Researcher: Euisuk Chung
Multiplexed imaging plays a crucial role in medical imaging by reducing scan time, minimizing motion artifacts, and decreasing physiological drift. This is especially beneficial for simultaneously monitoring coexisting drugs, cell populations, and functional readouts. Effective multiplexing strategies enhance quantification and improve biological interpretation by lowering variability between scans.
We propose a multiplex ultrasound imaging based on the distinct vaporization dynamics (VD) from perfluorocarbon nanodroplets (PFCnDs) with different optical absorbers. The vaporization and recondensation of laser-activated PFCnDs are recorded and analyzed using plane-wave ultrasound. We utilize non-negative matrix factorization (NMF) to separate unique VD signals from each PFCnDs.
NMF effectively separated signals within mixed samples and produced spatial component maps that correspond to the prepared compositions and their locations. The estimated component abundances align well with the known mixture ratios with low cross-talk, allowing interpretable images from a single-wavelength acquisition. The VD-based dictionary enabled consistent component recovery across multiple trials, indicating that dynamics-driven unmixing can be a practical method.