Determine subjects that potentially failed segmentation, based on region-wise data. The data can be anything, but there must be one numerical value per subject per region.
qc.from.regionwise.df(
rdf,
z_threshold = 2.8,
verbosity = 0L,
num_bad_regions_allowed = 1L
)data.frame, the region data. The first column must contain the subject identifier, all other columns should contain numerical data for a single region. (Each row represents a subject.) This can be produced by calling group.agg.atlas.native.
numerical, the cutoff value for considering a subject an outlier (in standard deviations).
integer, controls the output to stdout. 0=off, 1=normal, 2=verbose.
integer, the number of regions in which subjects are allowed to be outliers without being reported as potentially failed segmentation
named list with entries: 'failed_subjects': vector of character strings, the subject identifiers which potentially failed segmentation. 'mean_dists_z': distance to mean, in standard deviations, per subject per region. 'num_outlier_subjects_per_region': number of outlier subjects by region. 'metadata': named list of metadata, e.g., hemi, atlas and measure used to compute these QC results.
Other quality check functions:
qc.for.group(),
qc.from.segstats.table()