Table 1 Key stages in OptiMouse
Stage | Goal | Comments |
|---|---|---|
Preparation | Define arena boundaries | The user must define the region of interest for analysis and provide actual arena dimensions. |
Preparation | Conversion of raw video data to MATLAB files | Once arenas are defined, the user can initiate conversion Conversion can be run in batch mode, allowing definition of multiple arenas and then running conversion for all, in a single operation. |
Detection | Determine optimal detection settings | The user may simply accept the default algorithm and initiate detection; however, in most cases, detection of body and nose positions can be significantly improved by small adjustments of the detection algorithm parameters and, often, by definition of multiple algorithms. In this stage, the movie is browsed to find detection settings that minimize errors in the detection of body and nose positions. The goal is to define a minimum number of algorithms so that at least one is appropriate for each frame of the movie. If multiple settings are defined, one of them must be specified as the default. |
Detection | Running detection | During detection, OptiMouse finds mouse positions in each frame according to the settings defined by the user. Like conversion, detection can be time consuming and thus the option for a batch mode is provided. Detection settings can be defined for multiple sessions, and then run in a single batch operation. Note that the output of the detection stage, the position files, can be immediately used for analysis manually or by using the OptiMouse analysis interface. |
Reviewing (optional) | Testing, and if needed, correcting, of the detection settings; annotation | If the stage is bypassed, then the default setting will be applied to all frames. The reviewing stage allows overriding of the default algorithm positions. Individual frames or groups of frames can be assigned with any of the non-default settings defined by the user, or set manually. OptiMouse provides multiple tools to browse the video or to locate individual frames with specific attributes. The impact of reviewing on the final position data depends on the number of errors associated with the default detection setting. A major effort has been made in OptiMouse to provide tools to efficiently locate problematic frames. Nevertheless, this stage may be time consuming. During reviewing, it is also possible to add annotation to the video. |
Analysis | Derive meaningful behavioral parameters from the position data | Several analysis options are provided at this stage via the graphical user interface. Additionally, task-specific analyses can be implemented by analyzing the Results file. In this stage, the user can define zones (within the arena) and analyze position data with respect to these zones. |