Abstract
Many animals orient using visual cues, but how a single cue is selected from among many is poorly understood. Here we show that Drosophila ring neurons—central brain neurons implicated in navigation—display visual stimulus selection. Using in vivo two-color two-photon imaging with genetically encoded calcium indicators, we demonstrate that individual ring neurons inherit simple-cell-like receptive fields from their upstream partners. Stimuli in the contralateral visual field suppressed responses to ipsilateral stimuli in both populations. Suppression strength depended on when and where the contralateral stimulus was presented, an effect stronger in ring neurons than in their upstream inputs. This history-dependent effect on the temporal structure of visual responses, which was well modeled by a simple biphasic filter, may determine how visual references are selected for the fly's internal compass. Our approach highlights how two-color calcium imaging can help identify and localize the origins of sensory transformations across synaptically connected neural populations.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 12 print issues and online access
$209.00 per year
only $17.42 per issue
Buy this article
- Purchase on Springer Link
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
References
Bisley, J.W. & Goldberg, M.E. Attention, intention, and priority in the parietal lobe. Annu. Rev. Neurosci. 33, 1–21 (2010).
Knudsen, E.I. Fundamental components of attention. Annu. Rev. Neurosci. 30, 57–78 (2007).
Squire, R.F., Noudoost, B., Schafer, R.J. & Moore, T. Prefrontal contributions to visual selective attention. Annu. Rev. Neurosci. 36, 451–466 (2013).
Maunsell, J.H.R. Neuronal mechanisms of visual attention. Annu. Rev. Vis. Sc. 1, 373–391 (2015).
Reynolds, J.H. & Chelazzi, L. Attentional modulation of visual processing. Annu. Rev. Neurosci. 27, 611–647 (2004).
Wimmer, R.D. et al. Thalamic control of sensory selection in divided attention. Nature 526, 705–709 (2015).
Zhang, S. et al. Selective attention. Long-range and local circuits for top-down modulation of visual cortex processing. Science 345, 660–665 (2014).
de Bivort, B.L. & van Swinderen, B. Evidence for selective attention in the insect brain. Curr. Opin. Insect Sci. 15, 9–15 (2016).
Borst, A. Fly visual course control: behaviour, algorithms and circuits. Nat. Rev. Neurosci. 15, 590–599 (2014).
Silies, M., Gohl, D.M. & Clandinin, T.R. Motion-detecting circuits in flies: coming into view. Annu. Rev. Neurosci. 37, 307–327 (2014).
Tang, S. & Guo, A. Choice behavior of Drosophila facing contradictory visual cues. Science 294, 1543–1547 (2001).
Zhang, K., Guo, J.Z., Peng, Y., Xi, W. & Guo, A. Dopamine-mushroom body circuit regulates saliency-based decision-making in Drosophila. Science 316, 1901–1904 (2007).
Sareen, P., Wolf, R. & Heisenberg, M. Attracting the attention of a fly. Proc. Natl. Acad. Sci. USA 108, 7230–7235 (2011).
Seelig, J.D. & Jayaraman, V. Feature detection and orientation tuning in the Drosophila central complex. Nature 503, 262–266 (2013).
Seelig, J.D. & Jayaraman, V. Neural dynamics for landmark orientation and angular path integration. Nature 521, 186–191 (2015).
Ofstad, T.A., Zuker, C.S. & Reiser, M.B. Visual place learning in Drosophila melanogaster. Nature 474, 204–207 (2011).
Neuser, K., Triphan, T., Mronz, M., Poeck, B. & Strauss, R. Analysis of a spatial orientation memory in Drosophila. Nature 453, 1244–1247 (2008).
Träger, U., Wagner, R., Bausenwein, B. & Homberg, U. A novel type of microglomerular synaptic complex in the polarization vision pathway of the locust brain. J. Comp. Neurol. 506, 288–300 (2008).
Held, M. et al. Microglomerular synaptic complexes in the sky-compass network of the honeybee connect parallel pathways from the anterior optic tubercle to the central complex. Front. Behav. Neurosci. 10, 186 (2016).
Omoto, J.J. et al. Visual input to the Drosophila central complex by developmentally and functionally distinct neuronal populations. Curr. Biol. 27, 1098–1110 (2017).
Dana, H. et al. Sensitive red protein calcium indicators for imaging neural activity. Elife 5, e12727 (2016).
Chen, T.W. et al. Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature 499, 295–300 (2013).
Jenett, A. et al. A GAL4-driver line resource for Drosophila neurobiology. Cell Rep. 2, 991–1001 (2012).
Kvon, E.Z. et al. Genome-scale functional characterization of Drosophila developmental enhancers in vivo. Nature 512, 91–95 (2014).
Klapoetke, N.C. et al. Independent optical excitation of distinct neural populations. Nat. Methods 11, 338–346 (2014).
Laissue, P.P. et al. Three-dimensional reconstruction of the antennal lobe in Drosophila melanogaster. J. Comp. Neurol. 405, 543–552 (1999).
Wilson, R.I., Turner, G.C. & Laurent, G. Transformation of olfactory representations in the Drosophila antennal lobe. Science 303, 366–370 (2004).
Wang, J.W., Wong, A.M., Flores, J., Vosshall, L.B. & Axel, R. Two-photon calcium imaging reveals an odor-evoked map of activity in the fly brain. Cell 112, 271–282 (2003).
Li, H., Li, Y., Lei, Z., Wang, K. & Guo, A. Transformation of odor selectivity from projection neurons to single mushroom body neurons mapped with dual-color calcium imaging. Proc. Natl. Acad. Sci. USA 110, 12084–12089 (2013).
Hubel, D.H. & Wiesel, T.N. Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. J. Physiol. (Lond.) 160, 106–154 (1962).
Adelson, E.H. & Bergen, J.R. Spatiotemporal energy models for the perception of motion. J. Opt. Soc. Am. A 2, 284–299 (1985).
DeAngelis, G.C., Ohzawa, I. & Freeman, R.D. Receptive-field dynamics in the central visual pathways. Trends Neurosci. 18, 451–458 (1995).
Ringach, D.L., Hawken, M.J. & Shapley, R. Dynamics of orientation tuning in macaque primary visual cortex. Nature 387, 281–284 (1997).
Cai, D., DeAngelis, G.C. & Freeman, R.D. Spatiotemporal receptive field organization in the lateral geniculate nucleus of cats and kittens. J. Neurophysiol. 78, 1045–1061 (1997).
Perge, J.A., Borghuis, B.G., Bours, R.J., Lankheet, M.J. & van Wezel, R.J. Temporal dynamics of direction tuning in motion-sensitive macaque area MT. J. Neurophysiol. 93, 2104–2116 (2005).
Schmid, A.M., Purpura, K.P. & Victor, J.D. Responses to orientation discontinuities in V1 and V2: physiological dissociations and functional implications. J. Neurosci. 34, 3559–3578 (2014).
Schmid, A.M., Purpura, K.P., Ohiorhenuan, I.E., Mechler, F. & Victor, J.D. Subpopulations of neurons in visual area v2 perform differentiation and integration operations in space and time. Front. Syst. Neurosci. 3, 15 (2009).
Pfeiffer, K. & Homberg, U. Organization and functional roles of the central complex in the insect brain. Annu. Rev. Entomol. 59, 165–184 (2014).
Fisher, Y.E., Silies, M. & Clandinin, T.R. Orientation selectivity sharpens motion detection in Drosophila. Neuron 88, 390–402 (2015).
Moran, J. & Desimone, R. Selective attention gates visual processing in the extrastriate cortex. Science 229, 782–784 (1985).
Treue, S. & Martínez Trujillo, J.C. Feature-based attention influences motion processing gain in macaque visual cortex. Nature 399, 575–579 (1999).
Longordo, F., To, M.S., Ikeda, K. & Stuart, G.J. Sublinear integration underlies binocular processing in primary visual cortex. Nat. Neurosci. 16, 714–723 (2013).
Bittner, K.C. et al. Conjunctive input processing drives feature selectivity in hippocampal CA1 neurons. Nat. Neurosci. 18, 1133–1142 (2015).
Dong, D.W. & Atick, J.J. Temporal decorrelation: a theory of lagged and nonlagged responses in the lateral geniculate nucleus. Network 6, 159–178 (1995).
Mysore, S.P. & Knudsen, E.I. Reciprocal inhibition of inhibition: a circuit motif for flexible categorization in stimulus selection. Neuron 73, 193–205 (2012).
Machens, C.K., Romo, R. & Brody, C.D. Flexible control of mutual inhibition: a neural model of two-interval discrimination. Science 307, 1121–1124 (2005).
Pfeiffer, K., Kinoshita, M. & Homberg, U. Polarization-sensitive and light-sensitive neurons in two parallel pathways passing through the anterior optic tubercle in the locust brain. J. Neurophysiol. 94, 3903–3915 (2005).
Collett, T.S. How ladybirds approach nearby stalks: a study of visual selectivity and attention. J. Comp. Physiol. A Neuroethol. Sens. Neural Behav. Physiol. 163, 355–363 (1988).
Cruz-Martín, A. et al. A dedicated circuit links direction-selective retinal ganglion cells to the primary visual cortex. Nature 507, 358–361 (2014).
Wertz, A. et al. Single-cell-initiated monosynaptic tracing reveals layer-specific cortical network modules. Science 349, 70–74 (2015).
Ruta, V. et al. A dimorphic pheromone circuit in Drosophila from sensory input to descending output. Nature 468, 686–690 (2010).
Nern, A., Pfeiffer, B.D. & Rubin, G.M. Optimized tools for multicolor stochastic labeling reveal diverse stereotyped cell arrangements in the fly visual system. Proc. Natl. Acad. Sci. USA 112, E2967–E2976 (2015).
Aso, Y. et al. The neuronal architecture of the mushroom body provides a logic for associative learning. Elife 3, e04577 (2014).
Viswanathan, S. et al. High-performance probes for light and electron microscopy. Nat. Methods 12, 568–576 (2015).
Peng, H., Ruan, Z., Long, F., Simpson, J.H. & Myers, E.W. V3D enables real-time 3D visualization and quantitative analysis of large-scale biological image data sets. Nat. Biotechnol. 28, 348–353 (2010).
Acknowledgements
We thank J. Strother, I. Negrashov and A. Wong for technical assistance; T. Wolff for critical input on the anatomy; B. Dickson for sharing reagent information; B. Sharp, K. Hibbard, J. McMahon, G. Ihrke and others at Janelia Fly Facility and PTR for fly husbandry; J. Hasseman, G. Tsegaye, H. Dionne and M. Ramirez for molecular biology assistance; J. Macklin and R. Patel for technical discussions; E. Nielson for art work and animation; J. Seelig, M. Reiser, E. Gruntman and S. Wu for useful discussions; E. Gruntman and K. Longden for comments on the manuscript; and members of Vivek's lab for discussions and comments on the manuscript. This work was supported by the Howard Hughes Medical Institute.
Author information
Authors and Affiliations
Contributions
Y.S., D.S.K. and V.J. conceived the study. Y.S. designed and performed all in vivo imaging experiments and data analysis. A.M.H. performed modeling with input from Y.S. and V.J. Y.S., A.N. and R.F. performed anatomical characterization and functional connectivity experiments. H.D. provided critical suggestions on reagents. Y.S., A.M.H. and V.J. wrote the manuscript with comments from all authors. E.R.S., V.J., L.L.L., K.S. and D.S.K. steer the Genetically-Encoded Neuronal Indicator and Effector (GENIE) team, which initiated this project.
Corresponding authors
Ethics declarations
Competing interests
D.S.K., E.R.S., L.L.L. and K.S. are co-applicants for a patent on materials and methods related to jRGECO1a and other red GECI variants (application number US 14/974,483). The other authors declare no competing interests.
Integrated supplementary information
Supplementary Figure 1 Anatomy and functional connectivity suggest that TB neurons are part of an excitatory feedforward pathway from the AOTU to the BU composed of parallel channels.
a, Different channels of Fig. 1e. The signal from the pre-synaptic marker synaptotagmin expressed in a subset of TB neurons (SS02927 split Gal4) is seen only in the BU, consistent with these neurons being part of a feedforward pathway from the AOTU to the BU.
b, An additional example of MCFO using SS02927 split-Gal4 to label a subset of TB neurons.
c, Examples of MCFO with sparse TB neuron labeling using R76B06-Gal4 as a driver line. A total of more than 60 flies with different labeling densities were examined. For examples of high-density labeling, see Supplementary Fig. 2a, b, e. For quantification of projection patterns, see Results and Online Methods.
d, Average responses showing that optogenetic activation of TB neurons with CsChrimson (yellow bar) induces slow and multi-phasic GCaMP6f responses in ring neurons. Each trace represents a fly (4 repeats for each fly); the mean response is overlaid with a dark thick line.
e, Variations of baseline activity in ring neurons, shown for different flies. Each trace represents a fly (6 flies, 4 repeats for each fly); the mean response across flies is overlaid with a dark thick line. Note that the activity does not return to baseline and that the elevation can sometimes persist for several seconds. F0 is computed at the beginning of an experimental run, each of which comprised 4 repeats.
Scale bars in a-c: 20 μm.
Supplementary Figure 2 Anatomical study showing a microglomerular organization of TB neuron outputs in the BU.
a-b, Maximum intensity projection (MIP) and one cross-section (Section) of MCFO images of TB neurons (using R76B06-Gal4 driver line) in BU of one hemisphere in two example flies ((a) and (b), respectively). TB neuron axons are organized into distinct micro-glomeruli. Individual glomeruli appear to be innervated by single TB neurons (see (f) for an apparent exception).
c-d, MCFO images of ring neurons (using R56H10-Gal4 driver line) in one BU. (c) and (d) are from two example flies; left panels show MIPs, and right panels show individual cross-sections. Ring neuron dendrites are also organized into distinct glomeruli. Based on their uniform labeling color, the majority of labeled glomeruli appear to be innervated by a single labeled ring neuron. However, some glomeruli showed labeling with more than one color even in single confocal sections, suggesting the presence of dendrites from more than one cell (see arrow in (c)). Note that ring neuron glomeruli appear hollow in the middle (more clearly seen in single sections), in contrast to the generally more compact structure of the TB neuron glomeruli.
e-f, MIP of MCFO images of a population of TB neurons (using the R76B06-Gal4 driver line) in one hemisphere, including examples of dense (e) and sparse (f) labeling. The white arrowheads in (f) point to the terminals of an atypical TB neuron that appear to innervate three glomeruli in BU. The arrow in (e) points to a terminal with an unusually extended, perhaps multi-glomerular, shape.
g-h, MIP of MCFO images of a population of ring neurons (using R56H10-Gal4 driver line) in one hemisphere, including examples of dense (g) and sparse (h) labeling. The white arrows in (h) point to the ring neuron that appears to branch and innervate several glomeruli in BU.
i, MIP of immunostaining against GFP and DsRed in a transgenic fly expressing red and green GECIs in TB, ring, and MT neurons, where they overlap in the AOTU and BU.
Supplementary Figure 3 Optimizing spectral separation in two-color two-photon imaging.
a, Schematic of in vivo two-color two-photon imaging setup, with parameters of optical elements labeled. A 532 nm DPSS laser beam is split into two to drive two DMDs for visual stimulation. A narrow-band green channel filter (511/20) was used to isolate the GCaMP signal from contamination by the 532 nm visual stimulation light. A 641/75 filter was chosen for the red channel, as discussed in (b) and (c). We also found a non-coherent emission from the mode-locked laser that would severely contaminate the signal in the red channel. This emission can be filtered by a long-pass filter (715LP).
b, Two-photon excitation and fluorescence emission spectra of jRGECO1a and GCaMP6f, (adapted from Dana H. et al., eLife, 2016, Ref. 20). The green line is the 532 nm laser for visual stimulation. The emission spectra of GCaMP and RGECO overlap substantially, and GCaMP emission is relatively strong around the peak of RGECO emission. Filtering out GCaMP signal by using a more red-shifted and narrower bandpass filter would also decrease the SNR in the RGECO channel. The excitation spectra show that around 1000 nm, the brightness (and thus SNR) of RGECO increases monotonically with the increase of wavelength. Increasing the excitation wavelength thus compensates for the decrease of SNR in the RGECO channel from emission filtering, and by this means we achieved usable SNRs and spectral separation between GCaMP and RGECO.
c, Labeled GCaMP and RGECO fluorescence emission spectra under non-optimized (first row) and optimized (second row) imaging conditions. Red and green channel PMT values in tissues only expressing GCaMP (middle column), and with both GCaMP and RGECO expressed and intermingled (right column), under the two imaging conditions. Each dot represents a single time point, and the dots are color-coded according to their glomeruli. Dots from the same glomeruli form a cluster. High levels of contamination lead to strong correlation between the two channels (upper middle).
d, Exemplar traces of neural processes expressing only GCaMP nonetheless show strong bleed-through from the green channel into the red when imaged under sub-optimal conditions. Under 1000 nm excitation and 625/90 filtering, GCaMP signal contaminates the jRGECO1a channel, and ΔF/F0 values in the red channel match those of GCaMP6f, which can exceed the maximum ΔF/F0 available for jRGECO1a. By contrast, under 1020 nm excitation and with 641/75 filtering, no signal was seen in the red channel (e). Note that the red channel should not have any signal, since the imaged neural processes expressed only GCaMP.
e, ΔF/F0 signals recorded in green and red channels in neurons expressing only GCaMP6f shows minimal bleed-through of signal from the green indicator into the red channel.
Supplementary Figure 4 Contralateral suppression is the dominant effect in BU.
a, Distribution of summated ring neuron responses to ipsilateral and contralateral bars, together with exemplar traces; each dot is a glomerulus. While the majority of the imaged glomeruli show contralateral suppression—positive response to an ipsilateral bar and negative response to a contralateral bar (lower right quadrant with blue shade and exemplar traces thereof)—there are also examples of contralateral enhancement (upper two quadrants).
b-c, Distribution of summated TB and ring neuron responses to ipsilateral and contralateral bars in different GECI combinations, as in (a); each dot is a glomerulus. Ipsilateral excitation and contralateral inhibition is the dominant response pattern, regardless of cell type and GECI combination. Note that we did not image the entire BU; instead, we chose a similar depth that optimized coverage of the dorsal BU (see Online Methods for details, (TB)G _(Ring)R: 17 BUs from 10 flies; (TB)R_(Ring)G: 15 BUs from 12 flies). Right panel in (b) is the same as (a). We found that responses (% ΔF/F0) to bilateral stimuli are the linear combinations of responses to ipsilateral and contralateral stimuli, regardless of the GECI combinations used. For (TB)G _(Ring)R: RBiring = 0.83 * RIpsiring + 0.30* RContraring -3.27, goodness of fit κ2 = 0.90; RBiTB= 0.89 * RIpsiTB+ 0.16* RContraTB-3.24, goodness of fit κ2 = 0.96. For (TB)R_(Ring)G: RBiring = 0.83 * RIpsiring + 0.23* RContraring -5.74, goodness of fit κ2 = 0.95; RBiTB= 0.83* RIpsiTB+ 0.34* RContraTB-2.97, goodness of fit κ2 = 0.95.
d, all selected glomeruli corresponding to Fig. 3e.
e, all selected glomeruli corresponding to Fig. 4e, f.
Supplementary Figure 5 Neuron-specific state-dependence.
a, Averaged images of the BU across frames during an UP state (epoch of elevated calcium activity not specific to stimulus presentation) and a normal state, with glomerulus boundaries shown in dashed lines. Clear differences (white arrow) can be seen in ring neurons but not TB neurons in the highlighted glomerulus (thick solid white line). Typically, state-dependence was limited to specific neurons (for example, a ring neuron here) of specific glomeruli.
b, Exemplar traces of TB and ring neurons; UP states in the ring neuron are highlighted within dashed boxes. Traces are taken from the highlighted glomerulus in (a).
c, Exemplar traces of TB and ring neurons from the same BU at the same time as in (b) but in a different glomerulus than (b); no state-dependence is seen.
Supplementary Figure 6 History dependence is strongest in bilateral stimuli.
a, Schematic of two-step stimulus sequence: a 1 s stimulus (either ipsilateral, contralateral, or bilateral bars) is followed by 1 s of darkness, and then a second stimulus (either an ipsilateral or contralateral bar) is presented for 1 s. Responses are grouped as X-Ipsi or X-Contra, where “X” designates the set of possible stimuli used in the first stimulus presentation (either ipsilateral, contralateral, or bilateral bars).
b, Average responses (mean ± s.e.m.) to X-Ipsi and X-Contra stimuli in (a) in an exemplar BU glomerulus.
c, Simulated responses generated by the circuit model in Fig. 6f-h capture the reduced history dependence (74%) in X-Ipsi compared to that in X-Bi (111%; Fig. 6h). While the model also predicts history dependence for X-Contra stimuli, rectification of the calcium signal prevents detection of history dependence in the data.
d, Trans-synaptic scatter plot of history dependence (see Online Methods for details) for X-Ipsi and X-Contra stimulus conditions, shown for all (gray) and selected (black) glomeruli (corresponding to the analogous plots in Fig. 4f-g). See Supplementary Tables 2 and 5 for statistical results.
e-f, Empirical cumulative distribution function of history dependence for X-Bi, X-Ipsi, and X-Contra conditions, shown for TB and ring neurons in all (solid line) and selected (dashed line) glomeruli in different GECI combinations. See Supplementary Tables 2 and 5 for statistical results.
Supplementary Figure 7 History dependence only involves the most recent stimuli.
a, Schematic of three-step stimulus sequence. A 1 s stimulus (either ipsilateral, or contralateral, or bilateral bars) is followed by 1 s of darkness, and then a second 1 s stimulus (either ipsilateral, or contralateral, or bilateral bars) is followed again by 1 s of darkness, and finally a third stimulus (bilateral bars) is presented for 1 s.
b, Exemplar grouped glomeruli responses (mean ± s.e.m.) show no clear history dependence of the third response based on the first stimulus. Responses are grouped as X-Ipsi-Bi, X-Contra-Bi, and X-Bi-Bi, where “X” designates the set of possible stimuli used during the first stimulus presentation (either ipsilateral, contralateral, or bilateral bars).
c, Simulated responses generated by the circuit model in Fig. 6f-h capture the lack of history dependence of the responses to the last stimulus (“Bi”) on stimuli presented two steps back (“X”).
d-e, Trans-synaptic scatter plot of history dependence (see Online Methods for details) of three step sequence as in (a), shown for all (gray) and selected (black) glomeruli in different GECI combinations. The population average does not exhibit obvious history dependence in either GECI combination. See Supplementary Table 4 for statistical results.
f-g, Cumulative distribution of history dependence of three step sequence as in (a), shown for all (solid lines) and selected (dashed lines) glomeruli in both GECI combinations. The population average does not exhibit obvious history dependence in either GECI combination. See Supplementary Table 4 for statistical results.
Supplementary Figure 8 Activity during the dark period is correlated with history dependence in ring neurons.
a, Schematic of the stimulus sequence. A 1 s stimulus (either ipsilateral, contralateral, or bilateral bars) is followed by a dark period of variable duration, and then a second stimulus (bilateral bars) is presented for 1 s.
b, Ring-neuron-specific persistent activity (arrowheads) in exemplar BU glomerulus during dark periods of varying duration (mean ± s.e.m.).
c, Zoomed-in traces from rectangle in (b).
d, Simulated ring neuron responses generated by the circuit model in Fig. 6f-h capture the rank-ordered activity shown in (c) (shown for 5s period of darkness).
e, Rank-ordered ring neuron responses to different stimulus sequences during the first stimulus (defined as the activity integrated over the period of the first stimulation) and during the dark period (defined as the activity integrated over the period beginning 1.5 s after the end of the first stimulus and ending at the beginning of the second stimulus) are reversed with respect to one another, irrespective of GECI combination.
f, Rank-ordered ring neuron responses during dark period and during second stimulus presentation (defined as the activity integrated over the period of the second stimulation) are consistent in ring neurons, irrespective of GECI combination.
g, The effect of stimulus history (see Online Methods for details) on ring neuron responses monotonically decreases as the duration of the dark period increases, irrespective of the GECI combination, while the effect on TB neuron responses fluctuates. ((TB)R _(Ring)G: 10 BUs from 8 flies; (TB)R _(Ring)G: 3 BUs from 3 flies).
Supplementary Figure 9 The properties of biphasic filters give rise to the observed suppression and history dependence in TB and ring neurons.
a, Biphasic filters consist of leading (l) and trailing (t) phases of opposite sign, separated in time. Each phase of the filter is described by an amplitude h and a latency τ.
b, Pearson’s correlation coefficient between model and data, shown for 38 TB neurons (dotted line) and 54 ring neurons (solid line).
c, Amplitudes of ipsilateral (solid markers) and contralateral (open markers) filters fit to 30 pairs of TB neurons (left) and ring neurons (right). Bold markers correspond to the filters shown in bold in Fig. 5a.
d, Peak latencies of ipsilateral (solid markers) and contralateral (open markers) filters fit to TB neurons (left) and ring neurons (right). The presentation times of bar stimuli are shown shaded in gray. Bold markers correspond to the filters shown in bold in Fig. 5a.
e, Contralateral suppression observed in the data, compared to leading amplitudes of biphasic filters (defined in (g)).
f, History dependence observed in the data, compared to trailing amplitudes of biphasic filters (defined in (g)).
g, Change in contralateral suppression (CS) and history dependence (HD) from TB to ring neurons, compared to the change in size of the contralateral filter (relative to the ipsilateral filter). See Supplementary Note for the quantification of the changes.
h, Leading amplitudes of the biphasic filter give a simple estimate of contralateral suppression, while trailing amplitudes give a simple estimate of history dependence. As in Figs. 3 and 4, we measure the extent of contralateral suppression in terms of the integrated response to an ipsilateral stimulus (AIpsi; green) versus a bilateral stimulus (ABi; purple), and we measure the extent of history dependence in terms of the integrated response to a bilateral stimulus when preceded by a contralateral stimulus (ABi|Contra; blue) versus an ipsilateral stimulus (ABi|Ipsi; green).
Supplementary Figure 10 Contralateral suppression does not depend on the location of the ipsilateral stimulus.
a, Schematic showing excitatory field mapping with a single stimulus (left, designated as “Single stimulus” below) and excitatory field mapping in the presence of a contralateral stimulus (right, designated as “With second stimulus” below).
b, Excitatory fields of paired TB and ring neurons. Note strong suppression in ring neurons but not in TB neurons. Dashed white lines indicate the cross section used in (d).
c, Summated excitatory fields of TB and ring neurons across several glomeruli in one BU (24 glomeruli segmented in this example). Note strong suppression in ring neurons but not TB neurons. Dashed white lines indicate the cross section used in (e).
d, Cross-section along the dashed white lines in (b), showing large-field suppression of the field in the presence of a contralateral stimulus.
e, Cross-section along the dashed white lines in (c), showing large-field suppression of the field in the presence of a contralateral stimulus.
f, Scatter plots of the ratio of summated ΔF/F0 over the area of the excitatory field mapped in the presence of a second stimulus versus that mapped with a single stimulus. This scatter plot shows post-synaptically enhanced contralateral suppression over different locations, irrespective of the GECI combination used. ((TB)G _(Ring)R: 9 BUs from 8 flies; (TB)R_(Ring)G: 2 BUs from 2 flies; P < 0.001, see Supplementary Table 6 for exact statistical results).
g, Cumulative distribution of the ratio shown in (f). This distribution shows post-synaptically enhanced contralateral suppression over different locations, irrespective of GECI combination. ((TB)G _(Ring)R: 9 BUs from 8 flies; (TB)R_(Ring)G: 2 BUs from 2 flies; P < 0.001, see Supplementary Table 6 for exact statistical results).
Supplementary Figure 11 Predicting responses to competing stimuli.
a, Schematic showing a simulated stimulus condition in which a left and right bar are each presented for the same fixed duration (black arrows), but separated by a variable inter-stimulus-interval (ISI; defined as positive when the left bar precedes the right bar, and vice versa). After a variable delay (defined relative to the end of the second stimulus, whether a left or right bar), a pair of left and right bars are presented simultaneously.
b, Two simulated neurons (left, upper, and right, lower) are each described by a pair of biphasic ipsilateral (black) and contralateral (gray) filters (see Supplementary Note). The overlap of each stimulus presentation with the corresponding filter (measured as the area under the ipsilateral filter swept out by each ipsilateral stimulus, and similarly the area under the contralateral filter swept out by each contralateral stimulus) determines whether its contribution to the overall response is excitatory (red) or inhibitory (blue). In this example, a left bar was presented less recently inside the left neuron’s RF relative to the presentation of a right bar inside the right neuron’s RF.
c, The response of each neuron is determined by the total integrated area swept out by all stimuli. This area has a larger positive value for the left neuron (upper) than for the right neuron (lower). This response value can be computed for various combinations of ISI and delay, as shown in the triangular phase plots on the right. The black point marks the specific combination of ISI and delay exemplified in panel (b).
d, The difference in response between the two neurons (computed by subtracting right from left) determines regions of parameter space for which the left neuron responds more strongly than the right (red regions), or the right neuron responds more strongly than the left (blue regions).
e, The difference in response shown in (d) can be used to extract a consistent rule for determining which stimulus (left or right) will be selected when both stimuli are presented simultaneously. We measure this selection by determining which side (left or right) will produce a larger response to bilateral bars, based on the timing of left and right bars that were previously presented within the RF. For small delays, a larger response is evoked on the side where the stimulus was presented more recently inside the RF (cool colors), while for large delays, a larger response is evoked on the side where the stimulus was presented less recently inside the RF (warm colors). The dotted line indicates the transition between the two regimes.
f, g, Monophasic filters produce a single regime in which responses are stronger for stimuli that have been presented more recently inside the RF (in contrast, biphasic filters produce two qualitatively different regimes). Note that on shorter timescales (brief stimuli presented with short ISI and short delays), a stronger response can be observed for stimuli that have been presented less recently inside the RF (see Supplementary Note).
Supplementary information
Supplementary Text and Figures
Supplementary Figures 1–11, Supplementary Tables 1–8 and Supplementary Note 1. (PDF 5400 kb)
Two-photon photoactivation reveals a visual pathway into the central complex.
Three-dimensional (3D) reconstruction of the ME-AOTU-BU-EB pathway revealed by two-photon photoactivation of PA-GFP, first in the BU and then in the AOTU. (MOV 1508 kb)
Two-photon photoactivation highlights TB and ring neurons.
Three-dimensional reconstruction of TB and ring neurons revealed by two-photon photoactivation of PA-GFP in the BU. (MOV 312 kb)
3D reconstruction of the brain of a transgenic fly expressing GCaMP6f and jRGECO1a in MT, TB and ring neurons.
Confocal imaged Z stack and three-dimensional reconstruction based on immunolabeling against red (anti-DsRed) and green (antiGFP) fluorescent proteins in a transgenic fly expressing GCaMP6f in MT and ring neurons and jRGECO1a in TB neurons (see Online Methods for details). The movie starts with Z stack and 3D reconstruction of one side of the brain including the AOTU and BU and connected neurites, then zooms into the BU with Z stack and 3D reconstruction based on imaging at higher resolution using a high-magnification objective. (MOV 6369 kb)
Two-color two-photon calcium imaging from TB and ring neurons in the BU.
Two-color imaging of calcium responses to visual stimulation in the BU with GCaMP6f labeling TB neurons (shown in green) and jRGECO1a labeling ring neurons (shown in magenta). (MOV 9315 kb)
Two-color two-photon calcium imaging reveals enhanced dynamic stimulus selection from TB to ring neurons.
Two-color imaging of calcium responses to small vertical bar stimuli (as in Fig. 3 and Fig. 4) in the BU in one glomerulus (contour highlighted in white), with jRGECO1a labeling TB neurons and GCaMP6f labeling ring neurons. Visual stimulation, TB axon and ring dendrite images (shown in pseudo-color) and calcium traces of segmented glomeruli are shown synchronously. (MOV 23675 kb)
Ring neuron responses are modeled by convolving a pair of biphasic temporal filters with spatiotemporal stimuli.
In this animation of Fig. 6i, j, we model the responses of a bilateral pair of ring neurons (bottom row) to a bar stimulus presented unilaterally, and then bilaterally (top row). Each neuron's response is described by an ipsilateral (ipsi) and a contralateral (contra) filter (middle row; Fig. 5a). At each time point, the overlap of past stimuli with the corresponding filters (measured as the area under the ipsilateral filter swept out by each ipsilateral stimulus, and similarly the area under the contralateral filter swept out by each contralateral stimulus) determines whether their contribution to the overall response is excitatory (red) or inhibitory (blue). The response of each neuron at each time point is determined by the total integrated area swept out by all past stimuli. (MP4 4247 kb)
Rights and permissions
About this article
Cite this article
Sun, Y., Nern, A., Franconville, R. et al. Neural signatures of dynamic stimulus selection in Drosophila. Nat Neurosci 20, 1104–1113 (2017). https://doi.org/10.1038/nn.4581
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/nn.4581
This article is cited by
-
Lineages to circuits: the developmental and evolutionary architecture of information channels into the central complex
Journal of Comparative Physiology A (2023)
-
Olfactory navigation in arthropods
Journal of Comparative Physiology A (2023)
-
Dopamine promotes head direction plasticity during orienting movements
Nature (2022)
-
A neuronal ensemble encoding adaptive choice during sensory conflict in Drosophila
Nature Communications (2021)
-
Excessive energy expenditure due to acute physical restraint disrupts Drosophila motivational feeding response
Scientific Reports (2021)
