1.

Introduction

Optical coherence tomography (OCT) is a well-known imaging diagnostic modality widely used for noninvasive imaging of soft biological tissues both in vivo and in vitro with high spatial resolution (3 to $5\mu \mathrm{m}$) and up to a few millimeters probing depth.^1–4 After being significantly improved, OCT has generated major interest as a tool for clinical diagnostics.^5,6 By using spatial and/or temporal analysis of the dynamic speckle patterns generated by moving red blood cells, various OCT modifications for visualization of subcutaneous blood vessels’ distribution in human skin in vivo have been suggested, including speckle variance OCT (svOCT),⁷ optical microangiography (OMAG),⁸ correlation map OCT (cmOCT),⁹ and double correlation OCT (dcOCT).¹⁰ It has also been demonstrated that the dcOCT approach can be used for visualization of molecular diffusion within the skin tissues in vivo^11,12 and for direct imaging of the electro-kinetic response of biological tissues ex vivo.¹³

In the current paper, we present the results of further OCT studies of the electro-kinetic response of biological tissues influenced by a low-frequency electric field at normal conditions and upon the topical application of optical clearing agents (OCAs). The tissues’ optical clearing is widely described elsewhere^14–18 and is based on refractive index matching between tissues’ structural compounds and OCA diffused into the tissues and tissue dehydration due to the osmotic properties of OCAs. Topical application of OCAs enhances light penetration depth and significantly increases the OCT image contrast.^19–23 Therefore, following the results of recent studies of the interaction of low-frequency electric fields with biological tissues by OCT,^13,24 we apply a 50% glycerol solution in water as an OCA to enhance the image contrast and possibly observe the spatial distribution of the electric field within the tissues. Bearing in mind that hyperosmotic OCAs may induce dehydration and corresponding alteration of tissues’ morphological and optical properties,¹⁷ in addition to OCT imaging, measurements of thermal profiles and impedance at normal conditions and upon optical clearing have been done.

2.

Method and Materials

For the experimental setup (Fig. 1), a standard swept-source OCT (OCM1300SS, Thorlabs Inc., Newton) operated in a polarization-sensitive mode without phase retardation has been used to acquire both two-dimensional (2-D) and three-dimensional (3-D) images of ex vivo biological tissues. The OCT experimental system consists of the swept-source engine, imaging module, and imaging probe. The swept source has a central wavelength of 1325 nm with a bandwidth of $\sim 100\mathrm{nm}$, a scanning rate of 16 kHz, and an output power of the probing light of 12 mW. The system is capable of acquiring a 3-D volume of $1024\times 1024\times 512\text{pixels}$ (i.e., up to $10\mathrm{mm}\times 10\mathrm{mm}\times 3\mathrm{mm}$) containing 1024 images within $\sim 40\mathrm{s}$, with respective axial and lateral resolutions of $\sim 13$ and $\sim 25\mu \mathrm{m}$. A function generator (GFG 2100, ISO-TECH, Corby, England) has been used to apply an alternating current (ac) of fixed frequency and voltage to tissue by means of two stainless steel electrodes. The ends of the electrodes were separated from each other by $\sim 1.5\mathrm{cm}$, as presented in Fig. 1.

Fig. 1

Schematic presentation of the experiment. The standard optical coherence tomography (OCT) probe (OCM1300SS, Thorlabs Inc., Newton) is placed above the surface of tissue sample to acquire OCT images from the area between two electrodes embedded into the tissue at $\sim 0.5\mathrm{mm}$ depth and separated by $\sim 5$ to 6 mm from each other.

To quantify the effect of the influence of a low frequency-electric field in the tissue samples, we use the dcOCT approach developed earlier.¹⁰ Within the framework of the dcOCT approach, the similarity between images is accessed by sequential applications of the Wiener filter and by performing cross-correlation. The idea behind dcOCT is that a difference between the regions of low correlation corresponding to motion or tissues’ alterations induced by the external electric field and the regions of high correlation correspond to static structures. To quantitatively assess the relative strength of the applied electric field in tissue, the mean of the cross-correlation values is evaluated.¹³

Figure 2 schematically shows this procedure with application to the obtained OCT images of the ex vivo tissue samples.

Fig. 2

Schematic presentation of double correlation OCT (dcOCT) approach. Step 1 corresponds to $I(x,z)$ image acquisition by conventional OCT. Step 2 shows the Wiener filtering of the obtained images. Step 3 represents the cross-correlation procedure between the Wiener filtering images $I_{\mathrm{w}}(x,z)$ obtained at $t$ and $t+1$ time intervals and the generation of correlation map defined by Eq. (2).

In the first step of the double correlation approach, by utilizing the adaptive Wiener filtering procedure, the background noise is suppressed and the OCT image artifacts (such as blur, physical offsets of tissue boundaries) are removed. Creation of the Wiener filter $I_{\mathrm{w}}(x,z)$ requires calculation of the filter’s coefficients by obtaining autocorrelation functions of the OCT signal $P_S[I(x,z)]$ and noise $P_N[I(x,z)]$. In the frequency domain, the Wiener filtering image is defined as^25,26

Here, $P_N[I(x,z)]$ is estimated by calculation of the local variance using 2-D correlation.²⁶ Following substraction of an estimate of the noise from the original OCT images, the cross-correlation between two grids is calculated as^10,13

Here, $\mathcal{F}[I_{\mathrm{w}}(t)]$ is the Fourier transform of $I_{\mathrm{w}}(t)$ and $\overline{\mathcal{F}[I_{\mathrm{w}}(t+1)]}$ is the complex conjugate of the Fourier transform $\mathcal{F}[I_{\mathrm{w}}(t+1)]$ of $I_{\mathrm{w}}(t+1)$; $u$ and $v$ are spatial frequencies in the $x$ and $z$ directions, respectively. $M$ and $N$ are the maximum number of pixels in the $x$ and $z$ directions, and $t$ is the time interval for image acquisition.

It should be pointed out that in cross-correlation analysis, the size of the grid should be carefully selected since it is a trade-off between the processing time and the final quality of the outcome.⁹ If a grid is too big (e.g., $40\times 40\text{pixels}$), blurring and a loss of structural signal will occur.⁹ For a small grid (e.g., $5\times 5\text{pixels}$), the background noise will have a significant impact on the structural signal, resulting in decorrelation. In this study, a grid size of $7\times 7$ was used to quantify the effect of influence of a low-frequency electric field in a fresh tissue sample.

Finally, the relative magnitude of influence of the low-frequency external electric field on the biological tissue is assessed as¹³

To speed up the computations, image analysis was performed on NVIDIA graphics processing units (GPUs) utilizing a compute unified device architecture parallel computing platform. The entire 3-D OCT volume of $1024\times 1024\times 512\text{pixels}$ containing 1024 images was processed on dual Tesla M2090 GPUs in $\sim 30\mathrm{s}$.¹⁰ The particular details of implementation are beyond the scope of this paper and are described elsewhere.¹⁰

Chicken breast pectoralis (6 to 7 weeks old), obtained from a primary supplier, were used in the experiments. A thermal camera (FLIR i3, FLIR Systems Inc., Wilsonville) was used to determine the rise in temperature of the tissue due to the exposure of an electric field. The camera provides a thermal image quality of $60\times 60\text{pixels}$ with a field of view of $12.5\mathrm{deg}(H)\times 12.5\mathrm{deg}(V)$, and a thermal sensitivity of 0.15°C. During the experiment, the thermal camera and OCT probe were placed to acquire thermal profiles and OCT images from the area between the electrodes. The measurements of resistance between the electrodes across the tissue sample were done by using a standard multimeter (M2005, AVO International, Dover, England).

3.

Results and Discussion

The results of the dcOCT imaging approach with and without topical application of OCA are presented in Figs. 3 and 4, respectively. In both figures, 2-D correlation images $C(x,z)$ are obtained at the same time intervals, i.e., 0, 1, 4, 60, 120, 200, 450, and 600 s.

Fig. 3

Two-dimensional (2-D) dcOCT images $C(x,z)$ of fresh chicken breast (pectoralis) ex vivo obtained during exposure with 10 V and 1   Hz alternating current at 0, 1, 4, 60, 120, 200, 450, and 600 s after topical application of 50% water–glycerol solution: images from (a) to (h), respectively. Scale bar corresponds to $250\mu \mathrm{m}$.

Fig. 4

2-D dcOCT images $C(x,z)$ of fresh chicken breast (pectoralis) ex vivo obtained during exposure with 10 V and 1 Hz alternating current at normal conditions at same time intervals as presented in Fig. 3. Scale bar corresponds to $250\mu \mathrm{m}$.

Figure 5 shows the relative magnitude of the influence of the electric field on tissue as a function of time, obtained by Eq. (4) for the images presented in Figs. 3 and 4.

Fig. 5

Relative magnitude of influence of the electric field on tissue as a function of time: triangles represent the results for tissue sample without optical clearing and circles represent the results for tissue sample topically exposed by optical clearing agent (OCA). The solid and dashed lines show the results of best fitting for tissue sample without and with topical application of OCA, respectively.

As one can see, the fluctuations induced by the electric field within the biological tissues are exponentially increased in time (see Fig. 5). The relative magnitude of influence of the electric field on biological tissue becomes lower when the tissue sample is exposed with the optical clearing that provides an observation of higher correlations between 2-D OCT images. It can be conceived that at low frequency, when an electrical field is applied, the charged ions move actively in the tissue sample exposed with the optical clearing agent rather than in a tissue sample under normal conditions.²⁷

Figures 6 and 7 show the evolution of thermal profiles measured at the surface of the chicken breast sample during the exposure of 10 V–1  Hz ac electric current without and with topical application of OCA. In both cases, the thermal IR camera was focused at the surface of the sample between the electrodes. Figures 6(a)–6(d) and 7(a)–7(d) present thermal profiles taken at 0, 5, 7.5, and 10 min, respectively.

Fig. 6

Thermal profiles of fresh chicken breast (pectoralis) ex vivo exposed with 10 V and 1 Hz alternating current at normal conditions, i.e., without an optical clearing. (a), (b), (c), and (d) represent thermal profiles taken at 0, 5, 7.5, and 10 min, respectively. Triangles 1 and 2 show the electrode positions under the tissue surface.

Fig. 7

Thermal profiles of tissue exposed with 10 V and 1 Hz alternating current taken after topical application of 50% glycerol solution in water at same time intervals as in Fig. 6.

As one can see, at normal conditions the temperature at the tissue surface reaches a maximum value of $\sim 26°\mathrm{C}$ in the area between the two electrodes (see Fig. 6), while the rise of temperature upon optical clearing is 1 deg less (see Fig. 7). It should also be pointed out that the kinetics of temperature changes is different for the samples with and without optical clearing exposure. Apparently, optical clearing slightly cools the biological tissues, thus reducing the temperature induced by the electric current by a few degrees.

The changes of temperature within the tissue sample induced by the electric current are likely attributed to variations in tissue resistance. Therefore, Fig. 8 shows the results of resistance measurements taken in parallel with temperature monitoring.

Fig. 8

Results of resistance measurements across the tissue sample as a function of time corresponding to the thermal profile measurements (see Figs. 6 and 7). Triangles and circles are, respectively, tissue with and without exposure of an OCA. The solid and dashed lines show the results of best fittings.

The resistance of the tissue sample measured at normal conditions is mostly constant and is reduced in $\sim 7.5\min$ due to the temperature increase (see Fig. 8). When OCA is topically applied, the resistance of the high water-content tissue²⁸ slightly increases (see Fig. 8). This is not necessarily in contradiction with the findings of this study. Topical application of OCA onto the surface of the tissue samples temporarily pushes water from the subsurface area out of the sampling volume, producing dehydration of topical tissue layers.^14–18 Therefore, the resistance measured at the subsurface area, i.e., in the area of main localization of current flows, increases. Eventually, in $\sim 7.5\min$, due to rehydration of upper tissue layers from in-depth layers, a decrease in resistance is observed (see Fig. 8). Many applications of bioelectrical impedance spectroscopy are based on the assumption of considerable differences between resistivity for high water-content tissues, but due to large confidence intervals, these differences are difficult to quantatively observe.^27,28 However, with the current dcOCT approach, the current flow as well as the temperature and resistance variations can be assessed with and without application of OCA.

4.

Conclusions

The electro-kinetic response of tissues has been observed and quantitatively evaluated by the dcOCT approach, utilizing consistent application of an adaptive Wiener filtering and a Fourier domain correlation algorithm. The results show that fluctuations induced by the electric field within the biological tissues are exponentially increased in time. We demonstrate that in comparison to the measurements of the resistance and temperature profiles at the surface of the tissue samples, the dcOCT approach is an order higher in sensitivity to the changes associated with the tissues’ electro-kinetic response. We also found that topical application of optical clearing reduces the tissues’ electro-kinetic response and cools the tissue, thus reducing by a few degrees the temperature induced by the applied electric current. We anticipate that the dcOCT approach can find a new application similar to bioelectrical impedance spectroscopy for monitoring the electro-optical properties of biological tissues, such as human skin, and their variations resulting from a transdermal drug or health-care products diffusion.