Morphology and synapse topography optimize linear encoding of synapse numbers in Drosophila looming responsive descending neurons

  1. Anthony Moreno-Sanchez
  2. Alexander N Vasserman
  3. HyoJong Jang
  4. Bryce W Hina
  5. Catherine R von Reyn
  6. Jessica Ausborn

  • Curated by eLife

    eLife logo

    eLife Assessment

    This valuable study provides a detailed picture of the synapse distributions for a set of visual projection neurons and their downstream partners, in combination with multi-compartmental modelling fitted to electrophysiological data. The model reveals interesting consequences of synapse topography for neuronal computation. The analysis, however, seems incomplete as the authors only analyze passive models of these spiking neurons, and do not attempt to connect their analysis to the bigger picture at the behavioral level.

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Synapses are often precisely organized on dendritic arbors, yet the role of synaptic topography in dendritic integration remains poorly understood. Utilizing electron microscopy (EM) connectomics we investigate synaptic topography in Drosophila melanogaster looming circuits, focusing on retinotopically tuned visual projection neurons (VPNs) that synapse onto descending neurons (DNs). Synapses of a given VPN type project to non-overlapping regions on DN dendrites. Within these spatially constrained clusters, synapses are not retinotopically organized, but instead adopt near random distributions. To investigate how this organization strategy impacts DN integration, we developed multicompartment models of DNs fitted to experimental data and using precise EM morphologies and synapse locations. We find that DN dendrite morphologies normalize EPSP amplitudes of individual synaptic inputs and that near random distributions of synapses ensure linear encoding of synapse numbers from individual VPNs. These findings illuminate how synaptic topography influences dendritic integration and suggest that linear encoding of synapse numbers may be a default strategy established through connectivity and passive neuron properties, upon which active properties and plasticity can then tune as needed.

Article activity feed

  1. eLife

    eLife Assessment

    This valuable study provides a detailed picture of the synapse distributions for a set of visual projection neurons and their downstream partners, in combination with multi-compartmental modelling fitted to electrophysiological data. The model reveals interesting consequences of synapse topography for neuronal computation. The analysis, however, seems incomplete as the authors only analyze passive models of these spiking neurons, and do not attempt to connect their analysis to the bigger picture at the behavioral level.

  2. eLife

    Reviewer #1 (Public review):

    Summary:

    This study makes use of the EM reconstruction of the fly brain to investigate the morphology and topography of the synapses between retinotopic, loom-sensitive visual projection neurons (VPNs) and downstream descending neurons (DNs). The authors analyzed the distribution of synapses on the dendritic trees of DNs and performed multi-compartmental modelling to study the implications of the synaptic arrangements for neuronal integration of input signals.

    Until recently, it has been unclear how spatial information is passed from retinotopic loom-sensitive neurons to descending neurons because the axons of the VPNs terminate in small optic glomeruli with no apparent topographic organization. It has recently been shown that synaptic weight gradients of VPNs connecting to DNs are the main mechanisms that allow for directed behavioral output (Dombrovski et al.). This study now goes one step further to determine if precise synapse location on the dendritic tree contributes further to the information processing. The study suggests that (1) none of the VPNs investigated show a retinotopic organization of synapses on DN dendrites. (2) Synapses of single VPNs are locally clustered. (3) Initial EPSPs at the synaptic location have, as expected, varying amplitudes but the amplitudes are passively normalized and only cover a small range when measured at the SIZ. (4) A near random distribution of synapses allows for linear integration of synaptic inputs when only a few VPNs are activated.

    Strengths:

    This study provides a detailed picture of the synapse distribution for a set of VPN and DN pairs, in combination with multi-compartmental modelling fitted to electrophysiological data. The data and methods are clear. The findings are overall interesting. The computational pipeline, which should ideally be made publicly available, will allow the community to make similar analyses on different neuronal classes, which will facilitate the detection of more general mechanisms of dendritic computation.

    Weaknesses:

    - In my opinion, we need more detail on the electrophysiological data and the fitting of the multi-compartmental model, which is the foundation of large parts of the study.
    - The study shows that the synapses of an individual VPN are locally clustered and suggests this as evidence for clustering of synapses of similar tuning (as has been shown previously in other systems). I am not fully convinced by the arguments here, since synapses of a single neuron are by necessity not randomly distributed in space.
    - As written, it was in parts unclear to me what the main hypotheses and conclusions were - e.g., how would a retinotopic distribution of synapses on dendritic trees contribute to information processing? Are the model predictions in line with the presumed behavioural role of these neurons?

  3. eLife

    Reviewer #2 (Public review):

    Summary:

    This article investigates the distribution of synapses on the dendritic arbors of descending neurons in the looming circuit of the fly visual system. The authors use publicly available EM reconstruction data of the adult fly brain to identify the positions of synapses from several types of visual projection neuron (VPN) to descending neuron (DN) connections. VPN dendrites are retinotopically organized, and axons from different VPN populations innervate distinct optic glomeruli. Yet the authors did not find any retinotopic organization of the synapses in the VPN-DN pairs they analyzed. They then constructed passive electrical models of the DNs with their structures extracted from the EM reconstructions. They focused on two specific DNs and parameterized their models by conducting whole-cell recordings within a voltage range below spiking threshold. Simulation of these passive models showed that irrespective of the location of a synapse, EPSPs became very similar at the spike initiation zone. This is consistent with the idea of synaptic democracy where EPSPs at far away synapses have higher amplitude compared to those nearer to the spike initiation zone so that they all attenuate to similar amplitudes while reaching there. The authors found that activating synapses from individual VPNs have the same effect as activating a random set of synapses. They conclude that despite some clustering of VPN synapses at small scale, they are distributed randomly over the dendritic arbor of DNs so that their EPSP amplitude encode the number of activated synapses, avoiding sublinearity from shunting effect.

    Strengths:

    - Experimental confirmation of the location of the spike initiation zone in the DN arbors is interesting and may provide better understanding of signal processing in these neurons.
    - Passive parameters obtained through electrophysiological recordings are useful.
    - These morphologically detailed single neuron models, if made available publicly, will be beneficial for building more complete models to understand the fly visual circuit.
    - The authors have complemented the work of Dombrovski et al by analyzing the distribution of synapses in more detail from EM data for a different set of neurons.

    Weaknesses:

    DNs are upstream of motorneurons, and one would expect, as demonstrated by Dombrovski et al, that specific DNs being activated by input from specific regions of the visual field will activate motoneurons so that the fly moves away from a looming object.

    The current work analyzed the synapse distribution on two DNs that do not seem to have such role, and emphasize the lack of retinotopy. However, it is not clear why one would expect retinotopy in synapse location on the dendritic arbor. The comparison with mammalian visual circuits is not appropriate because those layers are extracting more and more complex visual features, whereas Drosophila DNs are supposed to drive motoneurons to generate suitable escape behavior.

    - The authors do not suggest the functional roles of these DNs in controlling the movement of the fly. They argue that the synapse distribution and the passive electrotonic structure of these neurons are optimized to make the composite EPSP encode the number of activated synapses, but do not explain why this is important.

    - Although DNs are spiking neurons, the authors limit their work to the subthreshold passive domain. If the EPSP at the spike initiation zone crosses spiking threshold, will encoding the number of synapses in EPSP amplitude still matter? Will it matter either if the composite EPSP remains subthreshold?

    - The temporal aspect of the input has been ignored by the authors in their simulations. First, it is not clear all the synapses from a single VPN should get activated together. One would expect a spike in a VPN to arrive at different synapses with different time delays depending on their electrotonic distance from the spike initiation zone and the signal propagation speed in the neurites.

    A looming stimulus should be expanding with time, but from the description of the simulations it does not seem that the authors have tried to incorporate this aspect in their design of the synaptic activation.

    - The suggestion in the abstract that linear encoding of synapse number is default strategy which is then tuned by active properties and plasticity seems strange. Developmentally active properties do not get inserted into passive neurons.

    - Much of the analysis (Figures 4, 5, 12) show relationships with physical distance along dendrite. In studying passive neurons it is more informative to use electrotonic distance which provides better insight.

  4. eLife

    Author response:

    We thank the editors and reviewers for their valuable feedback and are committed to addressing their suggestions in a revised manuscript. We appreciate the reviewers’ recognition of the value of our findings, including the insights into the consequences of synaptic topography and the investigation of spike initiation zones in DNs, which further advance our understanding of signal processing. Our studies offer broader insights into synaptic organization and its significance for dendritic integration in an ethologically relevant context.

    We particularly appreciate the reviewer's suggestion to elaborate on the electrophysiological properties of DNs and to consider the electrotonic distance in our analysis. We also thank the reviewers for highlighting points that need clarification. In short, our models suggest that DNs effectively distribute synapses to maintain linear encoding of synapse numbers when multiple synapses are coactivated. This supports the results of an earlier study suggesting that synapse number gradients encode the location of an approaching stimulus in these neurons (Dombrovski et al., 2023).

    We also agree with the reviewers that the temporal activation of synapses is highly relevant for this system. However, we have focused on synaptic topography because the characterization of temporal patterns of VPN activity is currently lacking in the field. A more detailed investigation of temporal dynamics is therefore beyond the scope of this study.

    With the publication of the reviewed preprint, we have now made the computational pipeline and models available on GitHub (https://github.com/AusbornLab/VPN-DN-synapse-normalization).

    Reference

    Dombrovski M, Peek MY, Park J-Y, Vaccari A, Sumathipala M, Morrow C, Breads P, Zhao A, Kurmangaliyev YZ, Sanfilippo P, Rehan A, Polsky J, Alghailani S, Tenshaw E, Namiki S, Zipursky SL, Card GM. 2023. Synaptic gradients transform object location to action. Nature 613:534–542. doi:10.1038/s41586-022-05562-8

  5. Version published to 10.7554/elife.99277.1 on eLife
  6. Version published to 10.7554/elife.99277 on eLife
  7. Version published to 10.1101/2024.04.24.591016v1 on bioRxiv