Optical coherence tomography (OCT) is a recently developed optical imaging technique that uses low coherence interferometry to perform high resolution, cross-sectional imaging in biological systems. While in vitro studies have been performed to demonstrate the feasibility of performing optical biopsy in human tissues, key technologies must be developed to extend this technique to in vivo internal organ systems. These advances include improvements in image acquisition speed, and the development of an OCT compatible catheter-endoscope. A fast scanning OCT system has recently been constructed. This system employs a high power (200 mW) chromium doped forsterite laser as the low coherence source and a piezoelectric fiber stretcher to induce reference arm optical path length delay. The fast scanning system acquires OCT images with an acquisition rate of four images per second, an axial resolution of 15 micrometers, and a signal to noise ratio of 112 dB. When incorporated with the recently constructed OCT compatible catheter-endoscope, this system is capable of obtaining high resolution endoscopic diagnostic images of tissue microstructure in vivo.

Optical imaging through highly scattering media is an active field of research. Recently, a novel method of ultrafast correlation interferometric reflectometry (UCIR) was proposed to convert the time propagation of pulses reflected from the objects into a corresponding coherence-domain interference pattern. This new approach permits femtosecond single-shot registration of reflections from objects within a medium using a CCD array which can be used for image through strongly turbulent media. In this paper, we describe the use of UCIR for imaging objects through a moving scattering medium. An 120 fs single-shot interferometric image of a slide is achieved with a resolution of 25 micrometers after backscattering through fast rotating scattering medium at 2.5 X 10³ rpm.

A new technique is demonstrated for noninvasive optical imaging for biological, medical, and non medical systems based on the polarization characteristics of the backscattered light. Two versions of the technique are demonstrated: one suitable for surface imaging and the other for subsurface imaging.

Breast imaging using computed optical tomography (COT) is an advanced form of medical optical imaging (MOI). X-ray tomography techniques are documented. COT is somewhat documented, but implementation of the concepts to create a practical, clinically useful imaging device has only recently been accomplished. The first step, building a data collection device incorporating a near infrared laser to acquire data in a reasonable time and with no biological hazard, has been accomplished. The luxury of straight-line propagation of x- rays is not applicable to COT. The next step, developing a reconstruction algorithm using data acquired from a turbid media to produce a clinically useful image in an acceptable time, has been accomplished and is being refined. Clinical evaluation of the effectiveness of the computed tomography laser mammography (CTLM) system is underway at two locations.

Continuous-wave ultrasonic modulation of laser light has been used to image objects buried in tissue-simulating turbid media. The ultrasonic wave focused into the turbid media modulated the laser light passing through the ultrasonic field. The modulated laser light collected by a photomultiplier tube reflected primarily the local mechanical and optical properties in the focal zone. Modulated signal was estimated using diffusion theory. The dependence of the ultrasound-modulated optical signal on the off-axis distance of the detector from the optic axis was studied. The mechanisms of ultrasonic modulation of light were discussed.

Since an ultrasound field can easily be focused in tissue, ultrasound modulated optical imaging (UMOI), has been recently proposed. It is based on the hypothesis that, at the ultrasound focal point, the compression and rarefaction of tissue will generate a modulated optical signal which is related to the local optical and/or ultrasound properties. The mechanisms of ultrasound-light interaction in a highly scattering medium are not well described to date. Feasibility studies on UMOI are presented in this paper based on optical properties changes, diffraction pattern and speckle pattern changes caused by focused ultrasound in 5 cm breast tissue. We used diffusion theory to calculate the portion of incident fluence entering the ultrasound focal zone and the portion of modulated fluence from the focal zone reaching the detector. Effects of optical property and refractive index changes are investigated. Based on the statistic distribution of the speckle pattern, the probability function of detecting modulated signal has been derived. To verify the theory, an experiment system has been developed which includes an He-Ne laser, a focused ultrasound generator, a photomultiplier detector and bandpass filter, an rf lock-in amplifier, and a spectrum analyzer. Transmission measurements on a travamulsion^TM aqueous phantom of 30 optical mean free path thickness can be obtained with the detector on axis with the laser beam, in which we believe only ballistic photons are detected.

This study investigates the potential use of combining near infrared (NIR) diffuse light and ultrasound methods for early detection of small breast tumors. A novel imaging system which combines these two methods has been introduced. A NIR phased array positioner capable of providing the line-of-sight of a small heterogeneity in a scattering medium is incorporated into an ultrasound imaging system. This combined system has been tested using models consisting of few mm emulated tumors buried in a scattering phantom. In each test, the combined system was mechanically translated to scan for a small object. At each scanning position, optical amplitude and phase data and an ultrasound image were acquired. The object line-of- sight information was coded into ultrasound images after scanning. Ultrasound image without and with line-of-sight information were evaluated by two radiologists. ROC curves using ultrasound and combined methods are measured and the areas under the average ROC curves of ultrasound and combined methods are 49% and 82%, respectively. Thus the combined method affords a significant improvement in detection sensitivity and specificity compared with conventional ultrasound in this phantom study.

Laser optoacoustic imaging experiments in biological tissues in vivo are presented along with the theoretical signal analysis procedure. The laser optoacoustic imaging system (LOIS) can operate in the reflection mode with emphasis on high z-axial (in-depth) resolution (up to 10 - 20 micrometers). Two examples of LOIS applications for non- invasive in vivo medical diagnostics are presented and discussed: (1) characterization of layered structure of port- wine stains, and (2) measurements of skin melanoma thickness. Potential miniaturization of LOIS for endoscopy applications is also discussed. The z-axial profiles of laser-induced pressure were shown to contain diagnostic information on location, dimensions and optical properties of tissue layers. Time-resolved signals detected by piezoelectric transducers were corrected for distortions such as diffraction and acoustic attenuation that occur upon pressure wave propagation in tissue. Wavelet transform applied to signals of laser- induced acoustic emission yielded high contrast pressure profiles with substantial signal-to-noise ratio.

Laser optoacoustic imaging is a promising diagnostic technique for early breast cancer detection. Capability of laser optoacoustic imaging for visualization of small spherical tumor phantoms located within the bulk collagen gels was studied. The experiments were performed with breast phantoms made of optically turbid collagen gel. Optical properties of the phantom resembled the optical properties of human breast at the wavelength of irradiation, 1064 nm ((mu) _a equals 0.11 1/cm, (mu) _s' equals 2.92 1/cm). Gel spheres with a higher absorption coefficient, (mu) _v equals 0.75 1/cm were used to simulate tumors. The experiments demonstrated the capability of laser optoacoustic imaging to detect and localize 2-mm 'tumors' at a depth of up to 60 mm within 100-mm thick breast phantoms. Laser optoacoustic images of the phantom tumors were reconstructed from experimentally measured pressure profiles. The optoacoustic images were compared with images obtained with x-ray mammography and ultrasonography. Comparative study revealed experimental conditions and phantom structure for which the laser optoacoustic imaging outperformed both the x-ray mammography and the ultrasonography. The results suggest that the laser optoacoustic imaging may occupy an important niche in breast cancer diagnostics, particularly, for diagnosis of small tumors in radiologically dense and acoustically homogeneous breast tissues.

We developed a continuous wave (cw) light imaging probe which includes 9 light sources and four pairs detectors (each pair has one 850 nm filtered detector and one 760 nm filtered detector). The light sources are controlled by a computer and the signals from the detectors are converted and processed in the computer. There are 16 measurement sections and total detection area is 9 cm multiplied by 4 cm which can be scanned every 8 seconds. The detector-source uses 2.5 cm spacing. In this study, we present the noise, drift, detectivity and spatial resolution test results of the imager. Changes of oxygenation and blood volume in about 2 cm depth from the surface of brain model can be detected. The temporal resolution is 8 seconds and spatial resolution is about 2 cm. The detectivity of OD changes can reach 0.008. With this cw imaging probe, we measured motor function in motor cortex area, visual function in occipital area, and cognitive activity in frontal forehead area of the human brian when the subjects are stimulated by moving fingers, viewing a flashing light and doing an analogy test, respectively. The experimental results show that the cw imaging probe can be used for functional images of brain activity, base upon changes of oxygenation and blood volume due to the stimulus.

Calcification is a key marker that signals the possible presence of cancer in human breast. Calcium carbonate particles hidden in highly scattering media were optically imaged to demonstrate the feasibility of using time-resolved transillumination shadowgram approach for identifying calcification formation for breast cancer screening.

Localization of absorbers/fluorochromes deep within a highly scattering large body of tissue such as the human breast can be affected most precisely by photon diffusive waves (yodh, physical Review) that are in and out of phase amplitude modulated sources in the MHz region. The detector of in-phase and anti-phase signal establishes a null plane. Such a null is highly sensitive to perturbations by extremely small objects, of the order of the size of 70 microliters volume and containing as little as 20 picamole of an absorber (ICG). In order to enlarge the search field for phased array, the relative amplitude of the rf (200 Mhz or 50 MHz) is modulated by a second lower frequency signal (60 Hz), the phase transition plane moves back and forth at the lower frequency. The phased array system is applied to the detection the blood concentration and its oxygenation/deoxygenation in human breast pendant within a 'breast box' 40 cm X 15 cm X 15 cm, containing a breast matching intralipid and 1 cm³ tumor phantom used as a blood volume, blood oxygenation calibrator. The breast is located by a pair of soft compression plates during the phased array scan. The sensitivity of the system is estimated by the detectable blood concentration and the corresponding oxygenation/deoxygenation of the blood in the 1 cm³ tumor phantom. We have done over 40 experiments for human breast test.

An effective projection imaging technique is proposed that allows the detection of realistic optical inhomogeneities in highly diffusive media and the discrimination between scattering and absorption contributions. The method is based on information derived from time-resolved transmittance measurements. Scattering inhomogeneities are discriminated using maps of the transport scattering coefficient, as provided by best fitting of the diffusion theory to experimental data. Absorption variations are effectively classified by time-gating on the tail of the transmittance pulse. Images were constructed from matrices of time-resolved transmittance measurements performed with a mode-locked dye laser and an electronic chain for time-correlated single photon counting. Data were collected from realistic tissue phantoms containing cylindrical inhomogeneities (1-cm height and 1-cm diameter) embedded in a 5 cm thick turbid slab. The optical coefficients of the inclusions were varied separately ((mu) _s' equals 5 divided by 20 cm^-1 and (mu) _a equals 0.025 divided by 0.4 cm^-1) with respect to the background values ((mu) _s ' equals 10 cm^-1 and (mu) _a equals 0.1 cm^-1). The overall acquisition time was less than 15 min. In all the experimental conditions considered, the method discriminated efficiently the scattering from the absorption contribution when either of the coefficients or both of them were inhomogeneous.

In order to reconstruct images of heterogeneities embedded deep within tissues, multiple-pixel measurements of frequency domain photon migration are necessary. Typically, these measurements are performed at multiple locations on the periphery of the sample using individual fiber optics. However, measurements using multiple fiber optics are both tedious and time consuming, especially as one increases spatial measurement resolution. In our research, we acquire multi-pixel measurements by employing a gain-modulated image intensified CCD camera and near infrared milliwatt laser diodes to monitor the propagation characteristics of sinusoidally modulated light as it passes through tissue-like phantoms containing 0.5% intralipid solutions. Our results show that images of millimeter sized absorbing objects located greater than 1 cm inside an ideal tissue phantom an be successfully obtained. In addition, we have been able to generate fluorescent images locating portions of micromolar concentrations of indocyanine green embedded within an intralipid solution contained within a glass cylindrical vessel (10 mm by 4 mm diameter).

A gain-modulated image intensifier with a CCD array was developed and used to simultaneously acquire multi-pixel modulation amplitude and phase images of the interference pattern within 0.5% intralipid solution and resulting from two 5 mW near infrared sources modulated 180 degrees out-of-phase at frequencies between 10 and 120 MHz. Gain modulation was achieved by modulating the voltage between the image intensifier's photocathode and multichannel plate input. Homodyning the gain modulation and source signals enabled images of the interference pattern to be obtained with total exposure times on the order of 3 seconds or less. A strongly absorbing, embedded heterogeneity could be localized in two dimensions by its deflection of the interference plane between two out-of-phase sources.

The use of fluorescent and phosphorescent dyes has long been established in microscopy and bio-assays for rapid and accurate biochemical spectroscopy. In this work, we explore extension of these dyes for in vivo tissue spectroscopy. Using a theoretical model for the time-dependent light propagation through scattering media, we investigate the coupling of light propagation and re-emission kinetics for florescent and phosphorescent dyes delivered deep within tissues. Our results show that the use of phosphorescent dyes can provide spectroscopy or localization of specific tissue compartments, but may be information-poor for localized spectroscopy or imaging. While re-emitted fluorescence signals contain contributions owing to both light scattering and kinetics, they do contain the necessary information for localized spectroscopy.

In the near-infrared spectral region (700 - 900 nm) light penetrates a few centimeters into tissues and hemoglobin dominates the absorption. Consequently, in vivo near-infrared tissue absorption spectroscopy becomes difficult for endogenous compounds of biological interest other than hemoglobin. Exogenous chromophore detection by fluorescence spectroscopy has the potential to provide enhanced sensitivity and specificity for in vivo optical tissue spectroscopy, facilitating the study of many important metabolites in tissues other than hemoglobin. We have performed measurements of the dc fluorescence intensity generated by a fluorophore (rhodamine B) homogeneously dissolved inside a highly scattering tissue-simulating phantom (aqueous suspension of titanium-dioxide particles). The phantom was prepared with optical coefficients (absorption and reduced scattering) similar to those of tissue in the near-infrared; these coefficients were measured with a frequency-domain spectrometer. Measurable dc fluorescence intensity signals from 1 nM rhodamine concentrations inside the phantom are reported. Furthermore, we were able to resolve changes in rhodamine concentration on the order of 1% using the dc fluorescence intensity. This dc fluorescence sensitivity is characterized experimentally at two concentrations (55 and 360 nM) and over a range of source-detector separations. Other aspects of the sensitivity are discussed over a large range of concentrations using a fluorescence photon migration model.

A series of fluorescent surface images were obtained from physical models of localized fluorophores embedded at various depths and separations in tissue phantoms. Our random walk theory was applied to create an analytical model of multiple flurophores embedded in tissue-like phantom. Using this model, from acquired set of surface images, the location of the fluorophores was reconstructed and compared it to their known 3-D distributions. A good correlation was found, and the ability to resolve fluorophores as a function of depth and separation was determined. In parallel in in-vitro study, specific coloring of sections of minor salivary glands was also demonstrated. These results demonstrate the possibility of using inverse methods to reconstruct unknown locations and concentrations of optical probes specifically bound to infiltrating lymphocytes in minor salivary glands of patients with Sjogren's syndrome.

Optical properties of microbubbles (EchoGen, Sonus Pharmaceuticals, WA) have been evaluated using a 200 MHz phase modulation system. Studies have shown that the mean optical pathlength has been increased at 754 nm by an average of 1.5 mm (std. 0.46 mm) when EchoGen was administered to 11 Fischer rats at the dose of 100 (mu) l or 150 (mu) l. As a comparison, the mean pathlength has been decreased at the same wavelength by an average of 4.2 mm (std. 3.66 mm) when indocyanine green (ICG) was administered at the dose of 100 (mu) l. The mean pathlength change at 816 nm is very small and the possible reasons are discussed. Potential applications of using microbubbles to provide a co-registration of ultrasound and light images of growing tumors are discussed.

In this paper, a reconstruction algorithm for fluorescence yield and lifetime imaging in dense scattering media is formulated and implemented. Two frequency domain radiation transport equations based on the diffusion approximation are used to model the migration of excitation and emitted photons. In the forward formulation, a finite element approach, which is specially effective for complex geometries and inhomogeneous distribution of medium properties, is adopted to obtain the required imaging operator and the simulated detector responses related to the photon fluxes on the boundary. Inverse formulation is derived based on the integral form of two diffusion equations. The technique is demonstrated by reconstructing spatial images of heterogenous fluorophore distribution and life time using simulated data obtained from homogeneous and complex (i.e., MRI breast map) media containing objects with fluorophore and with and without added noise.

An experimental study of detectability in steady-state optical image reconstruction through varying the single target size (15 mm - 4 mm in diameters) at three different locations (center, middle and near boundary surface) and a through evaluation of our finite element based image reconstruction algorithm to distinguish multiple targets are performed in this paper. The multi-target experiments consisted of several interested geometric and contrast combinations including two targets with the same contrast (2:1 between the target and the background) at three different separation distances, two targets with different contrasts (2:1 and 4:1) at two different separation distances and three targets with the same and different sizes (20 mm, 15 mm and 8 mm in diameters) and contrasts (2:1, 4:1 and 8:1) respectively. The reconstruction algorithm used along with a few imaging enhancement methods including total variation minimization, dual meshing and spatial low pass filtering are discussed. Quantitative measures of image quality including the size, location and shape of the heterogeneity are used to quantify the analysis. The results show that near 22:1 ratio (tissue thickness relative to detectable anomaly size) can be obtained and multiple targets can be correctly resolved using dc data with an 86 mm diameter circular tissue-like phantom.

There are two reasons that might be attributed to the difficulty for the imaging problem in optical tomography, and in inverse problems in general. Firstly, the problem is mostly underdetermined. Secondly, the inverse problem is highly ill- conditioned due to the diffusive nature of the photons. We introduce Bayesian optimization that provides a method to incorporate a priori knowledge in the inversion and we show with the concept of nullspace that the Bayesian prior probability generalizes conventional regularization by introducing a prior model. Reconstruction results of test objects from simulated data and a reconstruction example on a head model show that use the nullspace gives considerable improvement.

The success of time-resolved imaging of an abnormal site embedded in thick tissue may rely on one's ability to quantify the absorption coefficient of the target as a specific spectroscopic signature. This task is particularly complicated when the scattering properties of the target differ from those of the surrounding tissue. Using data obtained from time- resolved transillumination experiments of abnormally absorbing and differentially scattering objects embedded in a tissue- like phantom, we show how a new deconvolution algorithm enables us to quantify the optical properties of the target. The algorithm is based on a photon random walk theory that expresses different time-dependent point spread functions to calculate the diffusive and absorptive contrasts obtained in time-of-flight measurements.

Optical inhomogeneities embedded in a turbid medium are characterized not only by their absorption and reduced scattering coefficients, but also by their index of refraction relative to the background medium. Although in diffusion theory it is impossible to separate the index of refraction from the absorption and reduced scattering coefficients in an infinite homogeneous medium, application of boundary conditions for an inhomogeneity adds enough information to separately determine these optical properties. A mismatched index of refraction affects diffuse photon propagation in two ways: photons travel at a different speed inside the inhomogeneity, and photons entering and leaving the inhomogeneity are influenced by Fresnel reflections at the surface of the object. We have integrated these two effects into the analytical solution to the diffusion equation for a cylinder in an infinite medium. Theoretical results are compared with experimental data, and the effect of index of refraction mismatch is evaluated for different combinations of optical properties.

We consider the problem of emission tomography in a highly- scattering medium probed by diffusing waves. Experimental validation of an algorithm that gives a direct solution to the image reconstruction problem is reported.

In this paper, we present a Born-Type iterative algorithm for reconstruction of absorption and diffusion coefficient distributions of a heterogeneous scattering medium. This method is derived based on the integral form of the diffusion equation for the photon flux. It takes into account the nonlinear nature of the problem by using an iterative perturbation approach. Within each iteration, the forward problem (update of the total field and Green's function) is solved by the finite element method (FEM), and the inverse problem (update of the medium properties) is obtained by a regularized least squares method. This method has been used to reconstruct 'pathologies' embedded in an inhomogeneous test medium simulating a normal female breast from frequency domain data. The test medium is constructed by assigning optical coefficients according to an MR derived anatomical map. Our simulation results show that the algorithm is computationally practical and can yield qualitatively and quantitatively correct absorption and scattering distributions of embedded objects from simulated data with up to 5% additive noise in the simulated measurement data.

A novel inverse algorithm, which combines a 2D matrix inversion with a 1D Fourier transform inversion, is designed for obtaining an image of 3D hidden objects in scattering media. The existence of 2D boundaries, on which source- detector pairs are located around, violates the condition for using a 3D Fourier transform inverse imaging. This effect has been handled in our approach. This method greatly reduces the computational burden, compared to standard 3D matrix inversion methods. The result of image of hidden objects using time- resolved simulated data is presented.

Human head has anatomically and optically layered structure (skin, skull, gray and white matters), so that light propagates 3-dimensionally in the head taking complicated paths. Revealing the light paths is necessarily required to solve inverse problems and realize 3-dimensional optical tomography. In order to simulate the intracephalic light propagation, computational model for 3-dimensional finite elements method (FEM) was constructed. It had a hemispherical geometry and a layered distribution of scattering and absorption coefficients. The model consisted of 12528 elements, and the dimension of model, scattering and absorption coefficients of each element were assigned according to those of neonatal head. Impulse source was assumed, and boundary and initial conditions were appropriately given. Time-dependent photon diffusion equation was solved for fluence rate. And 3-dimensional distribution of fluence rate and time-dependent light propagation were then obtained. Light paths between source and detectors were also calculated. These results were visualized using computer graphics techniques.

Experimentally acquired time-resolved and continuous-wave diffuse transmittances of near-infrared light through cylindrical solid phantoms are presented. Homogeneous and inhomogeneous tissue-like phantoms (68 mm diameter, epoxy resin) were used. The transmittances were acquired at 18 degree intervals around the phantoms. Experimental results are compared with results of two dimensional and three dimensional finite-element simulations. The influence of absorbing boundaries on the shape of time-resolved and continuous-wave transmittances has been investigated for slab and cylindrical geometry. Analytical solutions of the photon diffusion equation with reflecting and absorbing boundary conditions were applied to estimate the optical properties of the material. A new index, time-resolved homogeneity index, is introduced for quick and simple inhomogeneity detection prior to image reconstruction.

A novel imaging algorithm for diffusion/optical tomography is presented for the case of the time dependent diffusion equation. Numerical tests are conducted for ranges of parameters realistic for applications to an early breast cancer diagnosis using ultrafast laser pulses. This is a perturbation-like method which works for both homogeneous a heterogeneous background media. Its main innovation lies in a new approach for a novel linearized problem (LP). Such an LP is derived and reduced to a boundary value problem for a coupled system of elliptic partial differential equations. As is well known, the solution of such a system amounts to the factorization of well conditioned, sparse matrices with few non-zero entries clustered along the diagonal, which can be done very rapidly. Thus, the main advantages of this technique are that it is fast and accurate. The authors call this approach the elliptic systems method (ESM). The ESM can be extended for other data collection schemes.

The optical properties of controlled size latex particles suspended in water have been investigated by using two different time-resolved transmittance set-ups. Least-mean square fitting between experimental data and analytical solution to diffusion approximation equation has given values of optical parameters in good agreement with the predictions of the Mie theory. In this way, the validity of these predictions was checked in the investigated experimental conditions and the data analysis confirms that time-resolved transmittance can be a reliable technique to measure the optical properties of scattering solutions.

The measurement of absorption coefficients in small volumes of turbid media using endoscope-compatible probes potentially has many applications including the assessment of concentrations of PDT and chemotherapy drugs in tissue. Methods for determination of absorption coefficients in small volumes for which the diffusion approximation is invalid have not previously been developed. Our aim its to develop a technique for measuring small changes in absorption based on broad wavelength, cw measurements. Our primary approach is to develop phenomenological algorithms based on the results of Monte Carlo simulations of light transport. The algorithms are tested using tissue phantoms to which known amounts of an absorber have been added. Results indicate that for some light excitation and collection geometries the average pathlength for the collected light does not depend on the scattering properties for the range of scattering parameters expected in tissue.

We used frequency-domain measurements with immersion in a matched scattering medium to measure the optical properties of tissue. This approach reduces the effects of boundaries and geometry on the measurements. The diffusion equation, typically used as the basis for determining tissue optical properties, is not valid near a boundary. By immersing the tissue in a scattering medium with nearly the same scattering and absorption coefficients, the boundary effects are greatly reduced and the measurement conditions nearly approximate that in an infinite medium. Also, the measurement is made as a differential against a true homogeneous medium for which the scattering and absorption coefficients can be obtained quite accurately. We demonstrated this approach, using frequency- domain measurements to determine absorption and scattering coefficients of living tissue immersed in diluted intralipid, Ropaque, or milk. Previous approaches to determine optical properties in the frequency-domain used multiple path lengths or multiple frequencies. Immersion is a third frequency-domain approach.

The vast majority of 'non-invasive' measurements of human tissues using near infrared spectroscopy rely on passing light through the dermis and subdermis of the skin. Accurate knowledge of the optical properties of these tissues is essential to put into models of light transport and predict the effects of skin perfusion on measurements of deep tissue. Additionally, the skin could be a useful accessible organ for non-invasively determining the constituents of blood flowing through it. Samples of abdominal human skin (including subdermal tissue) were obtained from either post mortem examinations or plastic surgery. The samples were separated into a dermal layer (epidermis and dermis, 1.5 to 2 mm tick), and a sub-cutaneous layer comprised largely of fat. They were enclosed between two glass coverslips and placed in an integrating sphere to measure their reflectance and transmittance over a range of wavelengths from 600 to 1000 nm. The reflectance and transmittance values were converted into average absorption and reduced scattering coefficients by comparison with a Monte Carlo model of light transport. Improvements to the Monte Carlo model and measurement technique removed some previous uncertainties. The results show excellent separation of reduced scattering and absorption coefficient, with clear absorption peaks of hemoglobin, water and lipid. The effect of tissue storage upon measured optical properties was investigated.

Tissue optical properties are necessary parameters for prescribing light dosimetry in photomedicine. In many diagnostic or therapeutic applications where optical fiber probes are used, pressure is often applied to the tissue to reduce index mismatch and increase light transmittance. In this study, we have measured in vitro optical properties as a function of pressure with a visible-IR spectrophotometer. A spectral range of 400 - 1800 nm with a spectral resolution of 5 nm was used for all measurements. Skin specimens of two Hispanic donors and three caucasian donors were obtained from the tissue bank. Each specimen, sandwiched between microscope slides, was compressed by a spring-loaded apparatus. Then diffuse reflectance and transmittance of each sample were measured at no load and at approximately 0.1 and 1 kgf/cm². Under compression, tissue thicknesses were reduced up to 78%. Generally, reflectance decreased while the overall transmittance increased under compression. The absorption and reduced scattering coefficients were calculated using the inverse adding doubling method. Compared with the no-load controls, there was an increase in the absorption and scattering coefficients among most of the compressed specimens.

A near infrared spectrometer has been constructed which is capable of performing spatially resolved diffuse reflectance measurements in the wavelength range of 0.9 - 1.6 micrometer. In this technique, broadband light is delivered by an optical fiber to a point on the tissue surface and diffusely reflected light is collected by 300 micrometer fibers located at 15 distances ranging between 1.0 to 10.0 mm from the source. The light from the detector fibers is imaged with a monochromator onto an InGaAs photodiode array. Wavelengths can be selected by automated scanning of the monochromator grating. A diffusion theory model fit to the reflectance versus distance data has been used to estimate the absorption and scattering coefficients ((mu) _a and (mu) _s') of phantoms and tissue under analysis. Reflectance measurements have been performed on tissue simulating water-based phantoms as well as in vivo on different skin locations. The absorption coefficient of skin was found to have a spectral structure similar to that of water. Unexpected spectral features in the scattering coefficient of skin were observed which may be a result of not considering the layered structure of skin in the current model. The temporal stability of the system has been demonstrated on tissue-simulating phantoms and human volunteers, indicating that the reflectance measurement may be suitable for in vivo monitoring of physiologically induced changes in the absorption and scattering coefficients.

Most instrumentation used to measure tissue optical properties non-invasively employ data analysis algorithms which rely upon the simplifying assumption that the tissue is semi-infinite and homogeneous. The influence of a layered tissue architecture on the determination of the scattering and absorption coefficients has been investigated. Steady state reflectance data for a two-layered tissue architecture were generated using a Monte Carlo code. These were analyzed using a diffusion theory model to calculate the scatter and absorption coefficients. The estimated tissue optical properties were different from those for either layer, and for certain choices of input optical properties would result in physically impossible results. The sensitivity and specificity of the estimated optical properties to changes in input optical properties were calculated for different layered geometries. For typical optical properties of skin the results suggest that a change in the absorption coefficient of 100% will result in an apparent change in scatter coefficient of 15% and that a similar change only in scatter coefficient will result in an apparent change of 100% in absorption coefficient.

The tissue spectrometer EMPHO allows measurements of absolute hemoglobinoxygenation values, noninvasive1y at any hemoglobin-perfused tissue, simply by applying visible light on the surface of theorgans under investigation. The hemoglobin oxygenation-algorithm is based on Kubelka-Munk-Theory fortackling both. absorbance and scattering phenomena. Broad-band tissue spectra ofbackscattered light serves as data basis for the analysis. In this study the algorithm was tested for measurements in highly scatteringmedia, in Intralipid©, where erythrocytes were added step by step. The hemoglobin concentration in the suspension varied from 0.01 to 1.0 [g hb Idl of suspension], which corresponds to the range of hemoglobin concentrations physiologically found in various types of tissue. The oxygenation was changed from 0 % to 100 % by using a hollow-fibre oxygenator. The costly study revealed that the algorithm works with high accuracy at a middlehemoglobin-concentration of 0.3 g/dl. The error of calculation was smaller than 2% of the absolute HbO value. The statistics proved that errors were larger at the highest and lowest values o hemoglobin concentration. It could clearly be shown that the errorcan be minimized to 1 % by application ofnew gold-standard hemoglobin spectra ofO % and 100 % oxygenation. Key words: tissue spectrometry, hemoglobin oxygenation, Kubelka-Munk Theory, light absorbance,light scattering, visible wavelength range.

The shift in optical absorption of hemoglobin upon binding of oxygen provides a basis for near-infrared monitoring of hemoglobin oxygen saturation, which is an important indicator of tissue oxygenation. Tumor oxygenation has long been studied, because hypoxic cells exhibit resistance to ionizing radiation therapy. The ability to measure noninvasively the oxygenation status of tumors and their response to oxygen modifiers is important in research and clinical settings. We have implemented a steady-state diffuse reflectance method of optical spectroscopy in scattering systems based on the theory of Farrell et al. (Med. Phys., 1992). In scattering phantoms containing erythrocytes, the method recovers the hemoglobin absorption spectrum (650 - 820 nm) and accurately monitors hemoglobin oxygen saturation. We have implemented a probe that individually positions several detection fibers normal to the surface of subcutaneous rodent tumors. Near-infrared absorption spectra reconstructed from diffuse reflectance measurements indicate a hemoglobin oxygen saturation of approximately 50% in R3230AC rat mammary adenocarcinomas when the anesthetized animal breathes room air. Administration of carbogen (95% oxygen, 5% carbon dioxide) via a nose cone produces a rapid and readily detectable increase in the saturation to 75% with no increase in tumor blood volume. Several methods of determining hemoglobin oxygen saturation from absorption spectra obtained by diffuse reflectance spectroscopy are compared, including singular value decomposition, which provides the ability to reconstruct the non-hemoglobin absorbing background without a priori knowledge of its structure or absolute magnitude.

The optical properties of scattering media determine the attenuation (A) and the transit time (<t>) of light diffusely reflected from the medium as well as the phase ((Phi) ) and modulation depth (M) of an intensity modulated light wave. Here it is described how the absorption coefficient can be derived from the ratio of changes in A, (Phi) and M. These ratios are approximately independent of the scattering properties. The changes in A, (Phi) and M can be induced either by small changes in the absorption coefficient of the medium or by tuning the wavelength over the absorption spectrum of medium. The application for the in vivo monitoring of haemoglobin and oxyhemoglobin concentrations in human tissue with near infra red light is discussed.

Functional evaluation of local hemoglobin concentration and hemoglobin oxygenation based on back scattering spectra from human skin in vivo have been obtained in visible range (502 - 628 nm) by a rapid microlightguide spectrometer (EMPHO II) with step 250 micrometer. Analysis of received results has shown that during local cooling there is two nearly simultaneous reactions: reduction of hemoglobin concentration and increase of hemoglobin oxygenation level. In a case when one has used previous heating of planning place for cooling, reduction of hemoglobin concentration is expressed higher by 22 - 33%.

To evaluate the reliability of EMPHO II in human skin measurement, hemoglobin oxygenation (HbO₂) and hemoglobin concentration (Hb_con) behaviors were measured under ischemia and congestion. Functional evaluation of HbO₂ and Hb_con of human skin at forearm or fingers of healthy volunteers had been obtained in visible range (500 - 628 nm) by a rapid microlightguide spectrophotometer (EMPHO II). In a first series of investigations, ischemia or congestion of the skin was induced by upper arm compression to 250 mmHg or 80 mmHg, respectively. The data show that HbO₂ decreased under conditions of ischemia but also congestion, while Hb_con increased enormously under congestion alone. In a second series of experiments, the local oxygen uptakes of the skin under various temperature conditions (5 to 45 degrees Celsius) were determined from the decrease of intercapillary oxygen content [d(Hb_con*HbO₂)/dt] which was induced by a stop of blood flow. We concluded that the measurement of intracapillary hemoglobin oxygenation of human skin by EMPHO II is reliable and stable under several conditions. Furthermore, our data suggested that the changes in HbO₂, Hb_con and O₂ uptake of the skin seem to be a very useful parameter which can quickly change when tissue hypoxia occurs by unbalance of O₂-demand and supply.

In the last four years near infrared spectroscopy (NIRS) has been used in cerebral functional activation studies to monitor changes in concentration of oxy-, deoxy- and total hemoglobin [(O₂Hb), (HHb) and (tHb) respectively] in response to different stimuli. Previous studies were performed with a 1 - 2 Hz temporal resolution and a poor signal-to-noise (S/N) ratio. The aim of this study was to investigate the response of the motor cortex region during a finger opposition task in single subjects using a novel continuous wave NIRS instrument with enhanced temporal resolution and S/N ratio. Six subjects performed a sequential finger opposition task with the right hand (20 s duration; 2 Hz). The optodes were positioned over the left motor cortex region using an inter-optode distance of 3.5 cm. The high S/N ratio and 0.1 s sampling time allowed clear monitoring of (O₂Hb) and (HHb) changes due to heart beat as well as to respiration. The contribution of the heart pulse to the total signal was less than 0.4%. As previously shown by others using pooled data, an increase of (O₂Hb) during the activation accompanied by a decrease of (HHb) was found in most subjects for every activation cycle. Our approach provides a better insight into the underlying physiological mechanisms.

Near infrared optical spectroscopy is becoming a useful method for monitoring regional cerebral oxygenation status. The method is simple, reliable and noninvasive and the information which it provides is clinically significant in managing a growing number of neurological ailments. Use of this technique has been described previously by numerous authors. In the present study, regional cerebral oxygen saturation was measured at rest in 94 subjects randomly elected from a diverse population of individuals. This sample consisted of 38 males and 65 females, with the age ranging from 18 - 70. There were 68 light-skinned individuals and 35 with darker skin comprising various ethnic and cultural backgrounds. Mean regional cerebral hemoglobin oxygen saturation was recorded as 67.14 plus or minus 8.84%. The association of the mean regional cerebral hemoglobin oxygen saturation in various group of individuals in relationship of their age, race, sex and skin color is examined.

The purpose of this paper is to evaluate the value of transcranial cerebral oximeter in combination with other monitoring techniques during the balloon occlusion test. In this study 22 patients underwent balloon occlusion testing and were monitored by neurological examination, electroencephalography, transcranial Doppler, and transcranial cerebral oximetry. Eighteen patients had an intracranial aneurysm and four patients had skull base meningiomas. Seventeen patients passed the test without any symptoms. One patient underwent extracranial-intracranial by-pass surgery after failing the first test and passed the second test after the treatment. Transcranial cerebral oximeter showed 10% or less decrease in rSO₂ in patients who passed the test and 10% or higher decrease in rSO₂ for more than one minute in patients who failed. Electroencephalography and cerebral oximetry are found to be the most dependable monitoring methods for predicting neurological deficits after balloon inflation. Transcranial cerebral oximeter has provided non- invasive monitoring with real-time quantitative information and it was very useful during the balloon occlusion test.

Neurological impairments following cardiopulmonary bypass (CPB) during open heart surgery can result from microembolism and ischaemia. Here we present preliminary results from monitoring cerebral hemodynamics during CPB with near infrared intensity modulated spectroscopy. In particular, the study had two main objectives: (1) to monitor the oxy- and deoxy hemoglobin concentrations and their changes during the CPB surgery and (2) to monitor the transport scattering coefficient ((mu) _s') of the brain especially during cooling and rewarming. A new method for the calculation of absolute absorption coefficients ((mu) _a) was also tested. This method is based upon the monitoring of attenuation and phase changes that are induced by variations in absorption. These variations can be generated either by alterations in the tissue oxygenation or by injecting a dye (indocyanine green) into the CPB circuit. Absolute oxy- and deoxyhemoglobin concentrations and their changes during the CPB were calculated. The preliminary results suggest that cooling of the brain does not significantly alter (mu) _s'.

Pulsed-photoacoustic spectroscopy (PPAS) is an in-situ technique used for quantitative monitoring of brain-tissue hemoglobin concentration and its oxygenation state. In contrast to most spectroscopic techniques that measure infrared absorption PPAS does not require knowledge of the pathlength of light traveling through tissues in order to determine the absorption coefficient and hence the concentration of absorbing species. The photoacoustic response (PAR) is produced by light absorption. Light scattering modifies the spatial distribution of the absorption. PPAS uses a pulsed, tunable optical source coupled to a 1 mm diameter fiber optic cable to transmit optical energy to the tissue. The fiber can be placed on the exterior surface or inserted into the tissue. An ultrasonic signal is produced by light absorbed in the tissue. Since the rate of conversion of laser light to heat is rapid and the laser pulse much shorter than the tissue thermal-diffusion length, the ultrasonic signal amplitude is proportional to the energy absorbed. Spectra of absorbing compounds can be obtained by measuring the variation in the acoustic signal with source wavelength. Our studies demonstrate that acoustic spectra obtained both in-vitro and in-vivo allows relative changes in the concentration of oxy- and de-oxyhemoglobin to b45e determined.

Optical tomography aims to image the distribution of optical properties in human bodies by measuring transmitted light at skin surfaces. Pervious calculations and experiments have been mainly performed on phantoms with simple geometries such as slabs and cylinders, but for optical tomography it is inevitable to fully understand light propagation through and perform experiments using phantoms with complicated structures in three dimensions. Therefore, we need stable and realistic solid phantoms for experimental studies toward the goal of optical tomography. In this study, we have fabricated two types of solid phantoms which optically and anatomically simulate human heads. One has a shape and structures of a part of human head above eye plane, and the other has a more simplified shape of hemisphere. These phantoms consisted of five layers which corresponded to five tissue types in human head; i.e., skin, skull, clear CSF layer, gray matter and white matter. Size and optical properties were given according to those of human neonatal head. After taking original shapes from MRI images, prototypes of five layers were fabricated by a rapid prototyping based photolithography. Epoxy resin with titanium oxide particles as scatterers and green dye as absorber was cast into the molds of the prototypes to make optical phantoms. Absorbers simulating inhomogeneities were also embedded.

Real time noninvasive head injury detection is needed in critical care facilities and triage site with limited resources. One tool missing right now is a small and fast noninvasive sensor which can help urgent care workers to (1) diagnose the location and severity of the injury, (2) to perform on site pre-hospital treatment if necessary, and (3) to make a decision on what kind of further medical action is needed. On the other hand, continuous monitoring of cerebral blood oxygenation is also needed in intensive care unit and in operation rooms. Pseudo-random modulation/near infrared sensor (PRM/NIR sensor) is developed to address these issues. It relies on advanced techniques in diode laser cw modulation and time resolved spectroscopy to perform fast and noninvasive brain tissue diagnostics. Phantom experiments have been conducted to study the feasibility of the sensor. Brain's optical properties are simulated with solutions of intralipid and ink. Hematomas are simulated with bags of paint and hemoglobin immersed in the solution of varies sizes, depths, and orientations. Effects of human skull and hair are studied experimentally. In animal experiment, the sensor was used to monitor the cerebral oxygenation change due to hypercapnia, hypoxia, and hyperventilation. Good correlations were found between NIR measurement parameters and physiological changes induced to the animals.

A simple and reliable solid tissue phantom, made of agar, intralipid and ink is described and characterized. Following a standardized preparation procedure, it is fast and easy to produce inhomogeneous structures with known optical properties and a high degree of repeatability. The proposed phantom is used to measure the edge-profile produced by abrupt variations of either the 'absorption or the scattering coefficient in a turbid slab. The samples were scanned using a system for time- resolved transmittance measurements, based on a mode-locked dye laser and an electronic chain for time-correlated single- photon counting. The photon time-distributions were then interpreted either with a solution of the diffusion equation or with a time-gating. The plots of the transport scattering coefficient, of the absorption coefficient, and of the intensity integrated over an early time-interval as a function of position can give insight into the problem of spatial resolution and the mechanism of imaging through diffusive media with different techniques.

Currently, methods for the detection of brain edema in patients or laboratory experiments are invasive or inconvenient for continuous monitoring. We have performed experiments on two models of brain edema to determine the validity of differential near infrared spectroscopy (NIR) as a real-time, low cost and noninvasive method of monitoring brain edema. As a chemical in-vitro model, we prepared serial dilutions of Liposyn III, a fat emulsion, to simulate varying degrees of brain water content. NIR light at two wavelengths (703 nm and 957 nm) was used to assess water content of Liposyn solutions. We demonstrated a strong relation between wavelength specific light interactance and water content, for (n equals 4) serial dilutions from 97.6% to 80.0% water, R² equals 0.985 plus or minus 0.017. Secondly an in vitro brain tissue model was developed to test the NIR method against wet-to-dry water content measurements. A total brain water content range of from 83.5 to 67.6 water was investigated (n equals 4). Using 695 nm and 957 nm NIR light, a correlation between NIR interactance and brain water content was again obtained, R² equals 0.957 plus or minus 0.027. Our preliminary results suggest differential NIR spectroscopy may serve as an accurate and useful technique for monitoring surface brain edema in clinical and laboratory settings.

Fiber optic evanescent wave Fourier transform infrared (FEW- FTIR) spectroscopy using fiberoptic sensors operated in the attenuated total reflection (ATR) regime in the middle infrared (IR) region of the spectrum (850 - 1850 cm^-1) has recently found application in the diagnostics of tissues. The method is suitable for noninvasive and rapid (seconds) direct measurements of the spectra of normal and pathological tissues in vitro, ex vivo and in vivo. The aim of our studies is the express testing of various tumor tissues at the early stages of their development. The method is expected to be further developed for endoscopic and biopsy applications. We measured in vivo the skin normal and malignant tissues on surface (directly on patients) in various cases of basaloma, melanoma and nevus. The experiments were performed in operating room for measurements of skin in the depth (under/in the layers of epidermis), human breast, stomach, lung, kidney tissues. The breast and skin tissues at different stages of tumor or cancer were distinguished very clearly in spectra of amide, side cyclic and noncyclic hydrogen bonded fragments of aminoacid residuals, phosphate groups and sugars. Computer monitoring is being developed for diagnostics.

Endovascular treatment of carotid cavernous fistula is done routinely in our institution. We have been monitoring these patients with transcranial cerebral oximetry. The transcranial cerebral oximeter is a reliable, low-cost, non-invasive device that provides real-time evaluation of regional brain oxygen saturation during and after endovascular treatment of cerebrovascular diseases. We used the INVOS 3100A (Somanetics, Troy, MI) in our study. We discuss seven patients with carotid-cavernous fistulas treated by endovascular balloon occlusion, each monitored continuously before, during, and after the procedure with transcranial cerebral oximetry. The cerebral oxygen saturation depicted was directly related to the side of the venous drainage of the fistula, with the brain oxygen saturation 15 - 20% higher on the side of the venous drainage. Following endovascular occlusion of the fistula, oxygen saturation gradually became equal on both sides. In our patients treated for carotid-cavernous fistula, we evaluated the sensitivity and usefulness of cerebral oximetry as an important non-invasive monitoring tool for the endovascular treatment of carotid-cavernous fistula.

The non-invasive determination of the depth of severe burns is an important problem whose solution would offer medical practitioners a valuable tool for diagnosing and treating severe burns. Burned tissue is essentially a turbid medium with spatial varying dynamics: light is multiply scattered by the tissue and layers of burned tissue are distinguished by the degree of blood flow. The dynamical properties of turbid media can be probed by monitoring the temporal fluctuations of scattered light speckles. Information on a system's dynamics is obtained from the temporal autocorrelation function of these intensity fluctuations. We have recently shown that the correlation diffusion equation (CDE) accurately predicts the temporal correlation function for turbid systems with spatially varying dynamics and that the dynamical properties of such systems can be imaged using standard reconstruction algorithms. In this contribution, we demonstrate the sensitivity of temporal field correlation measurements to variations of 100 micrometers in burn thickness and the potential applicability of the CDE for quantitation of burn thickness. Results are presented from burn phantoms and pig models. The combination of diffusing temporal light correlation with diffuse reflectometry for enhanced burn diagnosis is investigated.

The early photons that arrive at a collector through a large thickness of tissue have potential for imaging internal organ structure, function, and status with improved image resolution relative to late arriving photons which have been diffusely scattered. Calculation of early photon arrival at a collector after transport through large tissue thicknesses is difficult, yet a comparison of predicted versus measured transmission is at the heart of imaging algorithms. The path integral description of light transport offers an approach toward such calculations. The method describes the movement of photons as particles undergoing collisions in a scattering medium based on the Brownian motion formalism of Feynman and Hibbs. This paper presents a basic introduction to the path integral description of photon transport and discusses the constrained classical path for describing the most probable path of a photon and the unconstrained classical path for describing the group path of an ensemble of photons.

In this study we analyze the limits of the diffusion approximation to the Boltzmann transport equation for photon propagation in the human brain. Two dimensional slices through the head are obtained with the method of magnetic resonance imaging (MRI). Based on these images we assign optical properties to different regions of the brain. A finite- difference transport/diffusion code is then used to calculate the fluence throughout the head. Differences between diffusion and transport calculations occur especially in void-like spaces and regions where the absorption coefficient is comparable to the reduced scattering coefficient.

We have investigated the accuracy of the frequency-domain, diffusion equation Green's function in modeling optical signals in turbid media. Our optical measurements were conducted in strongly scattering media having absorption coefficients in the range 0.035 - 0.14 cm^-1 and reduced scattering coefficients in the range 5 - 22 cm^-1. The optical signal was both delivered to and collected from the sample by means of optical fibers. We have verified that the r-independent factor in the Green's function does not agree with the experimental data. This discrepancy shows that the photons launched in the medium by an optical fiber cannot be modeled by a point source located at a fixed position. These results have no influence on multi-distance measurement protocols (which only employ the r-dependent part of the solution), while they must be considered when calibration measurements on a known sample are performed.

We consider two different skin structure models. The first structure consists of epidermis, dermis/blood, and subcutaneous tissue. The second structure consists of epidermis/dermis, adipose tissue and muscle tissue. A new solution based on diffusion theory of the cw local diffuse reflectance from a three-layered skin tissue structure is presented. Comparisons with Monte Carlo simulations are carried out favorably. It is shown that the functional form of the radial dependence of the diffuse reflectance from multilayer and single layer models are identical. We use a modified expression originating from diffusion theory to fit the diffuse reflectance. We discuss the sensitivity of the local diffuse reflectance as a function of the optical properties of separate layers in both structures. Moreover, we investigate the influence on the local diffuse reflectance with changes in the optical properties corresponding to normal changes in tissue glucose concentration and blood volume. The necessity of multilayer models lies within their ability to provide a detailed description of the light-tissue interaction rate than their applicability to practical data analysis of the local diffuse reflectance measurements.

Near infrared light propagation in the adult head has been shown to be considerably affected by the inhomogeneity of tissue. This indicates that the contribution to the NIR signal by absorption perturbations in each layer of the head is different. In this study, the change in NIR signal caused by a local or global absorption perturbation in a particular layer of the head has been analyzed by Monte Carlo prediction. A slab model of the adult human head consisting of five layers is used and the absorption coefficient in the whole region of a layer is changed for global absorption perturbations and in a small region of 10 mm by 10 mm by 5 mm rectangular solid for the local absorption perturbation. The differential pathlengths, which is partial derivative of the attenuation, were calculated from the differences in detected light intensity caused by the perturbations. Since the position of the local absorption perturbation with respect to the optodes affects the NIR signal, the absorption coefficient is changed in a local region either just below an optode or below the middle point of the optodes and the resulting attenuation in NIR signals are calculated.

Multi-wavelength, frequency-domain measurements of phase-shift provides a self-calibrating measure of isotropic scattering coefficients. From these measurements we reconstruct particle size distribution and volume fraction for three latex suspensions that are concentrated, yet non-interacting. Extrapolation of the technique to non-spherical particles, such as titanium dioxide, illustrates preliminary success of the approach.

Using the finite-difference time-domain technique (FDTD), the scattering patterns from cells and organelles of arbitrary shape can be computed. With this method Maxwell's curl equations are discretized in space and time and the electric and magnetic fields are computed at all points within and around the cell. The cell is constructed as a dielectric object and the far-field scattering pattern, containing both amplitude and direction information is computed. Results are presented for three-dimensional cells containing different combinations of organelles, such as nucleus, cytoplasm, and mitochondria, to assess the effect of each on the scattering pattern. The computed scattering patterns indicate that small organelles such as mitochondria play an important role in scattering from cells and variations in the refractive index of the nucleus also affect the scattering characteristics.

Knowledge of the size distribution of scatterers in tissue is necessary for understanding the physical processes involved in light-tissue interaction. In this paper we propose and test a model of light scattering in tissue on a microscopic scale. We start from the hypothesis that tissue can be treated as fractal over a certain range of dimensions and proceed to derive a simple scaling law for particle sizes. To test the model, we use a number of sizes of randomly distributed spheres to approximate the fractal structure. Our results show that the fractal model yields credible estimates of the magnitudes of the optical scattering cross sections of tissue, as well as their angle and wavelength dependencies. The numerical data are used to estimate the sizes of the particles that contribute most to the total scattering and backscattering coefficients at a several wavelengths in the visible and near-infrared bands.

This report concerns with investigations on the spatial polarization anisotropy of a backscattered intensity pattern produced at a boundary plane between a scattering medium and surroundings under illumination of a focused laser beam. The investigations are conducted by means of Monte Carlo simulations based on a free pathlength distribution function of the fractal medium, a phase function of the Rayleigh-Debye scattering theory, and a negative-exponential decay of the propagating light by absorption. In this report, we demonstrate numerically the spatial polarization anisotropy of the intensity distribution produced at the boundary plane between the medium and the surroundings, a bow-tie intensity pattern for the co-polarization and a cloverleaf intensity pattern for the cross-polarization, and their dependence on the fractal dimension of the medium. Consequently, we can discuss the validity of our model and algorithm for the fractal medium as the scattering object by studying the dependences of the polarization properties of the multiply- backscattered light on the dimension, the scattering order, and the absorption in comparison with the case of the non- fractal medium.

In order to study the contribution of endogenous and exogenous fluorophores that participate in the global fluorescence signal, a Gaussian analyzing model which can be used for the extraction of important clinical information about tissue metabolic-structural characterization has been developed. This model is intended to the detection of lesions that are characterized by a low or masqued fluorescence emission of a specific fluorophore. The model described here in is based on experimental data of fluorescence spectra measured in phantoms simulating tissue absorption, diffusion, autofluorescence and induced fluorescence (with known concentrations of MTHPC). Each fluorescence peak in the measured spectrum is supposed to be situated in the center of a Gaussian distribution. So, the fitted fluorescence model is composed of the sum of all the Gaussians representing principal and secondary peaks appeared in the measured spectrum. Every Gaussian is characterized by three parameters: bandwidth (Bw), central peak position ((lambda) _p) and proportionality coefficient ((alpha) ). The built model can be then used to reconstruct the spectral shape of the fluorescence emission of a given phantom or tissue when only the proportionality coefficients are fitted and to separate effects due to fluorescence yield, tissue modulation, and fluorophores concentration. This approach permits us to identify tissue characteristics more accurately.

A novel method for correcting the fluorescence emission and excitation spectra is applied to native fluorescence spectra from normal and cancerous human breast tissues. The method effectively eliminates the distortions produced by internal light-absorption and allows a direct, real-time, correction without any iterative procedures. A simplified photon- diffusion model was used to develop the method. An analysis of both the true fluorescence spectra, and the diffuse reflectance spectra transformed into the ratio of absorption and reduced scattering coefficients, shows distinctive biochemical differences between cancerous and normal breast tissues. The fluorescence spectra feature a lower contribution of NADH and, possibly, collagen and elastin in cancerous tumor tissues as compared with normal tissues. The fluorescence spectra from cancerous tumors also show a lower degree of variability than the spectra from normal tissues. The corrected spectra from cancerous tumors show a greater similarity in their profiles than the non-corrected fluorescence spectra distorted by the internal light- absorption.

Improved models based on Kubelka-Munk theory, as well as improved algorithms, for the quantification of chromophore concentrations in tissues based on optical multi-wavelength measurements, recorded in the visible range, are presented. The usefulness of principal component analysis (PCA) and canonical correlation analysis (CCA) for model validation is investigated and proven by many examples. It is shown how PCA is able to detect the number of independent components with variable contribution. A subsequent CCA points out these components, which should be included in the model function. Furthermore, three different methods for estimating the model parameters are presented: a nonlinear least squares (NLLS) algorithm, an ordinary least squares (LS) algorithm and a mixed LS-total least squares algorithm wLS-TLS with appropriate weights. The wLS-TLS algorithm offers the best compromise since it is almost as efficient as ordinary LS and almost as accurate as NLLS fitting.

An optical topography algorithm is presented for a two dimensional stratified slab. Under the approximation of phase function, the Green function technique is employed to solve the time resolved propagation equation of ultrafast laser pulses within the slab. The initial condition of the Green function can be expressed with an explicit form and is correlated with the optical parameters and the incident angles. This leads to a straightforward numerical inverse algorithm to reconstruct the optical parameters simultaneously. The reconstruction algorithm shows promise in noninvasive detection of the structure of turbid medium.

A multichannel time resolved imaging system is being developed at University College London (UCL) suitable for optical tomography of the human breast and neonatal brain. The system utilizes the time correlated single photon counting technique, operating in reverse start-stop mode. The detectors are custom made multi-anode microchannel plate photomultipliers (Hamamatsu R411OU-05MOD). Signal processing is implemented using fast NIM logic, and the system is based upon the ORTEC 9308-D picosecond time analyzer in place of a conventional time to amplitude converter. The system is designed to acquire 32 channels of time-resolved data simultaneously, with sub-100 ps temporal resolution, and at count rates of at least 2 multiplied by 10⁴ photons per second per channel. This paper discusses the theoretical considerations which led to the final design, and describes the detector and electronic hardware on which the system is based.

Analysis of 3D shape and size of skin lesions and features obtained by analyzing the images of skin-lesions, may help early diagnosis of melanoma. Our efforts in non-invasive imaging of skin lesions have been in the direction of characterization of the lesion volumes with transillumination based optical imaging modality. Light photons after entering into the skin interact with matter for multiple scattering events. The backscattered radiation re-emerging from the skin forms two dimensional projection images of the transilluminated skin-lesion. These projection images are obtained through direct view and mirror reflections. The problem of solving for the volume of the lesion from these projections is a non-linear inverse problem. We attempt to solve this using dipole solutions to the diffusion equation. The objective here is to obtain the jacobian (voxel weights) of the system of equations and progressively solve linear inverse problem. An iterative nonlinear inversion method is proposed, which incorporates the salient features of the rigorous perturbation Monte Carlo method at much lower computational overheads, to compute the jacobian. The jacobian obtained based on diffusion theory is used in reconstruction of the lesions based on algebraic reconstruction techniques. a milk-gelatin phantom of tissue and embedded lesion is constructed in which the concentration of milk determines extent of scattering. The absorber is simulated by adding ink to the same mixture in an embedded capillary tube. The phantom is trans-illuminated using the Nevoscope. Reconstructions of the phantom based on diffusion theory are obtained. Our results show smoother reconstructions for the diffusion theory based method.

In a paper by the first three authors a new algorithm for optical tomography (OT) in the time domain was described and tested for the case of a uniform background medium (see current proceedings). The goal of this paper is to test this method for the case where the background medium is an anatomically accurate optical map of the breast tissue obtained on the basis of a segmented magnetic resonance (MR) image. The key innovation of this imaging algorithm lies in a new approach for a novel linearized problem (LP). Such an LP is reduced to a boundary value problem for a coupled system of elliptic partial differential equations, rather than a conventional form of an ill-posed integral equation. The solution of this system in turn is done very rapidly. Thus, this is a fast imaging algorithm.

The use of diffusion theory in calculations on photon waves necessitates a new look at boundary conditions, since the standard boundary conditions have been derived under static conditions. When the underlain process satisfies the transport equation, the proper boundary conditions are obtained by solving the Milne problem. This paper presents benchmark- quality values for extrapolation distances calculated by transport theory, for various values of absorption and three models of the phase function -- isotropic, linearly anisotropic and Henyey-Greenstein scattering. The results show that the static boundary conditions are perfectly adequate up to photon wave frequencies of 1 GHz or even more. Specifically, the quantity (Sigma) _trd, where (Sigma) _tr' equals (Sigma) _tr - ik, where (Sigma) _tr is the macroscopic transport cross section and k the wave number in the medium and d the linear the linear extrapolation distance, is essentially independent of frequency over this range. We have also examined the ratio of the diffusion length as given by transport theory to that given by diffusion theory itself. This is extremely insensitive to frequency, but for substantial absorption, using the diffusion theory result can lead to substantial errors in thick media, especially for Henyey-Greenstein scattering.

Theoretical and computer modeling using Mie theory, radiative transfer theory, diffusion wave correlation spectroscopy, Monte Carlo simulation technique were applied for human sclera optics analysis in a process of its optical enlightening caused by osmolitically active chemicals administration.

NIR spectroscopy and imaging are having an increasing interest as not invasive diagnostics, for the evolution experienced in the last years by both instrumentation technology and interpretation models. In the frequency domain, the informative content of the diffusive photons waves is primarily limited by the approximation of the reconstruction algorithm, according to the measurement precision of the whole diagnostic apparatus. In this work, the influence of optical fiber delivery and collection systems has been studied, to determine signal amplitudes and phase errors, according to refractive effects at the coupling areas, and to propagation effects as intermodal dispersion induced by bending and other mechanical instabilities of the fibers.

In this paper, we report our initial experience with multi- target imaging experiments. Our intent mimics that of our initial pass at single-target imaging; that is, to vary selected parameters over a modest range in order to demonstrate feasibility and quickly identify potential areas where performance degrades relative to single-target imaging studies. This work can then be supplemented with a more exhaustive and directed examination of certain imaging parameters in order to more fully identify the limitations of multi-target imaging in the laboratory. In order for this paper to be reasonably self-contained, we briefly and qualitatively review of our image reconstruction algorithm and our experimental procedures and parameters, the details of which have been discussed at length elsewhere.

An extended series of transillumination experiments has been performed in vitro on animal samples (bovine muscle, up to 30- mm-thick; chicken wing and quail femur, 12-mm-thick) and in vivo on the human hand (thickness, about 20 mm), using a pulsed light source (7 ns, about 10^-4 J/pulse, 10 Hz rep rate) from a collimated (1.2 m) Nd:YAG laser beam (1064 nm). A PIN photodiode connected to a digital oscilloscope was used to measure the maximum intensity of the beam pulse transmitted through the sample (i.e., no temporal discrimination of the output signal was attempted) while it was scanned across the source/detector assembly. One dimensional scans were performed on bovine muscle samples in which thin metallic test objects were embedded, in order to study the spatial resolution of the technique (for bovine muscle at 1064 nm, absorption and reduced scattering coefficients are reported to be about 1 cm^-1 and 3 cm^-1, respectively). The measured spatial resolution was as good as 3.6 mm in 30 mm of tissue thickness. In the two-dimensional scans of the chicken and quail sample, fat and bone tissues can be easily seen with good resolution, whereas imaging of the middle finger of a human hand shows cartilaginoid and bone tissue with 1 - 2 mm resolution. Hence, this simple collimated quasi-cw technique gives significantly better results for tissue imaging than pure cw transillumination. Use of (pulsed) light above 1000 nm and a high energy content per pulse are supposed to explain the positive experimental findings.

A new method of dental introscopy in-vitro is suggested by the authors. This method implies the usage of longitudinal tomography techniques and is characterized by non-invasive and non-harmful diagnostics features, as well as interactive regime of image reconstruction which lets an operator (doctor) to control the diagnostics process in real time. He-Ne laser emission is used for obtaining of the projections. By the means of longitudinal tomography, images of different sections of an object (tooth) can be reconstructed. An experiment was held by the authors in which 100 projections of a tooth (premolar) were obtained and images of 10 different sections were reconstructed. These images were later compared to real sections of the tooth. This experiment proved that optical longitudinal tomography can be successfully used for dental introscopy. Authors claim that optical tomographic methods can be used for diagnostics of other biological objects as well. Such objects are characterized by spatial geometrical anisotropy (tubular bones, phalanxes of fingers, penis, etc.). It is especially promising to use this method for children's dentistry. the authors discuss some features of the data acquisition system for optical longitudinal tomography. Reconstruction algorithms are described. The results of experimental reconstruction are presented and advantages of this diagnostics method are discussed.

A prototype system based on intensity-modulation spectroscopy (IMS) was produced with the goal of developing 'optoencephalography' as a new instrument for clinical application and for investigating human brain functions. This system can use dual wavelengths (787 and 827 nm) to simultaneously measure reflectances at 8 measurement positions on the human head. Using the system, we measured the changes in blood circulation and oxygenation changes caused by epileptic seizures and specific brain functions. The former measurements were made simultaneously with tests to determine the epileptic focus by using single-photon-emission computed tomography (SPECT) and electrodes set in the brian. Four measurement positions were fixed in each temporal region. The areas where cerebral blood flow increased, as observed by SPECT, corresponded to the positions where the regional cerebral blood volume (rCBV) increased, as measured by the IMS system. Furthermore, the timing of the epileptic seizures, as measured by the depth-electrodes, corresponded to the timing of the increase in rCBV measured by the prototype system. Our measurements of changes in blood circulation as a result of brain functions were made for motor functions to compare the differences between the right and left hemisphere in how they respond to specific functions. Four measurement positions were set in bilateral motor areas. Significant differences in blood circulation in connection with brain activities were observed between the right and left hemispheres.

In the diffusion equation approximation, the photon mean trajectory (PMT) and photon path deviation were calculated in relation to scattering medium optical characteristics and methods of measurements (time- or frequency-domain). It was shown that in macroinhomogeneity detection there exists an optimum combination between the spatial resolution (which is defined by photon path standard deviations from PMT) and the value of signal perturbation induced by inhomogeneities. One can evaluate this optimum by the scale matching of photon path rms deviations and macroinhomogeneity size. In turn, such a matching can be realized by the proper choice of the source and detector mutual positions and the values of temporal delay in time-domain or modulation frequency in frequency-domain measurements.

Near infrared spectroscopy, based on frequency domain methods, has been successfully applied to study turbid media, using amplitude and phase measurements. Oximetry based on phase measurements only may prove advantageous over amplitude phase measurements due to relative simplicity of use and calibration of systems with phase output only. Suggestions for hemoglobin saturation calculation, based on the diffusion approximation and phase information only, have been proposed in the past. There exist many factors though that may introduce inaccuracies in the data analysis and in the measurements, besides the inherent inability of such systems to account for absorption of water and chromophores besides oxy- and deoxy- hemoglobin. Such factors, as theoretical approximations, voltage to phase conversion and initial phase calibration, have to be carefully considered in order to produce consistent and reliable calculations. In the present work, simulation models, based on the solution of the diffusion equation in the frequency domain, have been developed, to study the performance of different approximations used for the calculation of hemoglobin saturation. A new approach for data analysis is presented and investigated. Methods for instrument calibration are discussed and an example, based on experimental measurements, is demonstrated.

A comparison among different techniques used for breast imaging has been carried out by using a Monte Carlo (MC) code to simulate confocal scanning on a diffusing slab containing absorbing inhomogeneities. The MC code evaluated the temporal point spread function (TPSF) for many positions of the source- receiver system with respect to the inhomogeneity. Each TPSF was fitted using the diffusion equation solution for a homogeneous slab. The information contained in each TPSF was summarized in three parameters: the absorption coefficient, the reduced scattering coefficient and the amplitude factor. An image was obtained plotting each of these parameters. To simulate the time gating technique the analytical functions obtained from the fit were used to evaluate the energy received within short gating times. To simulate the frequency domain technique the Fourier transform of the analytical function was calculated. The results obtained from simulations showed that a good estimation of the dimension of an absorbing inhomogeneity can be retrieved from the images obtained either from continuous wave domain, or time domain, or frequency domain measurements. The better contrast and the independence of the image quality on absorption properties of the diffusing medium indicates that the time gating technique is the one producing the best image quality.

A gain modulating framing camera and its application towards the study of real time cellular phenomena is described. Based on a unique operating principle, this framing camera can be modulated by over 90% at 1 GHz. The camera consists of an image converter with a pair of deflection electrodes and a rectangular aperture. Since a sinusoidal electric field is applied to the deflection electrodes, the photoelectron image- forming beam is continuously deflected and swept on the aperture. A bias is applied to center the sweep of the photoelectron beam on the edge of the aperture. The gain modulating with high depth can hence be accomplished. We are now constructing a fluorescence lifetime imaging microscope system employing this gain modulating frame camera based on the phase domain method. Such high depth modulation enables us to achieve frequency signals as low as 1 Hz in heterodyne operation. We describe examples of application of the system towards the observation of various cellular phenomena.

We present two cw methods for localizing a source of fluorescence buried in a medium with optical properties similar to those of tissue in the near infrared region. The first approach is based on the fact that, for small excitation beam diameters, the absolute intensity at a given depth in the medium depends on the diameter of the incident beam. For a well-chosen pair of beam diameters, the ratio of these intensities in a scattering medium depends uniquely on the depth from the surface of incidence. Thus, the ratio of the fluorescence resulting from sequential excitation using two beam diameters can be used to determine the depth at which the fluorescence originated. The second method is based on spatially resolved surface measurements of the diffuse fluorescence from the buried source. Using a form of the diffusion theory analysis of Farrell et al. (Med. Phys., 1992) for the spatially resolved diffuse reflectance from a pencil beam incident on a scattering medium, it is possible to reconstruct the depth of the source from the shape of the surface fluorescence profile. Preliminary experimental results obtained using a 1.0 cm diameter sphere containing the tumor localizing fluorophore Nile Blue A show that the spatially resolved measurement reports the location of fluorescent sources as deep as 4.0 cm with an accuracy of 0.4 cm or better.

The impact of background lumiphore in luminescence optical tomography is examined. To demonstrate its effects, numerical simulations were performed to calculate the diffusion-regime limiting form of forward-problem solutions for a specific test medium. Image reconstructions were performed using a CGD algorithm with a rescaling technique and positivity constraints. In addition, we develop a modification to the basic algorithm that makes use of the maximum possible concentration in order to estimate the background concentration, and show that it improves image quality when background lumiphore is present. We conclude that the usual measure of background lumiphore's effect, which is the target- to-background lumiphore concentration ratio, is not adequate to define the contribution from the background lumiphore. The reason for this is that image quality is also a function of target size and location. An alternative measure that we find superior is described. The results indicate that the improved algorithm yields better image quality for low target-to- background ratios.

Many applications of native fluorescence spectroscopy of intrinsic biomolecules such as Try, Tyr, Phe, NADH and FAD are reported on both the characterization and the discrimination of malignant tissues from the normal. In the field of diagnostic oncology, extensive studies have been made to distinguish the normal from malignant condition in breast, cervix, colon and bronchus. From the studies made by Alfano and co-workers, it was found that the emission at 340 and 440 nm under UV excitation have shown statistically significant difference between normal and malignant tissues. As tissues are highly complex in nature, it is worth to known whether the changes arise from cells or from other extracellular tissue components, so as to enable us to have better understanding on the transformation mechanism of normal into malignant and to go for an improved approach in the effective optical diagnosis. In this context, the present study addresses the question of whether there are differences in the native cellular fluorescence characteristics between normal and malignant epithelial cells from human larynx. With this aim, the UV fluorescence emission spectra in the wavelength region of excitation between 270 - 310 nm and the excitation spectra for 340 nm emission were measured and analyzed. In order to quantify the altered fluorescence signal between the normal and malignant cells, different ratio parameters were introduced.

In recent years researchers have made significant progress in understanding the physics of fluorescence in highly scattering materials such as tissues in the near-infrared. We have quantitatively verified a model which describes fluorescence in ideal (homogeneous and infinite) tissue-like media. Given the quantitative accuracy of this model, one can use measurements of the fluorescence of a tissue with a homogeneous distribution of fluorophore to obtain the quantum yield, lifetime of the probe, and the absorption and scattering coefficients of the tissue at the fluorescent wavelength. We demonstrate that this can be done with a simple measurement of the photon density as a function of source- detector separation at the excitation and emission wavelengths. To verify our approach we present the lifetime, quantum yield of the fluorescent probe (rhodamine B), and the absorption and scattering coefficients of the medium at the emission peak wavelength that are obtained by a fit of the model to experimental measurements.

We have developed a 64-channel time-resoled spectroscopy (TRS) system based on a time-correlated single photon counting (TCPC) method to achieve near infrared spectroscopy and/or optical computed tomography (CT) for clinical applications. This system employs advanced devices such as a high power picosecond light pulser (PLP), a high sensitive/fast photomultiplier tube (PMT) and a high speed signal processing circuit. The PLP offers an average optical power of around 0.25 mW, and the PMT has a quantum efficiency greater than 2% at 800 nm. The signal processing circuit is composed of miniaturized CFD/TAC modules and signal acquisition unit with 1 MHz ADC, and provides 64 independent TCPC circuits for time- of-flight measurement. System performance was estimated by measuring transmitted and reflected light passing through tissue-like phantom models simulating the human breast and infant head. We concluded that this system has the sufficient performance for optical CT utilizing time-of-flight measurement.

Human newborns can suffer from neuro-developmental abnormalities, when they are born as preterms. With near infrared spectroscopy (NIRS) it is possible to investigate any brain disease occurring together with these neuro- abnormalities. The specific absorption properties of haemoglobin and oxygenated haemoglobin in the near infrared region allow to measure the oxygenation status and several other variables. Local variations in cerebral blood volume (CBV) and blood oxygenation is important for a better understanding of these abnormalities.

The purpose of this study is to validate a new analysis technique to provide continuous data for cerebral blood flow (CBF) with near infrared spectrophotometer (NIRS, NIRO-500 Hamamatsu Photonics, Japan). In our mathematical model, we obtain two differential equations by considering the in-out balance of the cerebral tissue hemoglobin concentrations in the arterial and venous side. Solving these two equations, we finally obtain the blood flow into the arterial side and out of the venous side separately. To verify the mathematical model, we compared the results of our model analysis with the flow velocity (FV) at the middle cerebral artery, which was measured by transcranial Doppler, in eight volunteers who were exposed to 60 degree head-up tilt. In our results, CBF measured by NIRS was highly correlated to the FV (y equals 1.0003 x + 0.0583, R equals 0.920), where x and y are %flow velocity measured by transcranial Doppler and %CBF measured by NIRS, respectively. Our method provides continuous changes in the cerebral hemodynamics noninvasively in dynamic conditions. It is not necessary to assume that hemodynamics is static during measurements. This method is of use in monitoring the cerebral hemodynamics noninvasively and continuously.

We propose a new algorithm for optical computed tomography (CT) to quantify the absorptive substances in highly scattering media such as human tissues. Our algorithm uses the uniform medium in which scattering and absorption coefficients are equivalent to the average coefficients of the actual object to be measured. In other words, we use an imaginary reference. When the weight function and re-emissions are measurable or can be calculated for the imaginary reference, we can describe the inside structure of the non-uniform object using the deviation of the absorption coefficient from the average value. Since it is difficult to prepare a phantom of which exterior shape and inside structure are identical to that of the real objects, conventional methods cannot be free from the significant errors. The averages of the absorption and scattering coefficients are obtained by measurements such as time-resolved spectroscopy. The weight function and re- missions can be calculated using the average values by a Monte Carlo simulation or a finite difference method. The absolute absorption coefficient is obtained as the sum of the average and the deviation. The validity of our algorithm was confirmed by measuring a tissue-like phantom which contains three different absorbers. We evaluated reconstructed images and confirmed that the new method gives better accuracy in the quantitation of the concentration of absorbing substance and a smaller image distortion. Our results have significant implications for optical CT which quantifies the concentration of absorbing substances accurately without measuring the reference phantom.

We present a re-engineered frequency-domain tissue oximeter operating in the near-infrared spectral region. This instrument is based on the multi-distance measurement protocol, which we have implemented in our original design by multiplexing multiple light sources. The new instrument uses intensity modulated (110 MHz) laser diodes emitting at 750 and 840 nm. The laser diodes are coupled to glass optical fibers (600 micrometer core diameter). The average light intensity delivered to the tissue is about 3 mW. The multiplexing electronics are based on solid state switches that allow for acquisition times per point as short as tens of milliseconds. Our tests on phantoms and in vivo with the new oximeter have shown significant improvement in terms of stability, reliability, and reproducibility with respect to the original prototype. Furthermore, by using optical fibers we achieve a high versatility in the design of the measuring probe, permitting custom design for various tissue contours and different measurements. To verify the improved performance of the new oximeter, we have performed an in vivo test consisting of monitoring the hemoglobin saturation (Y) and concentration (THC) on the calf of 18 healthy volunteers during walking and running routines.

In the present study we evaluated a new 50 MHz single wavelength multi-source detector imaging system for noninvasive optical imaging of human brain function. The system is based on phase resolved technology and allows measurements in remission mode of changes in phase shift and light intensity in the phased array approach and the conventional phase modulation approach. To evaluate the potentials and limitations of the system with respect to functional brain imaging we assessed the sensitivity of imaging based on photon migration pattern and absorption changes in the two different approaches as a function of object depth in a homogeneous brian model under controlled conditions. Furthermore first results using this technique for noninvasive optical imaging of functional brain activation in human subjects are presented and discussed with respect to potentials and limitations of the respective parameter.

The noninvasive NIR detection technique can be used to detect the clinic-related parameter such as hemoglobin concentration and saturation. By now, many techniques have been demonstrated in both time-domain and frequency domain to obtain quantitative optical properties. In this paper, we demonstrate one kind of simple homodyne phase modulation system -- in- phase and quadrature phase (I&O) detection system, introducing its principle and constitution, presenting its performance through the tests of ink, intralipid and blood model.

The recent improvement in medical devices that are safe, economical and efficacious has led to significant interests in the near infrared (NIR) optical characteristics of strongly scattering medium such as human tissues, particularly breast tissue, brian and skeletal muscles. Since the frequency domain equipment involves lower peak powers, slower rise times and hence smaller bandwidths than the time domain circuits, they appear to be more economic and portable as medical device. We have developed a 50 MHz time-sharing 3-wavelength (754 nm, 790 nm, 830 nm) single sideband (SSB) phase modulation system, which measures the essential characteristics of light propagation in strongly scattering medium. Some new techniques, such as phase locked loop (PLL), dynode feedback are used to get high accuracy (better than 0.05 degrees in a 1 Hz bandwidth), high sensitivity (6 cm separation on forehead), low noise (less than 0.05 degrees), low drift (less than 0.01 degrees/hr within 11 hours) and low phase-amplitude cross-talk (less than plus or minus 0.038 degrees/dB). A blood model test has been given. Significant results of hemoglobin oxy/deoxygenation measurements re shown in this paper.

Near-infrared spectroscopy (NIRS) can detect changes in cerebral hemoglobin oxygenation in response to motor, visual or cognitive stimulation. This study explored potential improvements for functional human brain mapping with NIRS: (1) So far, only primary cortical areas, like motor cortex or primary visual areas were studied. We tested the feasibility of identifying an extrastriate visual motion area (MT) with single site NIRS. (2) The temporal resolution of commercial systems is on the order of seconds and their spectral resolution is poor. We tested the feasibility of the detection of cerebral hemoglobin oxygenation changes during visual stimulation at high temporal (100 ms) and spectral resolution (5 nm) using a whole spectrum approach (CCD-NIRS). (3) The spatial resolution of commercial systems is poor. In this study we used a 16 channel functional NIRS-imaging device to test the feasibility of mapping changes in cortical blood volume during visual stimulation (over primary and secondary areas). We show that (1) even conventional single site NIRS allows to identify secondary visual areas, (2) a CCD-NIRS system affords a high temporal (100 ms) and spectral (5 nm) resolution for the detection of changes in cerebral hemoglobin oxygenation during visual stimulation, (3) functional NIRS- imaging can localize focal blood volume changes over both primary and secondary cortical areas.