Troubleshooting: Data visualization
Tracks and Integrated Visualization
▪I defined a free interval, and in the visualization I get an interval every other sample
▪I plotted a variable in the Integrated visualization. I only get a few values on the timeline. Why?
▪Screen recording could not be started
▪There seems to be a delay between video and track
▪Video is choppy or does not play smooth
▪The points are too small (or too large) or the traces are too narrow (or too thick)
▪When I import UltraVox XT data, some calls are visualized as 1, others as 0.5. Why?
▪The Integrated Visualization shows “Not calculated”
▪The Data Preparation Report says “Failed” for some trials
▪The name of the trial is not fully visible
Heatmaps
▪A heatmap is plotted outside the arena, or merged with the heatmap of other arenas
▪In the Heatmaps screen the Group Mean button is not available
▪How can I merge heatmaps generated in different arenas?
I defined “In zone” for two adjacent zones. In the Integrated visualization, I see that the animal is in the two zones simultaneously
This happens when a Zone exit threshold has been defined that is greater than 0, and the two zones Center and Border are adjacent. In such a case, when the animal moves from Border to Center, it is at some point considered to be in Center, but if it is not yet outside the zone exit threshold of Border, it is also considered to be in Border.
To prevent this, re-define the two zones in such a way there is a gap between the two, its width being at least the same as the zone exit threshold.
I defined a free interval, and in the visualization I get an interval every other sample
Most probably you defined a free interval based on two values of external data or a hardware variable, for instance:
▪From Heart rate >= 300 to Heart rate >= 400
▪From Number of pellet drops >= 1, to Number of pellet drops >= 10
When the value of the (resampled) variable reaches the second (higher) threshold, the interval ends and EthoVision XT finds the next sample that matches the criterion defined in From (first threshold). If the current value is higher than the first threshold, a new free interval is defined. However, if the value at the next sample is also higher than the second threshold, the interval is ended, resulting in a free interval of one sample. The search is repeated in the next samples, generating the pattern shown in the first plot below.
In such cases, do not use a free interval. Instead, use a state variable that is based on the original external or hardware signal. You can find the state variable in the Analysis profile. For example, for the external data Heart rate, in the Analysis profile under External data choose Heart rate state. For a hardware signal, under Hardware choose Hardware state. There you can define the range of values you are interested in. In the example above about heart rate:
I plotted a variable in the Integrated visualization. I only get a few values on the timeline. Why?
This may happen for variables which require two or more consecutive data points to be calculated. Consider the following examples:
▪Your data contain missing samples. If one sample is missing, the variable is not calculated for one, two or more time points. For example, Distance moved is not calculated for two time points, and Turn Angle is not calculated for three time points.
▪You may have applied Minimal Distance Moved smoothing method. When using this method, if the distance between body points for time t and t+1 is lower then the Minimal Distance Moved, the body point at time t+1 is “moved” to the same coordinates as the point as time t. In this case Distance moved gets the value 0 (zero), but any variable that quantifies change in direction has no meaning here. Therefore, is not calculated for that sample.
If many samples in your track are filtered out that way, variables like Turn Angle, Meander and Heading only appear in short segments. Because the filter removes data points that do not represent movement, these few remaining segments are a more realistic measure of the subject’s behavior.
See also
▪How variables are calculated: Dependent Variables in Detail
I interpolated my data, but some variables including Elongation and Mobility still show missing values
Interpolation is only applied to missing body points to obtain valid x,y coordinates. When the subject is not detected, the subject’s contour and surface area remain unknown. This is why variables such as Activity, Elongation and Mobility still show missing values after interpolation.
Screen recording could not be started
When trying to make a screen recording, you get the following message:
This may be caused by the fact that older versions of the Noldus MediaLooks A/V filters are still installed, but the newest version was not installed.
1.Close EthoVision XT.
2.In the Windows Control Panel, choose Programs and Features, and remove old versions of Noldus MediaLooks A/V Filter.
3.Download the full installation ZIP file (see Install EthoVision XT), extract and browse to Drivers and tools > PreRequisites > MediaLooks. Copy Noldus Medialooks ScreenCapture 2.0 x64.msi to your EthoVision XT computer and run that file.
4.Follow the installation, then start EthoVision XT.
There seems to be a delay between video and track
The track is lagging behind the video or the other way round.
This issue can only occur when you track live and save video.
Solution:
1.Choose Acquisition > Edit Tracks.
2.Choose the trial from the list at the top.
3.Play the video until the subject and the track are well in view.
tip Choose a frame when the animal walks fast so you can see exactly in which sample the nose point overlaps with the subject’s nose in the video.
4.In the Playback Control window, click the Offset button.
This button is only available when you tracked live and saved video.
5.In the Synchronize video window that opens, enter a number of frames (negative or positive) to correct the discrepancy between track and video. Click OK when ready.
For example, enter -1 to move the video one frame backward relative to the track.
This issue can occur in computers with two graphics cards: a primary (integrated) card and a secondary card, usually more advanced and powerful than the integrated one. The issue is solved when you choose which card EthoVision XT uses.
See The camera image is not shown in the Arena Settings or Integrated Visualization
Video is choppy or does not play smooth
This may occur if you play video at high playback speed like 16x in the Integrated Visualization. The effect is worse if the secondary graphics card (GPU) which takes care of the playback is less powerful.
The points are too small (or too large) or the traces are too narrow (or too thick)
Solution: in various parts of EthoVision XT you can change the thickness of the lines and the size of the dots.
1.Click Show/Hide (at the top-right corner of the screen) and select Detection Features.
2.Adjust the Line thickness and Dot size.
Note:
▪What you select in the Detection Settings screen are also selected in the Acquisition screen, and the other way round.
▪What you select in one of the following screens is also selected in the other screens: Track Editor, Track Visualization, Integrated Visualization, Manual Scoring screen
▪The line thickness is also applied to the body contour in Detection Settings and Acquisition screen. It is not available in the Manual Scoring screen.
When I import UltraVox XT data, some calls are visualized as 1, others as 0.5. Why?
This is due to the type of resampling selected in the external variable of your Analysis profile. Consider the following example. The first plot shows the original signal imported from UltraVox XT and the second plot shows the same signal after resampling:
For the second call at 25.5, the time between the “1” an “0” in the original signal is so short that it fits in one sample interval.
Two points in the same sample interval result in downsampling. Now open the Analysis profile, and open the External data variable. See External Data
▪If you choose Mean for downsampling, the value A in the resampled signal will be the average of 1 and 0, that is 0.5.
▪If you choose Total for downsampling, the value in the resampled signal will be 1 + 0 = 1. After selecting Total, the signal is as follows:
So, choosing Total should solve the issue.
The Integrated Visualization shows “Not calculated”
One or more trials are corrupt, or something went wrong with editing the tracks. See The results table contains “?” in many cells.
The Data Preparation Report says “Failed” for some trials
This message occurs in experiments with Deep learning based body point detection. In some cases EthoVision XT fails to review the tracks and fix subject identity swaps. For example, when the two subjects spent most of the trial time in close contact. See a note in Prepare the data in multi-subject trials
In most cases, this occurs when a Data profile is active that does not include that trial. Make a Data profile that includes that trial and visualize the data. See Filter tracks
The name of the trial is not fully visible
The name of the trial on the toolbar of the Integrated Visualization screen is partially hidden. This may occur when the display is set to a high resolution. Reduce the screen resolution in Windows.
A heatmap is plotted outside the arena, or merged with the heatmap of other arenas
For example:
Solution: In the Heatmap Settings pane, click Colors and under Smoothing select a lower value.
Result:
See also
In the Heatmaps screen the Group Mean button is not available
Probably you are working with a DanioVision experiment. The Group Mean button is disabled in the experiments set as DanioVision.
How can I merge heatmaps generated in different arenas?
For example, you acquired tracks in Arena 1 and Arena 3 with subjects that belong to the same treatment group, and you would like to create one heatmap based on the data of both arenas.
Unfortunately merging heatmaps from different arenas is not possible within EthoVision ZXT. However, with simple JavaScript code you can re-calculate the coordinates in the tracks and export them to an external applications that creates heatmaps. Here below we describe a possible solution:
1.In the Arena Settings, define a Point of interest (see Draw a point) in each arena. For example, the south-west corner of the open field. This point will be the new reference point of the coordinates, so it must represent the same physical location in all arenas.
If the arenas includes already a zone or point, like the shelter of PhenoTyper, and the zone has the same position and orientation in all arenas, then you do not need to define a Point of interest.
2.In the Analysis profile, add two JavaScript continuous variables. See JavaScript continuous
The first variable is for recalculating the x-coordinate. Add the following text (replace “Point” with the name of you point or zone).
const g_zone = "Point";
function Start()
{
}
function Stop()
{
}
function Process()
{
var pt1 = GetCenter();
var pt2 = GetPointPoi(g_zone);
if (pt1)
{
var xc = pt1.x - pt2.x;
SetOutput(xc);
}
else
{
SetOutputMissing();
}
}
3.For the second variable, repeat the step above, this time replace x with y.
4.Export the two variables (Analysis > Export > Raw Data). These contain the x,y coordinates expressed relative to the point in the arena.
5.Import the data in an application that creates heatmaps, like Excel Power Maps, R or MatLab.