Quantitative hemodynamic imaging: a method to correct the effects of optical properties on laser speckle imaging

Abstract

Significance: Studying cerebral hemodynamics may provide diagnostic information on neurological conditions. Wide-field imaging techniques, such as laser speckle imaging (LSI) and optical intrinsic signal imaging, are commonly used to study cerebral hemodynamics. However, they often do not account appropriately for the optical properties of the brain that can vary among subjects and even during a single measurement. Here, we describe the combination of LSI and spatial-frequency domain imaging (SFDI) into a wide-field quantitative hemodynamic imaging (QHI) system that can correct the effects of optical properties on LSI measurements to achieve a quantitative measurement of cerebral blood flow (CBF).

Aim: We describe the design, fabrication, and testing of QHI.

Approach: The QHI hardware combines LSI and SFDI with spatial and temporal synchronization. We characterized system sensitivity, accuracy, and precision with tissue-mimicking phantoms. With SFDI optical property measurements, we describe a method derived from dynamic light scattering to obtain absolute CBF values from LSI and SFDI measurements. We illustrate the potential benefits of absolute CBF measurements in resting-state and dynamic experiments.

Results: QHI achieved a 50-Hz raw acquisition frame rate with a $10\times 10\mathrm{mm}$ field of view and flow sensitivity up to $\sim 4\mathrm{mm}/s$. The extracted SFDI optical properties agreed well with a commercial system ($R^2\ge 0.98$). The system showed high stability with low coefficients of variations over multiple sessions within the same day ($<1\%$) and over multiple days ($<4\%$). When optical properties were considered, the in-vivo hypercapnia gas challenge showed a slight difference in CBF ($-1.5\%$ to 0.5% difference). The in-vivo resting-state experiment showed a change in CBF ranking for nine out of 13 animals when the correction method was applied to LSI CBF measurements.

Conclusions: We developed a wide-field QHI system to account for the confounding effects of optical properties on CBF LSI measurements using the information obtained from SFDI.

Keywords: cerebral hemodynamics; diffuse optics; laser speckle imaging; spatial frequency domain imaging; tissue optics.

Fig. 1

Overall schematic of the QHI system. (a) Hardware setup of QHI with LSI and SFDI components [AL, aspheric lens; AcL, achromatic lens; CL, condenser lens; D, diffuser; DMD, digital micromirror device; F, appropriate notch (for SFDI) and line filters (for LSI) at LSI wavelength, IL, imaging lens with spacers; L, plano-convex lens; LED, light emitting diode; LPDM, long-pass dichroic mirror; P, linear polarizer (cross-polarized between illumination and detection arms); and SPDM, short-pass dichroic mirror]. (b) Triggering scheme of the system where the computer sends a serial trigger to activate an automatic Arduino triggering scheme for the LED and the DMD. The electrical output of the DMD was then used to trigger the camera. The exposure time was set to 10 ms with a 20 ms period per full cycle, allowing for an effective 50 Hz raw frame rate. The Arduino triggering scheme can be deactivated with another serial trigger from the computer.

Fig. 2

Flow chart of the method of obtaining absolute flow value using the Brownian diffusion coefficient $D_b$ from measured $K$, ${\mu }_a$, and ${{\mu }_s}^{\prime }$. Here, a $\beta$ factor was approximated using a static silicone phantom.

Fig. 3

Flow phantom experiment to characterize the sensitivity of the LSI arm. (a) A schematic of the silicone flow phantom with an embedded glass capillary pipet tube where a solution of 1% intralipid was flowed at different speeds using a syringe pump. A representative flow map of the captured FOV was also shown with the chosen ROI (black rectangles) for static and flow regions. (b) Time-course data of the experiment where the flow speed was increased at each 1-min interval (top table). The time course of ${\mathrm{SFI}}_{\beta \text{-corrected}}$ within the flow region showed an increasing pattern with increasing speed between 0 and $\sim 4\mathrm{mm}/\mathrm{s}$. A plateau pattern was observed at higher speeds ($>4\mathrm{mm}/\mathrm{s}$). For the static region, ${\mathrm{SFI}}_{\beta \text{-corrected}}$ showed no change during the experiment.

Fig. 4

Correlation (top) and Bland-Altman plots (bottom) between QHI and a commercial ReflectRS system from Modulim of four static silicone phantoms. Here, we compared the median values obtained from the chosen ROI. Only 660 nm and 850 nm wavelengths were compared, as 780 nm is not available with the ReflectRS. For the correlation plot, $R^2$ values were also listed for the linear fit.

Fig. 5

Measurements of QHI precision were performed for six trials (5 min each) within the same day on the same static silicone phantom (i.e., phantom 1). The mean and standard deviation are plotted here. The recorded values for ${\mu }_a$, ${{\mu }_s}^{\prime }$, and $K$ and the calculated coefficient of variation across all trials were shown in Table S1 in the Supplementary Material.

Fig. 6

Measurements of QHI precision over five days. The median values within the selected ROI for each day are plotted and further used in the analysis. The recorded values for ${\mu }_a$, ${{\mu }_s}^{\prime }$, and $K$ and the calculated coefficient of variation across all five days are shown in Tables S2–S4 in the Supplementary Material. $K$ is higher for the phantom with higher ${\mu }_a$ (phantom 1 versus phantom 2) and lower for the phantom with higher ${{\mu }_s}^{\prime }$ (phantom 1 versus phantom 3). With higher ${\mu }_a$ and ${{\mu }_s}^{\prime }$ (phantom 4), $K$ is similar to that of phantom 1. This agrees with the findings of Mazhar et al.

Fig. 7

CBF dynamics slightly differentiate between $D_b$ and SFI measurements for a mouse undergoing a hypercapnia gas challenge. (a) Representative images of $D_b$ and SFI during hypercapnia with the chosen semicircular ROI (black) on the left hemisphere. Scale bars are 1 mm. b) CBF measurements normalized to $t=0$ where both SFI and $D_b$ analysis was performed. Median data are chosen from the left hemisphere to extract the time course of SFI and ${\mathrm{D}}_b$ values. The data is further filtered using a moving filter of 200 frames ($\sim 4\mathrm{s}$). (c) A difference of $-1.5$ to 0.5% between the normalized $D_b$ and SFI curves is observed throughout the gas challenge experiment. (d) Optical properties derived at 633 nm showing a slight decrease in both ${\mu }_a$ and ${{\mu }_s}^{\prime }$ across the time course.

Fig. 8

Baseline CBF measurements from 13 mice (2% isoflurane for 1 min) highlighted the effect of optical properties on the interpretation of CBF measurements using LSI. (a) Median values of data obtained from the left hemisphere were used to generate time course data for the 1-min imaging period. Both the pulsatile CBF waveform and breathing-related motion artifacts could be resolved. On the representative image on the right, the scale bar is 1 mm. (b) Mean values of the SFI time course during each imaging period were then used to rank the animals in ascending order (blue to red circles). Animal IDs (1 to 13) were then assigned based on this ranking. Here, the $D_b$ values differed from SFI in ranking for nine out of 13 animals.

Fig. 9

On one representative animal, $D_b$, when compared to SFI map, showed a spatial difference in CBF measurements but not in temporal dynamics. (a) Representative normalized maps of $D_b$ and SFI highlight the spatial difference in estimated CBF. Individual pixels in each image were normalized relative to the respective center pixel (*). A percentage difference map was generated using $\%\text{Difference}=(D_{b,\text{norm}}-{\mathrm{SFI}}_{\text{norm}})/{\mathrm{SFI}}_{\text{norm}}\times 100$. $D_b$ showed increased blood flow in large vessels (red arrows) compared with SFI. Scale bars are 1 mm. (b) Within the chosen semicircular ROI, temporal dynamics (i.e., rCBF) did not show notable differences between SFI and $D_b$ measurements among regions with decreasing ($<-5\%$), minimally changing ($-5$ to 5%), and increasing ($>5\%$) spatial differences. The power spectrum of both $D_b$ and SFI (not shown here) showed negligible differences in the pulsatile waveforms.