Knowledge Base

Complex Data Selections

Combining Selection Boxes

Selection boxes can be combined in a variety of ways to create complex data selections.

Basic Rules for Combining Selection Boxes

Boxes in a Linear Sequence (AND Logic)

When you line up boxes in a single sequence, the data in the Result box (and thus selected for analysis) satisfy all conditions set by the boxes (AND logic).

Example: Select the samples in the center of the open field (Nest; second box), AND only for subjects with the value of the independent variable Treatment = Injected (Filter; third box).

Boxes in Parallel (OR Logic)

When you split a sequence in two or more branches which end in the same Result box, the data in the Result box (and thus selected for analysis) satisfy either one condition or another set in those branches (OR logic).

Example: Select all samples when the animal was either Moving (Nest, top box) OR Mobile (Nest; bottom).

Multiple Filters

To Filter by Two or More Values of a Variable

Specify those values in one Filter box.

Example: A Morris water maze experiment contains a number of trials of different type, either training or probe. Training trials are marked with t1, t2, etc. according to their position in a sequence, using the user-defined variable Trial type. You want to analyze the tracks for type t1 and t2, not probes (p).

1. Create a Filter box corresponding to the filtering variable.
2. Specify the values of the variable.
3. Insert the Filter box in the sequence.

To Filter by Two or More Variables

Create one box per variable. Connect them between the Start box and the Result box using the logic you require. See Basic Rules for Combining Selection Boxes.

Example: Your experiment contains a number of tracks recorded for two groups of subjects, Young and Aged, and on a variable number of days after the treatment. You want to calculate the statistics for the Young subjects, and for Day 1 after the treatment. Since the selection criteria depend on different variables, place the Filter boxes one after the other; which comes first does not matter.

Multiple Nesting Criteria

To Nest over Multiple Values of the Same Variable

• For variables like In Zone, specify those values in one Nest box.
• For variables with more than 2 states, like Body Elongation State, Activity State, and Mobility State, create one Nest box per value, and combine the resulting boxes with OR logic.

Example 1: With the Variable In Zone

You want to visualize the track segments when the subject was entirely in one of the open arms of the Elevated Plus Maze. Each open arm has been defined as a zone.

1. Create a Nest box corresponding to the variable; in this example, In Zone.
2. Specify your selection.
3. Insert the Nest box in the sequence.

Analysis is done on the samples collected when the center point, nose-point, and tail-base of the subject were all in either open arm 1 or open arm 2 of the plus maze.

• If the zones do not overlap, from the In the Following Zones list select When in Any of the Selected Zones (OR logic).
• If the zones do overlap, and you want to select the time when the subject was in the intersection of those zones, from the In the Following Zones list select When in All Selected Zones (AND logic).

Example 2: With the Variable Activity State

You want to select the track segments when the subject was either Active or Highly Active. It is assumed here that Activity State is defined with 3 possible states: Inactive, Moderately Active, and Highly Active.

1. Create a Nest box for the variable Activity State. Under Calculate Nesting For, choose Moderately Active.
2. Create a second Nest box for Activity State. Under Calculate Nesting For, choose Highly Active.
3. Combine the Nest boxes with OR logic. See Boxes in Parallel = OR Logic.

To Nest over Two or More Variables

Create one Nest box per criterion. To refine nesting according to two or more nesting criteria, create the corresponding Nest boxes and connect them between the Start box and the Result box using the logic you require.

Example: You want to visualize the track segments when the animal was completely in the center of the open field AND its body was stretched. The center of the open field was defined as a zone.

If you want to nest over multiple values of the same variable (for example, nest over Highly Mobile AND Mobile of the variable Mobility), create one Nest box per value and connect them using the logic you require (AND, OR, or Multiple Results).

Order of Selection Boxes in a Sequence

In some instances the order in which you place selection boxes in the Data Selection sequence may be important.

Filter Boxes

• If two or more Filter boxes refer to different variables (for example, one to filter trials, another one to filter drug doses), then the order you place the Filter boxes in does not matter.
• If two or more Filter boxes refer to the same variable, then selection will only work if the elements selected in the second box are also selected in the first box.

Example: If you filter male subjects in one Filter box and female subjects in the second Filter box, the selection contains no data.

Nest Boxes

The order you place the Nest boxes relative to Filter boxes does not matter.

Source: EthoVision XT 19 Help - Complex Data Selections, Noldus Information Technology