fmrilss: Least Squares Separate (LSS) Analysis for fMRI Data

This package implements efficient least squares separate (LSS) analysis for functional magnetic resonance imaging (fMRI) data. LSS is used to estimate trial-by-trial activation patterns in event-related fMRI designs.

Main functions

• lss: Main function for performing LSS analysis

• lss_naive: Naive LSS implementation for reference

• project_confounds: R implementation for projecting out confounds

• project_confounds_cpp: Fast C++ confound projection

• lss_beta_cpp: Vectorized C++ LSS beta computation

• get_data_matrix: Helper function for data extraction

Features

• Optimized C++ implementation using vectorized matrix algebra

• Memory-efficient projection without forming Q matrices

• Cholesky decomposition for numerical stability

• Fallback R implementation with QR decomposition

• Support for various design matrix configurations

• Robust numerical handling for edge cases

• OpenMP support for multi-core processing

Author

Your Name <your.email@example.com>