One of the key features of scanSTAT is the ability to detect and partially correct for a variety of artifacts which can produce decrements in image quality. By detecting images in which pixels have changed intensity from background noise to "brain-like" intensities (or vice versa), scanSTAT can identify artifacts arising from subject motion or scanner problems.
Tutorial/Demonstration - Artifact Detection
The Artifact Data Graph
You have the option of having scanSTAT attempt to correct for some kinds of artifacts. Select Statistics -> Statistical Options... and notice the check boxes by each of the artifact classes. Checking these boxes will tell scanSTAT that you want it to ignore any images in which the category checked exceeds the indicated threshold. For example, if the box next to "Spike" is checked and the threshold is set to 10%, scanSTAT will not consider any images in which 10% of the pixels have changed intensity from background to signal or vice versa. This may improve the statistical overlay maps.
|Motion artifacts are determined on the basis of the difference in the number of pixels that increase above the noise floor and decrease below it. The control parameter is the percentage of pixels that must change to flag a motion artifact|
|Spikes are detected on the basis of the number of pixels that increase above the noise floor on successive images. A large number (e.g., 25%) is usually appropriate|
|Other artifacts include any significant increase in the number of pixels above the noise floor compared to the first image in the series.|
scanSTAT monitors these artifacts during the statistical calculation. As calculation progresses, you may note that the black line under the image, which represents the paradigm reference vector, is sometimes sprinkled with colored spots. These represent images in which one or more of the artifact categories exceeds the threshold. Green marks represent images in which the increase/decrease ratio exceeds threshold, probably indicating subject motion. Red marks indicate spike artifacts. Grey bars extending over the whole height of the display area indicate images which were ignored in the statistical calculation. You can determine which artifact criteria will lead to an image being ignored by checking the box next to the category in the Statistical Options... screen under the Statistics menu.
For some real-world examples, have a look at Section 6B.
The Displacement category of artifact detection relies on a center-of-mass calculation for the image. When this center of mass shifts, the shift can be detected and quantified, resulting in a method of tracking the direction and degree of subject motion. In addition to using this statistic as a category of artifact detection, scanSTAT can represent it in a graphical form which can be useful for observing subject motion.
Tutorial: Using the Motion Bullseye
The Motion Bullseye display
The Motion display during calculation
Go back to previous section
Go to next section
Go back to main page