Dear Arian,
The IQR is the best measure of data quality before pre-processing. If the IQR is below 75%, you should consider whether you should include the data. 65% is probably a no go and 70% is a bit borderline and also depends on your sample. For ADHD data you will probably have a different threshold.
The sample homogeneity measure (in older versions the correlation coefficient, now the z-score) shows homogeneity after pre-processing and also reflects anatomical variation. A good quality scan may deviate because of large ventricles, but a not so good scan with some noise may show the anatomy as expected and will not show up in the homogeneity check.
While the IQR may allow you to set a hard threshold below which you will not accept quality, the homogeneity check will only give you an indication of which data are most off. This does not necessarily mean that you have to exclude these data, but you should check these scans carefully.
Best,
Christian
On Mon, 8 May 2023 14:35:23 -0400, Arian Klows <[log in to unmask]> wrote:
>Hi,
>
>I have a question related to the quality checking of data in CAT12.
>Previously, I used to assess data quality by reviewing the IQR values in
>the report after each preprocessing step. My usual threshold for selecting
>subjects was a 75% IQR cutoff.
>Here, for subject, PT-123, the IQR was 68.81% (see below).
>[image: image.png]
>
>
>Interestingly, when I performed the "check sample homogeneity of 3D data"
>analysis, this subject was not identified as an outlier. Instead, the
>analysis suggested that subjects 019, 022, 026, 060, and 126 should be
>excluded (detailed below).
>
>I am now wondering what the appropriate method is for excluding patients
>based on data quality. Could you provide some guidance on this?
>
>------------------------------------------------------------------------
>08-May-2023 13:53:59 - Running job #11
>------------------------------------------------------------------------
>08-May-2023 13:53:59 - Running 'Check sample homogeneity of 3D data'
>Corresponding xml-files were found.
>
>Compressed filenames sample 1:
>/bme/home/ak/MRI/Study/smri/mwp1PT_{001,002,003,004,005,006,007,008,009,010,011,012,013,014,015,016,017,018,019,020,021,022,023,024,025,026,027,028,029,030,031,032,033,034,035,036,037,038,039,040,041,042,043,044,045,046,047,048,049,050,051,052,053,054,055,056,057,058,059,060,061,062,063,064,065,066,067,068,069,070,071,072,073,074,075,076,077,078,079,080,081,082,083,084,085,086,087,088,089,090,091,092,093,094,095,096,097,098,099,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144}.nii
>
>
>These data have a mean correlation below 2 standard deviations.
>This does not necessarily mean that you have to exclude these data.
>However, these data have to be carefully checked:
>/bme/home/ak/MRI/Study/smri/mwp1PT_126.nii: 0.811
>/bme/home/ak/MRI/Study/smri/mwp1PT_060.nii: 0.811
>/bme/home/ak/MRI/Study/smri/mwp1PT_026.nii: 0.807
>/bme/home/ak/MRI/Study/smri/mwp1PT_019.nii: 0.779
>/bme/home/ak/MRI/Study/smri/mwp1PT_022.nii: 0.747
>08-May-2023 13:54:39 - Done 'Check sample homogeneity of 3D data'
>08-May-2023 13:54:39 - Done
>
|