prism-neuro 0.1.0

PeRmutation Inference for Statistical Mapping

Prism

Fast, modular, and extensible permutation-based statistical inference for neuroimaging.

Prism is a Python library for running fast, scalable, and fully nonparametric statistical analyses on brain imaging data. It replicates much of the core functionality of Anderson Winkler's PALM, but without MATLAB dependencies.

---

📚 Documentation

Full documentation, including API reference and usage guides, is available at:
https://josephiturner.com/prism/

For background and motivation, see the project manuscript.

---

🚀 Features

• Mass univariate GLM analysis with flexible contrast modeling
• Permutation-based testing including sign flips and blockwise shuffling
• Support for TFCE, FDR, FWE (Westfall–Young), and GPD-based p-value tail estimation
• Voxelwise map comparison tools for assessing spatial similarity
• CLI interface modeled after PALM
• Works directly with NIfTI files or NumPy arrays

---

🛠️ Installation

Prism is available on PyPI as prism-neuro:

pip install prism-neuro

Note: While the package is named prism-neuro on PyPI, you still use import prism in your code.

Development Installation

git clone https://github.com/josephisaacturner/prism.git
cd prism
pip install -e .

If you're on MacOS with Apple silicon, you may need to manually install a jax dependency:

pip install jax-metal

---

Minimal Example

import numpy as np
from prism.datasets.dataset import Dataset

Y = np.random.randn(100, 50)        # Brain data (samples x voxels)
X = np.random.randn(100, 2)         # Design matrix
C = np.array([1, -1])               # Contrast

dataset = Dataset(
    data=Y,
    design=X,
    contrast=C,
    output_prefix="prism_example",
    n_permutations=1000
)

results = dataset.permutation_analysis()

---

Contributing

We welcome contributions! To get started:

1. Fork this repository
2. Create a new branch: git checkout -b feature-name
3. Make your changes
4. Commit: git commit -m "Description"
5. Push and open a pull request!

---

📂 Project Structure

prism/
├── prism/                            # Core Python package
│   ├── data/                         # Brain templates or masks
│   ├── datasets/                     # Dataset class logic
│   ├── stats/                        # GLM and stat functions
│   ├── permutation_inference.py      # Main permutation logic
│   ├── preprocessing.py              # Data loading, masking, preprocessing
│   ├── tfce.py                       # TFCE implementation
│   ├── spatial_similarity.py         # Spatial map correlation engine
│   ├── prism_cli.py                  # Command-line interface
├── docs/                             # Markdown documentation (MkDocs)
├── notebooks/                        # Example Jupyter notebooks
├── tests/                            # Unit tests
├── assets/                           # Static images and logos
├── manuscript/                       # Project manuscript
├── README.md                         # You are here
├── requirements.txt                  # Dependencies
├── pyproject.toml                    # Build system metadata
└── LICENSE

---

License

Prism is released under the MIT License.