Predictive Clinical Neuroscience Toolkit

Predictive Clinical Neuroscience software toolkit (formerly nispat).

A Python package for normative modelling, spatial statistics and pattern recognition.

IMPORTANT

Deprecation warning

This is PCNtoolkit version 1.X.X, released originally in June 2025. Any scripts, models, and results created with version 0.X.X are not compatible with this and future versions of the toolkit.

To use the models created with versions 0.35 and earlier, please install the appropriate version using pip install pcntoolkit==0.35, or replace 0.35 with your desired version. The old version of the toolbox is also still available on GitHub.

Installation

pip install pcntoolkit

Documentation

See the documentation for more details.

Documentation for the earlier version of the toolbox is available here

Example usage

from pcntoolkit import {load_fcon, BLR, NormativeModel}

fcon1000 = load_fcon()

train, test = fcon1000.train_test_split()

# Create a BLR model with heteroskedastic noise
model = NormativeModel(BLR(heteroskedastic=True), 
                       inscaler='standardize', 
                       outscaler='standardize')

model.fit_predict(train, test)