Decoding olfactory bulb output: A Behavioural assessment of rate, synchrony, and respiratory phase coding

Fu, Xiaochen1; Reinert, Janine1; Reuschenbach, Josefine1; Mago, Adam1; Lindeman, Sander1; Fukunaga, Izumi 1

Published Jun 10, 2025 on Dryad. https://doi.org/10.5061/dryad.3tx95x6t8

Data files

Jun 10, 2025 version files 33.67 MB

Figure1.zip

33.57 MB
Figure2.zip

35.20 KB
Figure3.zip

12.16 KB
Figure4.zip

15.14 KB
Figure5.zip

5.62 KB
Figure6.zip

11.49 KB
Figure7.zip

9.02 KB
README.md

13.84 KB

Abstract

The olfactory system is a well-known model for studying the temporal encoding of sensory stimuli due to its rhythmic stimulus delivery through respiration. Sniff-locked activity is pervasive in the primary olfactory area, the olfactory bulb, and is considered critical to structuring the output of its computation. We tested the behavioural importance of these temporal features using simple closed-loop optogenetics embedded in custom behavioural paradigms. We found that mice perceive differences in evoked spike counts and discriminate between synchronous vs. asynchronous activations of the output neurons. Surprisingly, they failed to distinguish the timing of evoked activity relative to the sniff cycle. These results suggest that, beyond the first steps of olfactory processing, sniff rhythms play a more nuanced role, with greater reliance on the spike rate and synchrony for the neural encoding of the environment, consistent with gradual transformation of encoding format at successive stages of sensory processing.

Dataset DOI: 10.5061/dryad.3tx95x6t8

Description of the data and file structure

Data contained here are the individual data points plotted in the main figures of the manuscript “Decoding Olfactory Bulb Output: A Behavioural Assessment of Rate, Synchrony, and Respiratory Phase Coding” to appear in the journal iScience. The data are organized in the same structure as the figures.

Data and file structure:

Individual data points are available in double-precision numeric arrays in the .txt format, except for the electrophysiology (Neuropixels) data, which uses a Matlab format.

Common abbreviations:

auROC: area under receiver operator curve
LOT: lateral olfactory tract

Files and variables

File: Figure1.zip

“Fig1_NeuropixelData_Mouse1.mat”: Matlab data that contains (i) spiketimes in a structure “units.spiketimes” in sedonds; (ii) onset times for 8 odours used stored in a structure “odours(x).Onset”, in seconds, where x is the xth odour used; (iii) onset times for inhalation (“InhalationOnset”), in seconds; (iv) times of water droplets in seconds (“Water”) where, on a rewarded trials, 2 droplets of water were dispensed; (v) times of licks in seconds (“Licks”).
“Fig1_NeuropixelData_Mouse2.mat”: as above but for a different mouse.
“Fig1_NeuropixelData_Mouse3.mat”: as above but for a different mouse.
“Fig1_NeuropixelData_Mouse4.mat”: as above but for a different mouse.
“Fig1_NeuropixelData_Mouse5.mat”: as above but for a different mouse.
“Fig1C_EB”: Processed data in .txt; 2D colormap showing the firing rate relative to baseline in response to ethyl butyrate (EB) presentation. The values are in Hz.
“Fig1C_Eug”: 2D colormap showing the firing rate relative to baseline in response to eugenol (Eug) presentation. The values are in Hz.
“Fig1C_MS”: 2D colormap showing the firing rate relative to baseline in response to methyl salicylate (MS) presentation. The values are in Hz.
“Fig1E_DecoderAccuracy_Rate_Observed”: accuracy from support vector machine trained to discriminate the identities of odours presented where the input data was the firing rates of recorded units.
“Fig1E_DecoderAccuracy_Rate_Shuffle”: accuracy from support vector machine trained to discriminate the identities of odours presented where the input data was the firing rates of recorded units where the trial order was randomly permutated.
“Fig1E_DecoderAccuracy_Timing_Observed”: accuracy from support vector machine trained to discriminate the identities of odours presented. The input data was the latency to 2 action potential generations after the inhalation onset for recorded units.
“Fig1E_DecoderAccuracy_Timing_Shuffle”: accuracy from support vector machine trained to discriminate the identities of odours presented. The input data was the latency to 2 action potential generations after the inhalation onset for recorded units but with the trial order randomly permutated.
“Fig1E_Window_size”: window size used to sample the input for the decoder.

File: Figure2.zip

“Fig2D_SpikeProbability_1pulse”: 2D colormap corresponding to the probability of spikes with respect to the onset of one light pulse. First dimension = units, second dimension = time.
“Fig2D_SpikeProbability_3pulses”: 2D colormap corresponding the probability of spikes with respect to the onset of the first pulse in a train of 3 light pulses.
“Fig2D_time_axis”: time axis in ms for the above two plots.
“Fig2E_EvokedSpikes_pulse1”: number of spikes evoked following the first light pulse in a train of 3 light pulses.
“Fig2E_EvokedSpikes_pulse2”: number of spikes evoked following the second light pulse in a train of 3 light pulses.
“Fig2E_EvokedSpikes_pulse3”: number of spikes evoked following the third light pulse in a train of 3 light pulses.

File: Figure3.zip

“Fig3D_Accuracy_5pulse_Detection”: Accuracy of detecting a 5-pulse optogenetic stimulation, expressed as auROC over time.
“Fig3D_Trials_5pulse_Detection: trial numbers for the accuracy plot for the 5-pulse detection task.
“Fig3D_Accuracy_Odour1_Discrimination”: Accuracy of discriminating between two odours prior to the 5-pulse detection task, expressed as auROC over time.
“Fig3D_Trials_odour1: trial numbers for the accuracy plot for odour discrimination.
“Fig3E_Accuracy_3pulse_Detection”: Accuracy of detecting a 3-pulse optogenetic stimulation, expressed as auROC over time.
“Fig3E_Trials_3pulse_Detection: trial numbers for the accuracy plot for the 3-pulse detection task.
“Fig3E_Accuracy_Odour1_Discrimination”: Accuracy of discriminating between two odours prior to the 3-pulse detection task, expressed as auROC over time.
“Fig3E_Trials_odour1: trial numbers for the accuracy plot for odour discrimination.
“Fig3F_Accuracy_1pulse_Detection”: Accuracy of detecting a 1-pulse optogenetic stimulation, expressed as auROC over time.
“Fig3F_Trials_1pulse_Detection: trial numbers for the accuracy plot for the 1-pulse detection task.
“Fig3F_Accuracy_Odour1_Discrimination”: Accuracy of discriminating between two odours prior to the 1-pulse detection task, expressed as auROC over time.
“Fig3F_Trials_odour1: trial numbers for the accuracy plot for odour discrimination.
“Fig3G_Trial2criterion_1pulseDetection”: number of trials taken to reach the accuracy level of 0.8 (auROC) for the 1-pulse detection task.
“Fig3G_Trial2criterion_3pulseDetection”: number of trials taken to reach the accuracy level of 0.8 (auROC) for the 3-pulse detection task.
“Fig3G_Trial2criterion_5pulseDetection”: number of trials taken to reach the accuracy level of 0.8 (auROC) for the 5-pulse detection task.

File: Figure4.zip

“Fig4C_2vs1_Accuracy”: Accuracy of discriminating between 2 pulses vs. 1 pulse of optogenetic stimulation over time, expressed as auROC.
“Fig4C_2vs1_Blocks”: time axis for the above plot, in blocks of trials, where 1 block corresponds to 100 trials.
“Fig4C_3vs1_Accuracy”: Accuracy of discriminating between 3 pulses vs. 1 pulse of optogenetic stimulation over time, expressed as auROC.
“Fig4C_3vs1_Blocks”: time axis for the above plot, in blocks of trials, where 1 block corresponds to 100 trials.
“Fig4C_4vs1_Accuracy”: Accuracy of discriminating between 4 pulses vs. 1 pulse of optogenetic stimulation over time, expressed as auROC.
“Fig4C_4vs1_Blocks”: time axis for the above plot, in blocks of trials, where 1 block corresponds to 100 trials.
“Fig4C_5vs1_Accuracy”: Accuracy of discriminating between 5 pulses vs. 1 pulse of optogenetic stimulation over time, expressed as auROC.
“Fig4C_5vs1_Blocks”: time axis for the above plot, in blocks of trials, where 1 block corresponds to 100 trials.
“Fig4D_2vs1_OverallAccuracy”: Overall accuracy of discriminating between 2 pulses vs. 1 pulse of optogenetic stimulation across all days.
“Fig4D_3vs1_OverallAccuracy”: Overall accuracy of discriminating between 3 pulses vs. 1 pulse of optogenetic stimulation across all days.
“Fig4D_4vs1_OverallAccuracy”: Overall accuracy of discriminating between 4 pulses vs. 1 pulse of optogenetic stimulation across all days.
“Fig4D_5vs1_OverallAccuracy”: Overall accuracy of discriminating between 5 pulses vs. 1 pulse of optogenetic stimulation across all days.
“Fig4F_2vs1_Accuracy_VariableLatencies”: Accuracy of discriminating between 2 pulses vs. 1 pulse of optogenetic stimulation where the onset of the first pulse occurred at variable latencies relative to the inhalation onset (20, 50, 80 ms).
“Fig4F_2vs1_Latencies”: Latencies in ms relative to the inhalation onset for the above data.
“Fig4F_3vs1_Accuracy_VariableLatencies”: Accuracy of discriminating between 3 pulses vs. 1 pulse of optogenetic stimulation where the onset of the first pulse occurred at variable latencies relative to the inhalation onset (20, 50, 80 ms).
“Fig4F_3vs1_Latencies”: Latencies in ms relative to the inhalation onset for the above data.
“Fig4F_4vs1_Accuracy_VariableLatencies”: Accuracy of discriminating between 4 pulses vs. 1 pulse of optogenetic stimulation where the onset of the first pulse occurred at variable latencies relative to the inhalation onset (20, 50, 80 ms).
“Fig4F_4vs1_Latencies”: Latencies in ms relative to the inhalation onset for the above data.
“Fig4F_5vs1_Accuracy_VariableLatencies”: Accuracy of discriminating between 5 pulses vs. 1 pulse of optogenetic stimulation where the onset of the first pulse occurred at variable latencies relative to the inhalation onset (20, 50, 80 ms).
“Fig4F_5vs1_Latencies”: Latencies in ms relative to the inhalation onset for the above data.

File: Figure5.zip

“Fig5D_Accuracy_20Vs120ms”: Accuracy of discriminating optogenetic stimulations with latencies of 20 ms vs. 120 ms relative to the inhalation onset. The accuracies are expressed in auROC.
“Fig5D_Trials_20Vs120ms”: time axis for the above data, in trials.
“Fig5D_Accuracy_odour”: Accuracy of discriminating between two odours. The accuracies are expressed in auROC.
“Fig5E_LastSessionAccuracy_20msVs80ms”: Accuracy of discriminating optogenetic stimulations with latencies of 20 ms vs. 80 ms relative to the inhalation onset for the last session of training.
“Fig5E_LastSessionAccuracy_20msVs120ms”: Accuracy of discriminating optogenetic stimulations with latencies of 20 ms vs. 120 ms relative to the inhalation onset for the last session of training.
“Fig5E_LastSessionAccuracy_80msVs130ms”: Accuracy of discriminating optogenetic stimulations with latencies of 80 ms vs. 20 ms relative to the inhalation onset for the last session of training.
“Fig5E_LastSessionAccuracy_Odour”: Accuracy of discriminating between two simple odours for the last session of training.

File: Figure6.zip

“Fig6C_ThresholdTime_Ramp”: The time when the PSTH crossed a threshold height for varying threshold levels, in response to a ramp profile light stimulus.
“Fig6C_ThresholdTime_Rectangular”: The time when the PSTH crossed a threshold height for varying threshold levels, in response to a rectangular profile light stimulus.
“Fig6C_ThresholdValues”: Threshold values for the PSTH height used in the above plots.
“Fig6E_DetectionAccuracy_ramp”: Accuracy of detecting the ramp light stimulus over trials, expressed as auROC.
“Fig6E_trials_ramp”: time axis for the above plot, in trials.
“Fig6E_DetectionAccuracy_rectangular”: Accuracy of detecting the ramp light stimulus over trials, expressed as auROC.
“Fig6E_trials_rectangular”: time axis for the above plot, in trials.
“Fig 6F_DetectionAccuracy_mouse1”: psychometric curve for detecting ramp light stimulus at different amplitude for mouse1.
“Fig6F_LightIntensities_mouse1”: Light levels, in mW, for the above plot.
“Fig 6F_DetectionAccuracy_mouse2”: psychometric curve for detecting ramp light stimulus at different amplitude for mouse2.
“Fig6F_LightIntensities_mouse2”: Light levels, in mW, for the above plot.
“Fig 6F_DetectionAccuracy_mouse3”: psychometric curve for detecting ramp light stimulus at different amplitude for mouse3.
“Fig6F_LightIntensities_mouse3”: Light levels, in mW, for the above plot.
“Fig6H_DiscriminationAccuracy”: Accuracy of discriminating ramp light stimuli delivered at 20 ms relative to the inhalation onset vs. 80 ms.
“Fig6H_trials”: time axis for the above data, in trials.

File: Figure7.zip

“Fig7D_DiscriminationAccuracy_250ms”: Accuracy in discriminating synchronous vs. asynchronous activations task, expressed as auROC, where the interval between the staggered pulses was 250 ms.
“Fig7D_trials_250ms”: time axis for the above data, in trials.
“Fig7D_DiscriminationAccuracy_500ms”: Accuracy in discriminating synchronous vs. asynchronous activations task, expressed as auROC, where the interval between the staggered pulses was 500 ms.
“Fig7D_trials_500ms”: time axis for the above data, in trials.
“Fig7F_DiscriminationAccuracy_FixedLatency1”: Accuracy in discriminating synchronous vs. asynchronous activations task for different intervals (delta_t), where the onset of stimulus delivery was fixed relative to the inhalation onset.
“Fig7F_DiscriminationAccuracy_RandomLatency”: Accuracy in discriminating synchronous vs. asynchronous activations task for different intervals (delta_t), where the onset of stimulus delivery was random.
“Fig7F_DiscriminationAccuracy_FixedLatency2”: Accuracy in discriminating synchronous vs. asynchronous activations task for different intervals (delta_t), where the onset of stimulus delivery was returned to the fixed condition relative to the inhalation onset.
“Fig7F_delta_t”: Intervals (delta_t) for the above data.
“Fig7G_OverallAccuracy_FixedLatency1”: Accuracy in discriminating synchronous vs. asynchronous activations task for all intervals combined, where the onset of stimulus delivery was fixed relative to the inhalation onset.
“Fig7G_OverallAccuracy_RandomLatency”: Accuracy in discriminating synchronous vs. asynchronous activations task for all intervals combined, where the onset of stimulus delivery was random.
“Fig7G_OverallAccuracy_FixedLatency2”: Accuracy in discriminating synchronous vs. asynchronous activations task for all intervals combined, where the onset of stimulus delivery was returned to the fixed condition relative to the inhalation onset.