Knowledge Base

The Radial Arm Maze Test

Introduction

the radial arm maze test

The radial arm maze was designed to measure spatial learning and memory in rats (Olton and Samuelson, 1976; Olton, 1978, 1985). The original apparatus consists of eight arms, each about 4 feet long, and all radiating from a small circular central platform. At the end of each arm there is a food site, the contents of which are not visible from the central platform. The design ensures that, after checking for food at the end of each arm, the rat is always forced to return to the central platform before making another choice. For example, when investigating working memory, all arms are provided with a food reward and the animal should visit each arm only once. Correct performance of the radial maze task requires intact spatial memory abilities. Performance is affected by hippocampal impairment and a variety of pharmacological agents (e.g., Janitzky et al. 2011). Figure 8.1 Example of an 8-arm radial maze.

references

Janitzky, K, Schwegler, H., Kröber, A., Roskoden, T., Yanagawa, Y. & Linke, R. (2011). Species-relevant inescapable stress differently influences memory consolidation and retrieval of mice in a spatial radial arm maze. Behavioural Brain Research, 219(1), 142-148. Olton, D.S. (1978). Characteristics of spatial memory. In: Hulse S.H., Fowler H, Honig W.K. (Eds.), Cognitive processes in animal behavior, Hillsdale, NJ: Lawrence Erlbaum Associates, 341-373. Olton, D.S. (1985). The radial arm maze as a tool in behavioral pharmacology. Physiol. & Behav. 40, 793-797. Olton, D.S., & Samuelson, R.J. (1976). Remembrance of places passed: Spatial memory in rats. Journal of Experimental Psychology: Animal Behavior Processes, 2, 97-116. Wenk, G.L. (2004). Unit 8.5A Assessment of Spatial Memory Using the Radial Arm Maze and Morris Water Maze, in: Current Protocols in Neuroscience, John Wiley and Sons, Inc.

Physical setup

The following suggestions are specifically to optimize video tracking: The camera should have a good view of the entire region the animal can be in. The arena should fill the field of view. There must be maximum contrast between the background and the animal. If you use white animals in the maze, make sure the floors of the arms are dark and vice versa. Preferably, the floor of the room should be the same color (and matted) as the maze, contrasting with the animal. Lighting should be uniform and even with no shadows. If you have reflections in your image, EthoVision XT may confuse those reflections with your subject, and track the reflections rather than your animal. To prevent reflections, place indirect lighting above or around the maze. 'Globe' type bulbs twice the diameter as standard light bulbs are ideal. They should be close enough to the maze so that there is no direct line of sight between the bulbs and the camera lens. The light should reflect off the walls and ceiling, so that it only reaches both the lens and radial maze indirectly.

The Radial Arm Maze test in EthoVision XT

In EthoVision XT, you can set up a radial 8-arm maze experiment by using a predefined template in the guided setup. Create an experiment. For details, see Chapter 1 of this manual, or see Setup an Experiment in the EthoVision XT Help.

experiment settings

1. Make sure that the radial maze is connected to the EthoVision XT computer. To connect the radial maze to EthoVision XT, you must have the Trial and Hardware Control add-on and the USB-IO box. The procedure and the necessary cabling may differ between manufacturers of the radial maze. Please see the accompanying documentation.
2. Choose Setup > Experiment Settings.
3. Under Video Source, and Tracked Features, make sure that the options selected corresponds to your needs. To adjust the camera settings, click the video icon in the camera row.
4. If you want to use Deep learning technique to track the subject's nose, under Body point detection technique choose Deep learning. Note that in order to use this technique there are additional requirements. Among other things, you need a powerful graphics

card. See Deep learning: Requirements and Limitations in the

EthoVision XT Help. 5. Under Trial Control Hardware, select Use of Trial Control Hardware and click Settings. Select Noldus USB-IO box. 6. In the Device Configuration window, for TTL ports 1 to 4, under

Device type, select Custom Hardware. Next, type in Doors1&2 for TTL

Port 1, Doors3&4 for TTL Port 2, Doors5&6 for TTL Port 3 and

Doors7&8 for TTL Port 4.

arena settings

Choose Setup > Arena Settings > Open Arena Settings 1. 1. In the Grab Background Image window, click the Grab button to grab a background image of the radial maze from the camera image. 2. Click 1. Draw Scale to calibrate and calibrate your arena. For details, see Make the arena in the EthoVision XT Help). 3. Click 2. Select Shape and Draw Arena and check that the predefined arena has the correct shape and size. If necessary, adjust the contour of the arena to fit the radial maze. 4. If you used the Zone template, click 3. Select Shape and Draw Zones. Check that the zones have the correct shape and size. If necessary, resize or add new zones. 5. If you are using automated doors, click Arena - Hardware mapping in the Arena Settings window to assign the doors to the arena.

trial control settings

In the Trial Control Settings you can define conditions for the start and stop of the track.

Choose Setup > Trial Control Settings > Open one of the following:

Default (no max duration) - When you use this profile, tracking

starts automatically 2 seconds after the animal has been placed in the Center zone. Tracking stops manually (for example when you click the Stop trial button).

Track duration 10 min - When you use this profile, tracking starts

automatically 2 seconds after the animal has been placed in the Center zone. The track stops automatically when 10 minutes have elapsed since the start of the track.

Controlling automated doors

If you want to control the automated doors with EthoVision XT, you need to create new Trial Control Settings. An example of Trial Control Settings for a protocol as described in Unit 8.5A of Current Protocols in Neuroscience (2004) to test basic working memory is the following: Start condition. We assume that the animal is released in the Center zone with all doors closed. Tracking starts 3 seconds after the animal is put in the Center zone. Opening all doors simultaneously. You can create a Sub-rule with a Reference for this. In the Sub-rule, each door is opened. For each door, you need to create a separate Hardware Action box, with the corresponding Device name ('1&2') and appropriate Action ('Output 1 High' which opens door 1. The figure below shows part of the sequence of Sub-rule 'open beginning' in which doors 1-3 are opened. The figure below shows the Reference box to the Sub-rule. The reference is activated immediately after the Start trial command. Closing and then opening all doors. When the animal has visited the Goal zone of one of the arms and returns to the Center zone, all doors are closed and re-opened after 5 seconds. You can create a Sub-rule with a Reference for this. In the Sub- rule, you need to create a Zone transition Condition for each arm separately. Next, all these 8 Zone transition Conditions need to be connected to an Operator 'Any input True'. The figure below shows part of the sequence of the Sub-rule 'open close Center'. Here you see the Zone transition Conditions for arms 1 and 2. The Zone transitions Conditions of all 8 arms are connected to the Operator 'Any input True'.

Next, with a Hardware Action box, all 8 arms are closed. The

figure below shows an example of closing door 4. Put the Hardware Action box for all doors in a sequence similar to opening doors as described in step 1 above. Stop condition. The trial stops when the animal has visited the Goal zones of all 8 arms OR after 15 minutes have elapsed since the start of tracking. For the first criterion (the animal visits the Goal zone of all arms) you use an In Zone Condition for each Goal zone. All eight In Zone Condition boxes are connected to the Operator 'Num simultaneously True inputs'. The figure below is an example and contains only two of the eight In Zone Goal zone Condition boxes. For the second criterion (after 15 minutes) you need a Time Condition box as shown in the figure below. The two criteria must be combined with OR logic. The Time Condition box and the 'Num simultaneously True inputs' Operator box are connected to a 'Any input True' Operator box (see figure below. In this example sequence, an extra delay of 3 sec is included before the trial stops).

detection settings

Choose Setup > Detection Settings > Detection Settings 1. We assume you followed the procedure on page 17.

Advanced detection settings

1. Under Video, check that the sample rate is set to 25 or 30 samples/ second for multiple body-point tracking.
2. Click Advanced.
3. Under Method:

By default, Static subtraction is selected as the detection method

and Rodents / Default as the method for nose-tail detection. The

background in a radial maze test usually does not change much, so Static subtraction usually works well. Make sure the lighting is even and there are no shadows in the maze. If after running some test trials, you get nose-tail swaps, try using the tracking method Rodents / For occlusions. If you selected to use the Deep learning technique to track the

subject's nose, next to Deep learning click Define and select a box

around the subject. Make sure that the box includes the subject's nose. For details, see the EthoVision XT Help. 4. optional Select Track noise reduction. Under Smoothing, set Track noise reduction to On. In some cases better quality tracking can be obtained by reducing track noise during acquisition. This may especially be the case if you use Trial and Hardware Control. As an example, if the center point of an animal is detected in the center-zone of the maze, you want specific doors to open followed by the closing of doors as soon as the animal enters one of the arms. If the detected center point is moving rapidly because of noise, this may result in a doors opening and closing, every time the center point crosses the border of the zone. Track noise reduction may solve this problem. 5. Under Subject size, click Advanced and set a minimum subject size to prevent droppings from detected.

trial list

Enter your independent variables such as Animal ID, Treatment, Baited arms, Name of the experimenter, etc. (for blind trials, enter only the Animal ID). If you want, you can predefine all your trials here, or you can enter the independent variable values as you carry out the trials. You can also prepare the trials in Excel, randomize them and then paste the values into your Trial List. Furthermore, you can define a list of trials for batch acquisition. See Acquire a series of trials in the EthoVision XT Help.

Baited arms

To mark the trials with the arms that were baited, create an independent variable Baited arms of text type, and define the various

combinations as Predefined Values: for example, 1-3-6-8, 1-3-5-7, etc. For

each subject to be tested, assign the correct value in the appropriate cell of the Baited arms column. Figure 8.2 The predefined Trial List in the radial 8-arm maze template experiment.

Acquiring data

protocol

The following protocol tests for basic working memory (the subject learns not to re-enter a baited arm).

Habituation (pre-training)

In a typical radial 8-arm maze test, first a pre-training session is carried out in which the animal is allowed to freely visit all arms for 10 to 15 minutes per day. Also, the doors are regularly opened and closed to habituate the animal.

Training trials

In this phase, the animal is exposed to where the food is, and this is what the later tested memory is based on. Animals are trained one session per day, for approx. one week. Typically, all arms are baited in this phase. At the start of the trial, the animal is placed it the center of the maze, facing the same direction on every trial, everyday. Stop the trial when (a) all eight arms have been visited; (b) 10 minutes passed since the start of the trial; or (c) 2 minutes since the animal's last arm entry.

Performance

Among the variables commonly used for the analysis of the performance are (a) the number of errors in each session (entering an arm that has been visited previously counted as an error) and the total number of errors across eight sessions, and (b) the number of correct

choices in the first eight arm entries of each session. See Data analysis

on page 152.

Reference memory errors

While the previous protocol is primarily sensitive to impairments in working memory, an alternate protocol allows a disassociation to be achieved between working and reference types of memory. In this protocol, four arms are baited for the training phase. The same maze arms are baited each day and, across sessions, the animal learns to ignore the other arms, which never contain a reward. This is the reference memory component of the task. In the test phase, food items are placed in the same arms as they were on the memory trials that day. It is recorded when the animal gets 100% correct on his memory trials. When the animal enters an arm that is not baited, it is marked as an error. See Wenk, G.L. (2004). Current Protocols in Neuroscience, 8.5A.1-8.5A.12.

Data analysis

data preparation

Track editing

Choose Acquisition > Edit Tracks. You can fix tracking errors but normally you will not need to edit your data.

Track smoothing

Choose Acquisition > Track Smoothing Profile > open one of the two

Track Smoothing Profiles: No filter.

MDM 0.2 cm - In this Track Smoothing Profile, the Direct Minimal

Distance Moved filter is used, with an MDM of 0.2 cm. Use Smoothing when you want to eliminate small movements, such as body wobbling during locomotion, that might affect dependent variables such as total distance moved.

Selecting data

Choose Analysis > Data Profile. The template contains two data profiles: All Data - This data profile contains all data.

Treated vs. Control - This data profile contains two Results

containers with data for the treated animals and the control animals

(see also trial list on page 149). This way you create groups of tracks

to obtain group statistics for both groups.

analysis profiles

Default analysis profiles

The template experiment contains three analysis profiles:

Distance and Velocity - In this profile, the total Distance Moved

and the mean Velocity are calculated for the center-point.

In Zones - In this profile, the Frequency, Duration and Latency to

First for all zones (arms and goal zones, see also arena settings on

page 143) are calculated for the center-point. A visit to a previously chosen arm is considered a working memory error. Normal healthy young rats will perform this task almost perfectly every time, so will visit each arm and goal zone only once.

Zone transitions - In this profile, the Frequency of zone transitions

from the center zone to any arm is calculated for the center-point. For all statistics also the group means with standard errors are calculated.

testing working memory

Time taken to obtain all rewards

The most basic test conducted using the radial-arm maze is to bait the ends of all of the arms with the reward and record the time taken to obtain all the rewards. The following procedure applies to trials where all baited arms were visited. 1. In the Analysis profile, choose Target visits and errors. 2. In the Target zones box, select all the zones that were baited. 3. Under Calculate Statistics for, select Target first visits. 4. Click the Trial Statistics tab and select Latency to Last.

Incorrect arm choices

To obtain a reward, an animal must remember which arm(s) it visited previously and not re-visit those arms. 1. In the analysis profile, choose Target visits and errors. 2. In the Target zones box, select all the zones that were baited. 3. Under Calculate Statistics for, select Target revisits. 4. Click the Trial Statistics tab and select Frequency.

Percentage of correct choices

One of the parameters described in the Unit 8.5A of Current Protocols in Neuroscience (Wenk, 2004) to calculate performance of all groups is: The percentage of correct choices made in each test session in relation to the total number of arms entered. 1. In the analysis profile, choose Target visits and errors. 2. In the Target zones box, select all the zones that were baited. 3. Under Calculate Statistics for, select Target first visits and Target revisits. 4. Click the Trial Statistics tab and select Frequency. 5. You can calculate the percentage of correct choices with: Target first visits /(Target first visits + Target revisits)*100. Where the statistics given by EthoVision XT are shown in italics.

testing reference memory

The radial-arm maze may also be used to test reference memory by only baiting some arms of the maze. In training trials, the reward is placed consistently from trial to trial for the same animal, but its placement is varied from animal to animal. In the test trials, the time to acquire the reward and the number of incorrect choices (i.e., entering an unbaited arm or re-entering a baited arm) are analyzed.

Time taken to obtain the rewards

1. In the analysis profile, choose Target visits and errors.
2. In the Target zones box, select all the zones that were baited.
3. Under Calculate Statistics for, select Target first visits.
4. Click the Trial Statistics tab and select Latency to Last.

Number of entries into unbaited arms

1. In the analysis profile, choose Target visits and errors.
2. In the Target zones box, select all the zones that were baited.
3. Under Calculate Statistics for, select Not-target first visits and Not- target revisits.
4. Click the Trial Statistics tab and select Frequency.

note Re-entries into baited arms are considered working memory

errors. Because the arms chosen to be baited differ between animals, make copies of the dependent variable Target visits and errors and edit the arms that are considered targets. Rename the variables (right-click the variable name and select Rename) and give them logical names (e.g. Baited arms 1-3-6-8). In the analysis results, locate the variable that applies to specific subjects. To easily locate which subjects were assigned to which baited arms in your results table, create an independent variable in the Trial List, which specifies which arms were baited (see page 149). This variable

can then be included in the results table (in the Statistics and Charts

screen, choose Show/Hide > Independent Variable).

integrated visualization

Choose Analysis > Results > Plot Integrated Data. Select a trial from the list on the tool bar. Choose a data profile and an analysis profile from the tool bar.

heatmaps

Choose Analysis > Results > Plot Heatmaps. Next, click Plot Heatmaps. Use heatmaps to visualize the location of the subject during the test.

calculating statistics

Choose Analysis > Results > Statistics & Charts. Select an analysis

profile from the list on the tool bar, then click Calculate. Choose Analysis > Export > Statistics. Choose whether you want to export the statistics per trial, or the combined statistics of groups of trials. Figure 8.3 Heatmap of a 1-minute trial in a radial maze.

---

Source: EthoVision XT 17.5 Application Manual, The Radial Arm Maze Test