Rapid monitoring of cerebral ischemia dynamics using laser-based optical imaging of blood oxygenation and flow

Abstract

Imaging blood flow or oxygenation changes using optical techniques is useful for monitoring cortical activity in healthy subjects as well as in diseased states such as stroke or epilepsy. However, in order to gain a better understanding of hemodynamics in conscious, freely moving animals, these techniques must be implemented in a small scale, portable design that is adaptable to a wearable format. We demonstrate a novel system which combines the two techniques of laser speckle contrast imaging and intrinsic optical signal imaging simultaneously, using compact laser sources, to monitor induced cortical ischemia in a full field format with high temporal acquisition rates. We further demonstrate the advantages of using combined measurements of speckle contrast and oxygenation to establish absolute flow velocities, as well as to statistically distinguish between veins and arteries. We accomplish this system using coherence reduction techniques applied to Vertical Cavity Surface Emitting Lasers (VCSELs) operating at 680, 795 and 850 nm. This system uses minimal optical components and can easily be adapted into a portable format for continuous monitoring of cortical hemodynamics.

Keywords: (140.2020) Diode lasers; (170.0110) Imaging systems; (170.3880) Medical and biological imaging; (170.6480) Spectroscopy, speckle.

Fig. 1

(a) Spectra for 680 nm, 795 nm, and 850 nm VCSELs. In each case, measurements were taken at threshold current (black), peak power (red), and current sweep operation (SW) through a range optimized for each device. (b) Interferogram envelopes for three wavelengths in SW operation, showing similar coherence lengths in each case. (c) Speckle contrast for 680 nm VCSEL near threshold, showing high contrast dropping off as power is increased

Fig. 2

(a) Multiwavelength system schematic. The three wavelength VCSEL package is driven using two current sources, which are alternated along with a rapid switch synchronized by a camera trigger. Software and hardware control is accomplished with a customized frame grabber program. The trigger sequence is shown on the bottom right, with SW and SM waveforms. (b) and (c) Detail of multi-wavelength VCSEL package. In the center of (c), three 850 nm wafers are visible, as well as one multi-VCSEL wafer each for 795 nm and 680 nm wavelengths.

Fig. 3

Processed image maps of flow and oxygenation (see Media 1). (a) Baseline LSCI map, showing speckle contrast values. (b)-(d) Percentage change in reflectance at peak of ischemia, for 680, 850 and 795 nm illumination respectively. Scale is matched to (a). (e) Montage of changing flow values during ischemia using relative flow index. 0 s indicates onset of ischemia. (f)-(h) Montage of concentration changes during ischemia for HbR, HbO and HbT respectively. Units are 10⁻⁵ M. Scale is matched to (e).

Fig. 4

Time course plots of 3 different vessels showing changes in blood flow, HbR, HbO and HbT. Grayed region indicates duration of ischemic induction. Significantly stronger ischemic responses are seen in veins compared to arteries. Further, larger veins exhibit stronger response than smaller veins.

Fig. 5

Line scan averaging for flow quantification in noisy data. (a) Averaged raw image shows vessels as seen with 680 illumination. Selected vessel flow convention is indicated by arrows. (b) Single line scan ‘carpet’ shows pixel intensities over time. (c) FFT averaged over several carpets shows preferred components with slope proportional to flow speed (d) Estimated flow speed map, based on LSCI image and calibration from single vessel line scan. Yellow arrows indicate flow direction. (e) Correlation between speckle flow index and flow velocity established from line scan. Measurements on vessels 1-6 as indicated in (d) are shown.

Fig. 6

Classification of veins and arteries using principal component analysis. (a) Scatter plot of data projections on first two principal components, showing partition estimate. The final result is found to be independent of initial partition. (b) 3D scatter plot showing vessel data against vessel diameter, baseline flow speed and [HbO] changes. All units are normalized. Projections in x-y and y-z are shown to aid visualization. (c) Resulting vessel classification mapped onto calibrated LSCI flow map. Yellow vessels are arteries and cyan vessels are veins.