Defining, Creating and Managing Experiments

Quick Index

"Experiments" is the term used in XNAT to describe data that is gathered or measured from direct interaction with a subject. These types of data can range from a simple consent form signed by a subject at the beginning of a study to demographic data gathered about the subject, their medication listblood test results, behavioral assessments for cognitive psychology measures, to MR or PET sessions taken when the subject is in the scanner.

"Experiments" are differentiated from "Assessors," which are measurements that are derived from experiment data. For example, if a radiologist looks at imaging data and performs a rad read, that person has not come in direct contact with the Subject - they are examining and assessing the Subject's Experiment data. 

Each experiment corresponds to a data type in XNAT. Data types are managed by the XNAT Administrator (see: Managing Data Types), and are reflected in the UI as a paired set of input forms and report pages. Experiments for each subject are listed on the Subject's report page, like so: 

This subject has had two MR Sessions performed, each of which is listed as a single experiment, and is given an experiment label. 

Creating Experiments

There are several ways of adding experiment data to your subjects in XNAT, many of which are covered elsewhere in this documentation. 

Uploading Imaging Data

If you have a scanner on site, imaging data can be uploaded directly from the scanner via a DICOM C-Store connection. Otherwise, you will commonly be receiving imaging data via the Upload Applet or zipped archives of scan data, and using the ZIP uploader

Advanced users may use REST scripts to upload data and files

Creating Experiments

Users can add new experiments through the UI in a couple of different ways.

  • From any point in XNAT, you can navigate to "New > Experiment" in the top navigation. 
  • Or, from the subject page, you can select "Add Experiment" from the Actions box in the right sidebar. 
Adding experiments from the subject page is the preferred method, as it will automatically populate project and subject information in the Add Experiment form as shown below:
You can either scroll down and find the name of the experiment you want to add, or start typing its name in the Filter text box to locate it more easily. After selecting an experiment to work on, you will be taken to an HTML form for data entry. 

Notes on Data Entry:

  • By default, XNAT installations are configured to time out users after 15 minutes of inactivity for security purposes. This can be configured by changing the value for session-timeout in {Tomcat directory}/webapps/xnat/WEB-INF/web.xml and restarting Tomcat.

        <!-- Default to 15 minute session timeouts -->
  • Unfortunately, this time out can cause problems during data entry of long forms of information, where several dozen fields require information. XNAT does not recognize data entry as "activity" in the same way that a page action (refresh, form submit, navigating forward or backward etc.) is. Users may find that their session has inadvertently timed out during data entry, causing their data to not be saved. To prevent this, a Session Timeout Warning pop-up was introduced in XNAT 1.6, informing the user that their session is about to expire and giving them the option of renewing it. If they click Renew, the time left in the session will go back to 15 minutes, or whatever was specified in web.xml. Also, a top bar was added in XNAT 1.6 which has the time left in the session as well as a renew link. This link is useful in cases where someone is in the middle of entering a form and wants to leave their computer for a few minutes without being logged out. For example, if there are two minutes left in the session, the user does not have to wait for the Session Timeout Warning pop-up to renew their session before they can leave their computer for a few minutes.
  • Most data entry forms are designed to be nearly identical with original paper forms. However, as there are few standards in paper form design, there have been few standards implemented in form design or data validation in XNAT. Sadly, users should not be terribly surprised to find that different forms behave differently, depending on who coded them. We are addressing design standards in XNAT 2.0 development. 

Managing Experiment Data

By default, experiment data in XNAT is organized chronologically. However, to this point there is not native support for more sophisticated data management associated with longitudinal studies, where a single subject is studied over a long period of time. In a longitudinal study, a subject returns for multiple visits to have experiment data collected, and reporting on such variables as disease progression is done on a visit-by-visit basis.

This feature is currently in development and an early version of it is available in XNAT Marketplace as a module. The Project Protocols Module will provide support for visits and longitudinal studies in XNAT.