results_no_auto_errorcorrection.xls

These are the results from comparing the different T1 images collected from the first scanner session. They were -

8- channel projector off, adjusted fft scaling

8- channel projector off, default fft scaling

8- channel projector on, default fft scaling

Birdcage projector off, default fft scaling

The volumes were bias corrected using SPM. Then, the files were loaded into Caret and edited to include only the right hemisphere for segmentation. First, segmentation was run without error correction, and topological errors were recorded for each (which Caret reports at the end of each segmentation). Then, an xyz search for handles revealed a different way to look at topological defects.

Next, the same volumes were segmented again, this time using automatic error correction, with the same handle search as before. All results are recorded in the spreadsheet above.

It appears that the birdcage is not as clean as the 8-channel, and adjusted FFT might help. More birdcage samples would better determine its quality. Also, there is a slight difference with the projector on, but I suggest that it only represents the range of the 8-channel quality without adjusted FFT.

MGB 4/18/2008