Evaluating Skilled Prehension in Mice Using an Auto-Trainer

Summary

Method to assess the impact of training on motor skills is a useful tool. Unfortunately, most behavioral assessments can be labor intensive and/or expensive.We describe here a robotic method of assessing prehension (reach-to-grasp) skill in mice.

Abstract

We describe a method to introduce naïve mice to a novel prehension (reach-to-grasp) task. Mice are housed singly in cages with a frontal slot that permits the mouse to reach out of its cage and retrieve food pellets. Minimal food restriction is employed to encourage the mice to perform the food retrieval from the slot. As the mice begin to associate coming to the slot for food, the pellets are manually pulled away to stimulate extension and pronation of their paw to grasp and retrieve the pellet through the frontal slot. When the mice begin to reach for the pellets as they arrive at the slot, the behavioral assay can be performed by measuring the rate at which they successfully grasp and retrieve the desired pellet. They are then introduced to an auto-trainer that automates both the process of providing food pellets for the mouse to grasp, and the recording of successful and failed reaching and grasping attempts. This allows for the collection of reaching data for multiple mice with minimal effort, to be used in experimental analysis as appropriate.

Introduction

Methods to experimentally test a motor skill pre- and post- neurological injury as well as modulate the timing, amount, and type of motor training are important to translational research. Over the last decade, mice, because of the attendant ease of genetic manipulation, have become a popular model system in which to elucidate the mechanisms of motor learning pre- and post- injury. However, behavioral assays in mice have not been optimized in the same way that such assays have been for other mammals (especially rats). Further, there are important differences between the behavior of a mouse and a rat that strongly suggest training the two species in different manners^1,2.

Skilled prehensile movements use a hand/paw to place food in the mouth, to manipulate an object, or to use a tool. Indeed, reaching to grasp various objects in daily life is a fundamental function of upper limbs and the reach-to-eat act is a form of prehension that many mammals use. Many of the genetic, physiological, and anatomic changes underpinning prehensile skill acquisition have been well defined in the field³. In translating preclinical findings to clinical outcomes, one needs a relevant test that is efficient and reproducible. Studies of rodent and human reaching demonstrate that prehension behavior is similar in humans and in animals⁴. Accordingly, these similarities suggest that prehension testing can serve as a translational model for investigating motor learning as well as impairments and treatments of human disease. Therefore, evaluating prehension in mice can offer a powerful tool in translational research studying both health and disease states⁴.

Unfortunately, the prehension task in mice, even for a small-scale laboratory setting, can be laborious and time consuming. To alleviate this problem, we describe here an automated version of the prehension task. The described task requires mice to extend a single paw through the mouse's home cage frontal slot, pronate the extended paw, grasp the food pellet reward, and pull the pellet back to the cage interior for consumption. The resulting data is presented as either a prehension success or failure. This automation successfully records the data and reduces the burden and time with which researchers must engage the task.

Protocol

All methods described here have been approved by the ACUC (Animal Care and Use Committee) of the Johns Hopkins University.

1. Preparing mouse cages for use

1. Create a slotted opening with dimensions of 0.8 cm width and 7 cm height from the base at the front end of each cage, as illustrated in Figure 1. This slot serves as the opening through which the animal will reach.
   NOTE: The auto-trainer was designed for the use with the standard mouse cage dimensions (as shown in Figure 1) supplied by most animal research supply vendors. Further, the auto-trainer will easily support other cage-types.
2. Inside each individual cage, add a platform adjacent to the slot to allow the mice to stand and reach the presented pellets. Ensure that the platform is located above the cage litter floor, approximately 3 cm in height. Use Petri dishes affixed with superglue and capped by a metal sheet approximately 10 cm x 15 cm, however, any flat surface large enough for a mouse to stand on to reach from will suffice.
3. Create a vertical notch through the middle of the front of the cage measuring 0.8 cm across and 7 cm high that will allow a mouse to reach his paw out of the cage.
4. From a thin sheet of metal, (approximately 2 mm thick) cut a cage gate into rectangles measuring 5 cm x 10 cm to serve as a uniform opening through which the animal is to reach.
   NOTE: Mice may chew on plastic cages which would change the size of the opening. The mouse will reach through this 0.8 cm slot when the metallic cage gate is placed over the cage's slotted opening during testing using tape, maintaining the effective width of the slot between cages.
5. Cover each cage's slot with tape when its mouse is not being tested to prevent litter from being expelled from the cage.

2. Introducing mice to the reaching motion

1. Record each mouse's starting weight and calculate 85% of that value to find their goal weight, rounding up to 20 g if the result is less. Give them a feeding regime to bring them to and then maintain this goal weight.
   1. Give each mouse 2.5 g of pellets the first day and note any change in their weight 24 h later.
      NOTE: Weigh the mice once per day and expect a weight drop of 0.25-1 g per day.
   2. Change each mouse's daily feed as required, based on this initial change and ongoing changes in each mouse's weight, in order to induce gradual weight loss (less than 0.8 g lost per day) and then maintain the resulting goal weight. Vary between three to six 500 mg pellets (1.5 to 3.0 g) per day to be effective.
      NOTE: Mice remain on this diet to maintain their goal weight throughout the protocol.
2. When a mouse has reached its goal weight, introduce each mouse to the concept of coming up to the gated slot for a supplementary food pellet. Start a training session by placing a 45 mg pellet on the pellet surface, directly in front of the slot, and allow each mouse to retrieve it. Most mice will take to this feeding arrangement within 1-2 days.
3. Once the mouse associates an open slot with being fed, encourage them to reach with a paw, rather than the mouth.
   NOTE: This is the most complex step, taking 1-2 days, and instilling counterproductive behavior in mice by mistake is very easy; please refer to the discussion section for further information and advice.
   1. Using a pair of tweezers, hold a pellet in the same position the mouse has retrieved pellets previously. As the mouse begins to bite for the pellet, pull it away approximately half a centimeter such that the pellet is out of reach of its mouth.
      NOTE: A mouse at its goal weight will attempt to retrieve the out-of-reach pellet. Whenever the mouse extends a paw through the slot, reinforce that behavior by allowing it to eat the pellet. Some mice may exhibit a preference for one paw over the other when extending for food.
   2. While not instrumental to experimentation, record whether the left or right paw is preferred. This may potentially allow for higher overall success rates in the behavioral assay; alternatively, eliminate a variable by forcing each mouse to reach with the same paw.
      NOTE: Better results are obtained if mice use their preferred paw.
   3. As each mouse associates extending a paw with eating a pellet, further reinforce that behavior by withholding the pellet in response to attempts to retrieve the pellet with the mouth and tongue. Mice will start to comply with this arrangement over 2 to 3 days.
   4. Finalize the introduction of the desired paw reaching behavior by placing the 45 mg pellet just under 1 cm from the outer edge of the cage gate, such that the leftmost or rightmost point of the pellet (whether it is to the right or left of the cage slot from the investigator's perspective, respectively) is tangent to a line extending straight out from the edge of the cage gate's slot. Allow the mouse to attempt to retrieve the pellet, being vigilant to remove the pellet and prevent its consumption if the mouse should attempt by some other method than paw extension.
      NOTE: When a mouse consistently extends a paw to grab at and is able to touch the provided pellet, it is ready for testing using the auto-trainer described below and associated behavioral assay. The time from naïve introductions to being prepared will vary between mice; if there are stragglers that take more than two weeks to understand, they should be excluded from the data set.

3. Using the auto-trainer

NOTE: Please see Figure 1-3 and the discussion section for a full description of the hardware, software, and the physical actions of the auto-trainer.

1. Prepare for the training session.
   1. Calibrate the bait pellet sensor. Click the Run arrow in the LabVIEW interface and note the bait pellet sensor reading both with and without a pellet in place. Click the Stop button to stop this test run and change the bait pellet sensor target to a value between those two readings (Figure 3 and Table 2). Most lighting conditions provide a reading between 1 and 4.
   2. Place the modified mouse cage on the auto-trainer (Figure 2). Affix the cage gate ( Figure 1) and align the pellet to the edge of the slot as in the manual procedure.
2. Run the mouse's training session using the LabVIEW interface.
   1. Input information as required to record data about the training session (Figure 3 and Table 2).
      1. Click the Mouse ID field and type the filename of each training session using the computer's keyboard.
      2. Click the Total Pellets to Dispense During Routine field to control how many pellets are dispensed for a single experiment (usually 20 - 30). To do so, click the up and down arrows or input the number using the computer's keyboard.
      3. Click the Pause After Pellet Number field to set a 5 s pause after the indicated pellet is removed from the diving board. To do so, click the up and down arrows or input the number using the computer's keyboard.
      4. Click the Pause Length field to set a pause in between the time a pellet is removed from the diving board and the time a new pellet is dispensed. To do so, click the up and down arrows or input the number using the computer's keyboard
         NOTE: Usually 1 s is an appropriate pause time. If the mice are anxious after each pellet is dispensed, it is advisable to increase the pause length using the Pause Length field to 5 s.
      5. Manually record the distance at which the pellet is placed in the Reach Distance field. To do so, click the up and down arrows or input the number using the computer's keyboard
         NOTE: The Size of the Acceleration and Time Arrays is exposed for debugging purposes and may be ignored.
      6. Click the Folder to Contain Logs field to select the file location to save the collected data.
      7. Once the information fields have been filled out, click the Run button to begin the training session. The auto-trainer will dispense individual pellets and track whether they fall through the funnel until the total number of pellets has been dispensed, and the last pellet has either been retrieved or dropped by the mouse. The program will stop automatically at this point. If necessary, it can also be stopped prematurely by clicking on the Stop button.
   2. Once the software is set up, place the home-cage of the mouse to be tested on the pedestal and observe the mouse so that you might gauge whether the mouse has indeed learned to attempt the required novel reaching behavior. After clicking the Run button, allow the mouse to investigate the slot and its new, unfamiliar surroundings.
      NOTE: Similar to when introducing mice to the concept of reaching, expect some mice to be more compliant than others. Mice that have grasped the concept should try to reach within 5-10 min and will associate the movement of the auto-trainer with the presented pellet, as when they associate an uncovered slot with food in the initial stages of this protocol.

Results

In general, it is recommended that each training session consist of about 20-30 trials, which may be set by the user, run automatically by the auto-trainer and saved into a single log file per session and mouse. Each trial can be run consecutively, right after the other, with 2-5 s of pause. Mice trained on the auto-trainer exhibit an increase in skill over 10 training sessions.

To compare the utility of the auto-trainer to manu...

Discussion

Our auto-trainer evaluates forelimb reach-to-grasp (prehension) in an automated manner. To achieve this endpoint, many of the parameters designed for the mouse prehension task, including pellet placement, pellet size, and training criteria, have been iterated over several years and adapted from prior protocols^2,5,6. The advancement here is the automation of the task using a robot that allows home-cage housing. Home-cage housing ...

Materials

Name	Company	Catalog Number	Comments
ABS Filament	Custom 3D Printed	N/A	utilized for pellet holder, frame, arm and funnel
ABS Sheet	McMaster-Carr	8586K581	3/8" thickness; used for platform compononents, positioning stand guides and base
Adruino Mini	Adruino	A000087	nano version also compatiable as well as other similar microcontrollers
Bench-Top Adjustable-Height Positioning Stand	McMaster-Carr	9967T43	35 lbs. load capacity
Clear Acrylic Round Tube	McMaster-Carr	8532K14	ID 3/8"
Low-Carbon Steel Wire	McMaster-Carr	8855K14	0.148" diameter
Pellet Dispenser	Lafayette Instrument: Neuroscience	80209-45	with 45 mg interchangeable pellet size wheel and optional stand
Photointerrupter Breakout Board	SparkFun	BOB-09322 ROHS	designed for Sharp GP1A57HRJ00F
Reflective Object Sensor	Fairchild Semiconductor	QRD1113	phototransistor output
Servo Motor	SparkFun	S8213	generic metal gear (micro size)
Transmissive Photointerrupter	Sharp	GP1A57HRJ00F	gap: 10 mm, slit: 1.8 mm