Predictive Clinical Neuroscience software toolkit (formerly nispat).
Methods for normative modelling, spatial statistics and pattern recognition. Documentation, including tutorials can be found on readthedocs. Click on the docs button above to visit the site.
using the download here: https://www.anaconda.com/download
conda create -n <env_name> python==3.12
source activate <env_name>
Using pip:
pip install pcntoolkit
Using a local clone of the repo:
python -m pip install .
python -c "import pcntoolkit as pk;print(pk.__file__)"
On some older architectures, the installation of nutpie may fail due to a missing Rust compiler. If this is the case, you can install it manually using conda prior to installing PCNtoolkit:
# install nutpie
conda install -c conda-forge nutpie
# install PCNtoolkit
pip install pcntoolkit
Otherwise install it first from https://www.anaconda.com/
conda --version
conda create -y python==3.12 --prefix=/shared/conda/<env_name>
conda activate /shared/conda/<env_name>
git clone https://github.com/amarquand/PCNtoolkit.git
cd PCNtoolkit/
python -m pip install .
python -c "import pcntoolkit as pk;print(pk.__file__)"
For normative modelling, functionality is handled by the normative.py script, which can be run from the command line, e.g.
# python normative.py -c /path/to/training/covariates -t /path/to/test/covariates -r /path/to/test/response/variables /path/to/my/training/response/variables
For more information, please see the following resources:
- documentation
- developer documentation
- a tutorial and worked through example on a real-world dataset