ArchePy: Multi-Subject Archetypal Analysis for fMRI

ArchePy is a Python port of the Multisubject Archetypal Analysis (MS-AA) Toolbox originally developed in MATLAB at the Technical University of Denmark by Hinrich, Bardenfleth, Røge, Churchill, Madsen, and Mørup.

It implements both spatial and temporal MS-AA with optional GPU acceleration via CuPy and includes minimal helpers for working with fMRI NIfTI data.

Important

Academic / non-profit use only. ArchePy is distributed under the same license as the upstream MATLAB toolbox, which restricts use to academic and non-profit institutions. Commercial users must obtain a separate license from DTU. See the license page for full terms.

Quick example

import numpy as np
from archepy import Subject, multi_subject_aa

rng = np.random.default_rng(0)
subjects = [
    Subject(
        X=rng.standard_normal((200, 1000)),
        sX=rng.standard_normal((200, 1000)),
    )
    for _ in range(3)
]

results, C, cost, varexpl, elapsed = multi_subject_aa(
    subjects, noc=10, opts={"maxiter": 100, "rngSEED": 42}
)
print(f"Variance explained: {varexpl * 100:.1f}%")

For an end-to-end fit on real fMRI data, see the tutorials section.

Installation

pip install archepy           # CPU only
pip install archepy[gpu]      # with CuPy (CUDA 12.x)
pip install archepy[fmri]     # with nibabel
pip install archepy[all]      # everything

See the README for full installation details and dev setup.

Contents

User guide

Reference

Project

Indices