Description of the data and file structure

Data contained here are the individual data points plotted in the main figures of the manuscript currently under consideration at PLOS Biology (DOI to be added here upon acceptance/publication).

Data and file structure:

Individual data points on scatter are available in double-precision numeric arrays (both .mat and .csv formats).

Common abbreviations:

• MC: Mitral cells
• TC: Tufted cells
• ROI: region of interest
• sPlus: Rewarded trials (S+)
• sMin: Unrewarded trials (S-)
• ΔF/F: Normalized calcium fluorescence (see Methods)
• CaData: Calcium data (reported in ΔF/F)
• exc: excitatory
• inh: inhibitory

 

Materials and Methods

Animals

Tbx21-Cre (doi: 10.1038/nn.3407), B6J.Cg-Gt(ROSA)26Sortm95.1(CAG- GCaMP6f)Hze/MwarJ, also known as Ai95D (https://doi.org/10.1016/j.neuron.2015.02.022), and Ai14 (doi: 133-140. 10.1038/nn.2467 ) mice were originally obtained from the Jackson Laboratory (stock numbers: 024507, 028865 and 007914, respectfully). Lbhd2-CreERT2 mice were generated previously and are also available through Jackson Laboratory (stock number 036054; doi: 10.1523/JNEUROSCI.3076-20.2021 ). Tbx21-Cre::Ai95D and Lbhd2-CreERT2::Ai95D mice were generated by crossing parents homozygous for each transgene. Lbhd2-CreERT2::Ai14 mice were generated by crossing LBHD2-CreERT2 mice with Ai14 mice.

 

Surgery

All recovery surgeries were conducted in an aseptic manner. For the cranial window and headplate implantations, 9-11 week-old male mice were deeply anesthetised with isoflurane (3-5% for induction, 1-2% for maintenance; IsoFlo, Zoetis Japan). A craniotomy of approximately 1.5 x 1 mm was performed over the left olfactory bulb, and a custom cut glass window (thickness No. 1; Matsunami, Japan) was implanted. Once the window was sealed with cyanoacrylate (Histoacryl, B. Braun, Germany), a custom-made metal headplate (26 x 12 mm) was implanted posterior to the cranial window. Dental acrylic (Kulzer, Hanau, Germany) was then added to cover the exposed skull and to secure both the headplate and cranial window.

For experiments involving pharmacological infusion, an additional cannula (10 mm length; C315GS- 4/SPC, Plastics One) was inserted to target the left anterior piriform cortex (coordinate: AP -2.2 mm & ML -2.4 mm from bregma; DV -6.1 mm at 45-degree angle from brain surface).

Virus injection

To express GCaMP6f in adult-born granule cells, during cranial window and headplate implantations, Lbhd2-CreERT2::Ai14 mice were injected with AAV1-syn-GCaMP6f-WPRE-SV40 (Addgene 100837-AAV1; titre was 1.84 x 1013 GC/mL at the time of synthesis) in the left SVZ (200 μL; coordinate: AP 1.0 mm & ML -1.0 mm from bregma; DV -2.2 mm vertically) using Nanoject III (Drummond, 3-000-207).

Habituation

Male mice were habituated to head fixation on a custom-made running wheel. Thereafter, water access was restricted by removing water bottles from their home cages. The mice were habituated to receive water from the port at the experimental setup on the following two to three days until they learned to drink at least 1 ml at the setup. The body weight was recorded daily, to ensure that it stayed above 80% of the original weight. Lick responses were measured using an IR beam sensor (PM- F25, Panasonic, Osaka, Japan).

During calcium imaging sessions, respiration of the mice was recorded using a flow sensor (AWM3100V, Honeywell, NC) placed close to the right nostril.

Discrimination training

After habituation, the mice were trained to associate one odour stimulus with a water reward (S+ odour) and another odour stimulus with no reward (S- odour). Both S+ and S- odours were a binary mix of ethyl butyrate (Sigma-Aldrich; W242705) and methyl butyrate (Tokyo Chemical Industry; B0763), but mixed with different ratios based on the photoionization detector readings. Odour presentation was targeted to the left nostril since the right nostril was used to record the nasal airflow with a flow sensor. For the initial, easy, discrimination training, an 80/20 vs 20/80 ratio was used. When mice reached 80% behavioural accuracy, difficult discrimination training started, using 60/40 vs 40/60 odour mixtures. The correct response to S+ odours was to lick within a 3-second window after stimulus onset (anticipatory licks), while the correct response to S- odours was to refrain from licking. The water reward consisted of multiple drops, with a total of ~18 μL per trial. Olfactory stimuli were presented using a custom flow-dilution olfactometer (doi: Elife 8. 10.7554/eLife.43558). On each trial, odour was presented for 1 second and delay to the reward was 3 seconds from the odour onset. Inter-trial interval was approximately 20 seconds.

 

Long odour discrimination

The mice proficient at the difficult discrimination task were trained to discriminate between the same odour mixtures but with4 seconds of odour duration. The response window and reward timing on S+ trials was the same between the two paradigms.

 

Disengagement paradigm

In this paradigm, the same odour mixtures as the difficult discrimination were used, but the water reward was delivered every trial, approximately 15 seconds before the odour onset. The time window used for measuring the anticipatory licks was identical to that of the difficult discrimination paradigm. The first session was considered a transition session and excluded from imaging analysis.

 

Random association paradigm

In this paradigm, the water reward was presented in 50% of the trials, regardless of the odour identity. This decoupled the odour identity and reward, but kept mice engaged, as indicated by the anticipatory licks. The first session was considered a transition session and excluded from imaging analysis.

 

Pharmacological inactivation of anterior piriform cortex

For pharmacological inactivation of the anterior piriform cortex, muscimol (M1523, Sigma- Aldrich, MO) was infused (2 mM in Ringer; 500 nL at 100 nL/min) through the previously implanted cannula, using a Hamilton syringe (1 mL Model 7001KH PST-3 80100, Hamilton Company, Nevada USA), approximately 10 minutes before the start of the imaging sessions. Two days prior to the first muscimol infusion, Ringer solution (500 nL at 100 nl/min) was infused.

 

In vivo calcium imaging

All the calcium data presented in this manuscript were obtained from awake mice. Two-photon fluorescence of GCaMP6f and tdTomato were measured simultaneously with a custom-made microscope (INSS, UK) fitted with a 25x objective (Nikon N25X-APO-MP1300, 1.1 N.A.) or a 16x objective (Nikon N16XLWD-PF, 0.8 NA), and high-power laser (980 nm; Insight DeepSee, MaiTai HP, Spectra-Physics, USA) at depths 50–400 μm below the surface of the olfactory bulb. Images from a single plane were obtained at ~30 Hz with a resonant scanner. In each trial, 400 image frames were acquired, with 100 frames before odour stimulus to obtain a baseline. Each day, the stage coordinates were chosen relative to a reference location, which was determined by the surface blood vessel pattern. Fields of view were 512 μm x 512 μm for apical dendrites, 256 μm x 256 μm for tufted and mitral cell somata, and 128 μm x 128 μm and for adult-born granule cell gemmules. Calcium data during difficult discrimination and disengaged experiments were obtained from Tbx21-Cre::Ai95D mice (Figs. 1 and 4). 6 male mice were used for somata imaging. All 6 were used to image mitral cell somata, while a subset (3 mice) were used to image tufted cell somata. To obtain calcium data from different subcelluar compartments of mitral cells, we used Lbhd2-CreERT2::Ai95D mice. For long odour discrimination, random association, and muscimol infusion experiments, calcium data was obtained from both Tbx21-Cre::Ai95D and Lbhd2-CreERT2::Ai95D mice. Finally, Lbhd2-CreERT2::Ai14 mice were used in to record red (tdTomato, mitral cells) and green (calcium indicator, gemmules) fluorescent signals during adult-born granule cell imaging experiments . The “superficial” and “deep” levels were approximately 40-50 mm above and below from the transition in the red fluorescence density, respectively.

 

Data analysis

All data was analyzed offline using custom MATLAB (MathWorks, USA) routines. To calculate the behavioural accuracy, the number of licks during a 3 second window from final valve opening until reward presentation was counted for each trial (anticipatory licks). Correct response to the rewarded odour was a minimum of 2 anticipatory licks, and correct response to the unrewarded odour was less then 2 anticipatory licks. Behavioural accuracy was calculated as the percentage of correct trials from the total number of trials.

 

To calculate the sniffing frequency and speed of inhalation, the sniffing signal was first filtered (1 Hz high-pass and 30 Hz low-pass) and normalised (z-score). Inhalation peaks were detected using the findpeaks MATLAB function. Sniff onsets were determined by searching back in time from each detected inhalation peak to the point where the signal crossed a threshold value. The detected onsets and peaks were then used to calculate the frequency (as 1/inter-onset time) and speed of inhalation (as onset-to-peak time).

 

Image analysis

For each field of view, the imaging data was manually curated based on motion artifacts and drift over time. Data with motion artifacts and/or drift were motion corrected using the NoRMCorre toolbox ( https://doi.org/10.1016/j.jneumeth.2017.07.031) and, when unsuccessfully corrected, excluded from analysis. Regions of interest (ROIs) were manually drawn using ImageJ (NIH, Bethesda, USA) based on the average field of view from each imaging session and exported for usage in MATLAB. Average pixel value from each ROI was offset with a value from the darkest region in the frame (e.g. a blood vessel). To account for bleaching over the course of the imaging session, the mean pixel values for all trials were concatenated and detrended using the MATLAB function detrend, then reshaped back into an array (individual trials x frames) before relative fluorescence change was obtained. For each ROI, the change in fluorescence (ΔF/F) was calculated by subtracting the mean pixel value from the baseline period (1 second before odour stimulus onset) and dividing by the baseline value. Odour evoked responses were calculated as the mean fluorescence change during the odour stimulus presentation, and post-odour evoked responses as the mean fluorescence change between odour stimulus offset and reward presentation. For the ‘long odour’ and ‘random association’ experiments, the time windows to calculate the evoked responses were based on the difficult odour discrimination experiments. All visible ROIs from each field of view are included in the plots unless otherwise stated.

Lick-aligned average

Rewarded trials were analysed for each imaging session. Onsets of anticipatory licks were defined as the average time of the first two licks observed after the start of odour presentation. Rewarded trials were grouped into early vs late lick trials if the anticipatory lick onsets occurred before or after the median onset time, respectively. Within each group, calcium transients were aligned to the anticipatory lick onset time for each rewarded trial and averaged. Note that the reaction time considers only the timing of lick onsets and is distinct from the discrimination time which considers the time at which S+ vs. S- responses diverge.  

Normalised S+ vs. S- difference

For each trial, the average value of relative fluorescence change was calculated for the odour period (first 1 second after the odour onset) and the post-odour period (1-3 s after the odour onset). The normalized difference between S+ and S- response amplitudes for the odour period, as well as the post-odour period, is calculated as follows:

![A mathematical equation with black text

Description automatically generated](file:////Users/izumifukunaga/Library/Group%20Containers/UBF8T346G9.Office/TemporaryItems/msohtmlclip/clip_image002.jpg "A mathematical equation with black text

Description automatically generated")

where i denotes the trial index, n is the number of trials, x is the evoked fluorescence change in response to the rewarded odour, and y is the evoked fluorescence change in response to the unrewarded odour.

 

Figure 1:

• HeatplotData_MCTbx (for 2D colormap; aligned to start of odour; Mitral cells)\ - Trial average ΔF/F values over time for 428 ROIs; frame rate = 30 Hz (see Methods).
• HeatplotData_TC (for 2D colormap; aligned to start of odour; Tufted cells)\ - Trial average ΔF/F values over time for 150 ROIs
• MeanEvokedAmp_MCTbx (mean GCaMP6f transients; Mitral cells)\ - Mean odour and post-odour evoked ΔF/F values for 428 ROIs\ - Includes vector of significant divergence per ROI
• MeanEvokedAmp_TC (mean GCaMP6f transients; Tufted cells)\ - Mean odour and post-odour evoked ΔF/F values for 150 ROIs\ - Includes vector of significant divergence per ROI
• ScatterData_discrimination_MCtbx (Evoked calcium amplitudes for scatter plots; Mitral cells; odour = during 1 second odour presentation; post = post-odour period)\ - Mean odour and post-odour evoked ΔF/F values for 428 ROIs\ - Includes vectors with all values (all), only significant divergent values (sign), and only non-significant values (nonsign) vectors
• ScatterData_discrimination_TC (Evoked calcium amplitudes for scatter plots; Tufted cells)\ - Mean odour and post-odour evoked ΔF/F values for 150 ROIs\ - Includes vectors with all values (all), only significant divergent values (sign), and only non-significant values (nonsign) vectors

Figure 2:

• HeatplotData_MCTbx_LickAligned (for 2D colormap; aligned to lick onsets; Mitral cells)\ - Trial average ΔF/F values over time for 428 ROIs

Figure 3:

• HeatplotData_MC_LongOdour (for 2D colormap; aligned to odour onset; Mitral cells)\ - Trial average ΔF/F values over time for 210 ROIs
• ScatterData_longOdour (Evoked calcium amplitudes for scatter plots for the long odour experiments; Mitral cells)\ - Mean evoked ΔF/F values for 210 ROIs during the early (odour) and late (post) phase of stimulus presentation\ - Includes vectors with all values (all), only significant divergent values (sign; dark datapoints in figure), and only non-significant values (nonsign; light datapoints in figure) vectors

Figure 4:

• HeatplotData_MC_Disengaged (for 2D colormap; Mitral cells during disengagement paradigm)\ - Trial average ΔF/F values over time for 147 ROIs
• HeatplotData_MC_RandomAssoc (for 2D colormap; Mitral cells during random association paradigm)\ - Trial average ΔF/F values over time for 301 ROIs
• MeanEvokedAmp_disengaged_MC (mean GCaMP6f transients for Mitral cells during disengagement paradigm)\ - Mean odour and post-odour evoked ΔF/F values for 147 ROIs\ - Includes vector of significant divergence per ROI
• MeanEvokedAmp_randomassoc_MC (mean GCaMP6f transients for Mitral cells during random association paradigm)\ - Mean odour and post-odour evoked ΔF/F values for 301 ROIs\ - Includes vector of significant divergence per ROI
• ScatterData_disengaged (Evoked calcium amplitudes for scatter plots; disengagement paradigm; Mitral cells)\ - Mean evoked ΔF/F values for 147 ROIs during the early (odour) and late (post) phase of stimulus presentation\ - Includes vectors with all values (all), only significant divergent values (sign; dark datapoints in figure), and only non-significant values (nonsign; light datapoints in figure) vectors
• ScatterData_randomassoc (Evoked calcium amplitudes for scatter plots; random association paradigm; Mitral cells)\ - Mean evoked ΔF/F values for 301 ROIs during the early (odour) and late (post) phase of stimulus presentation\ - Includes vectors with all values (all), only significant divergent values (sign; dark datapoints in figure), and only non-significant values (nonsign; light datapoints in figure) vectors
• AnticipatoryLicks (Anticipatory licks for individual sessions of Discrimination, Random Association, and Disengagement paradigms).\ - Mean number of anticipatory licks (see Methods) per session\ - Discr: Discrimination paradigm; Diseng: Disengagement paradigm; Rand: Random association paradigm

Figure 5:

• HeatplotData_ROImatched_MC_Control (for 2D colormap; Mitral cells during control sessions; ROIs matched with corresponding muscimol sessions)\ - Trial average ΔF/F values over time for 123 ROIs
• HeatplotData_ROImatched_MC_Muscimol (for 2D colormap; Mitral cells during muscimol sessions)\ - Trial average ΔF/F values over time for 123 ROIs
• MeanEvokedAmp_ROImatched_MC_Control (Mean GCaMP6f transients; control sessions)\ - Mean odour and post-odour evoked ΔF/F values for 123 ROIs\ - Includes vector of significant divergence per ROI
• MeanEvokedAmp_ROImatched_MC_Muscimol (Mean GCaMP6f transients; muscimol sessions)\ - Mean odour and post-odour evoked ΔF/F values for 123 ROIs\ - Includes vector of significant divergence per ROI
• NormalizedDiff_ROImatched_MC_Control (Normalised difference in rewarded vs. unrewarded odours during control sessions; see Methods)\ - Difference between odour and post-odour evoked ΔF/F values in response to rewarded and unrewarded odours for 123 ROIs
• NormalizedDiff_ROImatched_MC_Muscimol (Normalised difference in S+ vs. S- odours during muscimol sessions; see Methods)\ - Difference between odour and post-odour evoked ΔF/F values in response to rewarded and unrewarded odours for 123 ROIs
• BehaviouralAccuracy (Behavioural accuracy during control vs. muscimol sessions).\ - Mean behavioural accuracy in percentage for all sessions

Figure 6

• HeatplotData_ApicalDendriteLBHD2 (for 2D colormap; Mitral cell apical dendrites imaged in the glomerular layer using Lbhd2-CreERT2::Ai95D mice)\ - Trial average ΔF/F values over time for 140 ROIs
• HeatplotData_MC_LBHD2 (for 2D colormap; Mitral cell somata from the Lbhd2-CreERT2::Ai95D mice)\ - Trial average ΔF/F values over time for 321 ROIs

Figure 8:

• HeatplotData_abGC_deep (for 2D colormap; deep adult-born granule cell gemmules)\ - Trial average ΔF/F values over time for 43 ROIs
• HeatplotData_abGC_superficial (for 2D colormap; superficial adult-born granule cell gemmules)\ - Trial average ΔF/F values over time for 28 ROIs
• meanEvokedAmp_abGC_deep (mean GCaMP6f transients from deep adult-born granule cell gemmules)\ - Mean odour and post-odour evoked ΔF/F values for 43 ROIs\ - Includes vector of significant divergence per ROI
• meanEvokedAmp_abGC_superficial (mean GCaMP6f transients from superficial adult-born granule cell gemmules)\ - Mean odour and post-odour evoked ΔF/F values for 28 ROIs\ - Includes vector of significant divergence per ROI
• NormlizedDiff_abGC_deep (Normalised difference in rewarded vs. unrewarded Evoked calcium amplitudes for deep adult-born granule cell gemmules; see Methods)\ - Difference between odour and post-odour evoked ΔF/F values in response to rewarded and unrewarded odours
• NormlizedDiff_abGC_superficial (Normalised difference in rewarded vs. unrewarded Evoked calcium amplitudes for superficial adult-born granule cell gemmules; see Methods)\ - Difference between odour and post-odour evoked ΔF/F values in response to rewarded and unrewarded odours

Supplementary Figure 1(figureSupp1)

• DiscriminationTime (Time in ms at which cumulative rewarded vs. unrewarded lick counts diverge significantly during discrimination paradigm)\ - Histogram data for discrimination time in ms

Supplementary Figure 2(figureSupp2)

• SniffingFrequency (Reciprocal of sniff intervals in Hz around the time of odour presentation)\ - Session average sniffing frequency in Hz.
• SniffingFrequencyCumHist (Cumulative histogram of all instantaneous sniff frequency during odour/post-odour period from individual sessions)\ - Cumulative histogram data for sniffing frequency in Hz
• SniffingVelocity (Time to peak in ms from inhalation onset to peak)\ - Session average sniffing velocity in ms.
• SniffingVelocityCumHist (Cumulative histogram of sniffing velocity for individual sessions)\ - Cumulative histogram data for sniffing velocity in ms

Supplementary Figure 3 (figureSupp3)

• SignificantDivergence_perFOV_MC (Frequencies of observing significant excitatory and inhibitory responses for Mitral cells)\ - Percentage of significantly divergent ROIs per field of view (FOV) during odour and post-odour periods; 17 FOVs
• SignificantDivergence_perFOV_TC (Frequencies of observing significant excitatory and inhibitory responses for Tufted cells)\ - Percentage of significantly divergent ROIs per field of view (FOV) during odour and post-odour periods; 6 FOVs

Supplementary Figure 5 (figureSupp5)

• NormalizedDiff_longOdour (Normalised difference in rewarded vs. unrewarded responses during discrimination paradigm with long odour; see Methods)\ - Cumulative histogram data for normalized difference data in ΔF/F
• NormalizedDiff_shortOdour (Normalised difference in S+ vs. S- responses during standard discrimination paradigm; see Methods)\ - Cumulative histogram data for normalized difference data in ΔF/F

Supplementary Figure 6 (figureSupp6)

• ScatterData_easyDiscrimination (Evoked calcium amplitudes for scatter plots; easy discrimination paradigm)\ - Mean odour and post-odour evoked ΔF/F values for 141 ROIs\ - Includes vectors with all values (all), only significant divergent values (sign; dark datapoints in figure), and only non-significant values (nonsign; light datapoints in figure) vectors

Supplementary Figure 7 (figureSupp7)

• ScatterData_MCdendrites_Discrimination (Evoked calcium amplitudes for scatter plots for individual apical dendrite ROIs during standard discrimination paradigm)\ - Mean odour and post-odour evoked ΔF/F values for 48 ROIs\ - Includes vectors with all values (all), only significant divergent values (sign; dark datapoints in figure), and only non-significant values (nonsign; light datapoints in figure) vectors