Knowledge Base

Dependent Variable Plots

Dependent variable plots show the values of the dependent variables selected in your active Analysis Profile on a time line. Dependent variables are displayed in different ways depending on their type:

• States like In Zone: colored bars indicating the state's duration.
• Point events like Rotation: vertical segments indicating the event's occurrence, with no duration.
• Continuous variables like Velocity: values are plotted as a time series, with one value per sample time.

Procedure

1. Define the variable and calculate statistics.
2. See Plot Integrated Data.

Scope of the Dependent Variable Plot

When the trial includes more than one arena or subject, the plot's header indicates which arena and subject it refers to.

Why Use Dependent Variable Plots?

Dependent variable plots are especially useful in the following scenarios:

• To check whether you set correct thresholds for state variables like Body Elongation, Movement, Mobility, and Activity.
  1. Open your Analysis Profile, double-click the variable's name and enter your best estimate for the thresholds.
  2. Plot integrated data. Play back your data, watch the video and check whether the states shown in the plot correspond with the states of the animals in the video. If necessary, repeat steps 1 and 2.
• To check that you have scored behaviors manually at a good level of accuracy.
  1. In your Analysis Profile, select the manually scored behaviors you want to check.
  2. Plot integrated data. Play back your data, watch the video and check whether the states shown in the plot correspond with the behavior of the animals in the video. If necessary, export the data to The Observer XT, where you can correct the errors.
• To view the behaviors of Behavior Recognition together with their probability values. This helps when you want to refine the definition of those behaviors based on their probability value.
• To check whether your Trial Control Settings do what you want them to do. You can visualize Trial Control events and Trial Control states.

View and Interpret Variable Plots

Effect of Data Selection on Plots

If the Data Profile contains nesting criteria, the selected intervals are shown with a white background. The non-selected intervals are shown semi-transparent, enabling you to see the excluded data in these intervals.

If the Data Profile contains time bins, these are not visualized.

Navigate Through the Data

Use the scrollbar at the bottom to run through the visualization. The hairline represents the current time in the trial and is always fixed in the middle of the plot area. For numerical variables, the current sample is highlighted with a circle. The current value of the variable is also shown near the variable name.

To move from the start to the end of the trial, you can also:

• Click anywhere in the plot, hold the left mouse button and drag the background.
• Hold the Shift key down and scroll with the mouse wheel. This way you zoom in/out while the cursor moves forward/backward by one time span. Note: Each wheel step corresponds to the time span chosen from the Show/Hide menu.

If part of the data plot is not visible at the current zoom level, you can either zoom out or move the plot line. Use the scrollbar at the right end of the plot area, or drag the background of the plot.

Tip: To visualize the whole vertical data range, zoom out the plot until you no longer see the vertical scrollbar.

Visualize a Single Variable Value

• For discrete variables (states or events): Each state or event is shown as a colored bar. When a variable includes multiple states like In Zone (with different zones being defined) or Mobility (with its states Immobile, Mobile, Highly Mobile), these are shown as parallel channels. Keep the mouse pointer above a colored bar. A tooltip is shown with the following information:
  • The name of the dependent variable.
  • Label: the name of the state or event.
  • Time: the sample time the mouse pointer points to.
  • Duration: the length of the state, in seconds. For point events like Rotation or Trial Control events, the duration is not shown.
• For numerical variables (including external data): Values are connected with lines. Keep the mouse pointer over the data plot. A tooltip is shown with the following information:
  • The name of the dependent variable.
  • The sample Time selected.
  • The value of the dependent variable at that time, indicated by a circle on the plot line.

Important: Although the plot lines may look continuous, the numerical variable values are not interpolated and do not become continuous. For data selection and analysis, the discrete samples are used.

Resize the Plots

• Resize the time window displayed: Point to the y-axis in the upper part of the plot so the mouse pointer turns to a double arrow. Drag the y-axis to the desired position to resize the time window.
• Resize the width of the plot: Point to the lower margin of the gray cell on the far left of the plot so the mouse pointer turns to a double arrow. Drag the margin to the desired position to have a larger or narrower plot.

Note: The data of continuous variables are stretched or condensed when you resize the plot, but the range of data does not change.

Adjust the Time Scale (X-Axis)

Do one of the following:

• Click the Zoom In or Zoom Out button on the toolbar, then click the time scale one or more times.
• Right-click the plot and select Zoom In or Zoom Out. Next, click the time scale one or more times.
• Press (Ctrl+.) and (Ctrl+,) to zoom in/out, respectively.
• Click the time scale, hold the Ctrl key down and turn the mouse wheel forward (to zoom in) and backward (to zoom out).

For the first two methods: To stop zooming in/out and return to the normal mouse pointer, click the No Zoom button on the toolbar, or right-click the plot and select No Zoom.

Notes

• Every time you zoom in/out, the plot scale is enlarged or reduced by a factor of 2.
• The shortest time that you can display is 0.1 seconds; the maximum time is seven days.
• When you zoom in/out by clicking the plot, the clicked time becomes the new current time.