Measuring Motivation Using the Progressive Ratio Task in Adolescent Mice

Abstract

Alterations in reward seeking are a hallmark of multiple psychiatric disorders including substance abuse and depression. One important aspect of reward seeking is ‘wanting’, which can be operationalized in both humans and rodents in tasks like the progressive ratio, in which an increasing amount of work is required to earn a given reward. Importantly, many disorders with reward seeking deficits are believed to have an important neurodevelopmental component, underscoring the importance of being able to study changes in motivation across the lifespan. While this task has been adapted for both adult and adolescent rats, in mice it has predominantly been used to assay motivational changes in adults. Specific concerns adapting this task from adult to adolescent mice include: 1) Optimizing a food restriction paradigm suitable for growing animals whose weights are naturally dynamically changing and 2) Identifying task conditions that allow younger and smaller mice to perform the task while minimizing the length of the behavioral shaping required to measure motivation at specific developmental dates. Towards that end, we now report a protocol for appropriate weight management for developing animals that require food restriction, and a protocol for behavioral shaping and progressive ratio testing in adolescent mice, including an assessment of whether the animals perform better with lever presses or nose pokes as the required operant response.

Basic Protocol 1:

Food restriction and weight management in the context of developing mice

Basic Protocol 2:

Operant box design, progressive ratio training, testing and data analysis in adolescent mice

INTRODUCTION:

Deficits in reward seeking, including impairments in motivation, are seen across multiple psychiatric disorders, and are a core symptom for diagnosis of major depression (Admon & Pizzagalli, 2015; Wang et al., 2021). To assess motivation in a mouse, researchers commonly use a progressive ratio (PR) task, where the effort needed to retrieve a reward increases across trials (Johnson et al., 2022). Most studies using the PR task have been conducted in adult mice. However, circuits supporting motivation develop across the lifespan, and it is important to be able to study this behavior longitudinally over the course of development, including in the adolescent period. Therefore, it would be advantageous to study reward seeking in an adolescent model.

Adaption of the PR task to adolescent mice requires consideration of differences between adolescent and adult animals. For example, adolescent mice may not have the strength to press a lever the way adult mice do, or the endurance to complete a comparable number of trials. Additionally, adolescent mice may have a harder time learning, and may require more training to learn the association between the operant behavior and the reward than their adult counterparts. Moreover, when the reward that is retrieved is a palatable substance, mice are typically food deprived to standardize satiety state across all animals during testing. Body weight percentages are easily calculated in adult mice, where their baseline weights remain relatively stable. By contrast, adolescent mice continue to grow during the testing period, and researchers must account for their dynamically changing weight when determining the extent of food restriction.

In our protocol, we utilize a classical operant-box, progressive-ratio paradigm, and modify the timeline of the training period to account for learning differences between adult and adolescent mice. In addition, we introduce two methods of maintaining appropriate body weights during food restriction. Our first method utilizes a standardized weight chart from Jackson Laboratories, while the second method introduces age-matched, non-food deprived adolescent mice to serve as weight controls to properly manage food restriction in the context of growing mice. Finally, we describe alternative operant actions, nose poking and lever pressing, that can be programmed into the test layout for more efficient training.

In this protocol, we separate our method into several key aspects: food restriction and weight management, where we describe daily weigh-ins and feedings (Basic Protocol 1); and operant box set-up and testing timeline, which describes alternative port functions for nose poking and lever pressing and additional training days (Basic Protocol 2). Our step-by-step testing protocol, coupled with appropriate statistical analysis, is a reliable and reproducible tool for investigating motivational differences in adolescent mice.

STRATEGIC PLANNING

Mice

The use and care of the animals must be conducted in agreement with the appropriate institutional and/or national review boards. Mice for testing can be either bred by the laboratory or obtained from a commercial supplier (e.g., Jackson Laboratories, Bar Harbor, ME, USA). Both sexes can be tested using the following protocols, with slight differences during food restriction that are highlighted below. Both commercially obtained, and in-laboratory bred C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME, USA; stock #000664) were used with these protocols. The following protocols are designed for adolescent mice and based on the length of training described in this protocol, the investigator should work backward to determine when to begin training to be performing the progressive ratio task (final 3 days of proposed schedule) during the desired age window of assessment.

For food restriction, baseline weights (prior to restriction) must be recorded. Mice should be food restricted at least one day prior to testing; it is useful to begin food restriction two or three days prior to testing so that mice are adequately motivated and so that observation of appropriate food amounts to maintain 85% to 90% body weight can be assessed before testing. Test mice and weight-control mice (for Alternate Protocol 1) should be housed separately, and single housing should be avoided in both groups. Each test mouse should be matched with a weight-control mouse of the same sex, age, and approximate baseline weight (for Alternate Protocol 1). Additionally, each test mouse can be sex-matched with its average projected growth as documented by a growth chart (e.g., Body Weight Information for C57BL/6J: https://www.jax.org/jax-mice-and-services/strain-data-sheet-pages/body-weight-chart-000664), which was the procedure we initially utilized (For Basic Protocol 1).

For our experiments, the use and care of animals was conducted in agreement with guidelines approved by the Institutional Animal Care and Use Committees at Columbia University and the New York State Psychiatric Institute.

Rewards

This protocol is designed to work with any palatable reward. The operant boxes used in our experiments have troughs, which can be filled with any liquid reward. Evaporated milk (e.g., Carnation brand, Nestle, Virginia, USA) was used in this experiment because it offers mice a calorie dense reward and is easy to clean. Other palatable rewards that have been used in the literature in adult mice include diluted sweetened condensed milk, sucrose (solution or pellets) or high fat-chow. Ideally the reward should be either nutritionally dense, or rewarding by taste (i.e., sweet reward).

BASIC PROTOCOL 1

Food Restriction and Weight Management in the Context of Developing Mice

Adolescent mice would naturally increase in body weight as the testing period progresses, complicating the question of how to calculate their percent bodyweight during food restriction. Here we propose two effective possible protocols: age-matched non-food restricted controls and the use of a growth chart. Calculation of percent baseline bodyweight requires knowledge the weight of animal on a given day (numerator) and the weight they would have been if they were not food restricted (denominator); this ratio is then multiplied by 100 to yield the percent baseline weight. In adults where growth has stabilized, the denominator reflects the weight of the animal before they began food restriction. Because adolescent mice would naturally gain weight at a rapid pace during the period of food restriction, using an adolescent animal’s own bodyweight prior to food restriction is not sufficient. Instead, it is necessary to extrapolate the weight gain for each mouse based either on the gain seen in their age and sex-matched not food restricted control or on a growth chart (such as the one provided for developing C57BL/6 mice by Jackson laboratory: https://www.jax.org/jax-mice-and-services/strain-data-sheet-pages/body-weight-chart-000664).

Materials:

Digital scale (accuracy to at least a 10^th of a gram, e.g., Ozeri ZK14-B Pronto Digital Multifunction Kitchen and Food Scale, San Diego, CA, USA)

Mouse chow (Prolab Isopro RMH 3000 Irradiated, PMI Nutrition International LLC, MN, USA)

Projected growth chart: (https://www.jax.org/jax-mice-and-services/strain-data-sheet-pages/body-weight-chart-000664)

1. Weigh mice daily prior to testing in a laminar flow hood using a digital scale.

2. Note weights for each individual mouse.

3. For each test mouse, compare the projected growth for a mouse of its age to its calculated weight.

   • Baseline weight + projected weight gain (growth chart) = Revised baseline weight.
   • This step can be done concurrently with Basic Protocol 2.
   • Ex. Mouse A (C57BL6/6J male mouse) has a baseline weight of 20.0 g at P25. At P26, Mouse A is projected to gain around 0.2 g (Jackson Laboratory Body Weight Information for C57BL/6J). At P26, Mouse A’s revised baseline weight is: 20.0 + 0.2 = 20.2 g.
4. For each test mouse, calculate body weight percentage using their calculated weight compared to their revised baseline weight.

   • [Calculated weight (test) / revised baseline weight (test)] x 100% = Body weight percentage.
     Ex. At P26, Mouse A weighs 18.2 g (calculated weight). Using the revised baseline weight from the previous step, Mouse A’s body weight percentage at P26 is: 18.2 / 20.2 × 100% = 90%.
5. Determine appropriate amount of food for each cage based off an average of the body weight percentages of the mice in that cage.

   1. If one mouse in a cage has a body weight percentage less than 85% and the rest of the cage is at an appropriate body weight percentage, separate the underweight mouse after testing and give full free access to food for an hour. Re-weigh and re-calculate body weight percentage. If still under 85%, separate the mouse overnight and give full free access to food. Re-weigh and re-calculate body weight percentage the following day. Mice that are chronically underweight despite continuous overnight full free access to food may need to be removed from the study.
   2. The appropriate food amount varies significantly among breeds, cohorts, and even cages. Use knowledge about weight gain/loss from previous days to inform determinations about the correct amount of food for each cage (it is helpful to start food restriction prior to testing for this reason).
   3. Male mice need about 0.2 g more food per day per mouse than female mice.
      Ex. At P26, Mouse A’s body weight percentage is 90% (as calculated in the steps above). The 3 other mice housed in the same cage are also at (or around) a body weight percentage of 90%. An appropriate amount of food to give would be 4 pellets (one per mouse) each weighing 2.4 g. If the body weight percentages stay at (or around) 90% at P27, 4 pellets of 2.4 g would continue to be given. On average, pellet weights are changed by 0.2 g for a change in body weight percentage of 5%, but this may need to be modified empirically for each experiment.
6. Weigh out food pellets using a digital scale.

   • Pellets for the same cage should each weigh the same amount (do not individualize pellets for mice within the same cage).

7. Place food pellets in each cage in the afternoon/evening after mice are done testing for the day.

Please see information about sample data in the Understanding Results section (below).

ALTERNATE PROTOCOL 1

Food Restriction and Weight Management of Developing Mice Without Projected Growth Chart: Utilization of Baseline Mice

The protocol using a growth chart is effective and more appropriate if there are constraints on the total number of animals that can be used. However, the growth chart only has weekly weight gain numbers, and since body weight needs to be calculated daily, there are limitations to its accuracy. Use of weight control mice provides a daily estimation of animal growth. However, accurate estimation with this method requires at least 2–3 age and sex-matched controls for every animal included in the study, which increases costs, space required, and animals needed. [*Copy Editor: We moved this from the end of the Basic Protocol 1 introduction here to serve as the AP1 intro. Please ask the authors if this is OK.]

1. Weigh mice daily prior to testing in a laminar flow hood using a digital scale. This should be done at approximately the same time each day.

2. Note weights for each individual mouse.

3. Compare the weights of the weight-control mice to their baseline weights.

   • Calculate weight (weight-control) – baseline weight (weight control) = Weight gained
   • Ex. Mouse B-Control (C57BL6/6J female mouse) has a baseline weight of 18.0 g. At P26, Mouse B-Control has a weight of 18.1 g. Mouse B-Control’s weight gain is: 18.1 – 18.0 = 0.1 g.
4. For each test mouse, add the weight gained by its matched weight-control mouse to its baseline weight.

   1. Baseline weight (test) + weight gained (weight-control) = Revised baseline weight.
   2. If weight gained is less than zero, do not add this number to test mouse baseline weight.
      Ex. Mouse B-Test has a baseline weight of 18.3 g. Using the calculated weight gain from above, at P26, Mouse B-Test’s revised baseline weight is: 18.3 + 0.1 = 18.4 g.
5. For each test mouse, calculate body weight percentage using their calculated weight compared to their revised baseline weight.

   • Calculated weight (test) / revised baseline weight (test) x 100 = Body weight percentage.
   • Ex. At P26, Mouse B-Test weighs 16.5 g (calculated weight). Using the revised baseline weight from the previous step, Mouse B-Test’s body weight percentage is: 16.5 / 18.4 × 100% = 90%.
6. Determine appropriate amount of food for each cage based off an average of the body weight percentages of the mice in that cage.

   1. If one mouse in a cage has a body weight percentage less than 85% and the rest of the cage is at an appropriate body weight percentage, separate the mouse after testing and give full free access to food for an hour. Re-weigh and re-calculate body weight percentage. If still under 85%, separate the mouse overnight and give full free access to food. Re-weigh and re-calculate body weight percentage the following day. Mice that are chronically underweight despite continuous overnight full free access to food may need to be dropped from the study.
   2. The appropriate food amount varies significantly among breeds, cohorts, and even cages. Use knowledge about weight gain/loss from previous days to inform determinations about the correct amount of food for each cage (it is helpful to start food restriction prior to testing for this reason).
   3. Male mice need about .2 g more food per day per mouse than female mice.
      Ex. At P26, Mouse B-Test’s body weight percentage is 90% (as calculated in the steps above). The 3 other mice housed in the same cage are also at (or around) a body weight percentage of 90%. An appropriate amount of food to give would be 4 pellets (one per mouse) each weighing 2.2 g. If the body weight percentages stay at (or around) 90% at P27, 4 pellets of 2.2 g would continue to be given. On average, pellet weights are changed by 0.2 g for a change in body weight percentage of 5%, but this may need to be modified empirically for each experiment.
7. Weight out food pellets using a digital scale.

   • Pellets for the same cage should each weigh the same amount (do not individualize pellets for mice within a cage).

8. Place food pellets in each cage in the afternoon/evening after mice are done testing for the day.

Please see information about sample data in the Understanding Results section (below).

BASIC PROTOCOL 2

Operant box design, progressive ratio training, testing, and data analysis in adolescent mice

This protocol details a classical operant-box set-up, and how to switch out levers for nose-poking ports. We also describe the timeline of operant training, and consecutive testing days that are intended to have mice exerting increasing levels of effort as the difficulty of the task increases. In this section, problems that may occur during operant training and testing are addressed, and we go into detail how we ensure mice learn to press or poke for reward. Behavioral outputs obtained from operant-box tasks are saved through the MEDPC program and then analyzed using custom MATLAB scripts to extract experimental outputs.

Strategic Planning:

Daily Programs

To appropriately train mice for the progressive ratio (PR) task, a sequence of trough training, continuous reinforcement (CRF), and random interval (RI) training days is required. Our testing consists of two days of trough training, two days of CRF 7 seconds, two days of CRF 5 seconds, one day of RI05, one day of RI10, one day of RI15, three days of RI20, and three days of PR. The details of each day are listed in the table below (Table 1). Trough training is designed to teach awareness about the presence of a reward and to teach an association between the sound of the dipper being lifted and the reward. Continuous reinforcement is designed to teach the association between the operant response and the reward. Random interval days are designed to increase lever pressing/nose poking and habituate mice to longer periods of time between rewards. This protocol uses the right lever or the nose poke port, but either left or right lever can be used.

Table 1.

Daily Training and Testing Parameters

Session Day	Session Timeout	Maximum Dippers (Rewarded)	Presses Required for Reward	Details/ Design	MEDPC Program
Trough Training (D1)	40 minutes	40	No requirement	Dipper remains up until a head entry is recognized at which point it will be lowered after 7 seconds and then raised again	DipTrain_Day1
Trough Training (D2)	1 hour	60	No requirement	Dipper remains up until ahead entry is recognized at which point it will be lowered after 7 seconds and then raised again	DipTrain_Day2
CRF 7s (D3 and D4)	1 hour	60	Yes; 1press	Dipper goes up after 1 lever press. Dipper remains up for 7 seconds before returning.	CRF_7s
Overnight (ON)	Up to 7 hours	Unlimited	Yes; 1 press	Dipper goes up after 1 lever press. Dipper remains up for 7 seconds before returning. This is an extended run for mice who have yet to learn to associate nose pokes/ levers pressing with reward.	Overnight
CRF 5s (D5 and D6)	1 hour	60	Yes; 1 press	Dipper goes up after 1 press. Dipper remains up for 5 seconds before returning	CRF_5s
RI05 (D7)	1 hour	60	Yes; 1 press	A reward is given an average of every 5 seconds if the lever is pressed at least once. Once a reward is earned, the dipper goes up for 5 seconds	RI05 or RI05_Noseport
RI10 (D8)	1 hour	60	Yes; 1 press	A reward is given an average of every 10 seconds if the lever is pressed at least once. Once a reward is earned, the dipper goes up for 5 seconds.	RI10
RI15 (D9)	1 hour	60	Yes; 1 press	A reward is given an average of every 15 seconds if the lever is pressed at least once. Once a reward is earned, the dipper goes up for 5 seconds.	RI15
RI20 (D10, D11, D12)	1 hour	60	Yes; 1 press	A reward is given an average of every 20 seconds if the lever is pressed at least once. Once a reward is earned, the dipper goes up for 5 seconds.	RI20
PR (D13, D14, D15)	2 hours, or after 3 minutes without press	No limit	Yes; exponential increase (2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048)	A reward is given after the lever is pressed 2 times, then 4 times, and so on. Once a reward is earned, the dipper goes up for 5 seconds	PR

To advance through the two trough training days mice have to retrieve 75% of rewards (30/40). Mice continue with trough training days until they have met this criteria, although in our experience two days is always sufficient. Once trough training days are completed, and mice have met the set criteria, we proceed to CRF 7s training, where lever pressing/nose poking is introduced. Mice have to make a single response in order to complete a trial (dipper is presented). Mice have up to one hour to complete the maximum amount of trials, which is 60 dippers. For the CRF 7s training days, we set two criteria. First, mice must complete at least 45 of the 60 trials within the 60-minute session. Second, of those trials completed, mice cannot miss more than 10 dippers. This shows that they are learning that an operant response results in reward delivery. This criteria is necessary in order to proceed to subsequent training days. If a mouse does not learn to associate the operant response with the reward (less than 45 rewards/dippers earned, or more than 10 of the 45+ rewards were missed), we implement an overnight (ON) training session. Overnight sessions are started the evening after a CRF 7s day if we have determined specific mice will need extended time to learn. Mice are given a small portion of food in the early afternoon (1.5 grams per mouse), which will ensure that mice are not underweight, but still hungry when placed back into the boxes. The overnight session works the same way as a CRF training day, where a single press or nose poke earns a reward, but the maximal session time is extended to run as long as needed (most frequently for the entire night or approximately 10–12 hours). This session is conducted in the evening hours when mice are more active, and the extended times allows for more opportunities to learn and receive reward. Overnight runs are usually done for only one night, but can be repeated after each CRF 7s day, and for one or two days after CRF 7s testing is completed if mice have still not learned the stimulus outcome association. This is encouraged if the cohort is relatively small and dropping mice from the cohort may underpower the experiment. Mice are typically dropped only when the majority of the cohort has learned to press or poke for reward, and the experimenter has exhausted 4 or more days of overnight testing.

Once mice have completed CRF 7s (and overnight training if needed), they can move on to CRF 5s training days. CRF 5s differs from CRF 7s in that the dipper-up time shortens from 7 seconds to 5 seconds (see Table 1). We typically require that mice retrieve at least 55 rewards to complete CRF 5s and move onto subsequent testing days; if mice do not, repeat CRF 5s testing until all animals reach this criteria.

After CRF 5s training is complete, mice move on to random interval training days. In random interval training there is a delay between the time when the mouse earned their previous reward and when mice can respond to earn the next reward, regardless of whether they make an operant response during the delay interval. For example, RI05 training describes a randomly selected delay interval with an average duration of 5 seconds. As the training days progress, the interval duration gradually increases until RI20, when the average delay interval is 20 seconds. This means that a mouse on average will have to wait for 20 seconds before an operant response will earn them a reward. Throughout the RI training days, mice should continue to earn most, if not all rewards, with very few missed dips, if any. By randomizing when an operant response will earn a reward, mice increase overall levels of operant responding. One should monitor the total number of operant responses made by mice across RI training days; regardless of the absolute number of responses, it should continue to increase across days for each individual mouse. If overall response rate is very low, but mice continue to earn the maximum number of rewards and limit missed dips, this could be a behavioral phenotype. However, mice who do not demonstrate high levels of pressing/poking by the final RI20 day are expected to do poorly in the progressive ratio test.

The progressive ratio test evaluates differences in effortful behavior and motivation. The test begins with a press/poke requirement of 2 (i.e., 2 presses or pokes to earn the first reward). This response demand increases exponentially from 2, to 4, to 8, and so forth, until a final ratio requirement of 2048 presses/pokes is needed for reward. As the ratio requirement increases, the time it takes to make the required number of responses also increases, and therefore the delay between rewards gets larger as the session goes on. Mice who do not press or nose poke at least once within 3 minutes will have their session end. In this protocol we repeat progressive ratio testing across three days to test stability of responses over time. However, operant responding often extinguishes across sessions, in which case it may make more sense to limit testing to one day.

Materials:

Reward (evaporated milk, Carnation brand, Nestle, Virginia, USA)

Operant box system (e.g., Med-Associates, NY, USA, or equivalent)

Single operant box set-up includes modular chamber, grid floor, waste pan, switchable liquid dipper with dipper cup and head entry port, lever panel, nose poke response key or secondary lever panel, back nose poke response key, LED house light, sound attenuating cubicle (medium density fiberboard), and a SmartCtrl Connection Panel.

Using the SmartCtrl Connection Panel, connect each component of the operant box (i.e., levers and nose ports, dipper, and light) to a single output port.

It is important to be consistent with which component connects to each control panel port. See MEDPC programs (Supporting Information) to match your pairings with the outputs in the program.

All boxes are then connected to a master power supply source and a computer.

We used the Interface Cabinet+Power Supply (Tabletop) paired with SmartCtrl Interface Modules (one per box), and a single USB Decode Card.

Cleaning wipes (Super-Sani-Cloth (55.00% isopropyl alcohol) or equivalent)

Computer: any Desktop PC will work

We placed the computer desktop, keyboard and mouse, and SmartCtrl Interface Cabinet on a sturdy desk that is easy to access.

USB for the purpose of transferring data to a secondary computer for analysis

Not necessary if you plan to do analysis on the same computer connected to the operant boxes.

MEDPC programs (Supporting Information)

Note: we’ve included a test program for the function of all components in the operant chambers, programs for dipper training days 1 and 2, CRF 7s and 5s training programs for lever pressing, RI training day programs for lever pressing, Progressive ratio programs for lever pressing, and an example of RI05 program for nose port specific training. All programs provided are written for ‘continuous’ mode in the operant boxes. Boxes can be switched from continuous mode to ‘toggle’ mode, and vice versa. Check that your boxes are in the correct mode prior to running programs.

Matlab (Version R2022a)

Please see script files in Supporting Information titled ‘Developmental_Progressive_Ratio_Task_Analysis’, ‘Raw_Data’, and ‘Analysis_Functions’

GraphPad Prism 9.4.1

Operant Box Preparation

• 1
  Fill troughs with evaporated milk and place under dippers.

  • Troughs should be full enough to ensure that the head of the dipper is fully submerged in evaporated milk throughout testing. In our set up, 45 mL per trough was sufficient.
• 2Turn on the computer and the program box. Turn on system using the switch. You should hear the boxes start and see the switch light turn on. Click on the MEDPC icon on your computer screen.
• 3
  Run a test trial on each operant box.

  1. Select the trial program you wish to use.
     In the case of our experiments, we used MEDPC program: ‘Testchambers’ (see program files in Supporting Information).
  2. Select the boxes you want to run for the day, and then run the program.
     This test program ensures that the lever/nose poke, head entry hole, lights, and dipper are all functioning normally.
     The program requires you to manually press the left lever, press the right lever or insert finger into the nose port, and insert finger into the head entry hole. The order goes:

     $$\mathit{\text{Left lever}}\to \mathit{\text{head entry hole}}\to \mathit{\text{right lever}}/\mathit{\text{nose port}}\to \mathit{\text{back head entry port}}.$$
     When pressing the levers or inserting a finger into the nose port and back head entry port, you should see the dipper go up and down (for each behavior). If this does not occur, restart the computer, and try again. Make sure that throughout the test program, the lights remain on, and that the dipper does not get stuck in the upright position.

     • If the dipper does get stuck, this is either due to improper placement of the trough, which you should readjust, or because the dipper is sticking to reward port due to dried milk. If this is the case, please turn off the boxes and clean the dipper areas thoroughly before retesting.
• 4
  Select the program you wish to run for the day. This can be done by clicking on each box’s respective cell (found on the MEDPC screen) and selecting the appropriate MEDPC program from your list. See list of MEDPC programs provided in Supporting Information.

  • For example, if you would like to run CRF 5s for your training day, please choose MEDPC option ‘CRF_5s.’

Once you have selected the appropriate program you wish to run in each box, you can add additional information that will be saved, such as animal I.D./ear tag, and a title for your experiment.

Testing

• 5The room should be quiet and dark throughout the test to minimize distractions for the mice.
• 6
  After adding any additional information, place mice into their respective boxes, click ‘Run’.

  • A window will pop up on the screen, where you can select which boxes to run, and then click ‘issue’.

  Note: if you do not require all boxes to run, only select the boxes in which you have mice.

• 7
  Record any observations you make regarding the boxes or mouse activity.

  • You can look into the port holes to view the mouse’s activity, but it is not advised to open the doors to the boxes during run time, as this will distract the mouse and skew results.

Data Collection and Clean Up

• 8After completion of the test, save the data output file.
  You can select which boxes to save, as well as which to abort. This option is helpful if any errors occur during testing, and you no longer wish to use an individual mouse’s data.

  • Data will save as a notes file unless you wish to save in another format. For our analysis, we kept the original formatting.
• 9If running multiple groups in one day, wipe down the inside of the boxes between each group.
• 10At the conclusion of the day’s testing, save data from the last run, turn off boxes using the switch on the program box, and remove remaining reward from the boxes and wipe down the inside of the boxes and the dipper with sanitizing wipes.

  Ensure that the dipper is well cleaned. Often, problems with dipper function can be attributed to a buildup of liquid reward.

Data Processing

• 11
  Prior to data collection, download Matlab scripts for data analysis. Three Matlab scripts have been provided with this manuscript:

  1. ‘Developmental_Progressive_Ratio_Task_Analysis’ is the master program in which the user specifies the path to the set of data they wish to analyze.
     This program will go through the pipeline of extracting the data from MedPC to Matlab format using ‘Raw_Data’ function.
     The master program also analyzes the extracted data using the ‘Analysis_Functions’.
  2. The ‘Raw Data’ program transforms the raw data from its MEDPC format to Matlab format.
     It is important to note that this program will only transfer formats for the files included in your path folder. You should create an analysis folder containing folders for each training day. Within each training day folder should be separate folders representing each mouse. This makes it easier to compartmentalize your data into fewer folders, and for the master program to analyze all training days in a single run.
  3. ‘Analysis_Functions’ computes the different behavioral variables of interest from the timestamps of behavioral events.
     It is important to keep record of which mice was trained to nose poke versus mice who were trained to lever press.

  The combination of the three scripts allows for the investigator to analyze large files, select specific training days to analyze, and organizes analysis outputs into a uniform list that is structured by mouse IDs.

• 12Once data has been collected, organize data sets by respective training days.
  Make appropriate folders that the Matlab script can direct to perform the analysis. In our setup, we make a folder for the specific cohort. Within this folder we had an Analysis folder containing folders for each training day. Within each training day folder we had folders for each animal. We placed MEDPC files into their respective animal folders.

  • The purpose of this organization is to allow Matlab to work in a loop that minimizes the input from the user. Instead of having to manually select and analyze each file individually, the program will analyze all mice for each day, and all days within the path folder.
  • On top of this, the programs work with this pathway set-up to format the outputs as a matrix, which facilitates the transfer of the data to GraphPad.
  • In addition to this, by creating individual folders for each mouse, the Matlab script provided will create an output final for every single mouse. This is very helpful for experimenters in the case of wanting to evaluate results from a given mouse in a cohort, rather than comparing results across all mice. It is advised to title individual animal folders by the appropriate mouse ID.
• 13
  After folders have been made, and data has been organized correctly, add your path to the Matlab script. Specifically, add the pathway up to the point of the ‘Analysis’ folder in each training day file.

  • This allows for the script to parse through all the files within the Analysis folder and organize data outputs by mouse.
• 14
  Once the pathway has been set, to direct the Matlab script to analyze a specific training day data, under ‘corepth’, copy and paste the path to your training day folder of choice.

  • If the folders were organized as suggested above, have your core path end at the folder for the day of choice.
  • Ex) if you want to analyze ‘Day 14 PR3’, add a core path that ends at this folder.
• 15After the core path has been set, click ‘Run’ in Matlab.
  Using the script provided, you will get output files titled ‘OPERANT DATA’. Open these files to view a 1×1 construct with several fields. Each field is for a specific behavioral output measured in the training day, such as session time and missed dips.

  • Each field is organized into a matrix of a single column, and each row represents a single mouse. By trial fields have an additional matrix for each mouse that you can open in a separate window in Matlab. This is especially helpful for progressive ratio days, to observe differences in press rates and latencies with each reward earned.
• 16After all data has been analyzed in Matlab, you can copy and paste data sets into graphing software and into Excel for further statistical analysis.

COMMENTARY

Background Information:

The progressive ratio task was developed as an operant task to measure motivation to pursue a reward (Hodos, 1961). This task was a modification on a previous test to measure motivation called the ‘obstruction technique’ (Warden, 1931) in which investigators applied an aversive obstruction, such as an electrified grid, between the animal and a palatable reward (i.e., food). It was believed that the maximal electric current the animal was willing to cross to retrieve the reward would imply a measurement of maximal motivation, which they termed as the mouse’s ‘breaking point’. However, while the obstruction technique was unable to generate reliable and stable breaking points, the progressive ratio task can produce stable break points, can be repeated across many days, and avoids harm to the animals caused by repeated shocks. Validation for the task as a measurement of motivation derived from the fact that changes in factors known to influence motivation (e.g., changes in appetitive nature of the reward or the satiety state of the animal) reliably affected the ‘breaking points’ measured in animals. Moreover, the automated and operant nature of the task confers substantial stability across days as well as task administrators.

There are multiple other tests that can be done to evaluate motivated behavior in mice, but most alternative experimental set-ups come with drawbacks for use in adolescence that are addressed in our protocol. For example, assessing motivation in an exploratory arena where mice must open doors of increasing weight to retrieve rewards (Spangenberg & Wichman, 2018), disregards the varying strength levels of mice across age, and would prove to be difficult in an adolescent study. Other tasks, such as one that requires that the mice jump large barriers to retrieve the reward, are useful, but again lack consideration for an adolescent mouse’s ability to perform those physical tasks. Moreover, because the task is relatively easy to learn, training is brief, and it is possible to run the task multiple times across the lifespan.

Critical Parameters:

The training and testing protocol used in this paper relies on a progressive/gradual shaping of behavior across successive sessions. It is therefore important to ensure that animals have achieved competency in each step of training before they move to the next step. During trough training, mice must have learned about the presence of a reward; this is shown by the receipt of 75% of available rewards (D1: approximately 30 of 40 rewards; D2: approximately 45 of 60 rewards). By the end of the constant reinforcement training period, mice should be receiving at least 40 of 60 rewards to demonstrate their grasp on the association between an operant response (lever press or nose poke) and a reward. Animals that continue to fail in meeting these benchmarks, should be removed from the study.

Because satiety has been established as a factor which effects motivation during the progressive ratio task, it is also important to ensure standardization of satiety across groups. Satiety state is generally standardized by maintaining an equivalent body weight percentage as well as standardizing the time since the mice last ate across groups. Both food restriction protocols described above allow for flexibility in feeding to appropriately control body weight percentage. Adolescent mice are still gaining necessary weight, which means that their development must be accounted for in the calculation of body weight percentage; this is controlled for using either the weight chart protocol or the alternate age-matched nonfood deprived control mice protocol. Mice should be at approximately 85–90% body weight, and this should be equivalent across groups. If mice fall below 85% baseline body weight, greater amounts of food should be given.

Mice vary in their activity levels throughout the day, so keeping the time of testing consistent is also important. Mice should be weighed, tested, and fed at approximately the same times each day. To keep distracting factors at a minimum, the testing room should always be dark and quiet. And it is also important to clean the boxes between sessions to decrease the distraction that the smell of another mouse might cause.

Troubleshooting:

See Table 2 (below).

Table 2.

Troubleshooting Guide for Training and Food Restriction

Problem	Possible Cause	Solution
A mouse is not meeting benchmarks during CRF (i.e., 45 operant responses or less, over 10 missed dips	Has not learned the association between an operant response and a reward	Run the mouse on an overnight session; if no improvement after session, remove mouse form cohort
Mice (as a group) are not performing well during RI and PR days (i.e., over 5 missed dips, not completing session before time limit)	1. Satiety is too high 2. Mice are over restricted; weak	1. Lower the bodyweights by giving less food 2. Increase average weight of pellets for cage(s)
A mouse (within a cage) is not performing well (i.e., over 5 missed dips, not completing session before time limit)	Satiety is too high, or over restricted	Examine body weight; temporarily separate mouse for appropriate food administration to lower or raise bodyweight
A mouse performs poorly during an overnight session (i.e., less than 200 presses within a seven-hour session)	Mouse is not interested in task or reward	1. Feed mouse after session; repeat ON session again in the next evening 2. If mouse fails to increase operant response in second ON session, remove mouse from remainder of experiment

Statistical Analysis

While it is important to track the performance of mice across training days to ensure appropriate behavioral shaping (see Figure 1), it is the data acquired during the progressive ratio days that is most important for assessing motivation. There are multiple ways to analyze the results obtained during the progressive ratio days, depending on the dependent and independent variables of interest. The first consideration is whether to average the data obtained by each animal across the 3 consecutive days of PR testing. We suggest first plotting the data by day (see Figure 2) and if the effects of the independent variable(s) of interest look consistent across days then it is appropriate to average across the PR days (see Figure 3). An advantage of this averaging is that some of the dependent variables (i.e., press rate, latency to press) should be assessed both as session averages and on a trial-by-trial basis, because the trial structure can vary quite a bit from trial to trial in the PR (see Figure 3). Averaging across PR days for each trial increases the number of mice represented in each of these trial intervals.

Figure 1. No differences between lever presses and nose pokes on task output variables across training days.

The number of (a) total presses/pokes, (b) rewards retrieved, (c) press/poke rate, (d) session time, (e) latency to retrieve reward, and (f) latency to initiate a session did not differ between adolescent animals required to make a lever press or a nose poke across the training days.

Figure 2. No differences between lever presses and nose pokes on PR task output variables.

The (a) breakpoint, as well as the number of (b) total presses/pokes, (c) rewards retrieved, (d) press/poke rate, (e) session time, (f) latency to retrieve reward, and (g) latency to initiate a session did not differ between adolescent animals required to make a lever press or a nose poke across the three progressive ratio test days.

Figure 3. Progressive ratio session averages and by trial assessments track small differences in lever pressing and nose pokes.

The averaged (a) breakpoint, (b) total presses/pokes, (c) rewards retrieved, (d) press/poke rate by trial, (e) press/poke rate, (f) session time, (g) latency to retrieve the reward, (h) latency to initiate a session, and (i) latency to initiate a session by trial did not significantly differ between adolescent animals required to make a lever press or a nose poke across the three progressive ratio test days nor within most trials in PR.

To initially compare the effect of a given independent variable across multiple PR testing days, plot a single output variable (e.g., session time) over each day, split by the independent variable of interest using a graphing software, such as GraphPad Prism. If statistical analysis is desired, use a two-way ANOVA to assess the main effect of the independent variables (i.e., day and treatment) on the dependent variable of interest, as well as if there is an interaction between these variables. This analysis can also be followed up with a post-hoc test such as a Bonferroni or Holm-Sidak to compare the effect of one independent variable on the dependent variable across levels of the other independent variable.

If the effects of your independent variable of interest look consistent across PR days, then another approach is to average your output variable across all PR days for each individual mouse and then compare this average output by your independent variable(s) of interest using a t-test or a one or two-way ANOVA (depending on the number of groups you want to compare within an independent variable or the number of independent variables, respectively).

Finally, to directly compare whether effort modulates the effects of an independent variable of interest on an output variable, we suggest comparing the effects of your independent variable on your output variable on both the low effort day (CRF) and high effort day (PR) using a repeated measures two-way ANOVA.

Understanding Results:

In the progressive ratio task, subsequent operant response requirements can increase in multiple ways. This includes a constant stepwise increase (i.e., y=2n) or steeper progressions, such as exponential increases (i.e., y=2ⁿ), where y is the requirement for presses, and n is the trial number. In the case of the proposed protocol, we went with an exponential progressive ratio, y=2ⁿ; the increase is therefore: 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, etc.

The construct of motivation can be subdivided into both response initiation and sustained action. The most classical measure of the latter from the progressive ratio task is the break point (Hodos, 1961). We define break point as the maximum operant responses (lever presses or nose pokes) that a mouse is willing to make to receive a single reward (Figure 2a, 3a). Breakpoint can be obtained by either reporting the last ratio requirement met in the PR test or the highest total operant responses a mouse exerted for a single reward trial. It is important to note that these two numbers will not necessarily be equivalent, especially when using a PR task where the number of presses required increases exponentially. In addition, break point calculations are not cumulative; you are comparing maximum effort from each trial to determine the greatest number of presses or pokes a mouse could do; it is not a summation of presses/pokes over the entire session. Break point measurements can only be assessed on PR task days, and in general, mice with reduced motivation will have lower break point averages across all PR days.

In comparison to breakpoint, which looks at the maximal number of lever presses made during a single PR trial, total lever pressing/nose poking is a measure of the cumulative operant responses made during a given session of the training and PR testing (Der-Avakian et al., 2015). Importantly, while breakpoint can only be calculated during PR, total operant responding can be calculated during all session types enabling comparisons of operant responding under high effort (e.g., PR) versus low effort (e.g., CRF) conditions. During the CRF days when only a single operant response elicits a reward, animals may make total press/poke numbers very similar to the total number of trials. The random interval days are designed to increase persistent pressing/poking and tolerance to intervals of non-reward by implementing a random delay in between the end of a reward (dipper goes down) and the time when the next reward can be earned (Figure 1a). Total pressing/poking is generally highest during PR days, likely because as rewards become more intermittent, animals become more persistent (Figure 2b). We also plot this data averaging pressing/poking across the three PR days (Figure 3b).

A related variable that can also be calculated is total rewards retrieved (Figure 1b, 2c, 3c). This parameter is a calculation of all rewards earned in a given session, minus any rewards that the animal did not retrieve. Rewards retrieved may also be expressed as a percentage of obtained (or missed) rewards across a session. As mice learn the set-up over each session, and more specifically when mice understand that lever pressing/poking earns them a reward, the number of rewards retrieved should be close to or equivalent to the number of rewards earned.

The press/poke rate, or the average number of presses/pokes in a given time period, can reflect both response initiation and sustained action (Figure 1c and 2d). During the PR days, however, rates should also be computed by trial (Figure 3d), in addition to being averaged across the session(s) (Figure 3f), because of the different requirements to earn a reward on each trial. Press/poke rates are useful measurements of the speed in which mice perform the task and may reveal differences in sustained activity during the session.

Relatedly, session time (Figure 1d, 2e, 3f) reflects the total time it takes an animal to complete a given session. For low effort, earlier training days, shorter session times reflect the shorter programmed delays between rewards. However, in the PR task, the session time relies more on the speed of operant responding, or lack thereof, as animals time out of a session after they become inactive for 3 minutes.

Latency to make a head entry upon earning a reward (latency to retrieve reward) and latency to initiate an operant response to earn a reward can be assessed to measure response initiation (Spangenberg & Wichman, 2018). In general, latency to retrieve a reward should be reduced across training days as mice learn to associate the sound of the dipper with the presentation of the reward (Figure 1e 2f and 3g). We illustrate average latencies to initiate operant responding during training (Figure 1f) as well as during PR days (Figure 2g and 3h). We illustrate this variable plotted both as an average across PR sessions (Figure 3h) and also as an average within sessions broken down by trial (Figure 3i).

There are multiple ways to interpret variations in the afore-described variables in the context of reward seeking. Lower total presses/pokes or decreased breakpoint on the PR days is frequently interpreted as reflecting decreased motivation. This might result from an inability to sustain responding under high effort conditions, or a delay in latency to initiate responding, or some combination of the two. Additionally, during PR testing, shorter session duration is typically also indicative of decreased motivation given that mice will time out of the task if they do not continue to press or poke. Shorter session duration generally correlates with lower breakpoint and fewer overall responses; however, this relationship depends on the mouse’s overall rate of responding. For this reason, session time, breakpoint, total responses and response rate should all be examined together to get a full picture of the behavioral performance of the mouse.

During training days, if a mouse retrieves few of the available rewards, seen in more missed dips/higher percentage of missed rewards, it is usually a sign that the mouse is satiated. To confirm this, look over the body weight percentages to see if mice are within an appropriate range. Comparatively, if weights are within range, and mice are still performing poorly in the PR task (i.e., greater amounts of missed dips), then this may reflect differences in hedonic reaction to a palatable reward. To test this, it would be beneficial to assess hedonic reaction in a separate task. In the case of a palatable liquid reward, one may use a gustometer/lickometer set-up; this will effectively measure hedonic response when the reward is freely available to the animal (Berridge et al., 2009; Berridge & Dayan, 2021).

Time Considerations:

Food restriction should begin at least two days prior to training. The training timeline proposed in this protocol stretches across fourteen days, including the overnight runs. Additional testing of cohorts with prior training in the operant boxes will not require early training days (Trough 1 and 2, and CRF 7s), omitting 4 days for retesting. Investigators should plan accordingly for 10–14 days of training and testing, and schedule to ensure their cohort will be of appropriate age(s) for testing on progressive ratio task days. Including food restriction days, this protocol requires approximately 16 days of hands-on time.

DATA AVAILABILITY STATEMENT:

The data, tools, and material (or their source) that support the protocol are available from the corresponding author upon reasonable request.