code written: 2020-01-12
last ran: 2020-01-14
website: http://rpubs.com/navona/SPINS_fiberMeasurements

Notes: This script summarizes the key output values from Slicer, for all tracts. As expected, we have data for all n=41 tracts (n=74 unique combinations with hemisphere). The data we have available is summarized below (collapsed across all sites):


Data summary.

Missing participants. In total, we only have data from n=407 participants, whereas n=422 were expected (a difference of -15.) The missing participants are SPN01_CMH_0007, SPN01_CMH_0050, SPN01_CMH_0106, SPN01_CMH_0107, SPN01_CMP_0219, SPN01_CMP_0220, SPN01_MRC_0044, SPN01_MRC_0049, SPN01_MRC_0058, SPN01_MRP_0132, SPN01_ZHH_0023, SPN01_ZHH_0038, SPN01_ZHP_0088, SPN01_ZHP_0091, SPN01_ZHP_0111; I need to follow up to understand why the pipeline failed (these participants do have DWI data that passed QC); I believe it’s a queue / Slurm error.

Missing tracts. The count variable in the table above indicates the number of participants with data for a given tract, and percent indicates corresponding percentage. We see that some tracts have data from far fewer participants than others. This is especially so for the SLF(s) and IOFF. The SLF(s) might be missing values given that they contain relatively few fibers on average; however, this is not the case for the IOFF. In total, we have data for 27553 tracts out of a possible maximum 30118, i.e., 91.4834982%.

Missing tracts by site. The following visualization suggests that missing values in some tracts may be driven site. I will need to investigate this:

Missing tracts by participant. Most participants have most tracts, as follows:

Tracts missing 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 38 58
Percent missing 0.00 1.35 2.70 4.05 5.41 6.76 8.11 9.46 10.81 12.16 13.51 14.86 16.22 17.57 18.92 20.27 21.62 22.97 24.32 25.68 27.03 28.38 29.73 51.35 78.38
Participant count 2 13 51 80 52 50 22 21 22 15 10 13 7 9 9 8 6 6 3 1 2 1 2 1 1

Visualization: Number of Fibers.

The following plot show the number of raw data for the number of fibers variable from the n=407 participants summarized above, separated by tract and hemisphere (n=74) and coloured by site / scanner. Outlier valies are apparent, especially in the SLF.