Handheld motion stabilized laser speckle imaging.

Laser speckle imaging (LSI) is a wide-field, noninvasive optical technique that allows researchers and clinicians to quantify blood flow in a variety of applications. However, traditional LSI devices are cart or tripod based mounted systems that are bulky and potentially difficult to maneuver in a clinical setting. We previously showed that the use of a handheld LSI device with the use of a fiducial marker (FM) to account for motion artifact is a viable alternative to mounted systems. Here we incorporated a handheld gimbal stabilizer (HGS) to produce a motion stabilized LSI (msLSI) device to further improve the quality of data acquired in handheld configurations. We evaluated the msLSI device using flow phantom experiments and using a dorsal window chamber model. For experiments, we quantified the speckle contrast of the FM (K) using the mounted data set and tested 80% and 85% of K as thresholds for useable images (K and K). Handheld data sets using the msLSI device (stabilized handheld) and handheld data sets without the HGS (handheld) were collected. Using K and K as the threshold, the number of images above the threshold for stabilized handheld (38 ± 7 and 10 ± 2) was significantly greater (p = 0.031) than for handheld operation (16 ± 2 and 4 ± 1). We quantified a region of interest within the flow region (K), which led to a percent difference of 8.5% ± 2.9% and 7.8% ± 3.1% between stabilized handheld and handheld configurations at each threshold. For experiments, we quantified the speckle contrast of the window chamber (K) using the mounted data set and tested 80% of K (K). Stabilized handheld operation provided 53 ± 24 images above K, while handheld operation provided only 23 ± 13 images. We quantified the speckle flow index (SFI) of the vessels and the background to calculate a signal-to-background ratio (SBR) of the window chamber. Stabilized handheld operation provided a greater SBR (2.32 ± 0.29) compared to handheld operation (1.83 ± 0.21). Both the number of images above threshold and SBR were statistically significantly greater in the stabilized handheld data sets (p = 0.0312). These results display the improved usability of handheld data acquired with an msLSI device.