News & Events


Mouse Meetings: A New Technique to Measure Social Interaction

Social interactions between mice have become a busy area of current research.  Mouse models now exist for schizophrenia, autism, and social anxiety, all of which present primarily or secondarily with abnormally low levels of social interaction (Crawley 2007; Seong et al., 2002).  Interactions involve decision-making based on perceptions, and require cognitive systems involved in reward, memory and emotion. Social interactions with a new animal typically show distinct behaviors that can change over time.  The authors of the current study sought to measure and encompass the complexity involved in a meeting between two unfamiliar mice, hoping to streamline data acquisition and improve characterization of behavioral mouse models.

Current techniques in the observation of mouse behavior typically require human observation and scoring of behaviors, paired with video recordings that are analyzed off-line.  However, the authors of the current study in Nature Methods believe this technique provides too many limitations: human observers cannot reliably analyze concurrent behaviors in more than one mouse, and there are too many behaviors that happen too quickly to be observed simultaneously.  Some can be missed even with review of video after the experiment, and the behavior of mice within close contact cannot be analyzed properly.  There is also a need to monitor behaviors as they change with time, as this could mark changes in the animals’ information processing. 

Chaumont et al. have developed mouse-tracking software called MiceProfiler, which can record mouse position, speed of mice, and resolves information about mice that are in close contact.  The software does this by using physics engines, software used in simulations and computer games that simulates physical systems and maintains tracking even during collision, when two mice are in very close contact.  After the study, the location of each of the mouse’s body parts is known for each time point.  MiceProfiler can automatically observe and categorize behaviors, identifying those sequences of behaviors that are consequences of one another.  This categorization is given automatically on a timeline, which can be scrolled through to identify labeled behaviors.  When compared with human observers, MiceProfiler automated software captured more behaviors. 

Using MiceProfiler, the authors detected changes in social interactions over time in nicotinic receptor knockout mice, and a previously unnoticed back-to-back posture, where newly introduced mice will not see each other, but will tolerate the other’s presence without directly interacting.  Having made observations and measurements previously unnoticed by the same laboratory, this technique shows promise for future research into social interactions, including in mouse models of psychiatric disorders.

What do you think?


What method do you use to score social interactions in mice? 

Do mouse models of decreased social interaction correlate usefully with human disorders? 

Can automated behavioral tracking systems replace human observation?

Further Reading:


de Chaumont F, Coura RD, Serreau P, Cressant A, Chabout J, Granon S, Olivo-Marin JC. (2012). Computerized video analysis of social interactions in mice.  Nat Methods. Mar 4;9(4):410-7. doi: 10.1038/nmeth.1924.

Crawley JN (2007) What’s wrong with my mouse?  Behavioral phenotyping of transgenic and knockout mice. Second edition. Hoboken, NJ: Wiley.

Seong E, Seasholtz AF, Burmeister M. Mouse models for psychiatric disorders (2002).  TRENDS in Genetics Vol.18 No.12. 643-650.

For a short demonstration of MiceProfiler:


Society for Neuroscience 2017 – Booth #1801 Learn More

The MedNR Forum

Notice to Customer Portal users

Customer Portal logins are now obsolete, please register with the new MedNR forum to participate