abcc_automated_QC

This repo has the code for automated_QC (structural QC as well as functional QC) performed on abcc data processed using the abcd-hcp-pipeline and file-mapped using DCAN Lab's file-mapper

Structural QC

The following are automated metrics for evaluating processed structural data quality:

Population Metrics from outlier detection:

• Subcortical segmentation volumes size – outliers based on 3 SD from the mean
  Usage:
  python3 ./subcortical_vol_outlier_detection.py ${output_base_dir} ${dseg_label_file} ${tsv_file}
  
  Returns:
  An updated tsv file with an additional column #subcortical_segmentation_vol_out, showing the number of outliers for each subject and session

• Cortical morphometry, split by region of interest – outliers based on 3 SD from them mean
  Usage:
  python3 ./Cortical_morphometry_surface_outlier_detection.py ${output_base_dir} ${pscalar_label_file} ${tsv_file}
  
  Returns:
  An updated tsv file with an additional column #cortical_morphometry_sulc_out, showing the number of outliers for each subject and session
  
  Parameters:
  

• output_base_dir: directory where you have all *_space-fsLR32k_sulc.pscalar.nii, *_space-ACPC_dseg.nii.gz files stored
  

• dseg_label_file: ./dseg_label.txt
  

• pscalar_label_file: Gordon.32k_fs_LR.dlabel.nii
  

• tsv_file: participants.tsv
  

Functional QC

Pconn File detection using pconn_outlier_detection.py

• Column for passing 5 minutes of data:(based on whether the pconns get produced)
  
• Column for passing 10 minutes of data:(based on whether the pconns get produced)
  10min_pconn

Population metrics from outlier detection

• Connectivity matrix for 5 min and 10 min pconns – outliers based on 3 SD from the mean Usage:
  python3 ./pconn_outlier_detection.py ${output_base_dir} ${tsv_file}
  
  Returns:
  An updated TSV file with the following additional columns:

• 5min_pconn: Pass/Fail based on file presence.

• 10min_pconn: Pass/Fail based on file presence.

• #pconn_out_5min: Number of outliers for each subject and session for the 5-minute connectivity matrix.

• #pconn_out_10min: Number of outliers for each subject and session for the 10-minute connectivity matrix.
  
  Parameters:
  

• output_base_dir: directory where you have all *_pconn.nii.gz files stored
  

• tsv_file: participants.tsv
  

Note - Any column that is nan/empty signifies that particular file did not exist.