Averaging and Conditional Averaging

Introduction

Averaging and conditional averaging tools give the user the possibility to create their own corridors of normality. The user will then be able to compare a single trial (or subject) with the corresponding corridor of normality for a chosen population of trials.

Whatever the type of averaging (standard or conditional), the creation of a corridor of normality requires several steps

  • Creation of events.
  • Creation and launch of the averaging script.
  • Display of the corridor at the same time as the curve of the subject/patient.

Standard averaging

The first step consists of creating the events and adding them on each trial you wish to use for creating the average. These events define the start and the end times of the period of interest during each trial. For example, if you want to create a corridor of normality of the knee during a jump, the start event will correspond to the beginning of the jump preparation phase and the end will correspond to the moment where the subject is totally stabilized. The second step consists of creating the averaging script. The “Average Script” dialogue box can be opened by selecting the “New Averaging Profile…” item in the “Calculation” Menu.

First, label the script by writing its name in the “Name” box. Next, define the group and the channel names where the data of the averaging script will be stored. For our example, let’s use the following names: “Jump Average” for the name field, “RefPoint” for the “group” field and “Average” for the “Channel” field. Now click on the “+” button A window will appear. Select the group and the channel of the data you want to normalize and the start and end event and click on OK twice.

 

Figure 1: Averaging configuration

Averaging calculations are not run automatically, and must be launched manually by right-clicking on the “Calculation” item of the “tree protocol” menu and by selecting the “Open Script Runner” item (Figure 2). A new window will appear. Enter the required information as appropriate, select the corresponding script and click on the “Run Scripts” button.

Figure 2: launch of the averaging script

A dialogue will appear to report on the execution of the script (successful or not: Figure 3).

Figure 3: State of the Averaging Script

Close this window by clicking on “OK” and click on “Cancel” to close the “Script Runner” window.

The last step consists of displaying the corridor of normality and the curve obtained with a subject/patient on the same graph. However, first we must insert the data corresponding to the corridor of normality into the workspace. Open the “select trial” dialogue box by clicking on the “File > Opens – Close Trials” command and select the trial corresponding to the corridor of normality. The trial will have the same name as the script (in our example Jump Average).

Once the trial is imported, create a graph by right-clicking on the “Graph” item in the “Protocol Tree” menu and name the graph (“Comparison between Corridor of normality and the patient curve” for example). Two data sets are usually plotted on this graph: the corridor of normality, and the curve of the patient from each individual trial.

Click on the “add” button (‘+’) in order to open the “Graph Definition” dialogue box. Select the “all open trials” option from the “Select” popup menu and tick the box “Time” as abscissa. Click on the “add” button of the Ordinate section and choose the following configuration: “RefPoint” for the group, “Average” for the Channel, “X” as Dim, a light colour (white for example; this colour will be use for the display of the corridor standard of deviation curve) and “confidence” as field. If you want to display the mean of the corridor of normality, a second ordinate can be added by clicking on the add button of the ordinate and by selecting the following configuration: “RefPoint” as group, “Average” as Channel, “X” as Dim, a light colour (white for example; this colour will be use for the display of the corridor standard of deviation curve) and “value” as field. Click on “OK” to close the window. An example of a completed dialogue is shown in Figure 4.

Figure 4: Corridor of normality graph configuration – part 1

The display of the patient curve required for the comparison needs to be set now. To open the “Graph Definition” dialogue box, click on the “+” button. Select the “Current trial” option from the “Select” popup menu and tick the box “Normalised Time” as abscissa. Select the “Start” and “End” events. Then configure the ordinate data as follow: “VectorAngle” as group, “Elbow” as Channel, “Value” as Dim, the red as colour and “value” as field. Click on “OK” twice to close all the opened windows. An example of a completed dialogue is shown in Figure 5.

Figure 5: Corridor of normality graph configuration – part 2

The corresponding graph item will then appear in the protocol tree menu (Figure 6).

Figure 6: New graph item

Select the panel where you want to display the graph and double-click on the item. The graph comparing the corridor of normality and the patient curve will then be displayed in the selected panel.

Figure 7: Comparison between Corridor of normality and the patient curve

Conditional Averaging

The conditional averaging works exactly the same way as the standard averaging with the difference that not all trials from the trial list will contribute to the average. Instead, only trials which satisfy a particular condition will be used. For example, if it is known that some trials come from left-handed subjects and some from right-handed subjects, you may wish to generate a separate average for each group.

To be able to do a conditional average the trials contained in list trial panel should first be categorised according to the conditions of interest.

As an example, we define two different categories (right and left) using the trial field dialogue box. Then we mark all trials according to whether the patient was right or left handed.

Figure 8: Categorization for the conditional averaging – step 1

Figure 9: Categorization for the conditional averaging – step 2

Once the categorization is done, all the other steps are similar to the standard average, except for the configuration of the “define portion” dialogue box (Figure 1) where the “condition” section has to be selected and filled in as shown in Figure 10.

The procedures for the execution of the script and for the display of corridor of normality are the same as for a standard averaging.

Please note that a conditional average script has to be created for each condition (in our example: one script for the right handed group and one for the left handed group).

Figure 10: configuration of the conditional option in the “Define Portion” dialogue box
Sidebar