Periodic Reporting for period 1 - MOUSEVILLAGE (The Mouse Village: a fully automated behavioral system for continuous social and cognitive testing)
Période du rapport: 2023-03-01 au 2024-08-31
The field of mental health drug discovery faces major challenges: many conditions lack effective treatments, and current medications are outdated. A promising solution is finding new biomarkers—measurable signs indicating a condition’s presence, progression, or treatment response. For preclinical research to succeed, these biomarkers need to be consistent across species, like humans and rodents.
However, rodent behavioral tests often differ from human diagnostics, and cognitive testing usually occurs in artificial, isolated settings. This overlooks the role of social interactions in behavior. While efforts are made to train rodents in tasks resembling human tests, these methods require lengthy training, extensive resources, and lead to high costs.
To overcome these issues, neuroscience and pharmaceutical labs are moving towards automation. Yet, most systems still involve human handling and high maintenance. Drawing from past research, we developed the Mouse Village, an open-source platform where animals live and train at their own pace, without human intervention.
The specific objectives of the project were:
Objective 1. 3D design the Mouse Village, composed of a system of connected cages with tubes equipped with antennas and a corridor giving access to the operant testing box where mice are trained in cognitively demanding task. Write all parts of the software suite that controls the behavior of the system and allows the automatic programing of the task training stages.
Objective 2. Build a few prototype systems and test them by training batches of mice of a few cognitively demanding tasks. Benchmark the automatic training and compare it with manual training in the same tasks.
Objective 3. Distribute the system with an open license system create, and distribute the Mouse Village (MV) – a new, fully-automated open platform where mice can train themselves around the clock in various cognitive tasks while freely interacting with other mice, without the need for human intervention. The MV will consist of interconnected cages leading to a touchscreen-based chamber that allows the animals to perform multiple cognitive tasks.
However, rodent behavioral tests often differ from human diagnostics, and cognitive testing usually occurs in artificial, isolated settings. This overlooks the role of social interactions in behavior. While efforts are made to train rodents in tasks resembling human tests, these methods require lengthy training, extensive resources, and lead to high costs.
To overcome these issues, neuroscience and pharmaceutical labs are moving towards automation. Yet, most systems still involve human handling and high maintenance. Drawing from past research, we developed the Mouse Village, an open-source platform where animals live and train at their own pace, without human intervention.
The specific objectives of the project were:
Objective 1. 3D design the Mouse Village, composed of a system of connected cages with tubes equipped with antennas and a corridor giving access to the operant testing box where mice are trained in cognitively demanding task. Write all parts of the software suite that controls the behavior of the system and allows the automatic programing of the task training stages.
Objective 2. Build a few prototype systems and test them by training batches of mice of a few cognitively demanding tasks. Benchmark the automatic training and compare it with manual training in the same tasks.
Objective 3. Distribute the system with an open license system create, and distribute the Mouse Village (MV) – a new, fully-automated open platform where mice can train themselves around the clock in various cognitive tasks while freely interacting with other mice, without the need for human intervention. The MV will consist of interconnected cages leading to a touchscreen-based chamber that allows the animals to perform multiple cognitive tasks.
1. Software: We have developed the Mouse Village version 2.0 new training system which includes a new software suite PyVillage running on a Raspberry Pi 5, a new software to run different cognitive tasks using a touchscreen, video tracking and water rewards. Moreover, the new Mouse Village 2.0 PyVillage incorporates a new GUI to program the schedule of the training of each individual mouse making it simple for experimentalists without programming skills to use the system. The latest version of the software can all be found in the repository https://github.com/BrainCircuitsBehaviorLab/village(s’ouvre dans une nouvelle fenêtre)
2. Hardware. The hardware used to run the Mouse Village has also been changed from a regular PC connected to an Arduino and a BPod board, to a Raspberry Pi 5 connected to BPod. In this way, the system has become more standardized and earlier to install as the Raspberry Pi guarantees homogeneity in the hardware from system to system and full compatibility with the PyVillage routine. In the past version MV 1.0 every installation in a different PC needed a different set of Python libraries, etc making the installation of all packages very cumbersome. With this new hardware implementation, the MV 2.0 leverages in the large number of Raspberry Pi extensions at a very affordable price (around 150€).
3. 3D design of the corridor. We have redesigned the corridor connecting the system of cages and the behavioral box. The new design allows for simpler maintenance and cleaning as the floor of the corridors can be easily detached and substitute by a fresh one in every cleaning. All pieces can be readily printed in a Fused Deposition Modeling (FDM) 3D printer available in many laboratories or research institutions. Moreover, we have designed a new layout for the four cages that make up the housing of the mice in the MV such that the system can all fit in a minimal space. All design file for 3D printing each piece are available at the open repository of the Training Village https://github.com/BrainCircuitsBehaviorLab/village(s’ouvre dans une nouvelle fenêtre).
4. Website and Documentation. We have created a website to distribute the system where all the information regarding installation, usage, maintenance, troubleshooting, etc is gathered in the same place. Moreover, the website is hosted by github pages, there is a built-in forum for discussion among users and developers (under the tab Issues, “https://github.com/BrainCircuitsBehaviorLab/village/issues). Together, we expect the website to provide a common place for the new versions of the code to be uploaded, where new extensions and modules will be appended and where end users will have an opportunity to ask questions and discuss problems with the developing team.
2. Hardware. The hardware used to run the Mouse Village has also been changed from a regular PC connected to an Arduino and a BPod board, to a Raspberry Pi 5 connected to BPod. In this way, the system has become more standardized and earlier to install as the Raspberry Pi guarantees homogeneity in the hardware from system to system and full compatibility with the PyVillage routine. In the past version MV 1.0 every installation in a different PC needed a different set of Python libraries, etc making the installation of all packages very cumbersome. With this new hardware implementation, the MV 2.0 leverages in the large number of Raspberry Pi extensions at a very affordable price (around 150€).
3. 3D design of the corridor. We have redesigned the corridor connecting the system of cages and the behavioral box. The new design allows for simpler maintenance and cleaning as the floor of the corridors can be easily detached and substitute by a fresh one in every cleaning. All pieces can be readily printed in a Fused Deposition Modeling (FDM) 3D printer available in many laboratories or research institutions. Moreover, we have designed a new layout for the four cages that make up the housing of the mice in the MV such that the system can all fit in a minimal space. All design file for 3D printing each piece are available at the open repository of the Training Village https://github.com/BrainCircuitsBehaviorLab/village(s’ouvre dans une nouvelle fenêtre).
4. Website and Documentation. We have created a website to distribute the system where all the information regarding installation, usage, maintenance, troubleshooting, etc is gathered in the same place. Moreover, the website is hosted by github pages, there is a built-in forum for discussion among users and developers (under the tab Issues, “https://github.com/BrainCircuitsBehaviorLab/village/issues). Together, we expect the website to provide a common place for the new versions of the code to be uploaded, where new extensions and modules will be appended and where end users will have an opportunity to ask questions and discuss problems with the developing team.
1. Enhanced Animal Welfare
The MV system significantly improves animal welfare, adhering to the 3Rs (Replace, Reduce, Refine) principles. It eliminates the need for transportation and handling by allowing mice to live in enriched environments with medium-sized groups (10-15 animals), fostering natural social behaviors. Unlike traditional setups, the MV lets mice forage for food and water by completing tasks at their own pace. A slight difference in water taste—acidic in home cages and sweetened with sucrose in task areas—motivates mice to enter the behavioral box and participate multiple times daily. Continuous remote monitoring provides real-time updates on weight, water intake, activity, and task performance, enabling timely intervention if needed. This self-paced approach reduces stress, increases successful training outcomes, and lowers the number of animals required. Long-term data collection also minimizes the size of experimental groups, setting new standards for animal care.
2. Boosting Laboratory Productivity
The MV enhances lab efficiency by generating 1,500-2,000 trials per day with minimal human effort. After setup, maintaining the system requires just 1-2 hours per week for cleaning, calibration, and manual weighing, plus about 15 minutes per day for remote monitoring. In contrast, manual training of three mice would demand 4-5 hours daily from a technician. The MV allows training of 10-12 mice over weeks with minimal manpower, making it a scalable solution for cognitive experiments.
3. Advancement in Long-Term Cognitive Testing
The MV enables groundbreaking cognitive testing over extended periods (e.g. 12 months), facilitating the study of cognitive decline and neurodegenerative diseases. Continuous monitoring captures changes as animals age, allowing researchers to track developmental stages and treatment effects. Unlike traditional tasks, which suffer from overhabituation or lengthy training, the MV supports versatile cognitive tests for consistent, long-term assessment. This capability improves the search for biomarkers and validation of animal model.
4. Improved Objectivity, Transparency, and Reproducibility
The MV automates behavioral experiments, reducing researcher bias and standardizing protocols. By quantifying every training step, it enforces blind testing and improves reproducibility, enabling other researchers to replicate the complete experimental procedures. This objectivity advances the reliability of findings across laboratories.
The MV system significantly improves animal welfare, adhering to the 3Rs (Replace, Reduce, Refine) principles. It eliminates the need for transportation and handling by allowing mice to live in enriched environments with medium-sized groups (10-15 animals), fostering natural social behaviors. Unlike traditional setups, the MV lets mice forage for food and water by completing tasks at their own pace. A slight difference in water taste—acidic in home cages and sweetened with sucrose in task areas—motivates mice to enter the behavioral box and participate multiple times daily. Continuous remote monitoring provides real-time updates on weight, water intake, activity, and task performance, enabling timely intervention if needed. This self-paced approach reduces stress, increases successful training outcomes, and lowers the number of animals required. Long-term data collection also minimizes the size of experimental groups, setting new standards for animal care.
2. Boosting Laboratory Productivity
The MV enhances lab efficiency by generating 1,500-2,000 trials per day with minimal human effort. After setup, maintaining the system requires just 1-2 hours per week for cleaning, calibration, and manual weighing, plus about 15 minutes per day for remote monitoring. In contrast, manual training of three mice would demand 4-5 hours daily from a technician. The MV allows training of 10-12 mice over weeks with minimal manpower, making it a scalable solution for cognitive experiments.
3. Advancement in Long-Term Cognitive Testing
The MV enables groundbreaking cognitive testing over extended periods (e.g. 12 months), facilitating the study of cognitive decline and neurodegenerative diseases. Continuous monitoring captures changes as animals age, allowing researchers to track developmental stages and treatment effects. Unlike traditional tasks, which suffer from overhabituation or lengthy training, the MV supports versatile cognitive tests for consistent, long-term assessment. This capability improves the search for biomarkers and validation of animal model.
4. Improved Objectivity, Transparency, and Reproducibility
The MV automates behavioral experiments, reducing researcher bias and standardizing protocols. By quantifying every training step, it enforces blind testing and improves reproducibility, enabling other researchers to replicate the complete experimental procedures. This objectivity advances the reliability of findings across laboratories.