modewise

Data-adaptive signal decomposition for multichannel recordings: fMRI first, and EEG, MEG and fNIRS by design. A Rust core with a Python package and a command line. Written for the Brain Kyb Lab at NTNU, and open to anyone.

modewise splits signals into oscillatory modes (MVMD, EMD and their relatives, wavelets), turns them into time–frequency maps, connectivity and band features, and tests them across a cohort. Analyses are pipelines you write in TOML or Python. Steps are cached by content, every run is recorded, and each method is checked against independent references.

Start in Python, Rust or on the command line. Operations and their parameters are listed in the operation reference.

Getting started

• Python
• Rust
• Command line
• For agents

User guide

• Pipelines
• Datasets
• Epochs and decompositions
• Regions
• Volumes
• Time–frequency maps
• Connectivity
• Cohort features
• Statistics

Examples

• Examples
• Case studies

Methods

• Empirical mode decomposition
• Wavelet multiresolution analysis
• Variational Latent Mode Decomposition (VLMD)
• Thesis acceptance

Reference

• Operation reference
• Python API
• Rust API