Knowledge Base

Basic Rules

For each sample in your track, EthoVision XT calculates a probability value for each behavior (see Behaviors Detected with Behavior Recognition). The sum of probability values is 100%.

All behaviors but Twitching and Hopping are mutually-exclusive states, that is, only one can be assigned to a sample. A behavior is assigned to a sample based on a number of decision rules that take into account the probability of that behavior relative to that of others, and the behaviors assigned to adjacent samples. See also the figure below.

• Twitching and Hopping have their own probability, but are not compared with that of the others. As a result, Twitching and Hopping are scored independent of other behaviors. Each may be assigned to a sample also when one of the other behaviors is scored for that sample. For example, Hopping when Walking is active.
• When two or more behaviors (excluded Twitching and Hopping) have similar probability values, and the same behaviors are not scored in adjacent samples, or when all behaviors have probability values below 25%, then none of the behaviors are scored for that sample. This results in gaps between scored behaviors when you visualize them in the Integrated Visualization.

Example

Below you see an example of how behaviors are scored based on their per-sample probability values. Top: X-Y chart of probability (here ranging from 0 to 1) of five behaviors (others are not shown for simplicity). Bottom: The resulting scored behaviors. In general, the behavior is assigned to a sample that has the highest probability (see for example Sniffing at the beginning of the time line). However, the state is smoothed based on the values in a number of samples before and after the current sample. For example, at 75 s Sniffing has a lower probability than Twitching, but the state Sniffing is still active. Note that Twitching is scored as a point event, with no duration, and is represented with a vertical segment.

The gap between scored states is caused by samples where two or more behaviors have similar probability values (for example at about 75.9 s, with Jumping and Walking having probabilities around 0.5) or the probability is below 0.25 for all behaviors. The example is taken from Rat Behavior Recognition data.

Behavior, "Not" Behavior and "Unknown"

Each behavior type, except Twitching and Hopping, has three complementary, mutually-exclusive states. For example, Grooming, Not Grooming and Unknown.

At any sample, the "Not" behavior is automatically scored if the corresponding behavior is not scored, unless Unknown is scored for that sample (see below).

Behavior "Unknown"

A behavior type is scored as Unknown in the following cases:

• At the beginning of the track if the Trial Control Settings rule does not include an additional Time condition that waits three seconds before the start of tracking, and in the last portion of the track (see Behavior Recognition: Data, Performance and Accuracy).
• When the nose and tail base points are not detected.
• In all cases when the subject is not found or samples are missed for more than 0.4 s.

Statistics of Unknown can be calculated for all behaviors except Twitching. The statistics are the same for all behaviors. For example, the duration of Unknown calculated for Grooming is equal to that of Unknown for Rearing supported.

Behavior Probability

Each behavior is associated with a probability, calculated per sample. Whether a behavior state is scored for a specific sample depends on a set of decision rules (for details see How Behaviors Are Scored in Behavior Recognition).

All behaviors but Twitching and Hopping are mutually-exclusive, therefore the sum of their probability values is 1. Twitching and Hopping have a probability too, but whether they are scored at a specific time does not depend on the probability of other behaviors.

How to Specify Behavior Probability

1. Click the Add button next to Behavior probability.
2. In the Behavior Probability tab, select the state(s) you would like to calculate the probability for.
3. Complete the procedure to add the variable. See Calculate Statistics: Procedure.

Application

You may want to filter the occurrences of a behavior based on their probability values. If the probability of that behavior is lower than a threshold at a specific sample, that sample is left out and the behavior state is re-calculated (see an example in Behaviors Detected with Behavior Recognition).

1. In the Analysis profile:
   • Add the behavior you are interested in.
   • Add the Behavior probability variable and, in the window that appears, select the same behavior.
2. Plot the integrated data to view the behavior detected and its probability together with the video. Check how those states correspond with the subject's behavior in the video.
3. To filter data based on probability, in the Analysis profile click the behavior variable, and select a probability value for Behavior probability threshold.
4. Plot the data again to see the result of filtering.

Merged Behavior

You can define a behavior state that includes two or more of the behaviors detected automatically. For example, to calculate the overall frequency of Rearing = Rearing supported + Rearing unsupported.

How to Specify Merged Behavior

1. Click the Add button next to Merged Behavior.
2. Under Select behaviors to merge, select which behaviors you would like to merge for analysis. For example, select Rearing supported and Rearing unsupported to analyze the total occurrences of Rearing.
3. Under Behavior probability threshold, select:
   • Default: Statistics will be calculated by using all samples scored as that behavior (no filtering based on the state's probability).
   • Probability greater than: Statistics will be calculated by using only the samples for which the probability of one of the merged behaviors is higher than a specific value. Select this value from the list. See the note below.
4. Under Behavior duration threshold, next to Exclude behavior instances shorter than, enter the minimal duration of the merged behavior to be considered for analysis.

   Note: The threshold is applied to the merged behavior, not the instances of single behaviors. For example, two instances of behaviors Rearing unsupported and Rearing supported follow each other in the time line. The first lasts one second, the other two seconds. When you merge the two behaviors and select a duration threshold of 3 seconds, the two behaviors are considered for analysis.

5. Complete the procedure to add the variable. See Calculate Statistics: Procedure.

Note

The Frequency of Merged behavior may not always be equal to the sum of the frequencies of the single behaviors. This can happen, for example, when an instance of Rearing supported is immediately followed by an instance of Rearing unsupported. Then the frequency of Merged behavior is 1, not 2.

Source: EthoVision XT 19 Help, Noldus Information Technology