Improvement of Decorrelation-Based OCT Angiography by an Adaptive Spatial-Temporal Kernel in Monitoring Stimulus-Evoked Hemodynamic Responses

Complex decorrelation-based OCT angiography (OCTA) has the potential for monitoring hemodynamic activities in a label-free, high-resolution, and quantitative manner. To improve the measurement dynamic range and uncertainty of blood flow, an adaptive spatial-temporal (ST) kernel was proposed for decorrelation estimation and it was validated through a theoretical simulation and experimental measurements. The ensemble size in the decorrelation computation was effectively enlarged by collecting samples of the phasor pair in both the spatial and temporal dimensions. The spatial sub-kernel size was adaptively changed to suppress the influence of bulk motion in the temporal dimension by solving a maximum entropy model. Using the flow phantom, it was observed that the decorrelation dynamic range presented an increase of ~49% and the uncertainty exhibited a decrease of ~40% and ~38% in the saturation and background limits, respectively. In monitoring the stimulus-evoked hemodynamic response, the extended dynamic range enabled an improvement of ~180% in the separability between different stimulation modes. Furthermore, the suppressed uncertainty and motion artifacts allowed a reliable temporal analysis of the hemodynamic response. The proposed adaptive ST-kernel will greatly promote the development of decorrelation-based quantitative OCTA in hemodynamic studies.