28 Apr nanoparticle flow velocimetry using multi-dimensional confocal imaging: measurement accuracy versus diffusion
This article presents a method for three-dimensional confocal microscopy to study the kinematics of nanometer-sized molecules and particles that are of great interest to a wide range of biological systems and cellular mechanics but have heretofore been obscured by limitations in measurement technology. The correlation-based velocity measurements from confocal laser scanning microscopy (CLSM) images are subject to random error due to the Brownian motion of nanometer-sized tracer particles and a bias error due to the formation of images by raster scanning. In our previous work, we developed a scanning laser image correlation-robust phase correlation (SLICR) and showed improved measurement accuracy by establishing an optimal spectral filter diameter combined with an analytical model of the CLSM measurement bias error (Jun et al., 2016). While the SLICR algorithm was designed for one-dimensional measurements, in practice, CLSM is mostly used for two or three-dimensional microscopic fields. The measurement errors significantly increase with the higher dimensionality in CLSM images. In this work, our new algorithm tackles these limitations twofold. First, we model and correct for the bias errors introduced by the effects of the volumetric laser scanning image acquisition. Second, we develop a new spectral filter using a phase masking technique that is optimized for the spectral content of CLSM images, without requiring a priori knowledge of displacement fields or flow tracer properties. We show that our method outperformed the standard cross-correlation (SCC) in reducing the random and bias errors and accelerated the convergence of ensemble correlation velocity measurements from CLSM images.
Figure 1: (a) Illustration of raster scanning pattern from CLSM, reconstructing 2D and 3D image coordinated by scanning mirrors and associated bias error in each component of the velocity measurement, (b) Phase-angle and phase quality planes showing varying degrees of SNR, and (c) algorithm for constructing the phase quality mask
Figure 2: (a) Convergence of SLICQ and SCC algorithms for images of 3 and 100nm flow tracer particles for the U component velocity measurement normalized by the expected value, (b) representative spatial cross-correlation types observed from the processed images, and (c) 3D velocity measurements with 3nm mCherry protein and 100nm particles.
No Comments