Fast brain fiber mapping with compressive hadamard-encoded scattered light imaging

Abstract

Fast, signal-efficient mapping of nerve fiber orientations is essential for resolving cortical microstructure, particularly in brain regions containing multiple crossing fiber populations. This work extends Computational Scattered Light Imaging (ComSLI) by combining Hadamard-encoded angular illumination with a compressive-sensing reconstruction framework. Hadamard multiplexing increases photon efficiency by illuminating multiple angular directions per exposure, while the compressive-sensing step exploits the sparse structure of angular scattering profiles to recover orientation-relevant information from fewer measurements. This enables further measurement compression, reducing the required number of illumination patterns to 25–12.5% of conventional angular sampling while preserving the scattering peak positions needed for fiber-orientation estimation. The combined approach addresses two central bottlenecks of LED-display-based ComSLI: limited illumination intensity and long acquisition time. It provides a pathway toward faster, compressed imaging of complex gray- and white-matter fiber architectures.