Stats and outliers

The group HTML report is a wrapper around helpers in pyfsviz.stats. Use them when you want CSVs, outlier lists, or plots without generating HTML.

FreeSurfer must be initialized: get_stats() runs asegstats2table and aparcstats2table. See Prerequisites.

Collect tables

from pyfsviz.stats import get_stats

stats = get_stats(["sub-001", "sub-002"], "reports/")
# stats["aseg"] -> Path to aseg.csv
# stats["aparc"] -> list of aparc CSV paths, including combined_aparc.csv

SUBJECTS_DIR must point at the tree that contains those subject folders. Optional measures (default area, volume, thickness) and hemis (default lh, rh) control which aparc tables are written.

Outliers

check_metrics() flags values more than sd_threshold standard deviations from the cohort mean, per region:

from pyfsviz.stats import check_metrics, summarize_outlier_subjects

from pathlib import Path

files = list(Path("reports").glob("*.csv"))
quality = check_metrics(files, sd_threshold=3.0)
outliers = summarize_outlier_subjects(quality)

Each outlier entry has subject_id, outlier_count, and findings (metric and region with the raw value).

Group comparisons

from pyfsviz.stats import gen_group_comparison_plots

groups = {
    "control": ["sub-001", "sub-002"],
    "patient": ["sub-101", "sub-102"],
}
figures = gen_group_comparison_plots(files, groups)

These are Plotly box plots (one per metric region) with a point per subject. The group HTML report embeds the same figures in one tab per stats table (aseg, LH Area, RH Thickness, …).

Distribution plots

from pyfsviz.stats import gen_metric_plots

plots = gen_metric_plots(files)

These are Plotly figures (the same family embedded in the group report). One figure is created per region in each stats table (aseg, lh_area_aparc, and so on).

For the HTML wrapper, see Group reports.

Feedback