Dynamic Laser Speckle Imaging

Abstract

ABSTRACTUtilizing a high-speed camera and recording back-scattered laser light at more than 20,000 frames per second, we introduce the first wide-field dynamic laser speckle imaging (DLSI) in which we are able to quantify the laser speckleintensity temporal auto-correlation function g2(τ) for every pixel individually to obtain a quantitative image of the dynamics of the light scattering particles in the sample. The ability to directly and quantitatively measure the intensity auto-correlation function allows us to solve the problem of how to quantitatively interpret data measured by laser speckle contrast imaging (LSCI), multi-exposure laser speckle imaging (MESI) and laser Doppler flowmetry (LDF). The intensity auto-correlation function is related to the field temporal auto-correlation function g1(τ), which has been quantitatively related to the dynamics of the light scattering particles including flowing red blood cells. The form of g1(τ) depends on the amount of light scattering (i.e. single or multiple scattering) and the type of particle motion (i.e. ordered or unordered). Although these forms of the field correlation functions have been established for over 30 years, there is no agreement nor experimental support on what scattering and motion regimes are relevant for the varied biomedical applications. We thus apply DLSI to image cerebral blood flow in mouse through a cranial window and show that the generally accepted form of g1(τ), is applicable only to visible surface vessels of a specific size (20 – 200μm). We demonstrate that for flow in smaller vessels and in parenchymal regions that the proper g1(τ) form corresponds with multiple scattering light and unordered motion which was never considered to be relevant for these techniques. We show that the wrong assumption for the field auto-correlation model results in a severe underestimation of flow changes when measuring blood flow changes during ischemic stroke. Finally, we describe how DLSI can be integrated with other laser speckle methods to guide model selection, or how it can be used by itself as a quantitative blood flow imaging technique.