Open Access
Add to BibSonomy
Abstract
To evaluate cerebral hemodynamics and spontaneous low-frequency oscillations (SLFOs) of cerebral blood flow in rat brain, we investigated an imaging method using a digital RGB camera. In this method, the RGB values were converted into tristimulus values in the CIE (Commission Internationale de l’Eclairage) XYZ color space, which is compatible with the common RGB working spaces. Monte Carlo simulation for light transport in tissue was then used to specify the relationship among the tristimulus XYZ values and the concentrations of oxygenated hemoglobin (CHbO), deoxygenated hemoglobin (CHbR), and total hemoglobin (CHbT) and cerebral tissue oxygen saturation (StO2). Applying the fast Fourier transform to each pixel of the sequential images of CHbT along the timeline, SLFOs of cerebral blood volume were visualized as a spatial map of power spectral density (PSD) at specific frequencies related to vasomotion. To confirm the feasibility of this method, we performed in vivo experiments using exposed rat brain during a cortical spreading depression (CSD) evoked by topical application of KCl. Cerebral hemodynamic responses to CSD such as initial hypoperfusion, profound hyperemia, and post-CSD oligemia and hypoxemia were successfully visualized with this method. At the transition to the hyperemia phase from hypoperfusion, CHbO and StO2 were significantly increased, which implied vasodilatation in arterioles and increased cerebral blood volume in response to CSD. In the wake of the hyperemic phase, CHbO and CHbT were significantly reduced to 25 ± 12% and 3.5 ± 1% of baseline, respectively, suggesting long-lasting vasoconstriction after CSD. In this persistent oligemia, StO2 significantly dropped to at most 23 ± 12% of the level before CSD, indicating long-lasting hypoxemia. The PSD value of SLFOs in CHbT for arteriole regions during CSD was significantly reduced to 28 ± 20% of baseline with respect to the pre-CSD level, which was correlated with the reduction in StO2. The results showed the possibility of RGB camera-based diffuse reflectance spectroscopy imaging for evaluating cerebral hemodynamics and SLFOs under normal and pathologic conditions.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Cortical spreading depression (CSD) is a wave of slow neuronal and glial depolarization propagating at approximately 2 to 5 mm/min over the cerebral cortex [1–3]. CSD is an important disease model for migraine [4] and is associated with various neurological disorders, such as seizure [5], ischemia [6, 7], stroke [8], and traumatic brain injury [9–11]. CSD initiated in vivo has been discussed in terms of the changes in optical intrinsic signals related to cerebral hemodynamics, such as changes in cerebral blood volume and the oxygenation state of hemoglobin, originating from neuro-vascular coupling in brain tissue [12]. Figure 1 shows a schematic diagram of spreading depression-induced sequential changes in cerebral blood flow (CBF) and the slow shift in extracellular local field potential (LFP). Three hemodynamic changes have been observed as responses to CSD [13, 14]. The first is hypoperfusion due to vasoconstriction that synchronizes with electrical depolarization, possibly induced by the elevation of extracellular potassium and vasoactive mediators released from neurons in parenchymal regions, astrocytes, and perivascular nerves during CSD. The second and most distinguishing change is profound hyperemia that is observed at or soon after the onset of DC shift of LFP. The third hemodynamic change is the longest-lasting attenuation of blood perfusion, which is called post-CSD oligemia. In this oligemic phase, tissue oxygen tension, which represents the balance between local oxygen supply and demand, is also persistently decreased below the pre-CSD baseline level [15]. To investigate the relationship between CSD and clinical disorders, evaluating the changes in hemodynamics of in vivo brain tissue is important.
Fig. 1 Schematic diagram of spreading depression-induced sequential changes in the extracellular slow LFP shift and CBF. Alphabetic characters represent deflection points at which the amplitude of CBF changes: onset of hypoperfusion phase, A; negative peak of hypoperfusion phase, B; positive peak of hypoperfusion phase, C; onset of post-CSD oligemia phase, D; bottom of post-CSD oligemia phase, E.
On the other hand, cerebral hemodynamics is always fluctuating due to various physiological factors. The power spectrum obtained from cerebral hemodynamics can be roughly divided into two components. The high-frequency component is related to heart beat and respiration. The low-frequency component is associated with vasomotion, which is spontaneous contraction and relaxation of arterioles (and in some instances venules), and is independent of heart beat and respiration. The frequency band of vasomotion can be divided into three different subcomponents based on the cause of the oscillation [16, 17]. The first subcomponent ranging from 0.04 to 0.15 Hz is called the myogenic component, which is associated with the activity of smooth muscle cells of arterioles [18]. The second one, which is called the neurogenic component, ranges from 0.02 to 0.04 Hz and is related to intrinsic neuronal activity [19]. The third and very low-frequency oscillations, known as the endothelial component, range from 0.003 to 0.02 Hz and represent the activity of endothelial cells in arterioles [18, 20]. Vasomotion is related to cerebral autoregulation, such as regulation of blood flow and vascular resistance, cancellation of the hypoxic region in the capillary plexus, and prevention and reduction of edema. In particular, 0.1-Hz vasomotion is correlated with cerebral vascular reactivity (CVR) [21–23], which is the change in CBF in response to a vasodilatory or vasoconstrictive stimulus. Reduction in CVR occurs in various cerebral diseases and dysfunctions, such as stroke, traumatic brain injury, and CSD. Although vasomotion is strictly a local phenomenon, the regulation of contractile activity of vascular smooth muscle cells is dependent on the complex interplay between vasodilator and vasoconstrictor stimuli from circulating hormones, neurotransmitters, endothelial-derived factors, and blood pressure. Therefore, evaluation of spontaneous oscillations in CBF may be a useful method for assessing risk and investigating different treatment strategies in neurological disorders, such as traumatic brain injury, seizure, ischemia, and stroke.
CBF during CSD in rodents has been investigated by laser speckle flowmetry [24], laser Doppler flowmetry [25], and the diffuse optical correlation method [26]. Diffuse reflectance spectroscopy (DRS) is also one of the most promising methods for assessing cerebral hemodynamics. DRS can be achieved simply with an uncomplicated optical system with a broad band light source, inexpensive optical components, and a spectrometer. Several approaches using a numerical simulation-based lookup table have been investigated for evaluating the absorption properties of biological tissues [27–31]. Various multispectral imaging systems using a filter wheel installed with several narrow-band optical filters have been utilized to visualize hemodynamic responses in rodent brain to cerebral focal ischemia [32], CSD [33–35], global hypoxia [36], and changes in the fraction of inspired oxygen [37]. Such a conventional multispectral imaging system is relatively time-consuming because the filter positions in the wheel have to be mechanically changed. This means that the imaging system sequentially captures each diffuse reflectance image at a different wavelength point. An acousto-optical tunable filter [38], the combination of a lenslet array with narrow-band filters [39], and a rapid modulation light emitting diode illumination system [40] have the potential to improve temporal resolution and the time lag of multispectral image acquisition at different wavelengths. Estimation of multispectral images from a red green blue (RGB) image acquired with a digital RGB camera is a promising method for rapid and cost-effective imaging. A multispectral imaging technique based on the Wiener estimation method has been applied to visualize cortical hemodynamics while varying the fraction of inspired oxygen [41]. However, the method is time-consuming for analyzing spectral data at each pixel of the multispectral image cube to reconstruct a set of images of oxygenated hemoglobin and deoxygenated hemoglobin.
In the present study, we investigated a simple and rapid imaging method for the oxygenated hemoglobin concentration (CHbO), deoxygenated hemoglobin concentration (CHbR), total hemoglobin concentration (CHbT), and tissue oxygen saturation (StO2) of in vivo exposed brain tissues based on DRS using a digital RGB camera. In the method, RGB values of each pixel of a color image were converted into tristimulus XYZ values in the CIE (Commission Internationale de l’Eclairage) XYZ color space, which is compatible with the common RGB working spaces. Monte Carlo simulation (MCS) for light transport in homogeneous tissue was introduced to specify a relationship among the XYZ values and CHbO and CHbR. CHbT and StO2 were also calculated from the estimated concentrations of oxygenated and deoxygenated hemoglobin. To confirm the feasibility of a method to evaluate CHbO, CHbR, CHbT, and StO2 of the cerebral cortex as hemodynamics parameters, we performed in vivo experiments using exposed rat brain during CSD. Furthermore, we extended this method for visualization of spontaneous low-frequency oscillations (SLFOs) in cerebral blood volume. We extracted the spontaneous low-frequency oscillation component from sequential images of the total hemoglobin concentration and reconstructed a spatial map of power spectral density (PSD) at specific frequencies.
2. Principle
2.1 Relationship between RGB values and hemoglobin concentration
RGB values of a pixel of an image of an exposed cerebral cortex acquired by a digital camera can be expressed as
where andare tristimulus values in the CIEXYZ color system. L1 is a transformation matrix to convert XYZ values to the corresponding RGB values. λ, Φ(λ), and Θ(λ) are the wavelength, spectral distribution of illuminant, and diffuse reflectance spectrum of the cerebral cortex, respectively. , , and are color matching functions in the CIEXYZ color system. The value of the constant κ that result in Y being equal to 100 for a perfect diffuser is given byIntegrals are executed over the visible wavelength range (400 to 700 nm). Assuming that the cerebral cortex contains oxygenated hemoglobin and deoxygenated hemoglobin, the diffuse reflectance obtained from the exposed cortical surface Θ can be expressed as [42]where, I0 and I are the incident and detected light intensities, respectively. p(μs’, l) is the path length probability function that depends on the scattering properties as well as the geometry of the measurements. μs’, μa, and l are the reduced scattering coefficient, absorption coefficient, and photon path length, respectively. Subscripts HbO and HbR indicate oxygenated hemoglobin and deoxygenated hemoglobin, respectively. The absorption coefficient of each chromophore is expressed as the product of its concentration C and extinction coefficient ε as μa = Cε. Therefore, the RGB values are expressed as the functions of CHbO and CHbR.2.2 Estimation of hemoglobin concentrations based on RGB images
Figure 2 shows the flow of estimation by the method. First, RGB values in each pixel of the image are transformed into XYZ values by multiplication with a matrix N1 as
in each pixel of the image. We assumed the following linear equations to estimate XYZ values from the measured RGB values. andFig. 2 Flow diagram of the process for estimating the concentration of oxygenated hemoglobin (CHbO), the concentration of deoxygenated hemoglobin (CHbR), the concentration of total hemoglobin (CHbT), and tissue oxygen saturation (StO2). (a) Preparation work for determining the multiple regression equations, (b) main process for deriving CHbO, CHbR, CHbT, and StO2 from the RGB value, and (c) two-dimensional spatial mapping of PSD by extracting low-frequency oscillation components at each pixel from time series CHbT images.
We measured RGB values for Color standard (ColorChecker, X-Rite Incorporated, Grand Rapids, MI, USA), which has 24 color patches and is supplied with data giving the CIEXYZ values for each patch under the specific spectral distribution of the light source and corresponding diffuse reflectance spectra. To statistically determine the coefficients of αi, βi, and χi (i = 0, 1, 2, 3), we performed multiple regression analysis. In the multiple regression analysis, the values of Xj, Yj, or Zj (j = 1, 2, …, 24) for each color patch were used as the response variable, whereas the measured values of Rj, Gj, or Bj for each color patch were used as the predictor variables. The regression coefficients αi, βi, and χi were used for the elements of a 4 × 3 matrix N1 as
Figure 3 shows the comparison between the estimated and expected values of X, Y, and Z obtained from the analysis. The estimated values well agree with the original values for each sample. The coefficient of determination r2 which is a statistical measure of how close the data are to the fitted regression line for X, Y, and Z were 0.98, 0.98, and 0.97, respectively, which indicates a good regression.Fig. 3 Comparison between the estimated and expected values of (a) X, (b) Y, and (c) Z obtained from the multiple regression analysis.
The values of X, Y, and Z are then transformed into CHbO and CHbR by multiplication with matrix N2. Specifying the elements of matrix N2 based on matrix L1 and Eqs. (2)-(6) is difficult because p(μs’, l) and l for each layer are usually unavailable. We simulated 450 diffuse reflectance spectra Θ(λ) in a wavelength range from 400 to 700 nm at intervals of 10 nm with MCS for light transport [43] in brain tissue. The simulation model used here consisted of a single layer of brain tissue, in which the absorption and scattering properties were uniformly distributed. In most biological soft tissues, the reduced scattering coefficient spectrum μs’(λ) can be approximated with the following power law function [44,45]:
where a and b are called the scattering amplitude and scattering power, respectively. The values of a and b are associated with microscopic geometric properties of the scattering medium such as scatterer density and size [44,45], respectively. The values of μa(λ) derived from the concentrations of CHbO and CHbR and those of μs’(λ) deduced by the combinations of a and b were used as inputs to the simulations of the diffuse reflectance spectrum Θ(λ). In this study, we used the reduced scattering coefficient spectrum published by [46] as the typical value. We fitted Eq. (12) to the typical scattering spectrum and obtained the values of 12 × 104 cm−1 and 1.38 for the scattering amplitude a and the scattering power b, respectively. Diffuse reflectance imaging with single or multiple wavelengths has suggested that during CSD, changes in light scattering occur due to cell deformation in tissue [33, 34]. Therefore, variations in light scattering properties should be considered in the MCS used for deriving the empirical formulae of CHbO and CHbR. To provide certain latitude in reduced scattering spectrum, five different values of 6.0 × 104, 9.0 × 104, 12 × 104, 15 × 104, and 18 × 104 cm−1 were calculated by multiplying the typical value [46] of a by 0.5, 0.75, 1.0, 1.25, and 1.5, respectively. In the same way, the five values of 1.24, 1.31, 1.38, 1.45, and 1.52 were calculated by multiplying the typical value [46] of b by 0.5, 0.75, 1.0, 1.25, and 1.5, respectively. Finally, the reduced scattering coefficients μs’(λ) of the brain tissue with the 25 different values were derived using Eq. (12).The absorption coefficients of oxygenated hemoglobin μa,HbO (λ) and deoxygenated hemoglobin μa,HbR (λ) were derived from the extinction coefficients of oxygenated hemoglobin εHbO (λ) and deoxygenated hemoglobin εHbR (λ) reported in literature [47], respectively, where the total hemoglobin concentration of a whole blood with a 44% hematocrit value is 2326 μM of hemoglobin [47]. The sum of the absorption coefficients of oxygenated hemoglobin and deoxygenated hemoglobin μa,HbO (λ) + μa,HbR (λ) = μa,HbT (λ) for CHbT = 116.3, 232.6, and 465.2 μM was calculated. All the three conditions of CHbT were used as input to the brain tissue in the simulation. Cerebral tissue oxygen saturation (StO2) was determined as StO2% = (CHbO/CHbT) × 100. It has been reported that long-lasting hypoxemia is occurred after CSD [11]. On the other hand, profound hyperemia will be observed after the onset of CSD, representing the increase in oxygenated hemoglobin. We used StO2 ranged from 0 to 100% in the MCS to consider not only normal physiological condition but also all possible hemodynamics responses to CSD. The six values of 0%, 20%, 40%, 60%, 80%, and 100% were chosen as StO2% for the simulation. In total, 450 diffuse reflectance spectra Θ(λ) over a wavelength range from 400 to 700 nm at intervals of 10 nm were numerically calculated. The XYZ values were then determined based on the simulated Θ(λ). The above procedures were carried out for the various combinations of CHbO, CHbR, a, and b to prepare the data sets of chromophore concentrations and XYZ values. Two regression equations for CHbO and CHbR were established by further multiple regression analysis with 450 data sets:
andThe regression coefficients γi and ωi (i = 0, 1, 2, 3) reflect the contributions of the XYZ values to CHbO and CHbR, respectively. Therefore, we used γi and ωi for the elements of a 4 × 2 matrix N2 asCalculation of CHbO and CHbR from XYZ values is thus expressed asOnce the matrices N1 and N2 are specified, the regression equations are applied to each pixel of the images of XYZ, and the two-dimensional distributions of oxygenated hemoglobin and deoxygenated hemoglobin are reconstructed as the images of CHbO and CHbR without MCS. The total hemoglobin concentration image is simply obtained as CHbT = CHbO + CHbR, whereas tissue oxygen saturation of hemoglobin is obtained as StO2% = (CHbO/CHbT) × 100. To visualize the intensity of the low-frequency oscillations in cerebral blood volume, we used sequential images of CHbT. By applying the fast Fourier transform to the time series signal of CHbT at each pixel of the sequential images along the time line, the two-dimensional spatial map of PSD in which each pixel has a frequency spectrum of oscillation in CHbT can be reconstructed.3. Experiments
3.1 Imaging system
Figure 4(a) shows a schematic diagram of the experimental system used in the present study. A white-light emitting diode (LA-HDF5010, Hayashi Watch Works Co., Ltd., Tokyo, Japan) illuminated the surface of the exposed cortex via a light guide. The angle of illumination was approximately 45° with respect to the cortical surface, to exclude specular reflection. Diffusely reflected light from the brain was captured with a 24-bit RGB CCD camera (DFK-31BF03.H, Imaging Source LLC, Charlotte, NC, USA) with a zoom lens to acquire an RGB image. A standard white diffuser with 99% reflectance (SRS-99-020, Labsphere Incorporated, North Sutton, NH, USA) was used as a reference material to calibrate the white balance of the camera and to control light guide output. The RGB images were acquired with a temporal resolution of eight frames per second and stored in a personal computer. The field of view of the system was 5.45 × 4.09 mm2 with 1024 × 768 pixels. The lateral resolution of the images was estimated to be 5.3 μm. We used the nominal value of spectral distribution of illuminant to calculate XYZ values. Therefore, we did not measure the spectrum at the light guide output directly.
Fig. 4 (a) Schematic diagram of the experimental system, and (b) representation of the rat skull with an RGB image of the exposed rat brain.
3.2 Animal experiments
Animal care and experimental procedures were approved by the Animal Research Committee of Tokyo University of Agriculture and Technology. Intraperitoneal anesthesia was initiated with α-chloralose (50 mg/kg) and urethane (600 mg/kg) in nine adult male Wistar rats (294 to 594 g). Anesthesia was maintained at a depth such that the rat had no response to a toe pinch. The rat’s head was placed in a stereotaxic frame. A longitudinal incision of approximately 20 mm in length was made along the head midline. The skull bone overlying the parietal cortex was removed with a high-speed drill to form an ellipsoidal cranial window (8.0-mm major axis and 6.0-mm minor axis), as shown in Fig. 4(b). The cranial window was bathed with normal saline. A burr hole (2-mm diameter) was drilled in the ipsilateral frontal bone as a site for applying 3 M KCl to the cortex. The KCl solution was applied to the cortical surface through the burr hole at 5 min after the onset of measurement and was washed out at 20 min after the onset of measurement. Therefore, the duration of KCl application was 15 min. CSD was induced by applying a 5 μl droplet of the KCl solution to the burr hole. The five rats were used to demonstrate the ability of the method for acquiring well-known hemodynamic changes due to CSD. Since the hemodynamic responses to CSD are not that fast, the RGB images were stored in a personal computer at five-second interval. The remaining four rats were used to confirm the possibility of the method for visualizing spontaneous low-frequency oscillation. To extract the oscillation signal at the specific frequency, data of high frequency resolution was required in this secondary experiment. Therefore, the RGB images were acquired with a temporal resolution of eight frames per second (fps) and stored in a personal computer at 0.125-second interval. The sequential RGB images measured were used to estimate the images of CHbO, CHbR, CHbT, and StO2 according to the process described in Fig. 2. By applying the fast Fourier transform to the time series signal of CHbT at each pixel of the sequential images along the time line, the two-dimensional spatial map of PSD in which each pixel had a frequency spectrum of oscillation in the vertebral blood volume could be reconstructed.
To evaluate the magnitude of signal S induced by CSD, we calculated the change in the signal based on the time series data. The signal at the beginning of the measurement was selected as a control Sc, which was subtracted from each of the subsequent signals S in the series. Each subtracted value, which demonstrated the change in the signal, S − Sc, over time, was normalized by dividing by Sc. The change in the signal is expressed as ΔS = {(S − Sc)/Sc} × 100. The above calculation was applied to the time series of CHbO, CHbR, CHbT, and StO2.
The extracellular LFP was recorded by a single Ag/AgCl electrode with a spherical tip (tip diameter: 1 mm). The recording electrode was placed on the cortex with care taken to avoid large blood vessels. An Ag/AgCl reference electrode (RC5, World Precision Instruments Inc., Sarasota, FL, USA) was placed in the neck muscle. The LFP signal was amplified at 1 to 100 Hz using a differential amplifier (DAM50, World Precision Instruments Inc.) under the DC mode and was digitized at 5 Hz using a data logger (midi LOGGER GL 900, GRAPHTEC, Corp., Yokohama, Japan) connected to a personal computer running GL900 application software.
3.3 Statistical considerations
A region of interest (ROI) was placed in each image for the analysis of time courses in CHbO, CHbR, CHbT, and StO2. Data are expressed as the mean ± SD. To compare whether the mean estimated value at each experimental condition differed significantly from the aggregate mean across experimental conditions, one-way repeated-measures ANOVA with Bonferroni’s correction was performed when comparing the relative change in the signal among the specific time points shown in Fig. 1: before the first CSD, A; initial hypoperfusion of the first CSD, B; hyperemia of the first CSD, C; bottom of oligemia after the final CSD, E. A P value of less than 0.05/4 was considered to be statistically significant.
4. Results and discussion
Figure 5 shows typical images of the exposed rat brain before CSD under normoxia for (a) the RGB color image, (b) CHbO, (c) CHbR, (d) CHbT, and (e) StO2. In Figs. 5(b), 5(c), and 5(d), the values of CHbO, CHbR, and CHbT, respectively, in the blood vessel region were higher than those in the parenchyma region, which indicates the difference in blood volume between the blood vessels and the parenchyma. The distribution of StO2 in the blood vessel region in Fig. 5(e) was probably due to the difference between arteries and veins. The values of CHbO detected outside the blood vessels in Fig. 5 reflect the oxygenated hemoglobin in capillaries of parenchymal region. The average values over the entire region were calculated to be CHbO = 347 ± 21 μM, CHbR = 33 ± 9 μM, CHbT = 379 ± 17 μM, and StO2 = 91 ± 2%, whereas those over the ROI on the parenchyma (white square in each image of Fig. 5(a)) were calculated to be CHbO = 256 ± 12 μM, CHbR = 56 ± 5 μM, CHbT = 309 ± 5 μM, and StO2 = 82 ± 2%.
Fig. 5 Typical images of exposed rat brain before CSD under normoxia for (a) the RGB color image, (b) oxygenated hemoglobin (CHbO), (c) deoxygenated hemoglobin (CHbR), (d) total hemoglobin (CHbT), and (e) tissue oxygen saturation (StO2). The scale bar in (a) represents 1 mm.
Figure 6 shows typical sequential images obtained before, during, and after CSD for (a) the RGB color image, (b) ΔCHbO, (c) ΔCHbR, (d) ΔCHbT, and (e) ΔStO2. The KCl solution was applied to the cortical surface through the burr hole at 5 min (60 sec) after the onset of measurement and was washed out at 20 min (1,200 sec) after the onset of measurement. The letter E in Fig. 5 indicates the electrode. The wave of change in hemodynamics spreading over the cerebral cortex can be observed in Figs. 6(b)-6(e). Each wave form moved from the left side to the right side of the image. The increase in ΔCHbR coincided with the decrease in ΔCHbO. The waves of ΔCHbR and ΔCHbO preceded the waves of increased ΔCHbT. The propagation speed of the wave of ΔCHbO was 5.02 mm/min, which corresponds to the typical propagation speed of CSD reported in the literature [1–3].
Fig. 6 Typical sequential images obtained before, during, and after CSD for (a) the RGB color image, (b) ΔCHbO, (c) ΔCHbR, (d) ΔCHbT, and (e) ΔStO2. Alphabetic characters E and ROI in (a) represent a recording electrode and region of interest, respectively. The scale bar in (a) represents 1 mm.
Figure 7 shows the typical time courses of (a) CHbT and (b) StO2 averaged over the area for the ROI in the parenchyma, as shown in Fig. 6, before, during, and after CSD. The signal of LFP was also compared with the time courses of CHbT and StO2 shown in Fig. 7. Note that the elevations in both CHbT and CHbO occurred repeatedly after topical application of KCl. The values of CHbT and StO2 immediately before CSD were 372 ± 14 μM and 54 ± 4%, respectively, which is comparable with the normal values reported in the literature [29, 39]. The average value of StO2 over all five samples was 65 ± 17% before CSD. This is lower than the typical arterial oxygen saturation, which ranges from 90% to 98%. The typical value of venous oxygen saturation is almost 65% [48]. The value of StO2 measured by this method represents oxygen saturation for the mixture of arterio-venous blood. Almost 75% of the total blood volume in the whole body is contained within the veins and venules [49], therefore 25% of it is contained within the arteries and arterioles. Assuming that the blood volume ratio of venules and arterioles in brain tissue is similar to that of the whole body, and that the oxygen saturation of arterial blood under the normal condition is 96%, the oxygen saturation of skin tissue is calculated to be 72.8%. This value is comparable with the average values of StO2 over all five samples obtained with this method before CSD.
Fig. 7 Typical time courses of (a) total hemoglobin (CHbT) and (b) tissue oxygen saturation (StO2) averaged over the area for the ROI on the parenchyma region (corresponding to the white square on the RGB image in Fig. 6(a)).
After application of KCl, the values of CHbO and CHbR were decreased and increased, respectively, which caused the decreases in CHbT and StO2. This initial phase of hemodynamic change coincided with the decrease in LFP signal. The values of CHbO and CHbT were dramatically increased immediately after the DC shift of LFP, whereas the value of CHbR decreased. After the profound increases in CHbO and CHbT, the value of StO2 dropped below baseline for up to 30 min, at least in this case. The time courses of CHbO, CHbR, CHbT, and StO2 after topical application of KCl were consistent with the well-known hemodynamic responses to CSD.
Figure 8 shows the values of (a) ΔCHbO, (b) ΔCHbR, (c) ΔCHbT, and (d) ΔStO2 averaged over the ROIs for all five samples at initial hypoperfusion, hyperemia, and post-CSD oligemia. The values of ΔCHbO, ΔCHbT, and ΔStO2 were decreased, whereas ΔCHbR was increased in the initial hypoperfusion phase. These hemodynamic changes observed at this phase are consistent with the vasoconstriction and tissue ischemia caused by CSD [12–14]. At transition to the hyperemia phase from hypoperfusion, the values of ΔCHbO and ΔStO2 were significantly increased, which implies vasodilatation in arterioles and increased cerebral blood volume or CBF in response to CSD [12–14]. In the wake of the hyperemic phase, the value of ΔCHbO and ΔCHbT were reduced to 27 ± 16% and 2 ± 4% of baseline, respectively, suggesting long-lasting vasoconstriction after CSD. In this persistent oligemia, ΔCHbR was increased (not statistically significant), and ΔStO2 significantly dropped to at most 26 ± 18% of the level before CSD, indicating long-lasting hypoxemia.
Fig. 8 Relative change in (a) oxygenated hemoglobin (ΔCHbO), (b) deoxygenated hemoglobin (ΔCHbR), (c) total hemoglobin (ΔCHbT), and (d) tissue oxygen saturation (ΔStO2) averaged over the ROIs for all five samples at initial hypoperfusion, hyperemia, and post-CSD oligemia. The error bars show the standard deviations (n = 9). *P < 0.05/9.
Figure 9 shows typical results obtained from the exposed cerebral cortex of a rat for (a) the time course of CHbT averaged over the ROI on arterioles over a 120-s period for before CSD and (b) the PSD. Oscillations in CHbT with approximately 1 cycle per second can be observed periodically in Fig. 9(a), corresponding to the peak of PSD arround 0.9 Hz shown in Fig. 9(b), which is consistent with the typical respiratory rate ranged from 50 to 100 breaths per minute for an anesthetized rat [50]. On the other hand, the PSD below 0.2 Hz shown in Fig. 9(b) reflects the low frequency component of CHbT in Fig. 9(a).
Fig. 9 Typical results obtained from the exposed cerebral cortex of a rat for (a) the time course of CHbT averaged over the ROI on arterioles over a 120-s period for before CSD and (b) the PSD.
Figure 10 shows typical results obtained from the exposed cerebral cortex of a rat before CSD for (a) a raw RGB image and the two-dimensional maps of PSD for low-frequency oscillation in CHbT over a 120-s period at (b) 0.05 Hz, (c) 0.1 Hz, and (d) 0.5 Hz. The arrows in the raw RGB image indicate the distribution of arterioles (bright red regions), venules (dark red regions), and parenchyma in the cerebral cortex. Arterioles and venules were identified by morphology, color, and the value of StO2.
Fig. 10 Typical results obtained from the exposed cerebral cortex of a rat before CSD for (a) the raw RGB image and two-dimensional maps of PSD for low-frequency oscillations in CHbT over a 120-s period at (b) 0.05 Hz, (c) 0.1 Hz, and (d) 0.5 Hz. The scale bar in (a) represents 1 mm.
Arterioles had a smaller diameter and fewer branching points at less acute angles. The images of the spectral power map for low-frequency oscillation at both 0.05 Hz and 0.1 Hz showed good contrast between arterioles and parenchyma or venules. On the other hand, the result at 0.5 Hz showed an extremely low power over the entire region of the image because the oscillation at 0.5 Hz is beyond the limits of the vasomotion-related frequency band. This implies the possibility that the method can distinguish the low-frequency spontaneous oscillations due to vasomotion from other physiological signals such as respiration.
Figure 11 shows typical results obtained from the exposed cerebral cortex of a rat for (a) a raw RGB image before CSD and the spectral power map of low-frequency oscillation in CHbT over a 120-s period for (b) before CSD, (c) during CSD (hyperemia), and (d) post-CSD oligemia. The image of PSD after CSD is darker than that before CSD. A decrease in PSD after CSD was localized in arteriole regions, suggesting suppression of vasomotion by CSD.
Fig. 11 Typical results obtained from the exposed cerebral cortex of a rat for (a) the raw RGB image before CSD and the spectral power map of low-frequency oscillations in CHbT over a 120-s period at 0.1 Hz for (b) before CSD, (c) during CSD (hyperemia), and (d) post-CSD oligemia. The corresponding images of StO2 are shown in (e) before CSD, (f) during CSD (hyperemia), and (g) post-CSD oligemia. The scale bar in (a) represents 1 mm.
Figure 11 also shows the images of StO2 for (e) before CSD, (f) during CSD (hyperemia), and (g) post-CSD oligemia. The values of StO2 in arterioles are higher than those in venules due to the fact that hemoglobin in arterial blood being much more oxygenated than in venous blood, as shown in Fig. 11(e). Therefore, the distribution of arterioles and venules can be clearly distinguished in the estimated image of StO2. After CSD, the tissue oxygen saturation was decreased over the entire region and was remarkable on arteriole regions (white arrows in Fig. 11(a)). Of note, the reduction in tissue oxygen saturation shown in Fig. 11(g) (lower right region of the image) corresponds to attenuation of low-frequency oscillations after CSD shown in Fig. 11(d).
Figure 12 shows the relative change in PSD averaged over the ROIs for four samples on (a) arterioles, (b) parenchyma, and (c) venules, for during CSD (hyperemia) and post-CSD oligemia, compared with before CSD. Although the relative changes in PSD for arterioles, parenchyma, and venules were slightly increased during CSD, we observed no significant differences between before and during CSD. After CSD, the relative change in PSD for only arterioles was significantly reduced to 28 ± 20% of baseline with respect to the pre-CSD level.
Fig. 12 Relative change in PSD averaged over the ROIs on (a) arterioles, (b) parenchyma, and (c) venules during CSD (hyperemia) and post-CSD oligemia, compared with before CSD. The error bars show the standard deviations (n = 4). *P < 0.05/3.
A low-frequency oscillation (0.1-Hz) in the diffuse reflectance signal has been associated with spontaneous fluctuations in CBF [3–5]. Reduction in spontaneous vasomotor activity is attributed to changes in vascular reactivity following CSD [5, 51]. A temporary impairment in cerebral autoregulation occurs immediately after CSD and persists for 35 min [52]. During post-CSD oligemia, CBF drops to approximately 70% of the control level, with a reduction in CVR [53]. Therefore, the reductions in tissue oxygen saturation and spontaneous low-frequency oscillations in the total hemoglobin concentration after CSD imply a transient impairment in cerebral autoregulation. The results shown in this study also demonstrated the possibility that this method can distinguish low-frequency spontaneous oscillations due to vasomotion and can be used to investigate the changes in CVR during CSD. Spatial heterogeneity and temporal changes in spontaneous low-frequency oscillations may be variable in different neurological disorders, such as neurotrauma, seizure, stroke, and ischemia.
Although oxygenated hemoglobin and deoxygenated hemoglobin are the main chromophores in the brain, light absorption by cytochrome c, cytochrome aa3, flavin adenine dinucleotide, and nicotinamide adenine dinucleotide also contributes to the diffuse reflectance in the visible wavelength region. Because our method uses the three broad band spectral responses of red, green, and blue channels, simultaneous evaluation of the other minor chromophores will be difficult. This method, which is based on diffusing reflection, integrates all information along the depth direction, and therefore, it does not have depth resolution. The diameter and depth of cerebral vasculature are usually variable among samples and could change due to the age of the rat. Their correct estimation is important for accurately evaluating the concentrations of oxygenated hemoglobin and deoxygenated hemoglobin in vascular regions.
It may be possible to estimate directly CHbO and CHbR from the RGB values without the CIEXYZ color space. In that case, it is necessary to know the spectral sensitivities of red, green, and blue channels of camera used. Generally, the spectral sensitivities of color camera are device-dependent values and some of them have not been published by manufacturers. On the other hand, the color matching functions of CIEXYZ color system has been published and available for calculating the XYZ values from the diffuse reflectance spectrum. Moreover, the XYZ values can be estimated from the measured RGB values based on the measurements of Color Checker. For this reason, we used the conversion from the measured RGB values to XYZ values.
The present study focused on evaluating changes in hemoglobin concentration and hemoglobin oxygen saturation qualitatively as a first step. We also believe that experiments with a tissue-mimicking phantom are needed to quantitatively validate the results. We have developed an agarose-based phantom that mimics the optical properties of biological tissues [37]. In the preliminary experiments, we attempted to change hemoglobin oxygen saturation in phantoms using Na2S2O4 solution. However, the measured diffuse reflectance spectra and XYZ values tended to exhibit unexpected fluctuations and maintaining a stable condition of the deoxygenated spectra at the specific hemoglobin oxygen saturation was difficult. We are now working on improving the phantom. Validation of the method with a phantom closer to realistic conditions of brain tissue is warranted in future work.
We demonstrated a simple imaging method with an RGB camera that can visualize physiological hemodynamic changes due to KCl-induced CSD. The method is advantageous because it can also be employed for color images captured by an RGB camera mounted with a stereoscopic microscope system as well as an intraoperative surgical microscope system. Therefore, this method may be useful for evaluating brain function and tissue viability during neurosurgery for cerebrovascular diseases. Application of our algorithm to super high-speed color cameras is promising for evaluating the fast intrinsic optical signals due to neuronal activities [40, 54]. These issues should be investigated in the future.
5. Conclusions
In summary, a method for imaging the oxygenated hemoglobin concentration (CHbO), deoxygenated hemoglobin concentration (CHbR), total hemoglobin concentration (CHbT), and tissue oxygen saturation (StO2) of in vivo exposed brain tissues based on DRS using a digital RGB camera was demonstrated in the present report. In vivo sequential recordings of all physiological parameters and the local field electrical potential signals of the exposed rat brain confirmed the feasibility of the method for evaluating cerebral hemodynamic responses to CSD, such as initial hypoperfusion, profound hyperemia, and post-CSD oligemia and hypoxemia. The current study also demonstrated that the method can be used to visualize the spatial map of spontaneous low-frequency oscillations in the cerebral hemoglobin concentration. The results showed the possibility that the method can be used for evaluating cerebral autoregulation functions and their effects on oxygen supply to cerebral tissue as well as important implications for future studies of normal and pathological regulation of CBF. The method may be useful for evaluating brain function and cerebral autoregulation during neurosurgery as well as for the diagnosis of several neurological disorders, such as neurotrauma, seizure, stroke, and ischemia. The resulting information promises to assist interpretation of these issues, which we intend to explore in future work.
Funding
Grant-in-Aid for Scientific Research (C) from the Japanese Society for the Promotion of Science (25350520, 22500401, 15K06105) and the US-ARMY ITC-PAC Research and Development Project (FA5209-15-P-0175, FA5209-16-P-0132).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References and links
1. A. A. P. Leão, “Spreading depression of activity in the cerebral cortex,” J. Neurophysiol. 7(6), 359–390 (1944). [CrossRef] [PubMed]
2. A. Gorji, “Spreading depression: a review of the clinical relevance,” Brain Res. Brain Res. Rev. 38(1-2), 33–60 (2001). [CrossRef] [PubMed]
3. G. G. Somjen, “Mechanisms of spreading depression and hypoxic spreading depression-like depolarization,” Physiol. Rev. 81(3), 1065–1096 (2001). [CrossRef] [PubMed]
4. M. Lauritzen, “Pathophysiology of the migraine aura. The spreading depression theory,” Brain 117(1), 199–210 (1994). [CrossRef] [PubMed]
5. M. Fabricius, S. Fuhr, L. Willumsen, J. P. Dreier, R. Bhatia, M. G. Boutelle, J. A. Hartings, R. Bullock, A. J. Strong, and M. Lauritzen, “Association of seizures with cortical spreading depression and peri-infarct depolarisations in the acutely injured human brain,” Clin. Neurophysiol. 119(9), 1973–1984 (2008). [CrossRef] [PubMed]
6. K. A. Hossmann, “Periinfarct depolarizations,” Cerebrovasc. Brain Metab. Rev. 8(3), 195–208 (1996). [PubMed]
7. K. Takano, L. L. Latour, J. E. Formato, R. A. Carano, K. G. Helmer, Y. Hasegawa, C. H. Sotak, and M. Fisher, “The role of spreading depression in focal ischemia evaluated by diffusion mapping,” Ann. Neurol. 39(3), 308–318 (1996). [CrossRef] [PubMed]
8. M. Lauritzen, J. P. Dreier, M. Fabricius, J. A. Hartings, R. Graf, and A. J. Strong, “Clinical relevance of cortical spreading depression in neurological disorders: migraine, malignant stroke, subarachnoid and intracranial hemorrhage, and traumatic brain injury,” J. Cereb. Blood Flow Metab. 31(1), 17–35 (2011). [CrossRef] [PubMed]
9. H. Oka, M. Kako, M. Matsushima, and K. Ando, “Traumatic spreading depression syndrome. review of a particular type of head injury in 37 patients,” Brain 100(2), 287–298 (1977). [CrossRef] [PubMed]
10. D. Torrente, R. Cabezas, M. F. Avila, L. M. García-Segura, G. E. Barreto, and R. C. Guedes, “Cortical spreading depression in traumatic brain injuries: Is there a role for astrocytes?” Neurosci. Lett. 565, 2–6 (2014). [CrossRef] [PubMed]
11. S. Sato, S. Kawauchi, W. Okuda, I. Nishidate, H. Nawashiro, and G. Tsumatori, “Real-time optical diagnosis of the rat brain exposed to a Laser-induced shock wave: Observation of spreading depolarization, vasoconstriction and hypoxemia-oligemia,” PLoS One 9(1), e82891 (2014). [CrossRef] [PubMed]
12. T. Bonhoeffer and A. Grinvald, [Brain mapping: The Methods], Academic Press, San Diego, 1996.
13. J. P. Dreier, “The role of spreading depression, spreading depolarization and spreading ischemia in neurological disease,” Nat. Med. 17(4), 439–447 (2011). [CrossRef] [PubMed]
14. C. Ayata, “Spreading depression and neurovascular coupling,” Stroke 44(6), S87–S89 (2013). [CrossRef] [PubMed]
15. H. Piilgaard and M. Lauritzen, “Persistent increase in oxygen consumption and impaired neurovascular coupling after spreading depression in rat neocortex,” J. Cereb. Blood Flow Metab. 29(9), 1517–1527 (2009). [CrossRef] [PubMed]
16. B. M. Bosch, A. Bringard, G. Ferretti, S. Schwartz, and K. Iglói, “Effect of cerebral vasomotion during physical exercise on associative memory, a near-infrared spectroscopy study,” Neurophotonics 4(4), 041404 (2017). [CrossRef] [PubMed]
17. Z. Zhang and R. Khatami, “Predominant endothelial vasomotor activity during human sleep: a near-infrared spectroscopy study,” Eur. J. Neurosci. 40(9), 3396–3404 (2014). [CrossRef] [PubMed]
18. A. Stefanovska, “Coupled oscillators. Complex but not complicated cardiovascular and brain interactions,” IEEE Eng. Med. Biol. Mag. 26(6), 25–29 (2007). [CrossRef] [PubMed]
19. R. Zhang, J. H. Zuckerman, K. Iwasaki, T. E. Wilson, C. G. Crandall, and B. D. Levine, “Autonomic neural control of dynamic cerebral autoregulation in humans,” Circulation 106(14), 1814–1820 (2002). [CrossRef] [PubMed]
20. H. D. Kvernmo, A. Stefanovska, K. A. Kirkebøen, and K. Kvernebo, “Oscillations in the human cutaneous blood perfusion signal modified by endothelium-dependent and endothelium-independent vasodilators,” Microvasc. Res. 57(3), 298–309 (1999). [CrossRef] [PubMed]
21. A. G. Hudetz, R. J. Roman, and D. R. Harder, “Spontaneous flow oscillations in the cerebral cortex during acute changes in mean arterial pressure,” J. Cereb. Blood Flow Metab. 12(3), 491–499 (1992). [CrossRef] [PubMed]
22. J. E. Mayhew, S. Askew, Y. Zheng, J. Porrill, G. W. Westby, P. Redgrave, D. M. Rector, and R. M. Harper, “Cerebral vasomotion: a 0.1-Hz oscillation in reflected light imaging of neural activity,” Neuroimage 4(3), 183–193 (1996). [CrossRef] [PubMed]
23. M. Guiou, S. Sheth, M. Nemoto, M. Walker, N. Pouratian, A. Ba, and A. W. Toga, “Cortical spreading depression produces long-term disruption of activity-related changes in cerebral blood volume and neurovascular coupling,” J. Biomed. Opt. 10(1), 011004 (2005). [CrossRef] [PubMed]
24. C. Ayata, H. K. Shin, S. Salomone, Y. Ozdemir-Gursoy, D. A. Boas, A. K. Dunn, and M. A. Moskowitz, “Pronounced hypoperfusion during spreading depression in mouse cortex,” J. Cereb. Blood Flow Metab. 24(10), 1172–1182 (2004). [CrossRef] [PubMed]
25. I. Sukhotinsky, E. Dilekoz, M. A. Moskowitz, and C. Ayata, “Hypoxia and hypotension transform the blood flow response to cortical spreading depression from hyperemia into hypoperfusion in the rat,” J. Cereb. Blood Flow Metab. 28(7), 1369–1376 (2008). [CrossRef] [PubMed]
26. C. Zhou, G. Yu, D. Furuya, J. Greenberg, A. Yodh, and T. Durduran, “Diffuse optical correlation tomography of cerebral blood flow during cortical spreading depression in rat brain,” Opt. Express 14(3), 1125–1144 (2006). [CrossRef] [PubMed]
27. N. Rajaram, T. H. Nguyen, and J. W. Tunnell, “Lookup table-based inverse model for determining optical properties of turbid media,” J. Biomed. Opt. 13(5), 050501 (2008). [CrossRef] [PubMed]
28. B. Yu, J. Y. Lo, T. F. Kuech, G. M. Palmer, J. E. Bender, and N. Ramanujam, “Cost-effective diffuse reflectance spectroscopy device for quantifying tissue absorption and scattering in vivo,” J. Biomed. Opt. 13(6), 060505 (2008). [CrossRef] [PubMed]
29. B. S. Nichols, N. Rajaram, and J. W. Tunnell, “Performance of a lookup table-based approach for measuring tissue optical properties with diffuse optical spectroscopy,” J. Biomed. Opt. 17(5), 057001 (2012). [CrossRef] [PubMed]
30. R. Hennessy, S. L. Lim, M. K. Markey, and J. W. Tunnell, “Monte Carlo lookup table-based inverse model for extracting optical properties from tissue-simulating phantoms using diffuse reflectance spectroscopy,” J. Biomed. Opt. 18(3), 037003 (2013). [CrossRef] [PubMed]
31. I. Nishidate, C. Mizushima, K. Yoshida, S. Kawauchi, S. Sato, and M. Sato, “In vivo estimation of light scattering and absorption properties of rat brain using a single-reflectance fiber probe during cortical spreading depression,” J. Biomed. Opt. 20(2), 027003 (2015). [CrossRef] [PubMed]
32. P. B. Jones, H. K. Shin, D. A. Boas, B. T. Hyman, M. A. Moskowitz, C. Ayata, and A. K. Dunn, “Simultaneous multispectral reflectance imaging and laser speckle flowmetry of cerebral blood flow and oxygen metabolism in focal cerebral ischemia,” J. Biomed. Opt. 13(4), 044007 (2008). [CrossRef] [PubMed]
33. A. M. Ba, M. Guiou, N. Pouratian, A. Muthialu, D. E. Rex, A. F. Cannestra, J. W. Y. Chen, and A. W. Toga, “Multiwavelength optical intrinsic signal imaging of cortical spreading depression,” J. Neurophysiol. 88(5), 2726–2735 (2002). [CrossRef] [PubMed]
34. S. Chen, Z. Feng, P. Li, S. L. Jacques, S. Zeng, and Q. Luo, “In vivo optical reflectance imaging of spreading depression waves in rat brain with and without focal cerebral ischemia,” J. Biomed. Opt. 11(3), 034002 (2006). [CrossRef] [PubMed]
35. C. Yin, F. Zhou, Y. Wang, W. Luo, Q. Luo, and P. Li, “Simultaneous detection of hemodynamics, mitochondrial metabolism and light scattering changes during cortical spreading depression in rats based on multi-spectral optical imaging,” Neuroimage 76, 70–80 (2013). [CrossRef] [PubMed]
36. S. Kawauchi, I. Nishidate, Y. Uozumi, H. Nawashiro, H. Ashida, and S. Sato, “Diffuse light reflectance signals as potential indicators of loss of viability in brain tissue due to hypoxia: charge-coupled-device-based imaging and fiber-based measurement,” J. Biomed. Opt. 18(1), 015003 (2013). [CrossRef] [PubMed]
37. I. Nishidate, T. Ishizuka, A. Mustari, K. Yoshida, S. Kawauchi, S. Sato, and M. Sato, “Evaluation of cerebral hemodynamics and tissue morphology of in vivo rat brain using spectral diffuse reflectance imaging,” Appl. Spectrosc. 71(5), 866–878 (2017). [CrossRef] [PubMed]
38. T. Arnold, M. De Biasio, and R. Leitner, “Hyper-spectral video endoscope for intra-surgery tissue classification using auto-fluorescence and reflectance spectroscopy,” Proc. SPIE 8087, 808711 (2011). [CrossRef]
39. A. Basiri, M. Nabili, S. Mathews, A. Libin, S. Groah, H. J. Noordmans, and J. C. Ramella-Roman, “Use of a multi-spectral camera in the characterization of skin wounds,” Opt. Express 18(4), 3244–3257 (2010). [CrossRef] [PubMed]
40. M. B. Bouchard, B. R. Chen, S. A. Burgess, and E. M. C. Hillman, “Ultra-fast multispectral optical imaging of cortical oxygenation, blood flow, and intracellular calcium dynamics,” Opt. Express 17(18), 15670–15678 (2009). [CrossRef] [PubMed]
41. K. Yoshida, I. Nishidate, T. Ishizuka, S. Kawauchi, S. Sato, and M. Sato, “Multispectral imaging of absorption and scattering properties of in vivo exposed rat brain using a digital red-green-blue camera,” J. Biomed. Opt. 20(5), 051026 (2015). [CrossRef] [PubMed]
42. A. A. Stratonnikov and V. B. Loschenov, “Evaluation of blood oxygen saturation in vivo from diffuse reflectance spectra,” J. Biomed. Opt. 6(4), 457–467 (2001). [CrossRef] [PubMed]
43. L. Wang, S. L. Jacques, and L. Zheng, “MCML-Monte Carlo modeling of light transport in multi-layered tissues,” Comput. Methods Programs Biomed. 47(2), 131–146 (1995). [CrossRef] [PubMed]
44. J. R. Mourant, J. P. Freyer, A. H. Hielscher, A. A. Eick, D. Shen, and T. M. Johnson, “Mechanisms of light scattering from biological cells relevant to noninvasive optical-tissue diagnostics,” Appl. Opt. 37(16), 3586–3593 (1998). [CrossRef] [PubMed]
45. D. Abookasis, C. C. Lay, M. S. Mathews, M. E. Linskey, R. D. Frostig, and B. J. Tromberg, “Imaging cortical absorption, scattering, and hemodynamic response during ischemic stroke using spatially modulated near-infrared illumination,” J. Biomed. Opt. 14(2), 024033 (2009). [CrossRef] [PubMed]
46. V. Tuchin, Tissue Optics: Light Scattering Methods and Instruments for Medical Diagnosis, 2nd ed. (SPIE Press, Bellingham, WA, 2007).
47. S. A. Prahl, “Tabulated Molar Extinction Coefficient for Hemoglobin in Water,” http://omlc.ogi.edu/spectra/hemoglobin/summary.html (1999).
48. P. van Beest, G. Wietasch, T. Scheeren, P. Spronk, and M. Kuiper, “Clinical review: use of venous oxygen saturations as a goal - a yet unfinished puzzle,” Crit. Care 15(5), 232 (2011). [CrossRef] [PubMed]
49. C. C. Pang, “Measurement of body venous tone,” J. Pharmacol. Toxicol. Methods 44(2), 341–360 (2000). [CrossRef] [PubMed]
50. J. Thomas and P. Lerche, [Anaesthesia and Analgesia for Veterinary Technicians, 4th ed.], Elsevier, (2011).
51. R. D. Piper, G. A. Lambert, and J. W. Duckworth, “Cortical blood flow changes during spreading depression in cats,” Am. J. Physiol. 261(1 Pt 2), H96–H102 (1991). [PubMed]
52. G. Florence, G. Bonvento, R. Charbonne, and J. Seylaz, “Spreading depression reversibly impairs autoregulation of cortical blood flow,” Am. J. Physiol. 266(4 Pt 2), R1136–R1140 (1994). [PubMed]
53. M. Lauritzen, “Long-lasting reduction of cortical blood flow of the brain after spreading depression with preserved autoregulation and impaired CO2 response,” J. Cereb. Blood Flow Metab. 4(4), 546–554 (1984). [CrossRef] [PubMed]
54. V. Tsytsarev, K. Premachandra, D. Takeshita, and S. Bahar, “Imaging cortical electrical stimulation in vivo: fast intrinsic optical signal versus voltage-sensitive dyes,” Opt. Lett. 33(9), 1032–1034 (2008). [CrossRef] [PubMed]
