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

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.
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. Inject buprenorphine SR subcutaneously at 0.001 mg/g immediately after surgery for analgesia.
31. 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).