Miniscope Recording Calcium Signals at Hippocampus of Mice Navigating an Odor Plume

Summary

This protocol investigates the brain-behavior relationship in hippocampal CA1 in mice navigating an odor plume. We provide a step-by-step protocol, including surgery to access imaging of the hippocampus, behavioral training, miniscope GCaMP6f recording and processing of the brain, and behavioral data to decode the mouse position from ROI neural activity.

Abstract

Mice navigate an odor plume with a complex spatiotemporal structure in the dark to find the source of odorants. This article describes a protocol to monitor behavior and record Ca²⁺ transients in dorsal CA1 stratum pyramidale neurons in the hippocampus (dCA1) in mice navigating an odor plume in a 50 cm x 50 cm x 25 cm odor arena. An epifluorescence miniscope focused through a gradient-index (GRIN) lens imaged Ca²⁺ transients in dCA1 neurons expressing the calcium sensor GCaMP6f in Thy1-GCaMP6f mice. The paper describes the behavioral protocol to train the mice to perform this odor plume navigation task in an automated odor arena. The methods include a step-by-step procedure for the surgery for GRIN lens implantation and baseplate placement for imaging GCaMP6f in CA1. The article provides information on real-time tracking of the mouse position to automate the start of the trials and delivery of a water reward. In addition, the protocol includes information on using an interface board to synchronize metadata describing the automation of the odor navigation task and frame times for the miniscope and a digital camera tracking mouse position. Moreover, the methods delineate the pipeline used to process GCaMP6f fluorescence movies by motion correction using the NorMCorre algorithm followed by identification of regions of interest with EXTRACT. Finally, the paper describes an artificial neural network approach to decode spatial paths from CA1 neural ensemble activity to predict mouse navigation of the odor plume.

Introduction

Although significant progress has been made in understanding neural circuits involved in olfactory navigation in head-fixed mice^1,2,3 and navigation strategies in freely moving mice^4,5,6,7,8, the role of neural circuits in ethologically relevant freely moving navigation of turbulent odor plumes is still unknown. This article describes monitoring neural activity by imaging Ca²⁺ transients in cells expressing the genetically encoded calcium sensor GCaMP6f in Thy1-GCaMP6f mice⁹ to study whether sequential neural dynamics of dorsal CA1 stratum pyramidale neurons in the hippocampus (dCA1) plays a role in odorant plume navigation. The methods provide information on imaged GCaMP6f fluorescence through a miniature epifluorescence microscope focused through a GRIN lens on dCA1^10,11,12. The methods explain how to monitor simultaneously spatial navigation and dCA1 neuron GCaMP6f calcium transients in mice performing an odor-plume navigation task where they received a water reward when they reached the spout delivering an odorant into an odor arena with a background laminar air flow^13,14. This article describes the methods required to achieve this task (Figure 1), including the stereotaxic surgery for the implantation of gradient-index (GRIN) lenses, the placement of a baseplate to secure the miniscope to the skull in a freely moving mouse, imaging with the miniature microscope and monitoring mouse movement with a high-speed digital camera, data preprocessing for removing motion artifacts and finding the regions of interest (ROIs), and preparation of datasets and artificial neural network training and prediction for decoding the X and Y positions of the mouse in the odor arena from changes in fluorescence in dCA1 ROIs⁷.

Miniscope recording of calcium signals in the CA1 region of the hippocampus of mice navigating an odor plume is relevant for understanding the computation of neural circuits involved with olfaction and spatial information in the complex behavior task of odor-plume navigation^2,14,15,16. The CA1 region of the hippocampus plays a role in spatial navigation and is crucial for creating a cognitive map of the environment for efficient navigation^17,18. Recording calcium signals with a miniscope is a valuable way to investigate the CA1 neurons that encode spatial information during odor plume navigation.

This technique combines the advantages of miniscope technology for recording GCaMP calcium signals with the well-established role of the CA1 hippocampus in spatial navigation to understand better how neural circuits drive complex behaviors¹⁹. Alternatively, approaches using 2-photon microscopy can record CA1 neurons^9,20, which requires a head-fixed mouse and restrains the possibility of freely moving to navigate an odor plume²¹. Local-field electrophysiological recordings of CA1 neurons allow the investigation of freely-moving mice navigating odor plumes²². Still, local field electrical signals impose limitations to estimating intracellular firing by isolating single-unit signals through spike sorting techniques. Miniscope signals allow the identification of ROIs associated directly with intracellular calcium signals in a reliable way^10,11 to investigate neural computations at single-cell resolution precisely. Miniscope technology provides a unique opportunity to better understand how the CA1 region encodes spatial information based on odor cues.

Furthermore, this technique investigates how specific neuronal populations process odor information for navigation and the relationship between neuronal activity patterns and decision-making during odor plume tracking. This method can contribute to a better understanding of how the brain processes odor and spatial information. While miniscopes offer a single-cell resolution for recording a freely moving mouse's brain, they require specialized surgery and data analysis expertise. In this paper, we provide a comprehensive protocol for helping researchers go through each step to investigate the neural mechanisms of odor-plume navigation.

The odor navigating task is a promising framework for studying neural coding and spatial odor cue memory in mice. The article's findings indicate that it is possible to decode the trajectory of the mouse navigating an odor plume based on neuronal ensemble calcium signals in dCA1. Understanding the role of dCA1 calcium signals in odor plume navigation is a crucial step to crack the neural circuit basis for odor-guided navigation in realistic environments^13,14.

Protocol

Studies were carried out in 3-6-month-old male and femaleThy1-GCaMP6f transgenic mice²³. All experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Colorado Anschutz Medical Campus in accordance with National Institutes of Health guidelines. The surgical procedures for GRIN lens implantation (Section 1) and the baseplate placement (Section 2) were adapted from previous works^{9,24,25,26,27,28,29}.

1. Stereotaxic surgery for implanting a GRIN lens implantation in the hippocampus

1. Wear sterile gloves, a headcover, a surgical mask, and a lab coat.
2. Use sterile conditions during all survival surgeries. Sterilize all surgical equipment by autoclaving.
3. Place sterile drapes around the surgical area.
4. Anesthetize the mouse by placing it in the induction chamber with 3% isoflurane for 10 min. Inject Buprenorphine Base in Polymer at 0.001 mg/g subcutaneously at the time of anesthetic induction.
5. Check whether the mouse stops moving and is under deep anesthesia, confirmed by pinching the hind paw to test the hind paw reflex.
6. Switch the isoflurane flow to the nose cone (Supplementary File 1).
7. Make sure the low-flow isoflurane anesthesia system for mice (Table of Materials) has the temperature pad set to homeothermic settings, and make sure the vapor source is set to externally compressed.
8. Place the anesthetized mouse in the digital stereotaxic instrument (Table of Materials). Place the front teeth of the mouse into the bite bar and place the nose cone in front of the nose to ensure the isoflurane flow and to lock the head into place. Adjust the level of anesthesia to between 2%-3% according to the paw reflex after pinching (Figure 2A).
9. Place the mouse ear bars into the ears and tighten them. Ensure the mouse's head is horizontal and does not move side-to-side or up and down. When the head is secure, ensure that lightly pressing downward on the skull does not cause the skull to slip out underneath the headbars.
10. Place the temperature probe in the rectum and tape the wire on the mouse's tail to avoid movement. Set the temperature of the pad to maintain the rectal temperature at 37 °C.
11. Add ophthalmic ointment to the eyes to prevent harmful air drying.
12. Shave the hair when the head is fixed and under the isoflurane flow provided by the nose cone to prevent movement. Shave the hair above the head with an electric shaver, then clear any remaining hairs off with hair remover and sterile cotton-tipped wood applicators (Figure 2B).
13. Make the operating field sterile by swabbing the scalp with a sterile cotton-tipped wood applicator with ethanol and betadine three times each.
14. Inject 0.1 mL of the local anesthetic lidocaine subcutaneously (S.Q.) under the skin between the eyes and ears with a 29 G needle. This will make a bubble. Wait for a few minutes before cutting the skin with scissors.
15. Using a tips-only technique, pull the skin up carefully using forceps and use a pair of small scissors to remove a circular portion of the skin. Absorb any blood with a sterile cotton swab.
16. Once the bleeding stops, clean the skull with hydrogen peroxide (5%) using a sterile cotton swab. Ensure that bone landmarks, including the sagittal suture and the bregmatic suture intersection to mark the bregma, are easy to identify after this procedure.
17. Confirm that the skull is flat relative to the stereotaxic apparatus by making sure that bregma and lambda sutures are exactly at the same Z coordinate. Perform this by placing a needle in the manipulator's arm and verifying if the tip of a needle is at the same Z coordinate when touching both sutures. Otherwise, adjust the angle of the head by moving the bite bar in the Z plate and repeating the procedure until the skill is in a flat position.
18. Attach a pipette to the micromanipulator holding a needle to apply tattoo ink. Make a mark on the top of bregma.
19. Zero the micromanipulator coordinates on bregma and move to the coordinateabove CA1 in the right hippocampus (medium-lateral +1.8 mm from bregma, anterior-posterior -2.4 mm from bregma)². This is the location chosen by Radvansky and Dombeck to study virtual odor gradient navigation in mice^18,30. Carefully drill a small permanent dent of 1 mm diameter with a drill (Table of Materials) at 10000 rotaions to make a permanent dent on the top of the target location (Figure 2C).
20. Use the dental drill to open a circular perforation of 1.5 mm diameter to allow the implantation of a 1 mm diameter and 4 mm length GRIN lens into the Brain (Table of Materials) (Figure 2D).
21. Puncture the dura and cortex by inserting a 23 G needle in the middle of the hole and moving the needle into the brain slowly (~0.1 mm/min) until it reaches the depth of 1.25 mm below the dura. Remove the needle slowly from the brain (~0.1 mm/min). Use sterile cotton swabs and saline to clean the blood.
22. Use a custom-made UCLA's GRIN lens holder constructed with two 1 mL micropipette tips cut to fit inside each other (GRIN lens holder, Table of Materials) (Figure 2E, F).
    1. Connect the GRIN lens holder to the micromanipulator and turn on the aspirator connected to the pipette to hold the GRIN lens. The air pressure produced by air suction holds the GRIN lens in place.
    2. Use a 4 mm length, 1 mm diameter GRIN lens. Implant the GRIN lens slowly into the cortex until it reaches a depth of -1.25 mm below the dura.
23. Place a drop of liquid tissue adhesive (Table of Materials) into the circular cranium perforation to seal the hole. Wait a few minutes to dry.
24. Prepare quick adhesive cement -biocompatible methacrylate resin- (Table of Materials) on the base of the GRIN lens to permanently seal it on the skull. Wait a few minutes and let it dry.
25. Turn off the aspirator to release and slowly remove the GRIN lens holder from the top of the GRIN lens that at this point is permanently attached to the skull.
26. Place the head bar on the top of the skull, centering the hole on the GRIN lens (Figure 2G-I) (Supplementary File 2 and Supplementary File 3).
27. Carefully place quick-drying adhesive cement (Table of Materials) around the base of the GRIN lens to seal the cranial window and attach the head bar to the skull. Place drying adhesive cement (Table of Materials) in the middle of the head bar and on the sides to keep it tightly fixed on the skull. Make sure it is completely dry. Ensure that no adhesive cement is placed on top of the GRIN lens; otherwise, it will permanently degrade the optical path and block the image.
28. Place a drop of low-toxicity silicone adhesive (Table of Materials) on top of the GRIN lens for protection from physical damage.
29. Turn off the Isoflurane
30. Take immediate post-operative care monitoring until the animal wakes from anesthesia.
    1. Take post-operative care until the full recovery from the surgery by monitoring the mice for signs of pain. Do not return the animal that has undergone surgery to the company of other animals.
    2. Briefly observe movement around the cage, eating and drinking, and normal reactions to handling to ensure they are not under pain or stress.
    3. If the animals show signs of severe neurological or tissue damage, humanely euthanize them immediately.

2. Baseplate placement for the miniscope

NOTE: The procedures for head-fixing the mouse start 2 weeks after the animal fully recovers from the surgery. The procedure for imaging dCA1 starts 3 weeks after the surgery after the animal fully recovers and the GCaMP6f signal becomes strongly visible. A baseplate is fixed on top of the GRIN lens for optical access of dCA1 GCaMP6f fluorescence through the GRIN lens using a miniscope. This protocol utilized the miniscope version 4 -V4 (Miniscope V4; Table of Materials).

1. Start head-fixing the mouse for 10 min by securing the head bar with pinch clamps to acclimate the mouse to being head fixed 2 weeks after surgery.
2. Head-fix the mouse 3 weeks after the surgery for placement of the baseplate.
3. Attach a miniscope to a 3D printed holder attached in a micromanipulator (Table of Materials) attached in a micromanipulator (Figure 3A, B).
4. Carefully remove the silicone adhesive (Table of Materials) from the top of the lens using fine tweezers and clean the surface of the GRIN lens using a lens wipe.
5. Secure the baseplate attached to the miniscope tightening the set screw on the side of the baseplate (Table of Materials) (Figure 3C, D).
6. Move the end of the miniscope until reaching 100 µm above the GRIN lens.
7. Use the sliders to configure the miniscope settings in the miniscope software (Table of Materials) following steps 2.8 to 2.10.
8. Click on the Power slider in the software to control the power and keep it around 10%.
9. Click on the Acquisition Rate slider and keep it around 10 Hz.
10. The v4 miniscope has an electrowetting lens that can be used for fine focusing. Click in the Focus slider to set it at 50% to keep the electrowetting lens in the middle of the focal range.
11. Use the software to view the top of the GRIN lens. It is possible to see the circular shape of the GRIN lens on the screen.
12. Bring the miniscope closer to the GRIN lens by adjusting the axial position with the micromanipulator and monitor the image with the computer.
    1. Once the top of the GRIN lens is visible, use fine focus up and down on the micromanipulator to reach the best focal plane to visualize cell flashes. Use the pseudo δF/F0 setting in the software to double-check the quality of the cell flashes.
    2. Set the focus to allow the maximum number of cells in the field of view with the highest attainable fluorescence intensity. At this point, ensure that the blood vessels are in focus.
13. It may happen that a mouse has no transient calcium signals -flashes. In this case, avoid the next steps (14-16) and verify the mouse again in the next weeks to double-check whether the transient calcium signals occur.
    NOTE: A decision can be made to euthanize the mouse if the transient calcium signals never occur 2 months after the surgery.
14. Once the focus is optimized, carefully build a wall of quick adhesive cement between the skull and the bottom of the baseplate. Take care not to cement the miniscope on the baseplate and be very careful not to get any cement on the top of the GRIN lens.
15. Once the cement is dry, remove the set screw and carefully remove the miniscope from the baseplate with the micromanipulator.
16. Use the base plate cover to protect the GRIN lens.

3. Construction of the odor arena

NOTE: This method delineates an automated odor arena based on the design of Connor et al.¹³ and Gumaste et al.¹⁴. The complete assembly can be found in the link provided in the Table of Materials (Supplementary File 4).

1. Construct a chamber with dimensions 50 cm (L) x 50 cm (W) x 25 cm (H) with 2 acrylic walls, an acrylic ceiling, a white expanded polyvinyl chloride (PVC) floor, and 2 unique walls at the front and rear that facilitate air flow (Figure 4A).
2. Install an air suction end at the rear of the arena with a tapered design and a computer fan attached to draw air out of the chamber. Use a physical knob to set the fan pulse width modulation (PWM) to regulate the rate of air flow.
   NOTE: A 3D printed honeycomb structure comprises walls for the front and rear of the odor arena to facilitate laminar flow of air while retaining a mouse inside¹⁰. The honeycomb wall at the front of the arena has an external flare for receiving large volumes of air in a non-disruptive manner.
3. Set up a water delivery system to deliver water reward via simple contraptions having a Nema 17 stepper motor coupled to a syringe. An A4988 microstepping driver allows precise control of the dispensed volume (https://github.com/dougollerenshaw/syringe_pump, Figure 4B).
4. Set up the air intake end with 4 sets of odor sources paired with water-delivery spouts, with each set positioned 10 cm apart along the x axis to define the 'lanes' along which an animal will navigate for a water reward (Figure 4C, D).
5. Install an odor delivery system managed by solenoid valves connected to tubes and odorant bottles (Figure 4D).
   NOTE: The valves are powered using a relay board in an arrangement which guarantees that either clean air or odorized air is flowing at all times, and that either one single lane or zero lanes can receive odorized air at any given time. A set of five 12 V solenoid valves are managed using a relay board containing 4 relays. When all 4 relays are set to 'off', the clean-air valve is open by default. When any single odor valve is opened, the clean-air valve is automatically closed. The relay board state is managed using digital outputs from the primary arduino controller. A consumer-grade aquarium air pump supplies air which is restricted to 20 mL/min using a manual air-flow regulator. Using a series of 1/16th inch inner-diameter tubes and splitters, the supply line delivers clean air each of the solenoid valves. Check-valves before and after the odorant guarantee the direction of air flow. The clean air line merges with the odorant lines prior to entering the arena to guarantee that odorized air is purged when the odor lines are closed.
6. Install a fast digital camera above the arena to monitor the animal behavior (Table of Materials).
   NOTE: Animal behavior is monitored at 60 Hz using a single fast digital camera mounted above the arena (Table of Materials). A 3.5 mm fixed focal length C-Series lens was mounted via a C/CS mount adapter, capable of capturing the entire arena (Focal Lens, Table of Materials).
7. Use a custom python code to manage the odor arena hardware (Table of Materials). The software integrates the camera and all hardware necessary for setting up experimental parameters and acquiring experimental data (Table of Materials).
8. Set up a PC connected to a Teensy 4.0 development board to provide the means for computer-mediated odor and water delivery (Odor Arena Hardware and Software, Table of Materials).
9. Set up the digital camera to export a clock signal when recording video frames. The signal is used for post-hoc synchronization with the miniscope using a USB interface board (Table of Materials), which records the sync-out signals from both systems.
   NOTE: During an experiment, the custom acquisition software also creates an events file that contains the important experimental events and the camera frame on which the event occurred. A timestamps file is also created to identify any dropped frames, a rare event (Odor Arena Software, Table of Materials).

4. Measuring air speed of the plume with a photoionization detector (PID) ( Figure 5)

NOTE: This method detects the time course of the odor plume through a PID that exposes the gaseous odorant to a high-intensity ultraviolet light that ionizes the odorant molecules. The device's output detects odorant molecules in the odor plume. This technique allows the estimation of the air speed in the odor arena by comparing the delay to detect the presence of odorants traveling through two locations using the PID.

1. Place a fast response miniature PID at two different distances. One location is close, and another location far -10 cm apart- from the odor source.
2. Change the GAIN switch in the front panel of the PID controller to position x5.
3. Change the PUMP switch in the front panel of the PID controller to position High.
4. Check the light-emitting diode (LED) status light showing the sensor (voltage) output in the front panel of the controller in the absence of odorants.
5. Switch the potentiometer OFFSET for zeroing the voltage output in the absence of odorants.
6. Turn on the odor valve in the odor arena.
7. Measure the delay in detecting the odor plume with the PID at each location after opening the valve. This procedure can be performed offline by recording simultaneously the output of the PID and the output of the valve recorded by the odor arena (odor on) with an interface board (Table of Materials).
8. Divide the difference in the delays in each location by the difference in the distance between the two PIDs to calculate the air speed of the plume.

5. Behavioral training mouse in the odor arena ( Figure 6)

NOTE: This section describes a behavioral task adapted from Findley et al.⁴. The mouse is water-restricted the day before to motivate seeking a water reward. The mouse navigates the odor plume (Figure 6B) towards a water spout located at the source of odor release to obtain water reinforcement (3 drops of 10 µL delivered at 1 Hz). During the training period, the mouse is maintained under the water restriction by having access to up to 2 mL a day. The body weight of the mouse is monitored during the water restriction period and should not be below 85% of the original body weight. The mouse receives approximately 1 mL of water per day during the training in the odor arena and is supplemented with an additional 1 mL of water per day in the cage after training. The mouse stays under water restriction for a maximum period of 72 h. Custom software (Table of Materials) detects the mouse's location in real-time (60 Hz) using a simple background subtraction and blob-localization technique. The user manually sets lane boundaries, the home boundary (the starting location for the mouse at the back of the arena), and the target boundary (near the odor source at the front). Additionally, the user can decide how the software utilizes these boundaries. For example, the user may deliver odors only when the mouse is behind the home boundary. For the mouse to receive a reward, the user may require it to remain within the odorized lane as it navigates to the target (the odor source). Once the mouse crosses the target boundary, it may receive a reward. During training, however, any of these requirements are adjusted by simply editing a 'yaml' file designed to be self-explanatory and user-friendly.

1. First, train the mouse to start trials by moving to the back of the arena (defined as the portion of the arena that is 40 cm away from the side where the air flows into the chamber). Wait until the mouse goes to the back of the arena, and then manually deliver odor and water in one random lane. Let the mouse find the source and drink the water.
   1. Repeat the procedure many times to create the association between odor and water (Figure 6A, B). Once the mouse learns to start trials, then use the automated software to deliver odors.
2. Train the mouse to do the two-lane odor navigation task using the automated custom software. In this task, randomly choose one of two odor ports to deliver odor and reinforce the mouse with water when it arrives at the water spout where the odorant is delivered. This protocol used the odorant isoamyl acetate diluted at 1% in mineral oil.
   NOTE: The mouse completes a session of about 20 trials of odor plume navigation in about 40 min. The mouse performs one session per day. The trained mouse achieves a percent of correct navigation above random choice at the end of the session (> 65% correct choices). The mouse should achieve the criterion of > 65% correct choices on the two-lane odor navigation task after 3 to 5 sessions of training.

6. Epifluorescence recording of a freely moving mouse in the odor arena

NOTE: The method describes recording the neuronal activity of stratum pyramidale (SP) cells in dorsal CA1 by imaging the genetically encoded calcium sensor GCaMP6f expressed in Thy1 mice⁹ by wide-field miniscope imaging during the two-spout odor plume navigation task (Supplementary Movie 1, and Supplementary Movie 2). A typical imaging session takes 40 min, allowing the mouse to complete approximately 20 trials of odor navigation. This technique records a mouse for several months.

1. Head fix the mouse, place the miniscope on the top of the baseplate using the micromanipulator and tighten the set screw.
2. Adjust the electrowetting lens to find the optimal focal plane with the largest number of cells with the highest fluorescence intensity.
3. Adjust the miniscope power to obtain the optimal dynamic range with a high signal-to-noise ratio without saturation. Attain this by imaging dorsal CA1 in Thy1-GCaMP6f mice with miniscope power set around 30% at an acquisition rate of 30 Hz.
4. Release the mouse inside the odor arena with the miniscope attached to the baseplate.
5. Start acquisition with the interface board to record the transistor-transistor logic (TTL) output of the digital camera located at the top of the arena and the TTL signal from the miniscope for later synchronization between behavioral and GCaMP6f video frames. The digital camera records at 60 Hz, and the miniscope records at 30 Hz.
6. Start recording the miniscope and behavioral movies and turn on the automated software for two spout odor navigation tasks.

7. Data preprocessing

NOTE: This method uses a MATLAB pipeline to process the data. The code is available on GitHub (Synchronization Software, Table of Materials). NoRMCorre³¹ is used for motion correction, and EXTRACT³² is used to find the ROIs with time-varying fluorescence signals reported as changes in fluorescence normalized by fluorescence between calcium transients (δF/F0).

1. Synchronize the odor arena metadata, the digital camera frames, and the miniscope frames using the sync signals recorded by the interface board by running the MATLAB code Synchronize_Files_JOVE.m, available on GitHub.
2. Perform motion correction of the synchronized miniscope frames using NoRMCorre (Motion Correction Software, Table of Materials).
3. Find the ROIs with time-varying δF/F0 signals using EXTRACT (ROI Extraction Software, Table of Materials).
4. Separate the data into trials (Synchronization Software, Table of Materials).
5. Label each trial either as a hit or a miss based respectively on the mouse's correct or wrong navigation behavior.
6. Use Behavior Ensemble and Neural Trajectory Observatory (BENTO)³³ to visualize the behavior and ROIs of each separate trial (Figure 7A, B, and Supplementary Movie 3).

8. Data analysis - Decoding spatial position from brain signals

NOTE: This method uses machine learning to decode the mouse's X and Y positions in the arena from the dCA1 ROIs⁷. The MATLAB code is available (Decoding Brain Signals Software, Table of Materials) at https://github.com/restrepd/drgMiniscope.

1. Input the EXTRACT.mat output file and the odor arena metadata file to the software drgDecodeOdorArenav2.m. The GitHub repository provides these two files for an example (Synchronization Software, Table of Materials):
   dFF_file='20220804_FCM22_withodor_miniscope_sync
   _L1andL4_ncorre_ext.mat';
   arena_file='20220804_FCM22withodor_odorarena_L1
   andL4_sync.mat';
   NOTE: drgDecodeOdorArenav2.m creates a dataset divided by within-trial data containing the ROI signals and metadata (X and Y positions of the mouse, odor spout location, water delivery, etc) for each trial. The code also analyzes decoding for data between trials. The code uses fitrnet to train an artificial neural network with the δF/F0 data for all ROIs for all trials but one to predict X and Y positions and predict the position of the trial left out using a leave one out procedure.
2. The neural network returns the X and Y positions as output. Use the trained neural network to make predictions. Input left out trials not used to train the network to make predictions of the X and Y positions of the mouse in the arena from the ROIs (Figure 8A, B).

Results

Using this procedure allows for visualizing and recording dCA1 GCaMP6f fluorescence transients in mice navigating the odor arena to find the source of odorants (Figure 6A,B, Supplementary Movie 1, and Supplementary Movie 2). The fluorescence images are motion-corrected with NoRMCorre, and EXTRACT is used to extract the ROIs. In addition, recording with an interface board allows for synchronization of the δF/F0 signals from the ROIs with odor and water delivery events in the odor arena (Figure 7A), as well as with movement of the mouse in the odor arena (Figure 7B, and Supplementary Movie 3). The representative result of the mouse navigating the odor plume includes a large number of calcium transients during the task (Figure 7A,B). In addition, it is possible to inspect how the calcium responses are aligned with the presence of odor and water reward (Figure 7A). The visualization of single trials with BENTO provides information about the calcium responses at different stages of the trial, including trial start, deciding, navigating, drinking, and returning to the back of the arena (Figure 7B). The method communicated valuable insights regarding the link between CA1 calcium responses and mice behavior during an odor-oriented navigation task.

The PID recordings can provide crucial information about the odor plume and the air speed of the plume. The representative result shows an increase in the PID response after opening the valve to release the odor plume inside the odor arena (Figure 5A). Furthermore, the protocol yields decoding of the X and Y positions of the mouse from the δF/F0 signals of the dCA1 ROIs (Figure 8). This technique predicts the spatial location of the mouse during the odor-plume navigation task based on the CA1 responses, which is relevant to better understanding how the CA1 neurons process odor and spatial information. Decoding the trajectory of the mouse from neuronal ensemble calcium signals in dCA1 is significant because it reveals how neurons in dorsal CA1 represent a cognitive map of odorant and spatial information to perform the complex task of odor plume navigation. The method has been expanded to different ROIs that behave exclusively as place cells and other cells that respond to the odor stimulus. Successful decoding of the trajectory of the mouse from neuronal ensemble signals can be confirmed by the strong correlation between the decoding prediction and the ground true X and Y positions of the mouse.

Figure 1: Study pipeline. (A) The stereotaxic surgery consists of implanting a GRIN lens in the CA1 layer of the hippocampus and a headplate on the skull for head fixing the mouse. (B) A baseplate is placed on the top of the GRIN lens to allow access optically the fluorescence of the CA1 neurons with a miniscope. (C) The mouse is trained on the odor plume navigation task. (D) Freely moving recording of the mouse behavior and CA1 neurons navigating the odor plume. Please click here to view a larger version of this figure.

Figure 2: Stereotaxic surgery. (A) Place the anesthetized mouse in the stereotaxic apparatus. Adjust the level of anesthesia between 2 to 3% accordingly with the pinching paw reflex response (B) Shave the hair above the head. (C) Carefully drill a small permanent dent with a drill to make a permanent dent on the top of the target location. (D) Use the dental drill to open a circular perforation of 1.5 mm diameter to allow the implantation of a 1 mm diameter and 4 mm length GRIN lens into the Brain. (E) Connect the GRIN lens holder to the micromanipulator and turn on the aspirator connected to the pipette to hold the GRIN lens. (F) Implant the GRIN lens slowly into the cortex until it reaches the depth of -1.25 mm below dura matter. (G) Illustrative post-mortem fixed head showing the head-plate and baseplate cemented on the skull with the implanted GRIN lens. (H) Post-mortem CT scan of the head illustrating the Head Bar on the top of the cranium and the GRIN lens implanted inside of the cranium. (I) Design of the head-plate to yield head-fixing of the mouse. The animal in this depiction is undraped to facilitate improved anatomical visualization. Please click here to view a larger version of this figure.

Figure 3: Miniscope base plate placement. (A) 3D printed miniscope holder coupled to a micromanipulator. (B) Miniscope attached to the holder. (C) Attaching the baseplate to the miniscope. (D) Tightening the set screw for fixing the baseplate to the miniscope. The set screw is released after cementing the baseplate to the mouse skull. Please click here to view a larger version of this figure.

Figure 4. Construction of the odor arena. (A) Odor arena transparent Chamber. There is a top digital camera for recording the mouse behavior. (B) Step motors coupled to syringes control the water delivery to reward the mouse. (C) View from the inside of the odor arena showing a honeycomb structure is used to yield a laminar flow and four odor delivery lines. (D) Odor delivery system, including tubes, valves, and odor bottles viewed from the outside of the odor arena. Please click here to view a larger version of this figure.

Figure 5: Odor plume recording. (A) Odor plume recorded with the photoionization detector (PID). Mean(bold blue)±Standard Deviation (light blue) of five PID traces of an odor plume propagating at 4.23 cm/s in the odor arena. An odorant source at 2 cm from the odor arena floor. The sensor head of the PID provides a voltage signal for the gas concentration of the odor plume. (B) Laser recording of the odor plume with unbounded flow at 20 cm/s. Please click here to view a larger version of this figure.

Figure 6: Mouse behavioral training. (A) Mouse learning to navigate toward an odor plume released on the right lane. The mouse learns to start a trial by going to the back of the odor arena, decide a side to navigate toward the odor plume, and drink a water reward. (B) Mouse learning to navigate toward an odor plume released on the left lane. The mouse is rewarded with odor if navigating to the correct lane. Please click here to view a larger version of this figure.

Figure 7: Preprocessing the data. (A) Calcium traces synchronized with the odor arena events for many trials. Each single trial starts with the odor delivery in red, and the rewarded trials end with a water delivery pulse in blue. The δF/F0 (unitless) calcium traces for each ROI are shown in black. Each line indicates an ROI. An interface board is used for recording the TTL outputs of the top camera on the odor arena and the miniscope camera for synchronizing the frames. NorMCorre is used for correcting the movement noise from the miniscope frames and EXTRACT is used to find the ROIs and extract the δF/F0 calcium traces. (B) Representative single trial of a mouse navigating a plume. The simultaneous visualization of the synchronized behavior (left panel) and δF/F0 calcium traces (right panel) from each ROI of a single trial are observed with BENTO. Please click here to view a larger version of this figure.

Figure 8. Decoding the position of the mouse from the CA1 signals. Decoding the X and Y position of the mouse from the CA1 ROIs. The decoding prediction is shown in blue and the mouse ground true position is shown in red. The predicted traces are strongly correlated with the ground truth. (A) Decoding the X positions from the ROIs (Pearson Correlation Coefficient = 0.88). (B) Decoding the Y positions from the ROIs (Pearson Correlation Coefficient = 0.88). Please click here to view a larger version of this figure.

Supplementary Movie 1: Representative example of synchronized behavioral and miniscope frames of a mouse navigating toward an odor plume in the left lane. (A) Behavioral frames of a mouse wearing a miniscope and navigating inside of the odor arena. (B) Miniscope frames of the mouse showing the raw calcium transients recorded through the GRIN lens. Please click here to download this Movie.

Supplementary Movie 2: Representative example of synchronized behavioral and miniscope frames of a mouse navigating toward an odor plume in the right lane. (A) Behavioral frames of a mouse wearing a miniscope and navigating inside of the odor arena. (B) Miniscope frames of the mouse showing the raw calcium transients recorded through the GRIN lens. Please click here to download this Movie.

Supplementary Movie 3: BENTO display of the processed data of mouse behavior and brain signals. Left panel: the mouse is navigating toward the right lane in the arena. The behavior annotations are shown in different colors. Right panel: δF/F0 calcium signals of the mouse navigating. Please click here to download this Movie.

Supplementary File 1: NoseconeRender.png. File for the 3D-printed nose cone to perform isoflurane anesthesia. Please click here to download this File.

Supplementary File 2: HeadbarRender.png. File for the head bar for head-fixing the mouse. Please click here to download this File.

Supplementary File 3: HeadbarTechnicalDrawing.png. File for the head bar for head-fixing the mouse. Please click here to download this File.

Supplementary File 4: ArenaFig_Draft2.tiff. Detailed layout of the odor arena. Please click here to download this File.

Discussion

This protocol meticulously outlines the steps to record place-cells and odor-responsive cells in the dCA1 area of the hippocampus of mice navigating an odor plume. The critical steps in the protocol include stereotaxic surgery, placement of the miniscope baseplate, construction of the odor area, checking the plume in the odor arena, behavioral training, miniscope recording of the freely moving mouse, data preprocessing, and data analysis. Additionally, the protocol explains the process of decoding the mouse trajectory from the dCA1 recordings.

A critical step in stereotaxic surgery is to follow the coordinates relative to bregma to place the GRIN lens at the correct location. A limitation of the method is the delay between the surgery and the time to start observing GCaMP6f signals, which can take 2-4 weeks. The mouse should be ready for use after this critical period. A difference between the current protocol and previous protocols^24,25,26,27 is using Thy1-GCaMP6f mice already expressing GCaMP6f in CA1 instead of injecting AAV-GCaMP6f virus in the hippocampus. It saves time during the surgery and does not require a wait time for the expression of the AAV virus in the brain. In addition, this protocol does not aspirate the brain and uses isoflurane instead of ketamine/xylazine, which provides a better control of the dose to prevent overdoses. A limitation of the method is the optical aberration of the GRIN lens, which limits the field of view. A crucial step for the baseplate placement is to keep the electrowetting lens at the center to prevent a limitation in adjusting the Z plate after cementing the baseplate on the head. A limitation of the construction of the odor arena is the system's complexity, which can take several months and may require the assistance of an engineer. A limitation of the behavioral training is that mice may prefer one side of the odor arena against another. One way to overcome this problem is to alternate trials rewarding the mouse for navigating toward the left and right lanes.

A critical step for checking the plume is to keep the PID needle aligned with the odor source to detect the plume's path. A crucial step in miniscope recording a freely moving mouse is to ensure that the mini-coax wire does not tangle during the task, which can be prevented with a commutator. Helium balloons can be used to prevent the mini-coax wire from coming in front of the mouse. It is critical for the data preprocessing to synchronize the TTL pulses of the digital top camera of the odor arena and the miniscope camera. For the EXTRACT procedure, it is recommended to use non-negative processing to extract the ROIs and δF/F0 signals better. EXTRACT yields a visual inspection of the traces for each ROI to exclude bad ones. It is critical to decode the X and Y positions of the mouse from the ROIs to have a large dataset with hundreds of epochs for better training the artificial neural network.

The significance of this freely moving recording method concerning existing head-fixed methods is studying the mouse behavior in an ethologically relevant context with the proper head movement for navigating a complex odor plume. This method is applicable to investigate the dynamic role of dCA1 neurons in complex odorant navigation. Furthermore, the procedure is not restricted to hippocampus or olfaction. Other potential applications of the technique include studying the role of different brain areas and sensory modalities, including possible applications in visual navigation by using LEDs to indicate the rewarded lanes. In addition, this method can be potentially applied in real-time closed-loop experiments in which decoding neural calcium from population triggers neurostimulation or sensory feedback^20,34,35,36.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Arduino Micro	Arduino	Micro
Biocompatible Methacrylate Resin	Parkell	S380	C&B-Metabond Adhesive Luting Cement
Decoding Brain Signals Software	CU Anschutz		https://github.com/restrepd/drgMiniscope
Dental Drill	Osada	LHP-6	AZ210015
Dental Drill Box	Osada	XL-230	30000 rotations per minute
Digital stereotaxic instrument	Stoelting	51730D	Mouse Stereotaxic Instument, #51904 Digital Manipulator Arm, 3-Axes, Add-On, LEFT
Drill Bit	FST Fine Science Tools	19007-05	Tip diameter 0.5 mm
Fast Digital Camera	Edmund Optics	BFS-U3-63S4C	FLIR Blackfly S
Focal Lens	Edmund Optics	C-Series	3.5 mm
GRIN lens	Inscopix	1050-004595	1 mm diameter and 4 mm length
GRIN lens Holder	UCLA		http://miniscope.org/index.php/Surgery_Protocol
Liquid Tissue Adhesive	3M	1469C	Vetbond Tissue Adhesive
Low-Flow Anesthesia System for Mice	Kent Scientific Corporation	SomnoSuite	https://www.kentscientific.com/products/somnosuite/
Low Toxicity Silicone Adhesive	WPI – World Precision Instruments	Kwik-sil
miniPID Controller	ASI – Aurora Scientific Inc.	Model 200B	Fast-Response Miniature Photo-Ionization Detector
Miniscope V4 Holder	UCLA	NA	https://github.com/Aharoni-Lab/Miniscope-v4/tree/master/Miniscope-v4-Holder
Miniscope V4	Labmaker	NA	https://www.labmaker.org/products/miniscope-v4
Miniscope Base Plate V2	Labmaker	NA	https://www.labmaker.org/products/miniscope-v4-base-plates-variant-2-pack-of-10
Miniscope DAQ-QT software	UCLA		https://github.com/Aharoni-Lab/Miniscope-DAQ-QT-Software/wiki
Motion Correction Software	CU Anschutz		https://github.com/restrepd/drgMiniscope
Odor Arena Hardware	Custom Made	3D Model	https://www.dropbox.com/scl/fo/lwtpqysnpzis32mhrx3cd/ADomsxyhxu42sqDmTBl2O6k?rlkey=b3l4809eradundt5l3iz0gq74& dl=0
Odor Arena Software	CUAnschutz		https://github.com/wryanw/odorarena
Odorant Isoamyl Acetate	Aldrich Chemical Co	06422AX	Diluted at 1% in odorless mineral oil
RHD USB Interface Board	Intan Technologies	C3100	Product discontinued. Alternatively use another equivalent board.
ROI Extraction Software	CU Anschutz		https://github.com/restrepd/drgMiniscope
Sutter Micromanipulator	Sutter Instrument Company	MP-285
Synchronization Software	CU Anschutz		https://github.com/fsimoesdesouza/Synchronization
Thy1-GCaMP6f mice	Jackson Laboratory	IMSR_JAX 028281	C57BL/6J-Tg(Thy1-GCaMP6f)GP5.12Dkim/J)