Protocol to test path integration in mice using an L-maze behavioral assay

Summary

Path integration (PI) is a component of navigation whereby an organism uses self-motion cues to estimate position and move toward a goal, without relying on external landmarks. PI deficits can predict cognitive decline and Alzheimer’s disease in humans. Here, we present a protocol to test PI in mice using a dry-land L-maze behavioral assay with three phases: habituation, training, and testing. This protocol may allow researchers to test for early disease in mouse models of neurodegeneration.

For complete details on the use and execution of this protocol, please refer to Graves et al.¹

Graphical abstract

Highlights

• •Instructions for habituation prior to behavioral testing in mice
• •Steps for testing path integration in mice using a dry-land L-maze assay
• •Guidance on using path integration in mouse models of neurodegeneration

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Path integration (PI) is a component of navigation whereby an organism uses self-motion cues to estimate position and move toward a goal, without relying on external landmarks. PI deficits can predict cognitive decline and Alzheimer’s disease in humans. Here, we present a protocol to test PI in mice using a dry-land L-maze behavioral assay with three phases: habituation, training, and testing. This protocol may allow researchers to test for early disease in mouse models of neurodegeneration.

Before you begin

Deficits in path integration (PI) are associated with dysfunction and/or loss of grid cells in the entorhinal cortex, a brain region particularly vulnerable to degeneration in Alzheimer’s disease (AD).² Disease-related changes in the entorhinal cortex can occur before clinical symptoms emerge.³ PI deficits can predict both hereditary and physiological AD risk.⁴ Thus, accurate testing for entorhinal cortex perturbations could lead to earlier intervention in diseases like AD.

PI ability in humans is tested using immersive virtual reality (iVR) simulations wherein participants navigate an environment using self-motion cues. In these tests, participants are typically guided in the iVR environment in a predetermined path with 2–3 stopping points along then way and then instructed to return to the starting point without guidance. PI is then reflected in the error distance between stopping point and starting point.^4,5,6

Testing PI in mouse models of brain aging or disease is a relatively new strategy for behaviorally assessing disease state. These studies have adapted the iVR simulations used in humans in various ways usually as a water-based L-shaped escape task.^7,8 Deficits in PI in mice using such L-maze tasks have been associated with grid cell dysfunction and phosphorylated tau burden, a key pathological hallmark of AD in humans.^5,7,8

This L-maze protocol allows for testing of PI in mice in a dry context, where water submersion is not ideal in instances of frailty and/or locomotor deficits.

Innovation

Other published works that include path integration tests in mice use water-based assays,^7,8 which can be difficult for mice with increased frailty and/or motor deficits. This L-maze protocol differs from others by being dry-land based, allowing for assessment of mice that would not otherwise tolerate water testing.

Institutional permissions (if applicable)

All animal experiments were reviewed and approved by the Mayo Clinic Institutional Animal Care and Use Committee and in accordance with the guidelines from the American Association for Laboratory Animal Science. To perform animal experiments described in this protocol, prior approval according to institutional or state legislation is required.

Preparation of the behavior room

Timing: variable

Proper setup of the behavior room is critical for L-maze testing success. Ideally, the room will be quiet, secluded, and a novel space to the mice being tested. The behavior room will need to be large enough to accommodate a computer station, a cart with singly housed mice during testing, and a curtained off section containing a circular Barnes maze table.

• 1.
  Fabricate an L-maze apparatus and start box.

  • a.
    Construct an L-maze apparatus with one short arm (∼30 cm) and one long arm (∼60 cm) affixed at a 90° angle with an open floor such that the mouse can travel within the L-maze on top of a Barnes maze table and access the table’s escape hole. The L-maze apparatus should be ∼7 cm wide and 20 cm tall.

    Note: These measurements fall in between what others have used for similar tests^7,8 and allow the apparatus to fit on top of a commonly sized Barnes maze table.

  • b.
    Construct a start box that is clear on all sides except the back side and is large enough to hold one mouse and small enough to fit in between the walls of the L-maze apparatus.

    CRITICAL: The start box must allow enough room light in to motivate the mouse to exit the box and explore for an escape.

    CRITICAL: The start box must have a recognizable back to orient the mouse during the testing phase when the start box is the only equipment on the Barnes maze table.

• 2.
  Setup behavior testing room.

  • a.
    Curtain off a section of the behavior room containing the Barnes maze table.

    CRITICAL: The curtain should encircle the Barnes maze table completely and hang 8–12 inches from the table’s edge. It should effectively block out all spatial cues from the rest of the room.

  • b.
    Place a camera with infrared vision above the Barnes maze table that captures the entire Barnes maze table and is compatible with Ethovision software.
  • c.
    Setup a bright light above the table with a handheld toggle to turn it on/off.
  • d.
    Clean the Barnes maze table and L-maze apparatus with hydrogen peroxide wipes to remove scent cues.
  • e.
    Setup a white noise machine to drown out any sound cues.

    Note: The L-maze will sit atop the Barnes maze table during the training phase of the assay but is not present on the Barnes maze table for the habituation and testing phases.

• 3.
  Setup up an L-maze experiment in Ethovision.

  • a.
    Track mouse head, center, and tail.
  • b.
    Create an arena and indicate the start box and escape hole positions as objects.
  • c.
    Create an experimental protocol with a conditional start that allows for time to transfer mice to the Barnes maze table (i.e.- 2 s after mouse is detected in arena) and a runtime of 90 s.
  • d.
    Set to detect when mouse enters escape hole.
  • e.
    Set to record 1 trial per mouse for the training phase. Repeat 5 times.
  • f.
    Set to record 1 trial per mouse for the testing phase.

Mouse handling

Timing: 30 min

Marking the mice for clear identification prior to the habituation phase minimizes in-test handling. This decreases stressful stimuli during the test that may confound results.

• 4.Ensure mice can be identified without extensive handling during the test.

Note: If mouse identification requires scruffing, use a sharpie to uniquely tail mark each mouse in a cage prior to bringing mice to the behavior room.

CRITICAL: Avoid scruffing the mouse repeatedly during training and testing to reduce stress that could interfere with performance.

CRITICAL: When transferring mice to and from the cage or table during testing, place the start box or beaker used to transfer the mouse into the cage and gently coax the mouse into it to further limit handling stress.

• 5.Transport mice to the behavior testing room in their home cages.
• 6.Bring enough clean cages to the behavior room to singly-house mice during testing.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Software and algorithms
Ethovision XT	Noldus	RRID:SCR_000441
GraphPad Prism	GraphPad	RRID:SCR_002798
Excel	Microsoft	RRID:SCR_016137
Other
Custom L-maze apparatus	Darren J. Baker1	N/A
Custom L-maze start box	Darren J. Baker1	N/A
Blackout curtains	Amazon	Commercially available
Custom Barnes maze table	Darren J. Baker9	N/A
White noise machine	Amazon	Commercially available
Camera (4.5–12.5 mm 1:1.2; IR ½” CS)	Computer	Commercially available

Step-by-step method details

Habituation of mice to the behavior room

Timing: 1 h

Adequate habitation of mice to the testing environment is crucial for successful behavioral testing. This phase allows mice to adjust to the testing room and to being singly housed in order to minimize the effects of stress in response to a new environment.

• 1.Singly-house mice in clean cages for 60 min in the room where testing will take place.

CRITICAL: Ensure that the lighting and sound within the room are as expected for testing.

• 2.Start a white noise machine in the behavior room.

CRITICAL: The white noise machine limits spatial sound cues in the environment during training and testing and helps maintain path integration test integrity.

Habituation of mice to the maze table

Timing: 1–2 min per mouse; no more than 30 min for the entire set

Habituating mice to the table on which L-maze training and testing occurs helps familiarize mice with the table and the exposes them to the possibility of escape. This ultimately improves efficiency in escape during the training phase.

• 3.Place mouse in an upside-down 500 mL clear glass beaker in the center of the Barnes maze table in the dark.
• 4.Toggle the overhead light on and allow the mouse to look around the space for 10 s.
• 5.Slowly slide the beaker and mouse to the edge of the table and allow them to explore a false escape hole for another 10 s.
• 6.Slowly slide the beaker and mouse around the edge of the table to the real escape hole and allow the mouse to escape. As soon as the mouse commits to escaping, toggle off the overhead light.
• 7.Allow the mouse to rest in the escape compartment for 30 s.
• 8.Return mouse to its testing cage.
• 9.Wipe clean the L-maze table and beaker used to transfer the mouse with hydrogen peroxide wipes to remove scent cues.
• 10.Repeat for each mouse.

Training with L-maze

Timing: 90 s per mouse; 5 rounds, each 15–20 min apart

The training phase of this protocol is where the mouse learns that escape is possible at a fixed location from the start box. Training takes place in the L-maze apparatus and forces the mouse to travel in a specified path from start box to escape hole.

CRITICAL: Ensure the white noise machine is on to limit sound cues in the environment.

• 11.Set the L-maze apparatus on the Barnes maze table with the end of the long arm over the escape hole on the Barnes maze table.
• 12.Take a background photo and set the arena in Ethovision for all training trials.
• 13.Load mouse into the start box.
• 14.Press start on the trial in Ethovision.
• 15.Load the start box with the mouse into the short arm of the L-maze apparatus and simultaneously toggle the lights on.

Note: The trial should be set to conditionally start a few seconds after the camera recognizes the mouse is the maze.

• 16.As soon as the mouse enters the escape hole, toggle off the room lights and stop the trial in Ethovision.

Note: If the mouse does not enter the escape hole within 90 s, allow the trial to end, then guide the mouse through the L-maze and into the escape hole, toggling off the lights once the mouse enters the escape hole.

• 17.Allow the mouse to rest in the escape hole for 30 s then transfer them gently back into the experimental singly-housed cage.
• 18.Wipe clean the interior of the L-maze and the start box with hydrogen peroxide wipes to remove scent cues.
• 19.Repeat 12–18 with the next mouse. Complete 5 rounds of training for the entire set of mice being tested.

CRITICAL: Keep training rounds <20 min apart for any individual mouse to limit long term memory reliance.

Testing path integration

Timing: 90 s per mouse

The testing phase of this protocol involves removing the L-maze apparatus from the testing table and allowing the mouse to travel from the same start position to the same escape hole position as during training, but without the forced L-maze path. During this phase, mice with high ability to path integrate and a desire to escape the bright lights/exposed nature of the testing table will travel in a straight line from start box to escape hole, ultimately scored as low “angular error”. Mice with low ability to path integrate but high spatial memory may travel in the shape of the L-maze that they were trained in during the training phase. Mice with low ability to path integrate and low spatial memory will travel in a random direction from the start box.

• 20.Continue running the white noise machine in the behavior room to drown out sound cues.
• 21.Remove the L-maze apparatus from the Barnes maze table.
• 22.Take a background photo and set the arena in Ethovision for all testing trials.
• 23.Load mouse into the start box.
• 24.Press start on the trial in Ethovision.
• 25.Place the start box with the mouse in the same spot on the open Barnes maze table as it was for each training trial and simultaneously toggle the lights on.

Note: The trial should be set to conditionally start a few seconds after the camera recognizes the mouse is the maze.

• 26.Stop the trial in Ethovision once the mouse enters the escape hole or 90 s elapses.
• 27.Place mouse back in home cage.
• 28.Wipe clean the Barnes maze table and the start box with hydrogen peroxide wipes.
• 29.Repeat 22–26 with the next mouse until all mice have gone through testing.

Analyzing path integration

Timing: variable

Analyzing path integration during the testing phase is done using the “heading to escape” statistic in the Ethovision software. This statistic is calculated as the average heading direction of the mouse relative the escape hole over a set distance traveled—typically, the first 10 cm of travel from the start box. A “0” in this statistic means the mouse traveled in a straight line from the box to the escape hole with perfect path integration ability.

• 30.
  Review the testing trials in Ethovision.

  • a.
    Quality control for proper mouse detection, especially at the beginning of trials.
• 31.
  Extract statistics.

  • a.
    Extract the “heading to escape” statistic for the first 10 cm of distance traveled by the mouse during the testing phase.

Note: This is the average direction the mouse traveled in relative to the escape hole such that “0” would indicate a straight line from start to escape and “perfect” path integration.

Expected outcomes

Mice with high ability to path integrate will display lower angular error (the absolute value of the “heading to escape statistic” in Ethovision). Mice with low ability to path integrate will display higher angular error. Mice with high spatial memory, but low ability to path integrate will display angular error that is consistent with the mouse traveling the same forward direction of the L-maze, rather than veering toward the most direct route to escape (typically, this is around an angular error of about 70°).

Limitations

One limitation that the authors have noted is a decline in the ability of mice to path integrate after the age of 9 months. This timeframe may vary from cohort to cohort, but it is worth noting that initial testing to determine control mice can indeed path integrate (i.e.-angular error <70°) is necessary before expanding to experimental mice. The authors also noted that male mice tend to have a higher ability to path integrate than their female counterparts. This could also vary from cohort to cohort and further necessitates initial validation and testing.

Another limitation is that a group of mice that has low ability to path integrate will quantify as having a large spread of angular errors, representative of random direction of travel during testing. This makes certain classical statistical analyses prone to needing large numbers of mice. Experimenters will need to run power calculations and perform testing with sufficient animal numbers in each group to overcome this. Alternatively, experimenters will need to analyze the spread of their data statistically in addition to analyzing the means.

An additional limitation is that this protocol differs from traditional PI testing in humans in two important ways: 1) it does not measure distance-error and 2) it contains a 15–20 min delay between training and testing rounds. Distance-error is a critical metric in human path integration tests that is measured by asking a test subject to estimate their position relative to the start at defined intervals during the test. Mice are unable to provide this information; thus, this protocol lacks the potential to measure distance-error and must rely solely on path error (analyzed as ‘heading to escape’ statistic in Ethovision software). The 15–20 min interval between training and testing rounds also differs from PI testing in humans and may require involvement of some memory processes beyond strictly working memory.

Troubleshooting

Problem 1

Mice may never leave the start box to explore the L-maze apparatus during training. This is an essential step (related to step 15).

Potential solution

Increase the lighting that the start box receives by ensuring that it is clear and/or topless. This will help motivate the mouse to explore for an escape. Additionally, try tapping lightly on the back of the start box to motivate the mouse to leave the box and explore the maze.

Problem 2

In some instances, Ethovision fails to accurately track a mouse, especially at the beginning of trials when there is movement due to loading of the mouse onto the table. This will invalidate the “heading to escape” statistic that angular error is derived from (related to step 30).

Potential solution

Manually review each testing trial and exclude the time points prior to the mouse leaving the start box and Ethovision accurately detecting the mouse. Additionally, re-running the video as another trial after tweaking detection settings and background corrections in Ethovision can resolve this issue.

Problem 3

Sometimes, mice of a control group that are expected to have high ability to path integrate, will display high angular error during testing. This is not entirely unexpected as individual variability in path integration ability exists and can be due to numerous known and unknown factors. Below are some potential factors that could lead to high angular error and suggestions for how to combat them (related to step 31).

Potential solution

• •Suspected age-related decline of path integration in a cohort: If possible, test mice at a younger age.
• •Suspected sex-related impacts of path integrations: Sufficiently power your experiment to split sexes and report them independently.
• •High variability in path integration results: Reduce stress in mice being tested by increasing habituation times, decreasing in-test handling, and/or habituating mice to the investigator and handling in the week or two prior to testing.

Problem 4

Large variability in calculated angular error due to single trial testing (related to step 31).

Potential solution

Perform three testing trials in mice with <15 min between trial and average angular error across the testing trials. If doing this, investigators will need to account for any spatial learning that occurs across testing trials.

Problem 5

The angular error statistic does not reflect the average heading direction of the mouse. This can occur due to oversampling of mouse position when a mouse moves slowly and/or has an unsteady gait (related to step 30).

Potential solution

• •Export a trace of the mouse’s path as an image from Ethovision and manually measure the angle formed between the mouse’s path and the direct escape path.
• •Alternatively, if this is a uniform issue across a testing cohort, decrease the location sampling frequency in the testing setup in Ethovision.

Problem 6

No path integration deficit observed in mice where one is hypothesized due to expected small effect size (related to step 31).

Potential solution

Increase the potential maximum angular error by increasing the size discrepancy between the long and short arms of the L-maze. Extending the long arm and/or decreasing the size of the short arm will give greater degrees of freedom to measure smaller differences in angular error.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Darren Baker (baker.darren@mayo.edu).

Technical contact

Technical questions on executing this protocol should be directed to and will be answered by the technical contact, Sara Graves (graves.sara@mayo.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

This study did not generate/analyze new datasets/code.

Acknowledgments

We would like to thank B. Wilbur for fabrication of the L-maze start box and the Mayo Clinic Division of Engineering Shop for fabrication of the L-maze apparatus. We would also like to thank T. Thao and R. Fierro Velasco for genotyping and animal support; the graphical abstract was generated in part using BioRender. This work was supported by the National Institutes of Health (R01AG068076 and U54AG079779), the Minnesota Partnership for Biotechnology and Medical Genomics (MNP #18.04), and the Glenn Foundation for Medical Research (all to D.J.B.).

Author contributions

S.I.G. designed, validated, and tested the protocol outlined in this work. S.I.G. wrote the first draft of this manuscript. S.I.G. (graves.sara@mayo.edu) is the technical contact for this work. D.J.B. (baker.darren@mayo.edu) directed and supervised the work, edited this manuscript, and serves as the lead contact.

Declaration of interests

D.J.B. has a potential financial interest related to this research. He is a co-inventor on patents held by Mayo Clinic and patent applications licensed to or filed by Unity Biotechnology and is a Unity Biotechnology shareholder. Research in the Baker laboratory has been reviewed by the Mayo Clinic Conflict of Interest Review Board and is being conducted in compliance with Mayo Clinic Conflict of Interest policies.