Table 2 Frequently asked questions concerning the key properties of OptiMouse

What programs must be installed on my computer to run OptiMouse?

 MATLAB with the image processing tool box. OptiMouse has been tested on MATLAB releases 2015b and 2016a and 2016b. OptiMouse is not compatible with some older versions of MATLAB which use different coding conventions for graphical interfaces.

Which operating systems can OptiMouse run on?

 OptiMouse was developed in Windows 8. However, it should be compatible with other operating systems that run MATLAB.

Do I need to know MATLAB to run OptiMouse?

 Running OptiMouse requires a very basic familiarity with MATLAB. At the minimum, it is required to set the path and call the program from the command line. More advanced data analyses, and modifications and extension of the code, naturally require MATLAB programming skills.

Which video files formats can OptiMouse use?

 OptiMouse was tested with the following formats: mp4, mpg, and wmv. However, any format supported by the MATLAB VideoReader object should be valid. The list of supported formats is obtained by typing “VideoReader.getFileFormats” on the MATLAB command line. On MATLAB 2016b running on a windows OS, this yields:

 .asx - ASX File

 .avi - AVI File

 .m4v - MPEG-4 Video

 .mj2 - Motion JPEG2000

 .mov - QuickTime movie

 .mp4 - MPEG-4

 .mpg - MPEG-1

 .wmv - Windows Media Video

Is OptiMouse suitable for real time processing?

 No. OptiMouse is an offline analysis software.

Is OptiMouse suitable for analyzing social interactions?

 No. OptiMouse is designed for analyzing the behavior of a single mouse.

Which types of behavioral tests is OptiMouse suitable for?

 OptiMouse is suitable for any test that involves a single mouse in a stationary arena. Examples of standard tests that fall into this category are place preference tests, open field behavior, plus mazes, three-chamber tests.

Does OptiMouse identify specific body regions?

 Yes. A key aspect in OptiMouse is the detection of body center and nose positions. Some of the detection algorithms in OptiMouse also detect the tail-end and tail-base as intermediate stages of nose detection. However, the coordinates of these positions are not used in other analyses.

Can OptiMouse detect particular postures, such as rearing and grooming, or gait patterns?

 No. OptiMouse does not provide automatic detection of body postures. Manual annotation on a frame by frame basis is possible in OptiMouse.

What detection algorithms are built into OptiMouse?

 OptiMouse includes several detection algorithms, each of which includes several parameters that can be modified by the user. All of them rely on a color contrast (after the movie has been transformed to grayscale) between the mouse and the arena. Most built-in algorithms employ “peeling” of the mouse image perimeter, which allows detection of the tail and then assists detection of the nose. A detailed explanation is provided in the user manual. In addition, OptiMouse allows incorporation of custom written detection functions.

Are there any constraints on the coat color of the mouse?

 Yes. The coat color must be distinct from the arena’s background. Ideally, the entire mouse should be darker or lighter than the arena. Thus, a black mouse on a white arena or vice-versa are the ideal scenarios. However, if the mouse contains small patches of a different color on its body, this should not significantly impair detection. Current detection algorithms in OptiMouse will not perform well with a two-colored black and white animal on a gray background.

I have an algorithm that works really well on my videos. Can I use it in OptiMouse?

 Yes. OptiMouse is designed to incorporate other detection algorithms. Custom algorithms are “declared” in one of the OptiMouse folders, and then are essentially incorporated into the user interface. Custom functions accept image data and other optional parameters, including user defined parameters. It is even possible to set user-defined parameters (which are not part of the current detection algorithm) graphically via the OptiMouse GUI. Custom written algorithms must, at minimum, return body and nose positions. See the user’s manual for a more detailed explanation of how to write and incorporate custom algorithms into OptiMouse.

How long will it take me to obtain results from a video file?

 The answer obviously depends on many factors. Under ideal conditions, which require minimal user intervention, the entire procedure to process a 10 minute video may require about 20–30 minutes. The actual values also depend on video frame rate, resolution, and computer processing speeds. Most of the time is spent on automated processing not requiring any user input. Such processing can be performed in batch mode, so that the actual user time (with minimal user intervention) is a few minutes.

Which stages require most user intervention?

 The two most time consuming stages are setting detection parameters and reviewing the video after detection has been performed. In both cases, the required time depends on the video quality. Videos in which the mouse is always easily separated from the arena facilitate both stages. Videos with variable conditions and poor image signal-to-noise ratios will require more tweaking of the detection settings. Setting detection parameters involves browsing the video and deciding on a number of detection algorithms. This typically requires a few minutes for each video (settings can be saved and applied to other movies if they share similar attributes, thus reducing the time required by the user). Reviewing of the movie can be a lengthy process and depends on the performance of the detection algorithms and the desired accuracy. The reviewing stage includes multiple tools to easily identify frames with erroneous detection, as well as frames in which the mouse is in particular parts of the arena, allowing the user to focus reviewing efforts on the frames that matter most.

Can behavioral data be synchronized with other data (e.g., electrophysiological data)?

 OptiMouse does not include a built-in synchronizing signal. However, the Results file contains a frame by frame account of various parameters such as body and nose position, body angle, speed, presence in a certain zone, and occurrence of annotated events. The Results file also contains a time stamp for every frame so that, if the first video frame is synchronized with other non-video data, all other OptiMouse values can be aligned as well.

Can I modify OptiMouse beyond the addition of custom detection algorithms?

 OptiMouse is stored in Github, with the hope that this will facilitate a community based development of the code. OptiMouse is written in MATLAB and individual code files (m-files) are annotated. The user manual provides a description of file formats, algorithms, and data conversions. The graphical user interface has been designed with the MATLAB GUIDE tool, which can be used to modify the existing interfaces and the code associated with various controls.

Which analyses are implemented in the Analysis interface?

 In the analysis stage there is a distinction between zone dependent and zone independent analysis. Without defining zones, OptiMouse can generate graphical displays of positions (as tracks or heatmaps), speeds, and body angles as a function of time. If events have been annotated during the annotation stage, their total occurrence during the session or their distribution in time can also be plotted. In addition, the Analysis GUI allows the definition of zones (of arbitrary number and shapes within the arena). Once zones are derived, the analysis of positions and events can also be shown as a function of zone entries.

Can I implement my own data analyses on positions detected by OptiMouse?

 Yes. The Position file contains frame by frame mouse positional data as well as user annotated events. The Results file also contains zone-related information. Both files are MATLAB data files (*.mat) which can be used for more complicated analysis. A detailed description of the Results file format is provided in the user manual. In addition, the user manual provides example code for the analysis of freezing episodes from the Results file.

Does OptiMouse support batch processing of position data files?

 No. However, OptiMouse contains an option for adding tags to individual files. These tags then allow grouping of files for statistical comparison of groups of files. See the manual for more details on the use of “experiment tags”.

How do I get more information about OptiMouse?

 OptiMouse includes a detailed user manual. For issues that are not covered in the manual, contact the corresponding author of this manuscript by email.