Free interval
Definition
A segment of track from sample A to sample B, where A and B are determined by a time, the value of another dependent variable, a Trial Control event, or the value of a signal from a hardware device.
Use cases
Frequent cases:
▪Group instances of behavior (e.g. Immobile) in a bout.
▪Filter instances of behavior (e.g. Not moving) longer than one second.
See Examples of Free intervals in the Analysis profile
Calculation
A Free interval is defined by a Start criterion and a Stop criterion. Depending on whether and how often the criteria are met within a track, a Free interval may result in zero, one or more instances of the interval within the same track.
note Like all intervals defined with Nesting, a Free interval is left-closed [A, B). That is, the start sample A, not the stop sample B, is included in the analysis.
Multiple instances of the same interval definition
When you define a Free interval, this may result in zero, one or more (up to 2000) instances of an interval depending on where the Start and Stop criteria are met.
An instance of a free interval starts at the sample time when the Start criterion is met, and stops at the sample time when the Stop criterion is met.
A new instance is defined if the Start criterion and Stop criterion are met again, after the time in the track that the last Stop criterion was met.
note Multiple instances of the same free interval definition cannot overlap. An interval is NOT defined when its start criterion falls within a previous instance of the same interval definition. For example, you define a Free interval that starts when the subject enters zone A, to 10 seconds after that time. If a new instance of In zone A occurs before the end of the interval (see the green bar marked with 2 in the figure below), that instance of In zone is ignored.
Similarly, if you define a Free interval that goes from track start to when the external data Heart rate > 400, then one interval is defined, from Track start to the time when Heart rate reaches 400 bpm for the first time. This because the “Track start” criterion is met once in each track, no matter how many times Heart rate reaches 400 bpm.
Multiple free intervals in the Analysis profile
You can define multiple intervals in the Analysis profile, by repeating the procedure below. The interval definitions are independent of each other, therefore the resulting intervals may be overlapping.
1.In the Analysis Profile, under Custom Variables, click the button next to Free Interval.
2.Under Start criterion select the starting point of the interval:
▪Time. To select a time. For example, Track start (that is, the first sample in a track) or an Elapsed time after the track start (format: H:mm:ss.fff). For example 0:01:00.00 for one minute after the start. note Track start corresponds to the first sample in the track (Recording time = 0.0), not the Trial time. For multiple-subject setups, Track start is the first sample of the animal detected first.
▪Dependent Variable. To select the time that a dependent variable has a specific value. For example, Distance moved >= 1 m, or frequency of In zone >=10. This list also includes behaviors scored manually and external data. See Free interval based on a dependent variable
▪Trial Control. To select the time that an event in your trial control procedure occurs. For example, when a condition becomes true, or a command to a hardware device is given. See Free interval based on a Trial Control event
▪Hardware. To select the time that a Hardware-related variable has a specific value. For example, when the number of food pellets dropped by a Pellet Dispenser is greater than 10. This option is only available when hardware devices are defined in your experiment. See Free interval based on hardware
3.Under Stop criterion select the end point of the interval. See above for most of the options.
If you choose Time, and under Interval stop you choose Elapsed time, you have two options: after track start to stop the interval at a specific time after the track start (that is, the first sample in the track); after start event to stop the interval a specific time after the event selected as start criterion. Note that this “start event” is not the same as the start time of the resulting interval, if you set some time before the event (for example, 10 s before the frequency of In zone becomes 1). The “start event” is when the Frequency of In zone becomes 1. However, the interval starts 10 s earlier.
4.Select Ignore last interval if incomplete to ignore the interval when the stop criterion is not met. If you do not select this option, and the stop criterion is not met, EthoVision XT defines an interval up to the end of the track.
5.A free interval may occur once or more in a track (see above). To analyze each instance separately, select Calculate statistics per interval and enter the number of intervals you want to view in the results (1 to 2000). To be sure you to view all intervals, enter a large number that exceeds a maximum possible number of instances per track. See Restrictions in EthoVision XT
To analyze all instances as one, de-select Calculate statistics per interval.
6.Complete the procedure to add the variable. See Calculate statistics: procedure.
Notes
▪Note the difference between Free Interval in the Analysis Profile (this topic) and Nesting over a Free interval in the Data Profile.
▪Define a Free Interval in an Analysis Profile when you want to analyze the interval itself, not other dependent variables within that interval. For example to calculate the duration of an interval.
▪Define a Free Interval in a Data Profile when you want to analyze any dependent variable within the resulting track segment. For example, to calculate the average velocity of the subject in the interval.
▪If a start criterion is not met in a track, EthoVision XT does not define the Free intervals in that track.
▪If a stop criterion is not met in a track, EthoVision XT does the following:
▪If you selected Ignore last intervals if incomplete, it does not define the interval.
▪If you did not select that option, it defines an interval up to the end of the track.
▪Free intervals in multiple-subjects setups. A free interval is calculated for each subject separately, based on the data (dependent variables, behavior etc.) for that subject. See Free interval based on multiple subjects
▪Free intervals and missing samples. A free interval is not interrupted if one, two or three consecutive missing samples occur. With four or more consecutive missing samples, the current interval is interrupted and a new one starts when the Start criterion is met.
▪Contiguous intervals. When two instances of a Free interval are contiguous, for example when the Interval Start criterion is supposed to be met immediately after the Interval Stop criterion for the previous instance is met (see Example 3 in Examples of Free intervals in the Analysis profile, where a track is split in segments of equal path length), one sample is always excluded between the end of an instance and the start of the next instance. For example, the first instance goes from time 0 (included) to sample 100 (excluded), then from sample 101 (included) to sample 250 (excluded). Samples 100 and 250 are not included in the analysis. The reason is that the statistic used to find the start of the interval is updated one sample after the end of the previous interval.
More details
▪Free interval based on a dependent variable
▪Free interval based on a Trial Control event
▪Free interval based on hardware
▪Free interval based on multiple subjects
▪Examples of Free intervals in the Analysis profile