion topography, we reconstructed all processes postsynaptic to the hair cell ribbon synapses (Fig. 2a, bottom), which are by definition utricular afferents, back to their soma locations in the utricular ganglion. Out of 944 ribbon synapses, 929 (98.4%) were apposed to afferent processes that could be followed out of the macula. A total of 105 utricular afferents were reconstructed, with an average of 3.0 ± 1.5 discrete hair cells contacting each afferent, and 3.4 ± 1.4 afferents contacting each hair cell (mean ± SD; Supplementary Data 1). In a 3D reconstruction of all afferents, there was a visible rostrocaudal gradient to their organization (Fig. 2d; note that this reconstruction is tilted slightly relative to a true horizontal view so that the afferent ganglion is visible beneath the central axon projections).

Afferents whose somata were rostrally located within the ganglion were innervated by hair cells in the rostral utricular macula, whereas afferents in the caudal portion of the ganglion were innervated by hair cells in the caudal macula. Consistent with previous reports²², each afferent received input from hair cells exclusively on one side or the other of the LPR. To infer directional tuning of each utricular afferent, we weighted its convergent hair cell vectors by their number of ribbon synapses. Here and throughout the manuscript, this anatomically inferred directional tuning will be referred to more simply as directional tuning, but we note that actual directional tuning might differ based on variations in synaptic weight, process morphology, and other biological variables that cannot be assessed with EM.

The resulting map revealed a sensory tuning topography in the utricular ganglion, with rostrally located afferents encoding rostral head tilts and caudally located afferents encoding caudal head tilts (Fig. 2d, e; Supplementary Movie 2). Quantifying hair cell to afferent connectivity revealed a strong correlation between the rostrocaudal position of each afferent soma and the hair cells that innervate it (Fig. 2f). However, unlike in the hair cell macula (Fig. 2b), afferents innervating hair cells lateral to the LPR are intermingled with those innervating hair cells medial to the LPR (Fig. 2g, note red arrows), both in this horizontal view and in sagittal and coronal views (Supplementary Fig. 2a, b). Nonetheless, these afferents with inferred contralateral tilt tuning formed a segregated axon bundle in the brainstem (Fig. 2d; Supplementary Movie 3). Afferent somata located near each other typically had more similar tuning than afferents that were further apart (Fig. 2h), although the strength of the relationship is less pronounced than in the hair cells themselves (cf. Figure 2c).

Central organization of directional tuning

We next asked how directional tuning was organized centrally in three utricular afferent target populations: the Mauthner escape circuit, the VOR circuit, and the vestibulospinal (VS) postural circuit. Our high-resolution reimaged territory in the brainstem allowed us to identify the postsynaptic targets of utricular afferents. These central brainstem targets were then reconstructed and characterized based on soma location and axon projection. Although most central vestibular neurons project axons outside the vestibular brainstem, into lower-resolution regions that were not reimaged, they could be reconstructed over long distances if their axons were myelinated¹⁶. Within these criteria, we characterized utricular input to the ipsilateral Mauthner cell, 4 commissural escape neurons projecting to the contralateral Mauthner cell, 23 VOR neurons, and 19 VS neurons. We note that additional commissural, VOR, and VS neurons presumably exist with insufficiently myelinated axons for adequate reconstruction, and therefore the set sampled here is likely biased towards earlier-born neurons. Additional neurons of less certain identity are described elsewhere²³.

The Mauthner cells are a specialized pair of reticulospinal neurons that trigger fast escape movements in response to multiple types of sensory input. Mauthner cells develop by 8 h after fertilization, excite primary (fast, early-born) spinal motor neurons, and drive fast escape movements by 1 dpf^24,25,26. We identified and reconstructed utricular inputs both onto the ipsilateral Mauthner cell and onto brainstem neurons whose axons cross the midline and appear to synapse onto the contralateral Mauthner cell. In total, 18 utricular afferents contacted the Mauthner cell lateral dendrite (example, Fig. 3a) with a total of 52 synaptic contacts. All utricular afferent synapses were clustered together tightly on ventral sub-branches of the Mauthner cell lateral dendrite (Fig. 3b), as reported²⁷. In addition, seven utricular afferents synapsed on a commissurally projecting neuron population that contacted the contralateral Mauthner (Fig. 3c), which we term “commissural escape” neurons (Supplementary Movie 3).

**Fig. 3: Escape circuits compute head movement direction from utricular input.**

Interestingly, no utricular afferents diverged to contact both the ipsilateral Mauthner cell and this commissural escape population. Utricular afferents presynaptic to the Mauthner cell were innervated by hair cells in the medial portion of the macula, whereas afferents that excited the commissural-projecting neurons were innervated by hair cells in the lateral portion of the macula, on the far side of the LPR. The inferred tuning vectors of all afferents contacting the Mauthner cell were averaged to yield an inferred tuning to tilt in the ipsilateral and rostral direction (Fig. 3c, d). Tuning vectors of commissural escape neurons pointed in the opposite direction, contralaterally and caudally (Fig. 3c, d; average difference between Mauthner and commissural neuron inferred tuning, 196°). Therefore, these commissural escape neurons are presumably similarly tuned to inertial movements as the contralateral Mauthner cell that they contact.

Vestibular stimuli can elicit locomotion^28,29, but escape behaviors have primarily been evoked with auditory or mixed auditory-vestibular stimuli that do not allow isolation of the vestibular component^30,31. We reasoned that the utricular afferents exciting the Mauthner cell are tuned to report ipsilateral head tilts and therefore contralateral translational movements, with the predicted consequence that a translational movement to the left should trigger escape bends to the left (Fig. 3e). This type of circuit would be useful to detect predator movement towards the zebrafish. While the water flow of a predator bow wave can be detected by lateral line circuits³², there is little known about whether the head deflection itself can also elicit escapes. To test this prediction, we delivered a large amplitude translational stimulus optimized for speed (~10 ms, >1 g) to freely moving zebrafish larvae while recording behavior at 508 frames/s. This stimulus evoked a classic short-latency escape response in 35% of trials (Fig. 3f; 28 escapes in 81 trials from 27 animals; escapes defined as C-bends within 10 ms of peak acceleration³⁰). Animals escaped in the direction predicted by the circuit in Fig. 3e on 64% of trials (18/28). However, due to sled limitations, this stimulus involved bidirectional movement to maximize the accelerative force. We therefore repeated this experiment with a second stimulus that was slightly slower but optimized to be unidirectional (~20 ms, 0.8 g). This stimulus elicited a similar frequency of escapes (28%: 56 escapes in 202 trials from 105 animals). Notably, successful escapes were strongly directionally biased: animals turned in the direction of peak acceleration on 95% (53/56) of escapes (Fig. 3g, shaded regions). Turns in the “wrong” direction occurred on trials where animals were accelerated rostrally or caudally (Fig. 3g). To test for utricular dependence, both stimuli were presented to the utricle-deficient rock solo fish line, an otogelin mutant^10,33. Utricle-deficient animals escaped about half as often as their sibling controls (Fig. 3h). Because experiments were carried out under infrared light, the remaining escapes were likely triggered by the lateral line system³². Based on these anatomical and behavioral results, we conclude that this utricular-activated escape circuit allows for computation of the direction of head deflections, such as occur during predator approach.

Next, we evaluated directional tuning in three vestibular nuclei. VOR neurons of the superior vestibular nuclei (SVN) and tangential nuclei collectively govern vertical and torsional eye movements. We reconstructed 12 neurons in the SVN, which inhibits the ipsilateral trochlear and oculomotor nuclei⁹, and 11 neurons of the tangential nucleus¹², which excites contralateral trochlear and oculomotor neurons (Fig. 4a, b). Of these VOR neurons, 22/23 had inferred tuning in the ipsilateral rostral or caudal directions (insets, Fig. 4a, b), consistent with their well-described roles in the VOR⁹. This inferred tuning also aligns with the angles of the anterior and posterior semicircular canals, which contribute to rotational VOR behaviors later in development⁹. Notably, the SVN and tangential neurons tuned for rostral vs caudal tilt also projected axons with largely distinct trajectories. SVN axons with inferred rostral tilt encoding traveled more laterally than those encoding caudal tilt (boxed inset, Fig. 4a), and presumably inhibit the motor neurons that drive downward eye rotation via the inferior rectus and superior oblique muscles³⁴. Tangential axons with inferred rostral tilt encoding traveled ventrally and are presumed to activate the activate the eyes-up pathway through motor neurons that control the superior rectus and inferior oblique, whereas those with inferred caudal tilt encoding traveled dorsally where they likely activate the eyes-down pathway through motor neurons of the inferior rectus and superior oblique^35,36,37 (Fig. 4b; Supplementary Movie 4).

**Fig. 4: Structure and tuning of central VOR and VS utricular targets.**

Similarly, we characterized the tuning of afferent inputs to the VS population, which is involved in postural control (Fig. 4c)³⁸. We extended our previously reported connectivity from myelinated utricular afferents³⁹ to include the unmyelinated afferents. The 19 VS neurons received input from a total of 61 afferents. In contrast to VOR nuclei, VS neurons primarily received input from more rostral tilt sensitive afferents (Fig. 4c). VS neurons were typically contacted by a greater number of distinct utricular afferents than VOR pathway neurons were (medians: 8 afferents per VS neuron, 25th−75th %ile, 3.5–15.5; 4 afferents per VOR neuron, 25th −75th %ile, 3–7; Wilcoxon–Mann–Whitney between the number of distinct utricular afferents contacting VS and VOR neurons, p = 0.036). Thus VS neurons are contacted by a large number of predominantly rostral tilt sensitive utricular afferents.

Little is known about subcellular organization of vestibular afferent input onto central neurons. The concentration of utricular inputs onto a portion of dendrite in the Mauthner cell (Fig. 3b) led us to use the high resolution of EM to examine whether utricular inputs were similarly concentrated in VOR and VS neurons. Dendrograms of utricular afferent input to VOR and VS neurons revealed that most utricular input is not evenly distributed across the dendritic arbor but instead arrives on a small subset of branches (examples, Fig. 4d). To quantify the concentration of synapses, we carried out Monte Carlo simulations of synaptic distribution with either randomly distributed synapses or locations weighted by distance from afferents (see Methods). We found that utricular afferent synapses were located more closely to each other than expected by random chance, likely related to the limited spatial range over which utricular afferents interact with vestibular dendrites (Fig. 4e, Supplementary Fig. 5). This result indicates that branch-specific computation may occur in vestibular nucleus neurons, perhaps in conjunction with localized cerebellar input⁴⁰.

These analyses revealed a rostrocaudal map of direction tuning in the utricular afferent ganglion, as well as patterns of direction tuning in central brainstem targets of the utricular afferents. We next considered the organization of temporal kinetics arising from distinct types of hair cells.

Temporal dynamics and developmental sequence of utricular hair cells

In addition to their directional tuning, hair cells are characterized as striolar or extrastriolar, based on morphological differences in soma shape and ciliary lengths^41,42,43,44. Striolar hair cells typically drive phasic, adapting, irregular-firing afferents with high-pass sensitivity, whereas extrastriolar hair cells drive tonic or tonic-phasic regular-firing afferents that are less dependent on stimulus frequency^{8,19,43,45,46,47}. Fish and frogs do not express the classical Type I striolar cell shape seen in amniotes, but striolar hair cells can still be recognized based on the length of the kinocilium relative to the tallest stereocilium^44,48. In the adult zebrafish utricle, striolar hair cells are estimated to have kinocilia and tallest stereocilium lengths around 5 µm, whereas extrastriolar hair cells have a kinocilium estimated at 6–8 µm and tallest stereocilium 2–3 µm¹⁷. We reconstructed the kinocilium and tallest stereocilium of each hair cell (Fig. 5a) and plotted the relationship between their lengths. In one group of hair cells, both kinocilium and stereocilium were >4.8 µm and the kinocilium to stereocilium length (K/S) ratio ranged from 1.1–1.7; these were identified as striolar (Fig. 5b). In a second group, the kinocilium was >4.8 µm but the tallest stereocilium was <5 µm, yielding K/S ratios from 1.75–3.3, which we identified as extrastriolar. These ciliary lengths are close to those in the adult zebrafish, suggesting that these are mostly mature hair cells. Finally, at the tail end of the ciliary length distributions, we identified hair cells in which the kinocilium was <5 µm; we classified these as immature, though they presumably exist on a continuum with established hair cells (Fig. 5b). Some of these immature hair cells were also characterized by less cytoplasm, with fewer mitochondria and vesicles, consistent with this identification.

**Fig. 5: The utricular afferent ganglion is organized in the mediolateral axis by developmental sequence and temporal dynamics.**

We marked the position of each cell by its kinocilium and found that striolar hair cells straddled the LPR while extrastriolar hair cells populated the rest of the macula, as expected^44,49 (Fig. 5c). Utricular hair cells differentiate in a temporal sequence, with roughly radial development from the center or striola to the periphery^{50,51,52,53,54}. We wondered whether hair cell ribbon counts, which generally increase during development⁵⁵, could establish additional signatures of this ongoing sequence of differentiation. Counts of the total number of ribbon synapses (example, Fig. 2a) showed that hair cells located centrally in the macula had the highest number of ribbons per cell, with fewer ribbons found in peripheral hair cells (Fig. 5c). Accordingly, striolar hair cells overall had the largest number of ribbons, followed by extrastriolar and immature hair cells (Fig. 5d). Even within the extrastriolar population, hair cells with more ribbons were located more centrally, and hair cells with fewer more peripherally, consistent with a radial growth pattern. Thus, both ciliary lengths and ribbon synapse number are consistent with a radial pattern of growth in the macula, providing proxies for developmental sequence.

Organization of the utricular ganglion by developmental sequence

We next asked whether similar proxies for developmental sequence could be identified in the utricular afferent ganglion neurons. Myelination is responsive to neuronal activity^56,57, and evidence suggests that it initiates first in early-born neurons^58,59. Of utricular afferents, 16 of 105 (15.2%) were myelinated throughout most of their central and peripheral processes; the remaining afferents were mostly or entirely unmyelinated (examples, Fig. 1d), similar to early developmental stages in other animals^60,61,62. Because the entire vestibular nerve is myelinated by adulthood^22,60, the myelinated afferents we observed are likely to be early-born, instead of a specialized category of afferents. Consistent with this idea, we found that afferent somata that are myelinated (Fig. 5e) and that receive the highest numbers of ribbon contacts from hair cells (Fig. 5f) occupy the most lateral edge of the nascent ganglion, the site of the earliest-born vestibular afferents^63,64,65. Later-born afferents are added more medially, forming a half-shell around the earliest-born somata^63,64,65. These observations are supported by quantification of mediolateral position (Fig. 5g) and the ribbon synapse count (Fig. 5h) in myelinated versus unmyelinated afferents. Thus, afferent myelination and soma position are proxies for developmental sequence.

Do early born hair cells preferentially connect with early born afferents? We found that myelinated afferents received over half of their input from striolar hair cells, and little input from immature hair cells (Fig. 5i). This was true whether quantified by the number of distinct hair cells providing input or the total number of synaptic ribbon connections (Supplementary Fig. 3d, e). In contrast, unmyelinated afferents predominantly received input from extrastriolar hair cells, and had a higher proportion of input from immature hair cells (Fig. 5i). At the level of individual neurons, afferents with laterally located somata tended to receive input from striolar hair cells, whereas medially located somata were the most likely to receive immature hair cell input (Fig. 5j).

We conclude from these results that early-born afferents, marked by early myelination, preferentially receive input from early-born, predominantly striolar hair cells, and occupy a lateral position in the ganglion. Later-born afferents contact later-born, predominantly extrastriolar hair cells, and occupy progressively more medial positions, forming a half-shell around the early-born afferents. At the same time, directional tuning is preserved in the rostrocaudal axis, with rostrally located afferents contacting rostrally located hair cells and vice versa (Fig. 2). Thus, the utricular ganglion is organized by directional tuning in the rostrocaudal axis and developmental sequence in the mediolateral axis. Moreover, because inferred afferent temporal dynamics are aligned with developmental order, the ganglion mediolateral axis is also organized from phasic, striolar-dominated afferents at the lateral edge to tonic, extrastriolar-dominated afferents more medially. Therefore, both the spatial and temporal tuning of vestibular signals are topographically organized in the utricular ganglion, though more loosely than at the macula.

Organization of central circuits by developmental sequence

Building on our observation that developmental sequence is an organizing principle of the utricular afferent ganglion and its connectivity with peripheral hair cells, we asked whether developmental sequence might also be linked to central connectivity. The downstream targets of vestibular afferents become functional at different times in development: Mauthner cells are the first to form connections and drive escape behavior, whereas VOR and VS neurons driving eye and postural movements are born later^38,66. Does the developmental sequence of utricular afferents predict their patterns of connectivity to brainstem targets? We mapped the hair cells contributing input to afferents that drove escape, VOR, or VS neurons (Fig. 6a). Central escape neurons (the Mauthner cell and the commissural escape neurons) received most of their input from afferents contacting striolar hair cells. In contrast, afferents exciting VOR and VS neurons received input from a much broader territory of the hair cell macula (Fig. 6a). We quantified hair cell input to these pathways by weighting each central synaptic contact by its afferent’s fraction of input from striolar, extrastriolar, or immature hair cells. Around two-thirds of utricular inputs to the escape circuits arose from striolar pathways, whereas VOR and VS neurons received a much larger portion of their inputs from extrastriolar regions (Fig. 6b). Similarly, the Mauthner cell and commissural escape neurons also received a higher proportion of their utricular inputs from myelinated afferents, in comparison to VS and VOR populations (Fig. 6a, c). Consistent with the interpretation that myelinated afferents are early-born and more mature, they also diverge to more postsynaptic target nuclei (Myelinated afferents diverging to Mauthner, SVN, tangential, and VS neurons: 9/16; to three targets, 3/16; to two targets, 1/16. Unmyelinated afferents diverging to all four targets, 0/90; to three targets, 9/90; to two targets, 25/90. Supplementary Data 2.) Therefore, early-born hair cells preferentially signal via early-born afferents to early-born brainstem populations, which in turn drive early-born spinal motor circuits for escape. Similarly, later-born hair cells excite later-born circuit elements for posture and oculomotor control.

**Fig. 6: Early developing afferent pathways with fast kinetics preferentially drive early developing central neurons for fast escapes.**

Developmental sequence may play a significant role not just across but also within neuronal populations. We found that three VS neurons whose axons travel more medially in the brainstem before joining the rest (VS_med) receive more input from afferents carrying immature hair cell information, whereas four VS neurons whose axons travel more ventrally (VS_vent) receive more input from afferents carrying striolar information, as compared to other VS neurons (Fig. 6d). Based on these inputs, VS subpopulations with distinct axonal trajectories are predicted to exhibit different temporal dynamics, with more phasic information carried by the VS_vent and more tonic by VS_med. Though the postsynaptic targets of these subpopulations of VS neurons are not known, this result may help identify circuits underlying later refinement of postural control¹⁴.

Collectively, our anatomical analyses demonstrate that directional tuning and developmental sequence pattern the entire vestibulomotor transformation (Fig. 7). The rostrocaudal axis of hair cell organization is largely preserved in the afferent ganglion, leading to distinct pathways that drive responses to pitch. Afferents encoding contralateral tilt are intermingled at the level of the ganglion but form spatially segregated pathways that underlie head direction computation in the escape circuit (Fig. 7a). At the same time, early-born hair cells contact early-born utricular afferents, which in turn preferentially drive escape behaviors mediated by fast, early-developing motor circuits. Later-born sensory pathways support postural and oculomotor behaviors mediated by a mixture of fast and slow muscles (Fig. 7b). These results also demonstrate that striolar signals, which are carried mostly by irregular-firing afferents with predicted high-pass, phasic encoding properties, establish connections with circuits driving rapid-onset reflexes via fast motor pools.

**Fig. 7: Directional tuning and developmental sequence are organizing principles of vestibulomotor connectivity.**

Discussion

Here we show that the utricular afferent ganglion, which carries gravity sensation into the brain, is organized along two axes: a rostrocaudal axis for directional tuning, and a mediolateral axis for development. Afferents with different directional tuning excite distinct brainstem populations. Further, we demonstrate that early-born afferents preferentially receive information from early-born and striolar hair cells, yielding a gradient from phasic to tonic signals in the ganglion. This developmental organization aligns vestibular signals with downstream motor circuits. Brainstem neurons governing the fastest motor circuits, which underlie early-developing escape behaviors, are excited mostly by utricular afferents with inferred phasic firing and early development. Oculomotor and postural circuits, which drive slower motor neurons, are excited preferentially by afferents with inferred tonic firing and late development. Collectively, these results demonstrate that the vestibular circuit is organized by both directional tuning and temporal dynamics to mediate transformation into motor outputs.

We demonstrate here that the utricular afferent ganglion is patterned both by directional sensitivity in the rostrocaudal axis, and by developmental sequence, correlating to temporal dynamics, in the mediolateral axis. As far as we are aware, this is the first demonstration of any topography in the vestibular afferent ganglia. A similar mediolateral gradient has been described for the nearby lateral line afferent ganglion, where early-born afferents are positioned more laterally than later-developing afferents⁶⁷. Early-born lateral line afferents also exhibit larger soma size, lower excitability, and more dorsal central projections than later-born afferents⁶⁸. Thus, they are well placed to mediate coarser, large-amplitude stimuli, similar to our observation that early-born utricular afferents carry striolar information with inferred phasic responses. The mechanisms shaping topographic patterning in these two sensory systems may have been templates for tonotopic organization in the auditory cochlear afferents that evolved later⁶⁹.

The link between developmental sequence and motor control has been shown most robustly in zebrafish^3,4,5,70 and fruit flies^6,7, suggesting that sequential developmental from fast to slow is an ancient principle of motor control. Building on these results, several classes of spinal neurons in mice have also been shown to differentiate by subtype according to their birth order^71,72. However, whether these subtypes are tied to different speeds of movement remains to be explored.

Our work demonstrates that developmental sequence also links sensory inputs to motor outputs. The vestibular system is tightly coupled to motor control, and therefore it seems plausible that its organization relies on related principles. Our data show that the Mauthner cell preferentially receives utricular input from striolar-driven afferents (65% of utricular input). Notably, these afferents are the earliest-born vestibular pathways; similarly, the Mauthner cell drives the escape reflex as early as 2 dpf via the earliest-born spinal motor and premotor neurons²⁶. Thus, the entire escape reflex arc runs from from early-developing hair cells, to early-developing afferents, to early-developing hindbrain reticulospinal neurons, to early-developing spinal elements and muscles. Both oculomotor and postural behaviors appear around 3–4 dpf in zebrafish^{12,15,73,74,75}, with continued maturation at later stages¹⁴. Correspondingly, the VS and VOR neurons develop after the Mauthner cell in zebrafish³⁸ (Goldblatt et al., 17th International Zebrafish Conference abstract, 2022), and these populations appear to develop at roughly the same time in amniotes⁷⁶. In mouse, VS neurons preferentially synapse with slow motor neurons⁷⁷, consistent with our observation that they receive more extrastriolar and later-developing input than escape circuits do. It is plausible that developmental sequence may be significant in circuit assembly beyond the vestibulomotor pathways examined here.

We find that VS neurons can be subdivided based on axon trajectory, with differing proportions of striolar, extrastriolar, and immature input across groups. The fact that the VS_med population, whose axons plunge directly towards the midline, get a higher proportion of utricular input from immature hair cells suggests that they might be later-born and participate in later-onset postural functions. Zebrafish refine their postural control over the pitch axis from 4 to 15 dpf by increasing use of fins and improved bout timing mediated in part by VS neurons^13,14,38. We speculate that these VS_med neurons may contribute to this refinement by preferentially connecting with fin motor neurons, which are located near the midline⁷⁸. In mammals, the slower-onset portion of postural reflexes is mediated in part by a subset of lateral vestibular nucleus neurons (LVN_C) which collateralize to contact pontine reticulospinal neurons⁷⁹. These similarities suggest that a VS population encoding tonic head movements could target multiple downstream targets for fine motor control. In contrast, the higher proportion of striolar inputs to VS_vent suggests a function in rapid postural control. Together, our findings suggest a relationship within the VS population between the developmental time course of sensory innervation, axon trajectory, and the speed of behavioral responses.

Although our reimaging did not extend to the oculomotor and trochlear nuclei, the dorsal trajectory of caudal-tilt sensitive (nose-up) tangential neurons corresponds well with the dorsal position of inferior rectus and superior oblique (eyes-down) motor neurons in these areas³⁷. Based on the earlier development of eyes-down motor neurons³⁶, we hypothesize that tangential and SVN pathways contacting these motor neurons may be earlier born as well.

The directional tuning of afferent inputs to the Mauthner cell and the commissural escape neurons was opposite in direction. Thus each Mauthner cell receives two forms of utricular inputs tuned to ipsilateral tilt: monosynaptic innervation from ipsilateral vestibular afferents and disynaptic innervation from the contralateral side. The identity of these commissural escape neurons receiving utricular input is not known. Based on their anatomy, they are not similar to excitatory spiral fiber neurons, whose axons target the Mauthner axon cap⁸⁰, nor to inhibitory commissural neurons underlying left-right choice in the auditory system, whose axons target the Mauthner cell body⁸¹. Instead, these commissural utricular neurons appear to target the Mauthner ventral dendrite (Fig. 3c). Vestibular commissural neurons in goldfish have previously been interpreted as inhibitory, because their activation diminishes the amplitude of the antidromically triggered spike^82,83. We speculate that these zebrafish commissural escape neurons are instead excitatory, because they carry signal that is similar to the predicted monosynaptic vestibular afferent inputs to that Mauthner cell. Alternatively, these commissural neurons might provide an inhibitory signal that sharpens lateralization, as described in the auditory and lateral line systems⁸¹. Further high-resolution imaging or physiology could resolve this question.

Additionally, because utricular inputs to the Mauthner cell and these commissural neurons are predominantly striolar, they are expected to be phasic and high-pass, with exquisite spike timing^45,84. Irregular-firing afferents carrying striolar signals are thought to be kinetically well suited to signal translational movements, which are typically brief, in contrast to the lower-frequency movements generated by tilt⁸⁵. In keeping with this prediction, we identified a vestibular-dependent escape behavior triggered by rapid translation. This behavior may be important for survival: a predator approaching from the right would cause a fast translational head displacement to the left. This would activate both the ipsilateral tilt tuned afferents on the right side and the contralateral tilt tuned afferents on the left side of the animal (Fig. 3e). Via the direct and indirect pathways described here, these would summate to activate the right side Mauthner cell, which triggers contraction of the left side of the body, leading to a correctly directed escape away from the predator.

We find that VOR circuit neurons in the tangential and SVN all receive inputs tuned to ipsilateral tilt, with subpopulations aligned to either rostral or caudal tilt directions for control of vertical and torsional eye movements. The shared ipsilateral sensitivity explains why these subpopulations were not differentiated in zebrafish imaging studies of vestibular tuning in the roll axis^86,87. Our anatomical data predict that VS neurons exhibit greater sensitivity to rostral tilt, consistent with our physiological results, in which VS neurons are preferentially sensitive to caudal translation³⁹. Some VS neurons are contacted by divergently tuned afferents in the rostrocaudal axis and are predicted to have correspondingly more complex tuning responses, as seen in mammals^88,89. These results align well with our observation that complex central tuning in VS neurons is constructed by feedforward excitation from afferents with differing tuning³⁹. We also discovered that vestibular afferents innervating the most lateral portion of the macula (contralateral tilt sensitive) form a separate axon bundle in the hindbrain (Fig. 2d and Supplementary Movie 3). This result implies differential targeting of this pathway, consistent with the commissural utricular neurons in the escape circuit. The contralateral tilt-sensitive pathway is also thought to drive local feedforward inhibition to amplify target tuning by a push-pull mechanism⁹⁰.

Our observation that VOR and VS neurons receive similar mixtures of striolar and extrastriolar input are consistent with physiological analyses suggesting that both populations receive similar amounts of synaptic input from irregular and regular firing vestibular afferents^84,91,92. In those analyses, irregular-firing (phasic) afferent inputs appeared to dominate, whereas in our anatomical analyses, striolar-driven afferents make up less than half the overall utricular input. We suggest that synaptic contacts onto VOR and VS neurons from irregular afferents are stronger in amplitude than those from regular-firing afferents, consistent with their morphologically larger synaptic contacts⁹³. However, irregular-firing inputs preferentially encode high-frequency head movements, which are less common⁹⁴. Therefore, their dominance in nerve stimulation experiments may reflect synaptic weights but not their overall influence on typical activity. Indeed, loss of striolar hair cells has little effect on oculomotor and basic postural behaviors⁹⁵, consistent with our observation that those pathways draw mostly on extrastriolar inputs. There are also likely to be significant variations across vertebrates in the balance between phasic and tonic vestibular signals, based on characteristics of movement⁹⁶.

Our analyses could not test whether developmental sequence governs connectivity, as opposed to the possibility that a genetically distinct class of early-born afferents preferentially connects to escape circuits. Future experiments manipulating circuit formation may shed light on this question. Our data also necessarily represent a single snapshot in time. Is the observed circuit, aligned with developmental sequence, likely to be maintained in adulthood? We suggest that it likely is, on the basis of the functional correspondence between early-born hair cells and early-born motor neurons which both display high-pass dynamics, while late-born hair cells and motor neurons display slower dynamics^97,98,99. However, the spatial organization of the ganglion may become distorted by continued development; further experiments will be needed to address this question. It will also be important to evaluate how later-born hair cells in the striolar region integrate into the circuit. If they drive activity in later-born central circuits, their afferents may change central temporal dynamics. Alternatively, if they integrate with early-born circuits as other striolar hair cells do, they may serve to augment existing signals.

Methods

All experiments on animals were approved by the Washington University Institutional Animal Care and Use Committee.

Ultrathin (60 nm thick) serial sections from a 5.5 dpf larval zebrafish were a generous loan from J. Lichtman and F. Engert. Using the published 18.8 × 18.8 × 60 nm³ per voxel and 56.4 × 56.4 × 60 nm³ per voxel reference map and reconstructions¹⁶, we re-imaged the right side of the fish, covering the utricular hair cells, utricular afferents (identified by their peripheral processes reaching the utricular macula), and a rostrocaudal extent of the brainstem that covered several major vestibular nuclei at 4.0 × 4.0 × 60 nm³ per voxel. The volume covered ~100 µm in the rostrocaudal axis, 150 µm in the mediolateral axis, and 100 µm in the dorsoventral axis, in an irregular shape designed to capture the afferent peripheral and central processes (Fig. 1b). Imaging was carried out at the Washington University Center for Cellular Imaging on a Zeiss Merlin 540 FE-SEM with a solid-state backscatter detector. WaferMapper software¹⁰⁰ was used to control an ATLAS scan engine for automated focus and acquisition¹⁰¹. The resulting images were aligned onto the original 56.4 × 56.4 × 60 nm³ per voxel dataset using affine and elastic transformations in FIJI’s TrakEM2 plugin^102,103, with custom support from UniDesign Solutions.

The entire image volume was hosted in a CATMAID instance^104,105. Vestibular circuit neurons and hair cells were reconstructed as skeletons, i.e. node points without volume fills. All utricular afferents were identified by stepping section by section through the entire anterior macula twice and marking every hair cell ribbon synapse. Ribbon synapses were identified by the characteristic large dark ribbon structure surrounded by vesicles (Fig. 2a). Every utricular afferent was followed as far as possible, in most cases to the corresponding soma in the vestibular ganglion. Only 1.6% (15/944) of processes adjacent to ribbons could not be followed to a soma due to the quality or ambiguity of the images.

Hair cell kinocilia and the tallest stereocilia were traced from the apical surface of each hair cell to their distal tips. The kinocilium was recognizable based on its distinctive structure (see Fig. 2a). Ciliary length was calculated as the sum of the Euclidean point-to-point distances. Positions of all cilia were plotted at the epithelial plane and a three-dimensional tuning vector for each hair cell was derived from the center of mass of all stereocilia to the kinocilium. Hair cell vector lengths were typically short in the dorsal-ventral axis relative to their extent in the other two axes (around one-third the normalized vector magnitude of the other two axes; Supplementary Fig. 1b), consistent with the mostly horizontal orientation of the utricular macula. Therefore, for the purposes of analysis we focused exclusively on their projection in the horizontal plane. Ganglion soma position was quantified in three dimensions for all analyses. During fixation, differential shrinkage caused a small tissue separation that is visible as a gap in the horizontal projection of the utricular ganglion reconstruction (Fig. 2g, upper right), but there was no loss of tissue. Tuning similarity was calculated as the cosine of the difference between the tuning directions of each pair of hair cells or afferents. Hair cell distance was determined by the 3D Euclidean distance between their kinocilia. Afferent distance was determined by the 3D Euclidean distance between their soma centers. Only hair cells on the medial side of LPR or afferents innervating the medial side of LPR were included for analysis of tuning similarity.

From the afferent somata in the utricular ganglion, afferent axons were then followed into the brain. A total of 105 afferents were successfully reconstructed by two experienced annotators (N.S. and M.W.B.) and all tracing was reviewed (M.W.B.). Central synapses were identified by close appositions, thickening of the presynaptic membrane, and clustered vesicles (e.g. Figs, 3a, 4a). The Mauthner cell was previously reconstructed, as were most of the VS neurons^16,39. Additional VS neurons were identified by reconstruction of utricular target neurons to the point that they joined up with previously traced myelinated axons. VOR neurons were identified based on their utricular input and their characteristic axonal projection patterns in the medial longitudinal fasciculus. Because much of the axonal projections lay outside the reimaged territory, only myelinated portions of axons could be reconstructed with confidence. Therefore, we were not able to follow some VOR axons all the way to the trochlear and oculomotor nuclei. Utricular commissural neurons were identified by their axons which crossed the midline and traveled to the contralateral Mauthner cell. We note that we have identified a large number of additional utricular target neurons in the brainstem that either do not fit into these categories (e.g., commissural neurons) or cannot be confidently identified due to the difficulty in extending their axons into lower-resolution territory. Therefore, the set of brainstem neurons analyzed here is likely to be strongly biased to early-born, or at least early-myelinated, and is not a complete description of all VS or VOR neurons. Nonetheless, the 19 VS neurons identified here is a large proportion of the 27 identified by retrograde labeling³⁸.

Behavioral data were acquired in 5–6 dpf larvae from wild-types or the line rock solo^AN66, an otogelin mutation³³. Animals were visually verified as having normal (+/−) or absent (−/−) utricular otoliths. Animals were free-swimming in a small dish with infrared transillumination and imaged at 508 frames/s with a HiSpec-1 2 g monochrome camera mounted on a Scientifica SliceScope with a 2X objective. The translational stimulus was delivered with an air-floated sled (Aerotech ABL 1500WB) and designed for large amplitude acceleration and jerk, to maximize responses of irregular otolith afferents¹⁰⁶. Due to the exploratory nature of this experiment, there was no attempt to separate larvae into responders and non-responders as has been done for acoustic stimuli³⁰, but animals were selected for behavioral tests based on whether they exhibited some response to dish tapping. Images were analyzed with ZebraZoom¹⁰⁷ version 1.28.12 (zebrazoom.org) to extract the smoothed tail angle and heading direction.

Synaptic clustering was measured as the average distances between synapses on the same dendritic arbor. Vestibular neuron arbor skeletons were first simplified for analysis by regularizing the distance between nodes to ~1 um. Each synapse was associated with the nearest skeleton node, and distances between synapses were calculated along the Euclidean length of the skeleton.

To understand how the observed mean distance between synapses compared to random distributions of synapses on the vestibular neuron arbors, we used several Monte Carlo models. For each model, the synapses of each vestibular neuron were redistributed across the arbor 100,000 times. The mean distance between synapses was measured the same way as the non-randomized mean distances was measured. The measure used for each of the 100,000 iterations was the average of the mean synaptic distances of 43 vestibular neurons observed for that iteration. Confidence intervals were calculated as the bounds containing 95% of the results of the 100,000 experimental randomizations.

The first model (unweighted distribution) compares the observed clustering of synapses to what would be expected if synapses of each vestibular neuron were randomly distributed across the entire arbor of that neuron. In this model, each vestibular neuron skeleton node had an equal probability of being randomly assigned to one of the synapses of that vestibular neuron.

In the second model (50 µm-from-afferent weighted distribution) we restricted nodes likely to be assigned to a synapse using the proximity of the nodes to the synaptic terminals of afferent axons. 105 afferent axons were skeletonized as above. We then counted the number of afferent synapse nodes within 50 µm of each vestibular neuron skeleton node (proximity score), subtracting 1 to remove the influence of synaptic afferents presynaptic to a given vestibular neuron node. We next shuffled synaptic locations while maintaining the likelihood of synaptic connectivity given this proximity score. For example, if a skeleton node with a score of 95 has a 7% chance of being a synaptic locus, whereas a skeleton node with a score of 60 has a 1% chance, then nodes with scores of 95 will be given a 7-fold greater chance of being assigned a synapse in the redistribution than nodes with scores of 60.

In the final, most restrictive model (<5 µm-from-afferent weighted distribution), we attempt to use proximity to afferent terminals to recapitulate the clustering we observe in the real data. Proximity scores were calculated as +1 for each vestibular neuron node 0 µm from a presynaptic node. The added score decreased linearly to 0 at 5 µm distance. The rest of the 5 µm model was the same as the 50 µm model.

Analyses and statistics were carried out in Igor Pro (Wavemetrics) or Matlab (Mathworks). Statistical tests were carried out as reported in text, two-tailed where relevant, and typically with nonparametric analyses due to the non-normal distribution of parameters.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The electron micrograph image data generated in this study is hosted online and publicly available at http://zebrafish.link/hildebrand16/data/vestibular_right. Source Data are provided with this paper. Within the Source Data, one folder contains the .json files needed to generate all 3D reconstructions plus instructions for use; an Excel (.xlsx) file contains data for most of the figure panels, as well as links to the online source of example images, and another folder contains central neuron reconstructions used in dendritic analyses in Fig. 5. Tabular quantification of hair cell to afferent and afferent to central target connectivity are provided (Supplementary Data 1, 2). Source data are provided with this paper.

Code availability

Code used in generating the Monte Carlo analyses and sensory tuning analyses are available at https://github.com/bagnall-lab/EM_analysis

References

Fetcho, J. R. & McLean, D. L. Some principles of organization of spinal neurons underlying locomotion in zebrafish and their implications. Ann. N. Y Acad. Sci. 1198, 94–104 (2010).
Article ADS PubMed PubMed Central Google Scholar
Liu, D. W. & Westerfield, M. Function of identified motoneurones and co-ordination of primary and secondary motor systems during zebra fish swimming. J. Physiol. 403, 73–89 (1988).
Article CAS PubMed PubMed Central Google Scholar
McLean, D. L., Fan, J., Higashijima, S., Hale, M. E. & Fetcho, J. R. A topographic map of recruitment in spinal cord. Nature 446, 71–75 (2007).
Article ADS CAS PubMed Google Scholar
McLean, D. L. & Fetcho, J. R. Spinal interneurons differentiate sequentially from those driving the fastest swimming movements in larval zebrafish to those driving the slowest ones. J. Neurosci. 29, 13566–13577 (2009).
Article CAS PubMed PubMed Central Google Scholar
Pujala, A. & Koyama, M. Chronology-based architecture of descending circuits that underlie the development of locomotor repertoire after birth. Elife 8, e42135 (2019).
Article PubMed PubMed Central Google Scholar
Mark, B. et al. A developmental framework linking neurogenesis and circuit formation in the Drosophila CNS. eLife 10, e67510 (2021).
Wreden, C. C. et al. Temporal cohorts of lineage-related Neurons perform analogous functions in distinct sensorimotor circuits. Curr. Biol. 27, 1521–1528.e4 (2017).
Article CAS PubMed Google Scholar
Eatock, R. A. & Songer, J. E. Vestibular hair cells and afferents: two channels for head motion signals. Annu Rev. Neurosci. 34, 501–534 (2011).
Article CAS PubMed Google Scholar
Straka, H. & Baker, R. Vestibular blueprint in early vertebrates. Front Neural Circuits 7, 182 (2013).
Article PubMed PubMed Central Google Scholar
Mo, W., Chen, F., Nechiporuk, A. & Nicolson, T. Quantification of vestibular-induced eye movements in zebrafish larvae. BMC Neurosci. 11, 110 (2010).
Article PubMed PubMed Central Google Scholar
Whitfield, T. T. et al. Mutations affecting development of the zebrafish inner ear and lateral line. Development 123, 241–254 (1996).
Article CAS PubMed Google Scholar
Bianco, I. H. et al. The tangential nucleus controls a Gravito-inertial Vestibulo-ocular reflex. Curr. Biol. 22, 1285–1295 (2012).
Article CAS PubMed PubMed Central Google Scholar
Ehrlich, D. E. & Schoppik, D. Control of movement initiation underlies the development of balance. Curr. Biol. 27, 334–344 (2017).
Article CAS PubMed PubMed Central Google Scholar
Ehrlich, D. E. & Schoppik, D. A primal role for the vestibular sense in the development of coordinated locomotion. Elife 8, e45839 (2019).
Article CAS PubMed PubMed Central Google Scholar
Riley, B. B. & Moorman, S. J. Development of utricular otoliths, but not saccular otoliths, is necessary for vestibular function and survival in zebrafish. J. Neurobiol. 43, 329–337 (2000).
Article CAS PubMed Google Scholar
Hildebrand, D. G. C. et al. Whole-brain serial-section electron microscopy in larval zebrafish. Nature 545, 345–349 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Platt, C. Zebrafish inner ear sensory surfaces are similar to those in goldfish. Hear Res. 65, 133–140 (1993).
Article CAS PubMed Google Scholar
Haddon, C. et al. Hair cells without supporting cells: further studies in the ear of the zebrafish mind bomb mutant. J. Neurocytol. 28, 837–850 (1999).
Article CAS PubMed Google Scholar
Fernandez, C. & Goldberg, J. M. Physiology of peripheral neurons innervating otolith organs of the squirrel monkey. I. Response to static tilts and to long-duration centrifugal force. J. Neurophysiol. 39, 970–984 (1976).
Article CAS PubMed Google Scholar
Si, X., Angelaki, D. E. & Dickman, D. Response properties of pigeon otolith afferents to linear acceleration. Exp. Brain Res 117, 242–250 (1997).
Article CAS PubMed Google Scholar
Yu, X. J., Dickman, J. D. & Angelaki, D. E. Detection thresholds of macaque otolith afferents. J. Neurosci. 32, 8306–8316 (2012).
Article CAS PubMed PubMed Central Google Scholar
Baird, R. A. & Schuff, N. R. Peripheral innervation patterns of vestibular nerve afferents in the bullfrog utriculus. J. Comp. Neurol. 342, 279–298 (1994).
Article CAS PubMed Google Scholar
Jia, Y. & Bagnall, M. W. Monosynaptic targets of utricular afferents in the larval zebrafish. Front Neurol. 13, 937054 (2022).
Saint-Amant, L. & Drapeau, P. Time course of the development of motor behaviors in the zebrafish embryo. J. Neurobiol. 37, 622–632 (1998).
Article CAS PubMed Google Scholar
Mendelson, B. Development of reticulospinal neurons of the zebrafish. I. Time of origin. J. Comp. Neurol. 251, 160–171 (1986).
Article CAS PubMed Google Scholar
Fetcho, J. R. Spinal network of the Mauthner cell. Brain Behav. Evol. 37, 298–316 (1991).
Article CAS PubMed Google Scholar
Kimmel, C. B., Schabtach, E. & Kimmel, R. J. Developmental interactions in the growth and branching of the lateral dendrite of Mauthner’s cell (Ambystoma mexicanum). Dev. Biol. 55, 244–259 (1977).
Article CAS PubMed Google Scholar
Gao, Y. & Nicolson, T. Temporal vestibular deficits in synaptojanin 1 (synj1) mutants. Front. Mol. Neurosci. 13, 604189 (2021).
Favre-Bulle, I. A., Stilgoe, A. B., Rubinsztein-Dunlop, H. & Scott, E. K. Optical trapping of otoliths drives vestibular behaviours in larval zebrafish. Nat. Commun. 8, 630 (2017).
Article ADS PubMed PubMed Central CAS Google Scholar
Burgess, H. A. & Granato, M. Sensorimotor gating in Larval zebrafish. J. Neurosci. 27, 4984–4994 (2007).
Article CAS PubMed PubMed Central Google Scholar
Canfield, J. G. & Rose, G. J. Hierarchical sensory guidance of Mauthner-mediated escape responses in goldfish (Carassius auratus) and cichlids (Haplochromis burtoni). Brain Behav. Evol. 48, 137–156 (1996).
Article CAS PubMed Google Scholar
Stewart, W. J., Nair, A., Jiang, H. & McHenry, M. J. Prey fish escape by sensing the bow wave of a predator. J. Exp. Biol. 217, 4328–4336 (2014).
Article PubMed Google Scholar
Roberts, R., Elsner, J. & Bagnall, M. W. Delayed otolith development does not impair vestibular circuit formation in zebrafish. J. Assoc. Res Otolaryngol. 18, 415–425 (2017).
Article PubMed PubMed Central Google Scholar
Straka, H., Fritzsch, B. & Glover, J. C. Connecting ears to eye muscles: evolution of a ‘simple’ reflex arc. Brain Behav. Evol. 83, 162–175 (2014).
Article PubMed Google Scholar
Rohregger, M. & Dieringer, N. Principles of linear and angular vestibuloocular reflex organization in the frog. J. Neurophysiol. 87, 385–398 (2002).
Article CAS PubMed Google Scholar
Greaney, M. R., Privorotskiy, A. E., D’Elia, K. P. & Schoppik, D. Extraocular motoneuron pools develop along a dorsoventral axis in zebrafish, Danio rerio. J. Comp. Neurol. 525, 65–78 (2017).
Article CAS PubMed Google Scholar
Schoppik, D. et al. Gaze-stabilizing central vestibular neurons project asymmetrically to extraocular motoneuron pools. J. Neurosci. 37, 11353–11365 (2017).
Article CAS PubMed PubMed Central Google Scholar
Hamling, K. R. et al. Synaptic encoding of vestibular sensation regulates movement timing and coordination. bioRxiv https://doi.org/10.1101/2021.07.05.451142 (2021).
Liu, Z. et al. Central vestibular tuning arises from patterned convergence of otolith afferents. Neuron 108, 748–762.e4 (2020).
Article CAS PubMed PubMed Central Google Scholar
Shin, M. et al. Multiple types of cerebellar target neurons and their circuitry in the vestibulo-ocular reflex. J. Neurosci. 31, 10776–10786 (2011).
Article CAS PubMed PubMed Central Google Scholar
Xue, J. & Peterson, E. H. Hair bundle heights in the utricle: differences between macular locations and hair cell types. J. Neurophysiol. 95, 171–186 (2006).
Article PubMed Google Scholar
Li, A., Xue, J. & Peterson, E. H. Architecture of the mouse utricle: macular organization and hair bundle heights. J. Neurophysiol. 99, 718–733 (2008).
Article CAS PubMed Google Scholar
Baird, R. A. & Lewis, E. R. Correspondences between afferent innervation patterns and response dynamics in the bullfrog utricle and lagena. Brain Res 369, 48–64 (1986).
Article CAS PubMed Google Scholar
Lewis, E. R. & Li, C. W. Hair cell types and distributions in the otolithic and auditory organs of the bullfrog. Brain Res 83, 35–50 (1975).
Article ADS Google Scholar
Songer, J. E. & Eatock, R. A. Tuning and timing in Mammalian type I hair cells and calyceal synapses. J. Neurosci. 33, 3706–3724 (2013).
Article CAS PubMed PubMed Central Google Scholar
Goldberg, J. M. Afferent diversity and the organization of central vestibular pathways. Exp. Brain Res 130, 277–297 (2000).
Article CAS PubMed PubMed Central Google Scholar
Fernandez, C. & Goldberg, J. M. Physiology of peripheral neurons innervating otolith organs of the squirrel monkey. III. Response dynamics. J. Neurophysiol. 39, 996–1008 (1976).
Article CAS PubMed Google Scholar
Platt, C. Hair cell distribution and orientation in goldfish otolith organs. J. Comp. Neurol. 172, 283–298 (1977).
Article CAS PubMed Google Scholar
Haddon, C. & Lewis, J. Early ear development in the embryo of the zebrafish, Danio rerio. J. Comp. Neurol. 365, 113–128 (1996).
Article CAS PubMed Google Scholar
Jiang, T., Kindt, K. & Wu, D. K. Transcription factor Emx2 controls stereociliary bundle orientation of sensory hair cells. Elife 6, e22268 (2017).
Warchol, M. E., Massoodnia, R., Pujol, R., Cox, B. C. & Stone, J. S. Development of hair cell phenotype and calyx nerve terminals in the neonatal mouse utricle. J. Comp. Neurol. 527, 1913–1928 (2019).
Article CAS PubMed PubMed Central Google Scholar
Denman-Johnson, K. & Forge, A. Establishment of hair bundle polarity and orientation in the developing vestibular system of the mouse. J. Neurocytol. 28, 821–835 (1999).
Article CAS PubMed Google Scholar
Sans, A. & Chat, M. Analysis of temporal and spatial patterns of rat vestibular hair cell differentiation by tritiated thymidine radioautography. J. Comp. Neurol. 206, 1–8 (1982).
Article CAS PubMed Google Scholar
Burns, J. C., On, D., Baker, W., Collado, M. S. & Corwin, J. T. Over half the hair cells in the mouse utricle first appear after birth, with significant numbers originating from early postnatal mitotic production in peripheral and striolar growth zones. J. Assoc. Res Otolaryngol. 13, 609–627 (2012).
Article PubMed PubMed Central Google Scholar
Lysakowski, A. Development of synaptic innervation in the rodent utricle. Ann. N. Y Acad. Sci. 871, 422–425 (1999).
Article ADS CAS PubMed PubMed Central Google Scholar
Koudelka, S. et al. Individual neuronal subtypes exhibit diversity in CNS myelination mediated by synaptic vesicle release. Curr. Biol. 26, 1447–1455 (2016).
Article CAS PubMed PubMed Central Google Scholar
Gibson, E. M. et al. Neuronal activity promotes oligodendrogenesis and adaptive myelination in the mammalian brain. Science 344, 1252304 (2014).
Article PubMed PubMed Central CAS Google Scholar
Almeida, R. G., Czopka, T., ffrench-Constant, C. & Lyons, D. A. Individual axons regulate the myelinating potential of single oligodendrocytes in vivo. Development 138, 4443–4450 (2011).
Article CAS PubMed PubMed Central Google Scholar
Almeida, R. G. & Lyons, D. A. On myelinated axon plasticity and neuronal circuit formation and function. J. Neurosci. 37, 10023–10034 (2017).
Article CAS PubMed PubMed Central Google Scholar
Dechesne, C. J., Desmadryl, G. & Desmemes, D. Myelination of the mouse vestibular ganglion. Acta Otolaryngologica 103, 18–23 (1987).
Article CAS Google Scholar
Sun, Y. J., Kobayashi, H., Yoshida, S., Shirasawa, N. & Naito, A. Regional differences in myelination of chick vestibulocochlear ganglion cells. Int. J. Developmental Neurosci. 31, 568–579 (2013).
Article CAS Google Scholar
Romand, R., Sans, A., Romand, M. R. & Marty, R. The structural maturation of the stato-acoustic nerve in the cat. J. Comp. Neurol. 170, 1–15 (1976).
Article CAS PubMed Google Scholar
Dyballa, S. et al. Distribution of neurosensory progenitor pools during inner ear morphogenesis unveiled by cell lineage reconstruction. Elife 6, e22268 (2017).
Zecca, A., Dyballa, S., Voltes, A., Bradley, R. & Pujades, C. The order and place of neuronal differentiation establish the topography of sensory projections and the entry points within the hindbrain. J. Neurosci. 35, 7475–7486 (2015).
Article CAS PubMed PubMed Central Google Scholar
Vemaraju, S., Kantarci, H., Padanad, M. S. & Riley, B. B. A spatial and temporal gradient of Fgf differentially regulates distinct stages of neural development in the zebrafish inner ear. PLoS Genet 8, e1003068 (2012).
Article CAS PubMed PubMed Central Google Scholar
Drapeau, P. et al. Development of the locomotor network in zebrafish. Prog. Neurobiol. 68, 85–111 (2002).
Article CAS PubMed Google Scholar
Alexandre, D. & Ghysen, A. Somatotopy of the lateral line projection in larval zebrafish. Proc. Natl Acad. Sci. USA 96, 7558–7562 (1999).
Article ADS CAS PubMed PubMed Central Google Scholar
Liao, J. C. & Haehnel, M. Physiology of afferent neurons in larval zebrafish provides a functional framework for lateral line somatotopy. J. Neurophysiol. 107, 2615–2623 (2012).
Article PubMed PubMed Central Google Scholar
Mann, Z. F. & Kelley, M. W. Development of tonotopy in the auditory periphery. Hearing Res. 276, 2–15 (2011).
Article Google Scholar
Kishore, S., Cadoff, E. B., Agha, M. A. & McLean, D. L. Orderly compartmental mapping of premotor inhibition in the developing zebrafish spinal cord. Science 370, 431–436 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Deska-Gauthier, D., Borowska-Fielding, J., Jones, C. & Zhang, Y. The temporal neurogenesis patterning of spinal p3-V3 interneurons into divergent subpopulation assemblies. J. Neurosci. 40, 1440–1452 (2020).
Article PubMed PubMed Central Google Scholar
Osseward, P. J. et al. Conserved genetic signatures parcellate cardinal spinal neuron classes into local and projection subsets. Science 372, 385–393 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Bagnall, M. W. & McLean, D. L. Modular organization of axial microcircuits in zebrafish. Science 343, 197–200 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Bagnall, M. W. & Schoppik, D. Development of vestibular behaviors in zebrafish. Curr. Opin. Neurobiol. 53, 83–89 (2018).
Article CAS PubMed PubMed Central Google Scholar
Easter, S. S. Jr. & Nicola, G. N. The development of eye movements in the zebrafish (Danio rerio). Dev. Psychobiol. 31, 267–276 (1997).
Article PubMed Google Scholar
Glover, J. C. Development of second-order vestibular projections in the chicken embryo. Ann. N. Y Acad. Sci. 781, 13–20 (1996).
Article ADS CAS PubMed Google Scholar
Basaldella, E., Takeoka, A., Sigrist, M. & Arber, S. Multisensory signaling shapes vestibulo-motor circuit specificity. Cell 163, 301–312 (2015).
Article CAS PubMed Google Scholar
Uemura, O. et al. Comparative functional genomics revealed conservation and diversification of three enhancers of the isl1 gene for motor and sensory neuron-specific expression. Dev. Biol. 278, 587–606 (2005).
Article CAS PubMed Google Scholar
Murray, A. J., Croce, K., Belton, T., Akay, T. & Jessell, T. M. Balance control mediated by vestibular circuits directing limb extension or antagonist muscle co-activation. Cell Rep. 22, 1325–1338 (2018).
Article CAS PubMed Google Scholar
Lacoste, A. M. et al. A convergent and essential interneuron pathway for Mauthner-cell-mediated escapes. Curr. Biol. 25, 1526–1534 (2015).
Article CAS PubMed PubMed Central Google Scholar
Koyama, M. et al. A circuit motif in the zebrafish hindbrain for a two alternative behavioral choice to turn left or right. Elife 5, e16808 (2016).
Korn, H., Oda, Y. & Faber, D. S. Long-term potentiation of inhibitory circuits and synapses in the central nervous system. Proc. Natl Acad. Sci. USA 89, 440–443 (1992).
Article ADS CAS PubMed PubMed Central Google Scholar
Triller, A. & Korn, H. Morphologically distinct classes of inhibitory synapses arise from the same neurons: ultrastructural identification from crossed vestibular interneurons intracellularly stained with HRP. J. Comp. Neurol. 203, 131–155 (1981).
Article CAS PubMed Google Scholar
Jamali, M., Chacron, M. J. & Cullen, K. E. Self-motion evokes precise spike timing in the primate vestibular system. Nat. Commun. 7, 13229 (2016).
Article ADS PubMed PubMed Central Google Scholar
Jamali, M., Carriot, J., Chacron, M. J. & Cullen, K. E. Coding strategies in the otolith system differ for translational head motion vs. static orientation relative to gravity. Elife https://doi.org/10.7554/eLife.45573 (2019).
Migault, G. et al. Whole-Brain calcium Imaging during physiological vestibular stimulation in larval zebrafish. Curr. Biol. 28, 3723–3735.e6 (2018).
Article CAS PubMed PubMed Central Google Scholar
Favre-Bulle, I. A., Vanwalleghem, G., Taylor, M. A., Rubinsztein-Dunlop, H. & Scott, E. K. Cellular-resolution imaging of vestibular processing across the larval zebrafish brain. Curr. Biol. 28, 3711–3722.e3 (2018).
Article CAS PubMed Google Scholar
Angelaki, D. E. & Dickman, J. D. Spatiotemporal processing of linear acceleration: primary afferent and central vestibular neuron responses. J. Neurophysiol. 84, 2113–2132 (2000).
Article CAS PubMed Google Scholar
Schor, R. H. & Angelaki, D. E. The algebra of neural response vectors. Ann. N. Y Acad. Sci. 656, 190–204 (1992).
Article ADS CAS PubMed Google Scholar
Uchino, Y. et al. Cross-striolar and commissural inhibition in the otolith system. Ann. N. Y Acad. Sci. 871, 162–172 (1999).
Article ADS CAS PubMed Google Scholar
Highstein, S. M., Goldberg, J. M., Moschovakis, A. K. & Fernandez, C. Inputs from regularly and irregularly discharging vestibular nerve afferents to secondary neurons in the vestibular nuclei of the squirrel monkey. II. Correlation with output pathways of secondary neurons. J. Neurophysiol. 58, 719–738 (1987).
Article CAS PubMed Google Scholar
Boyle, R., Goldberg, J. M. & Highstein, S. M. Inputs from regularly and irregularly discharging vestibular nerve afferents to secondary neurons in squirrel monkey vestibular nuclei. III. Correlation with vestibulospinal and vestibuloocular output pathways. J. Neurophysiol. 68, 471–484 (1992).
Article CAS PubMed Google Scholar
Sato, F. & Sasaki, H. Morphological correlations between spontaneously discharging primary vestibular afferents and vestibular nucleus neurons in the cat. J. Comp. Neurol. 333, 554–566 (1993).
Article CAS PubMed Google Scholar
Carriot, J., Jamali, M., Chacron, M. J. & Cullen, K. E. The statistics of the vestibular input experienced during natural self-motion differ between rodents and primates. J. Physiol. 595, 2751–2766 (2017).
Article CAS PubMed PubMed Central Google Scholar
Ono, K. et al. Retinoic acid degradation shapes zonal development of vestibular organs and sensitivity to transient linear accelerations. Nat. Commun. 11, 63–77 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Straka, H., Vibert, N., Vidal, P. P., Moore, L. E. & Dutia, M. B. Intrinsic membrane properties of vertebrate vestibular neurons: Function, development and plasticity. Prog. Neurobiol. 76, 349–392 (2005).
Article CAS PubMed Google Scholar
Menelaou, E. & McLean, D. L. A gradient in endogenous rhythmicity and oscillatory drive matches recruitment order in an axial motor pool. J. Neurosci. 32, 10925–10939 (2012).
Article CAS PubMed PubMed Central Google Scholar
Ampatzis, K., Song, J., Ausborn, J. & El Manira, A. Pattern of innervation and recruitment of different classes of motoneurons in adult zebrafish. J. Neurosci. 33, 10875–10886 (2013).
Article CAS PubMed PubMed Central Google Scholar
Gabriel, J. P. et al. Principles governing recruitment of motoneurons during swimming in zebrafish. Nat. Neurosci. 14, 93–99 (2011).
Article CAS PubMed Google Scholar
Hayworth, K. J. et al. Imaging ATUM ultrathin section libraries with WaferMapper: a multi-scale approach to EM reconstruction of neural circuits. Front. Neural Circuits 8, 68 (2014).
Article PubMed PubMed Central Google Scholar
Morgan, J. L., Berger, D. R., Wetzel, A. W. & Lichtman, J. W. The fuzzy logic of network connectivity in mouse visual thalamus. Cell 165, 192–206 (2016).
Article CAS PubMed PubMed Central Google Scholar
Samuel, A. et al. TrakEM2 software for neural circuit reconstruction. PLoS One 7, e38011 (2012).
Article CAS Google Scholar
Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
Article CAS PubMed Google Scholar
Schneider-Mizell, C. M. et al. Quantitative neuroanatomy for connectomics in Drosophila. Elife 5, e12059 (2016).
Saalfeld, S., Cardona, A., Hartenstein, V. & Tomancak, P. CATMAID: collaborative annotation toolkit for massive amounts of image data. Bioinformatics 25, 1984–1986 (2009).
Article CAS PubMed PubMed Central Google Scholar
Laurens, J., Meng, H. & Angelaki, D. E. Neural representation of orientation relative to gravity in the macaque cerebellum. Neuron 80, 1508–1518 (2013).
Article CAS PubMed Google Scholar
Mirat, O., Sternberg, J. R., Severi, K. E. & Wyart, C. ZebraZoom: an automated program for high-throughput behavioral analysis and categorization. Front. Neural Circuits 7, 107 (2013).
Article PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank Drs. Mohini Sengupta, Vamsi Daliparthi and Mr. Saul Bello Rojas for thoughtful critiques of the paper. We are grateful to Drs. Daniel Kerschensteiner, David Schoppik and members from the Schoppik lab for insightful comments. Serial-section EM imaging was performed in part through the use of Washington University Center for Cellular Imaging (WUCCI) supported by Washington University School of Medicine, The Children’s Discovery Institute of Washington University and St. Louis Children’s Hospital (CDI-CORE-2015-505 and CDI-CORE-2019-813) and the Foundation for Barnes-Jewish Hospital (3770 and 4642). This work is supported by funding through the National Institutes of Health (NIH) R01 DC016413 (M.W.B.), EY030623 (J.L.M.), EY029313 (J.L.M.), a Sloan Research Fellowship (M.W.B.), a Fellowship in Neuroscience from the Leon Levy Foundation and a NARSAD Young Investigator Grant from the Brain and Behavior Research Foundation (D.G.C.H.), an unrestricted grant to the Department of Ophthalmology and Visual Sciences from Research to Prevent Blindness (J.L.M.), and a Research to Prevent Blindness Career Development Award (J.L.M.). M.W.B. is a Pew Biomedical Scholar and a McKnight Foundation Scholar.

Author information

Authors and Affiliations

Dept. of Neuroscience, Washington University in St. Louis, St. Louis, MO, USA
Zhikai Liu, Yizhen Jia, Nicholas Slimmon & Martha W. Bagnall
Laboratory of Neural Systems, The Rockefeller University, New York, NY, USA
David G. C. Hildebrand
Dept. of Ophthalmology, Washington University in St. Louis, St. Louis, MO, USA
Joshua L. Morgan

Authors

Zhikai Liu
View author publications
You can also search for this author in PubMed Google Scholar
David G. C. Hildebrand
View author publications
You can also search for this author in PubMed Google Scholar
Joshua L. Morgan
View author publications
You can also search for this author in PubMed Google Scholar
Yizhen Jia
View author publications
You can also search for this author in PubMed Google Scholar
Nicholas Slimmon
View author publications
You can also search for this author in PubMed Google Scholar
Martha W. Bagnall
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Z.L., Y.J., and M.W.B. conceived the project, performed the experiments and analyzed the data. J.L.M. and D.G.C.H. carried out the alignment of the serial-section EM image data. J.L.M. designed and carried out the Monte Carlo simulations. Z.L., M.W.B., Y.J., and N.S. completed the reconstructions. Z.L. and M.W.B. wrote the manuscript with input from all other authors.

Corresponding author

Correspondence to Martha W. Bagnall.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Movie 1

Supplementary Movie 2

Supplementary Movie 3

Supplementary Movie 4

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Liu, Z., Hildebrand, D.G.C., Morgan, J.L. et al. Organization of the gravity-sensing system in zebrafish. Nat Commun 13, 5060 (2022). https://doi.org/10.1038/s41467-022-32824-w

Download citation

Received: 11 October 2021
Accepted: 18 August 2022
Published: 27 August 2022
DOI: https://doi.org/10.1038/s41467-022-32824-w

This article is cited by

Measuring Optokinetic Reflex and Vestibulo-Ocular Reflex in Unilateral Vestibular Organ Damage Model of Zebrafish
- Kang Hyeon Lim
- Hong Ki Kim
- June Choi
Journal of the Association for Research in Otolaryngology (2024)
Automated synapse-level reconstruction of neural circuits in the larval zebrafish brain
- Fabian Svara
- Dominique Förster
- Herwig Baier
Nature Methods (2022)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.