Cholinergic Nucleus 4 (Ch4) Normative Calculator

Convert an individual’s Ch4 gray-matter volume into a normative W-score — how far it deviates from the volume expected in cognitively healthy adults of the same age, sex, and head size — and see where the individual falls within the clinically normal reference distribution.

Research demonstrator · runs entirely in your browser
Research and educational use only. This is a demonstration of the normative model from a study in preparation. It is not a medical device, is not validated for clinical use, and must not be used for diagnosis or patient management. A valid W-score requires Ch4 gray-matter volume measured with the same imaging pipeline as the study (see Required imaging pipeline below). The model was calibrated by cross-validation and has not yet been externally validated. All values are computed locally — no data is transmitted or stored.

Patient input

Ch4 GMV is typically ~0.17–0.30 mL (1st–99th centile in the reference). Must be measured with the study pipeline.

Result

— —

Position in the clinically normal reference distribution

The bell curve is the calibrated normal reference (W ~ standard normal in controls). The shaded tail is the low-volume zone (W ≤ −1.645, the 5th centile). The patient’s value is marked.

Predicted Ch4
—
Percentile
—
Deviation
—

Required imaging pipeline for a valid W-score

The W-score is only meaningful if Ch4 gray-matter volume is measured exactly as in the study. A value from a different atlas, threshold, or software package is not comparable and will give a misleading result.

  1. Segmentation. Segment the T1-weighted volumetric scan with CAT12 v12.9 (r2577) running under SPM12 (MATLAB R2023b), using the default preprocessing pipeline including surface estimation.
  2. Ch4 localization. Use the left and right Ch4 labels of the Jülich-Brain maximum-probability atlas distributed with CAT12, in the space of CAT12’s normalization template (MNI152NLin2009cAsym).
  3. Gray-matter volume. Sum the modulated, spatially normalized gray-matter map (mwp1) within both labels, sampled on the 1-mm atlas grid, and multiply by the voxel volume — each voxel contributes in proportion to its estimated gray-matter content. Because the map is modulated, this is a native-space-equivalent volume, reported in millilitres (mL). This is the value entered above.
  4. Head size. Take total intracranial volume (TIV) from the CAT12 subject measures, in mL.
  5. Acquisition site. The model carried a site random intercept; for a scan from a site not in the training sample this intercept is set to zero — the calculator already does this, so no site input is needed.

Atlas: Zaborszky et al., NeuroImage 2008 (JulichBrain magnocellular basal-forebrain maps). Full details are in the study Methods.