2D source-detector arrays enhance spatial information in diffuse imaging

Diffuse optical imaging is a noninvasive technique for the study of highly-scattering media. One important example is near-infrared imaging of biological tissue, such as functional mapping of the brain,¹ optical mammography,² and optical oximetry in various tissues. The high sensitivity of diffuse optical imaging to changes in blood distribution and oxygenation within tissue means that the technique has significant potential for diagnostic and functional studies.³ Unfortunately, the strong diffusion of light in tissue reduces spatial information and complicates the quantitative measurement of tissue properties.

Previously-proposed methods of enhancing spatial resolution have used time-domain⁴ and frequency-domain^5,6 approaches. The issue of depth discrimination has been tackled by introducing off-axis detection,⁷ by applying two-layer or multilayer models,⁸ or by fully-fledged solutions of the inverse imaging problem.⁹ Here, we propose a multi-element phased-array method that extends the concept of two-element phased-arrays previously proposed,⁵ and that does not rely on any a-priori information about boundary conditions or spatial features of tissue inhomogeneities. This new approach achieves an enhanced spatial resolution¹⁰ and depth discrimination¹¹ with respect to a single-source/single-detector imaging scheme.

Let's consider an array of three collinear continuous-wave light sources and a single optical detector, as shown in Figure 1. We indicate the detected intensity associated with each source as I_i. Our phased-array approach consists of normalizing I_i by the background intensity (I_0i) and introducing individual amplitude (A_i) and phase (α_i) factors to yield a phased-array intensity (I_P-A) according to Equation 1.

(1)

The distance between source and detector (6cm) represents a typical source-detector distance used in optical studies of thick tissues.

The improvement in spatial resolution is illustrated in Figure 2, which shows linear scans of single-source and phased-array intensities within a turbid medium (milk with black India ink). Two black cylinders were equidistant (3.0cm) from the source array and the collection fiber.Figure 2 shows that the two cylinders are not resolved by the single-source intensity trace, but are resolved by the phased-array trace with two separate peaks associated with each cylinder.

To test the depth-discrimination potential, we used the experimental setup illustrated in Figure 3, which uses a superposition of three linear source arrays along different directions. We calculated the phased-array intensity associated with each linear array (according to Equation 1) and took the maximum to optimize sensitivity to directional structures. Three black cylinders were placed between the source array and the detector fiber, with their axes parallel to the plane of the source array and at different depths.Figure 4shows that the single-source intensity is equally sensitive to the top and bottom cylinders, while the intensity with the phased array on top (bottom) is more sensitive to the top (bottom) cylinder.

Enhancement in spatial resolution (Figure 2) and depth-discrimination capability (Figure 4) finds important applications in diffuse optical imaging. In particular, the approach presented here directly lends itself to implementations in optical mammography, where it is possible to employ a transmission geometry similar to the one used in these studies.

Detecting tumors, identifying localized tissue areas associated with specific functional activities, or assessing local changes in tissue metabolism are applications within reach of near-infrared imaging of tissue. While the intrinsic optical contrast associated with the blood spatial distribution and with its oxygenation is high, the spatial-information content of diffuse optical imaging is intrinsically limited by the diffusive nature of near-infrared light propagation in tissue. Our proposed phased-array approach has the potential to enhance the spatial resolution and to afford depth discrimination, thus improving the spatial information content of diffuse optical images. Our next steps will be to explore different sets of amplitude and phase coefficients in Equation 1, to implement a dual-source array and detector array, and to study the feasibility of applying these principles to a reflectance geometry.

This research was supported by the National Science Foundation (Award BES-93840) and by the National Institutes of Health (Grants DA14178 and CA95885).