A neural correlate of individual odor preference in Drosophila
- Organismic and Evolutionary Biology, Harvard University, United States
- Center for Brain Science, Harvard University, Cambridge, United States
- McGovern Institute, MIT, United States
- MIT Media Lab, MIT, United States
- Janelia Research Campus, Howard Hughes Medical Institute, United States
- Department of Biological Engineering, MIT, United States
- Koch Institute, Department of Biology, MIT, United States
- Howard Hughes Medical Institute, United States
- Department of Brain and Cognitive Sciences, MIT, United States
Figures
Figure 1 with 10 supplements
Idiosyncratic calcium dynamics predict individual odor preferences.
(A) Olfactory circuit schematic. Olfactory receptor neurons (ORNs, peach outline) and projection neurons (PNs, plum outline) are comprised of ~51 classes corresponding to odor receptor response …
Figure 1—figure supplement 1
Behavioral measurements and individual preference persistence.
(A) Behavioral measurement apparatus (adapted from Figure 1A of Honegger et al., 2020). (B) Odor preference persistence over 3 hours for flies given a choice between 3-octanol and air (n = 34 …
Figure 1—figure supplement 2
Average glomerulus-odor time-dependent responses.
Time-dependent responses of each glomerulus identified in our study to the 13 odors in our odor panel. Data represents the average across flies (olfactory receptor neuron [ORN], peach curves, n = 65 …
Figure 1—figure supplement 3
Individual glomerulus-odor responses.
Idiosyncratic odor coding measured in olfactory receptor neurons (ORNs) (left, 208 recordings across 65 flies) and projection neurons (PNs) (right, 406 trials across 122 flies). Each column …
Figure 1—figure supplement 4
Correspondence in calcium responses between lobes and trials.
(A) Scatter plots of max Δf/f attained over an odor presentation in a left-lobe recording vs. a right-lobe recording in the same fly (same data as presented in Figure 1—figure supplement 3). Plum …
Figure 1—figure supplement 5
Glomerulus responses and identification.
(A) Glomerulus odor responses measured in projection neurons (PNs) vs. those measured in olfactory receptor neurons (ORNs) (n = 65 flies). Points correspond to the odorants listed in Figure 1G. (B) …
Figure 1—figure supplement 6
Idiosyncrasy of olfactory receptor neuron (ORN) and projection neuron (PN) responses.
(A) Logistic regression classifier accuracy of decoding individual identity from individual odor panel peak responses. PCA was performed on population responses and the specified fraction of …
Figure 1—figure supplement 7
Calcium response correlation matrices.
Correlation between calcium response dimensions across flies measured in olfactory receptor neurons (ORNs, n = 65 flies) (top) and projection neurons (PNs, n = 122 flies) (bottom). Glomerulus-odor …
Figure 1—figure supplement 8
Calcium imaging principal component loadings.
(A, B) First 10 principal component loadings measured from calcium responses in olfactory receptor neurons (ORNs) (A, n = 65 flies) and projection neurons (PNs) (B, n = 122 flies). Loadings are …
Figure 1—figure supplement 9
Estimating latent calcium–behavior correlations.
(A) Schematic of inference approach to estimate the correlation between latent calcium (c) and behavioral (b) states (R2latent). This method can be applied identically to infer R2latent between Brp …
Figure 1—figure supplement 10
OCT-AIR preference prediction.
(A) Bootstrapped R2 of OCT-AIR preference prediction from each of the first five principal components of neural activity measured in olfactory receptor neurons (ORNs) (top, all data) or projection …
Figure 2 with 2 supplements
Variation in relative glomerular responses explains individual odor preference.
(A) PC 2 loadings of projection neuron (PN) activity for flies tested for OCT-MCH preference (n = 69 flies). (B) Interpreted PN PC 2 loadings. (C) Measured OCT-MCH preference vs. preference …
Figure 2—figure supplement 1
Measured preference vs. projection neuron (PN) activity-based predicted preference, split by training/testing set.
(A) Measured OCT-AIR preference vs. preference predicted from PC 1 of PN activity in a training set (n = 18 flies). (B) Measured OCT-AIR preference vs. preference predicted from PC 1 on PN activity …
Figure 2—figure supplement 2
Time-dependent preference- and odor-decoding.
(A) R2 of odor-vs.-air preference predicted by PC 1 of projection neuron (PN) activity as a function of time across trials (n = 53 flies). (B) R2 of odor-vs.-air preference predicted by PC 1 of …
Figure 3 with 2 supplements
Idiosyncratic presynaptic marker density in DM2 and DC2 predicts OCT-MCH preference.
(A) Experiment outline. (B) Example slice from a z-stack of the antennal lobe expressing Orco>Brp-Short (green) with DC2 and DM2 visible (white dashed outline). nc82 counterstain (magenta). (C) …
Figure 3—figure supplement 1
ORN>Brp-Short characterization and model predictions.
(A–C) Right vs. left glomerulus properties measured from z-stacks of stained Orco>Brp-Short samples from a training set of flies (n = 22): (A) Volume, (B) total Brp-Short fluorescence, and (C) …
Figure 3—figure supplement 2
Calcium and Brp-Short predictor variation.
(A) Histogram of average projection neuron (PN) Δf/f across all coding dimensions in flies in which OCT-AIR preference was measured (top) and OCT-AIR preference vs. average PN Δf/f (n = 53 flies) …
Figure 4 with 6 supplements
Simulation of olfactory circuits under developmental stochasticity.
(A) Antennal lobe (AL) modeling analysis outline. (B) Leaky-integrator dynamics of each simulated neuron. When a neuron’s voltage reaches its firing threshold, a templated action potential is …
Figure 4—figure supplement 1
Antennal lobe (AL) model raster plot.
(A) Action potential raster plot of olfactory receptor neurons (ORNs) in the baseline simulated AL. Rows are individual ORNs, black ticks indicate action potentials. Random shades of gold at left …
Figure 4—figure supplement 2
Antennal lobe (AL) model baseline outputs compared to experimental data.
(A) Distributions of model neuron firing rates by cell type across odors (transparent black points are individual neuron-odor combinations). Black lozenge symbols indicate the mean firing rate of …
Figure 4—figure supplement 3
Sensitivity analysis of aORN, aeLN, aiLN, aPN parameters.
Left, blue to red colormap: magnitude of parameter manipulation. Center, dark blue to yellow colormap: mean glomerular firing rate (Hz) responses of projection neurons (PNs) (DL1, DM1, DM2, DM3, …
Figure 4—figure supplement 4
Synapse counts vs. glomerular volume in the hemibrain and antennal lobe (AL) model.
(A) Left: scatter plot of total projection neuron (PN) input synapses within a glomerulus vs. that glomerulus’ volume from the hemibrain dataset. Solid line represents the maximum likelihood-fit …
Figure 4—figure supplement 5
Projection neuron (PN) response PCA loadings under various sources of circuit idiosyncrasy.
(A) Loadings of the principal components of PN glomerulus-odor responses as simulated across antennal lobe (AL) models where Gaussian noise with an SD equal to 0, 20, 50, and 100% of each synapse …
Figure 4—figure supplement 6
Classifiability of simulated idiosyncratic behavior under different sources of circuit idiosyncrasy.
Simulated projection neuron (PN) odor-glomerulus firing rates projected into their first three principal components. Individual points represent single runs of resampled antennal lobe (AL) models, …
Videos
Video 1
Example recording with automated tracking of an odor-vs.-air behavioral assay.
The recent positions of each fly (green line) are shown in different colors. Red bar indicates when the odor stream is turned on.
Video 2
Example recording with automated tracking of an odor-vs.-odor behavioral assay.
The recent positions of each fly (green line) are shown in different colors. Magenta and green bars at right indicate when MCH and OCT are respectively flowing into the top and bottom halves of each …
Video 3
Confocal image stack of expanded DC2>Brp-Short.
Magenta is nc82 stain, green is Or13a>Brp-Short. Frames are z-slices spaced at 0.54 µm. Image height corresponds to a post-expansion field of view of 107 × 90 µm (~2.5× linear expansion factor).
Video 4
Simulated antennal lobe (AL) connectivity matrices.
Left: glomerular density resampling. Each frame corresponds to the hemibrain connectome synaptic weights, rescaled according to a sample from the relationship between synapse count and volume …
Tables
Table 1
Calcium and Brp-Short – behavior model statistics.
| Behavior neasured | Neural predictor | Figure panel | n | β0 | β1 | R2 | p-Value |
|---|---|---|---|---|---|---|---|
| OCT vs. AIR | PN calcium PC 1 | Figure 2—figure supplement 1A | 18 | –0.26 | –0.079 | 0.16 | 0.099 |
| OCT vs. AIR | PN calcium average all dimensions | Figure 1—figure supplement 10I | 53 | –0.051 | –0.38 | 0.098 | 0.022 |
| OCT vs. AIR | ORN calcium PC 1 | Figure 1—figure supplement 10B | 30 | –0.29 | –0.053 | 0.23 | 0.007 |
| OCT vs. AIR | ORN calcium average all dimensions | Figure 1—figure supplement 10E | 30 | –0.032 | –0.71 | 0.25 | 0.005 |
| OCT vs. MCH | PN calcium PC 2 | Figure 2—figure supplement 1C | 47 | –0.058 | –0.081 | 0.15 | 0.006 |
| OCT vs. MCH | PN calcium DM2–DC2 (% difference) | Figure 2I | 69 | –0.032 | –0.0018 | 0.12 | 0.004 |
| OCT vs. MCH | ORN calcium PC 1 | Figure 1L | 35 | –0.14 | –0.027 | 0.031 | 0.32 |
| OCT vs. MCH | ORN Brp-Short PC 2 (train data only) | Figure 3—figure supplement 1I | 22 | –0.087 | 0.017 | 0.22 | 0.028 |
| OCT vs. MCH | ORN Brp-Short PC 2 (all data) | Figure 3F | 53 | –0.019 | 0.012 | 0.088 | 0.031 |
-
MCH, 4-methylcyclohexanol; OCT, 3-octanol; ORN, olfactory receptor neuron; PC, principal component; PN, projection neuron.
Table 2
Typical electrophysiology features of antennal lobe cell types, used as model parameters.
| Parameter | Olfactory receptor neurons | Local neurons | Projection neurons |
|---|---|---|---|
| Membrane resting potential | –70 mV (Dubin and Harris, 1997) | –50 mV (Seki et al., 2010) | –55 mV (Jeanne and Wilson, 2015) |
| Action potential threshold | –50 mV (Dubin and Harris, 1997) | –40 mV (Seki et al., 2010) | –40 mV (Jeanne and Wilson, 2015) |
| Action potential minimum | –70 mV (Cao et al., 2016) | –60 mV (Seki et al., 2010) | –55 mV (Jeanne and Wilson, 2015) |
| Action potential maximum | 0 mV (Dubin and Harris, 1997) | 0 mV (Seki et al., 2010) | –30 mV (Wilson and Laurent, 2005) |
| Action potential duration | 2 ms (Jeanne and Wilson, 2015) | 4 ms (Seki et al., 2010) | 2 ms (Jeanne and Wilson, 2015) |
| Membrane capacitance | 73 pF (assumed = projection neurons) | 64 pF (Huang et al., 2018) | 73 pF (Huang et al., 2018) |
| Membrane resistance | 1.8 GOhm (Dubin and Harris, 1997) | 1 GOhm (Seki et al., 2010) | 0.3 GOhm (Jeanne and Wilson, 2015) |
Key resources table
| Reagent type (species) or resource | Designation | Source or reference | Identifiers | Additional information |
|---|---|---|---|---|
| Genetic reagent (Drosophila melanogaster) | P{20XUAS-IVS-GCaMP6m}attP40 | Bloomington Drosophila Stock Center | RRID:BDSC_42748 | |
| Genetic reagent (D. melanogaster) | w[*]; P{w[+mC]=Or13a-GAL4.F}40.1 | Bloomington Drosophila Stock Center | RRID:BDSC_9945 | |
| Genetic reagent (D. melanogaster) | w[*]; P{w[+mC]=Or19a-GAL4.F}61.1 | Bloomington Drosophila Stock Center | RRID:BDSC_9947 | |
| Genetic reagent (D. melanogaster) | w[*]; P{w[+mC]=Or22a-GAL4.7.717}14.2 | Bloomington Drosophila Stock Center | RRID:BDSC_9951 | |
| Genetic reagent (D. melanogaster) | w[*]; P{w[+mC]=Orco-GAL4.W}11.17; TM2/TM6B, Tb[1] | Bloomington Drosophila Stock Center | RRID:BDSC_26818 | |
| Genetic reagent (D. melanogaster) | isokh11 isogenic line | https://doi.org/10.1073/pnas.1901623116 | Honegger et al., 2020 | |
| Genetic reagent (D. melanogaster) | GH146-Gal4 | https://doi.org/10.1073/pnas.1901623116 | Gift of Y. Zhong (Honegger et al., 2020) | |
| Genetic reagent (D. melanogaster) | w; UAS-Brp-Short-mStrawberry; UAS-mCD8-GFP; + | https://doi.org/10.7554/eLife.03726 | Gift of T. Mosca (Mosca and Luo, 2014) | |
| Antibody | Anti-nc82 (mouse monoclonal) | Developmental Studies Hybridoma Bank | DSHB:nc82; RRID:AB_2314866 | (1:40) |
| Antibody | Anti-GFP (chicken polyclonal) | Aves Labs | Aves Labs:GFP-1020; RRID:AB_10000240 | (1:1000) |
| Antibody | Anti-mStrawberry (rabbit polyclonal) | biorbyt | Biorbyt:orb256074 | (1:1000) |
| Antibody | Atto 647N-conjugated anti-mouse (goat polyclonal) | MilliporeSigma | Sigma-Aldrich:50185; RRID:AB_1137661 | (1:250) |
| Antibody | Alexa Fluor 568-conjugated anti-rabbit (goat polyclonal) | Thermo Fisher | Thermo Fisher Scientific:A-11011; RRID:AB_143157 | (1:250) |
| Antibody | Alexa Fluor 488-conjugated anti-chicken (goat polyclonal) | Thermo Fisher | Thermo Fisher Scientific:A-11039; RRID:AB_2534096 | (1:250) |
| Chemical compound, drug | 2-Heptanone | MilliporeSigma | CAS #110-43-0 | |
| Chemical compound, drug | 1-Pentanol | MilliporeSigma | CAS #71-41-0 | |
| Chemical compound, drug | 3-Octanol | MilliporeSigma | CAS #589-98-0 | |
| Chemical compound, drug | Hexyl-acetate | MilliporeSigma | CAS #142-92-7 | |
| Chemical compound, drug | 4-Methylcyclohexanol | MilliporeSigma | CAS #589-91-3 | |
| Chemical compound, drug | Pentyl acetate | MilliporeSigma | CAS #628-63-7 | |
| Chemical compound, drug | 1-Butanol | MilliporeSigma | CAS #71-36-3 | |
| Chemical compound, drug | Ethyl lactate | MilliporeSigma | CAS #97-64-3 | |
| Chemical compound, drug | Geranyl acetate | Millipore Sigma | CAS #105-87-3 | |
| Chemical compound, drug | 1-Hexanol | MilliporeSigma | CAS #111-27-34 | |
| Chemical compound, drug | Citronella java essential oil | Aura Cacia | Aura Cacia:191112 | |
| Software, algorithm | Python (version 3.6) | Python Software Foundation | RRID:SCR_008394 | |
| Software, algorithm | MathWorks, MATLAB pca documentation, 2018 | RRID:SCR_001622 |
