The effect of the layered surface tissues of the head on the measurement of brain oxygenation by near infrared spectroscopy (NIRS) has been investigated by both time-of-flight measurement and Monte Carlo simulation on models consisting of three or four separate layered homogeneous media. The clear cerebrospinal fluid (CSF) layer surrounding the brain has previously been shown to significantly affect light distribution, and in the head the brain surface is deeply folded with many CSF filled sulci. Therefore the most sophisticated model has four layers including a clear 'CSF' layer containing slots imitating sulci on the brain. Simpler models are also used and the light distributions in each are compared to examine the effect of the CSF layer. The mean optical pathlength of each model calculated from the temporal point spread function (TPSF) of the time-of-flight measurement agrees well with the Monte Carlo prediction. The fractional pathlength in each of the layers and the spatial sensitivity profile, which indicates the volume of tissue interrogated, are calculated by the Monte Carlo method. Results show that the light distribution in the head is highly affected by the existence of the clear CSF layer, and the optical pathlength and spatial sensitivity profile in the models with a clear layer are quite different from those without. However the presence of the sulci hardly affect the light distribution, the results from the sophisticated brain model with sulci being almost the same as those of the model with a simple CSF layer.

The propagation of light in biological tissues has been studied theoretically by Monte Carlo simulations and experimentally in vitro on bovine and porcine liver, human breast tissues and fat emulsions. The spatial intensity profile of a scattered collimated beam is measured at the output surface of the sample and compared to theoretical predictions. Measurements on fat emulsions and liver are in good agreement with theory. Absorption and reduced scattering coefficients can be obtained by finding a simulated profile matching our experiments. The precision of these values has been confirmed by the comparison between simulations and experiments on the temporal spread of light pulses. On the contrary the width of the spatial intensity profiles measured on breast tissues are systematically too large to be predicted by our simulations. The heterogeneity of the sample, not considered in simulations, could explain these differences.

We present a method that uses a single Monte Carlo simulation (modeling the light distribution in a turbid media) to obtain a set of results for different optical coefficients. This approach eliminates the repetition of simulations for each different parameter. It is based on an adequate formulation of the radiance, expressed as a function of two probability distributions. The first depends only on the phase function and the geometry, while the second one depends only on the optical coefficients. This procedure requires the storage of the number of collisions and the total pathlength of each photon. An example of the applicability of this method is shown, for which an important spare of computer time has been achieved. The main drawback of the method is a loss of convergence which limits its applicability. This effect is discussed as a function of the optical coefficients.

When performing direct-contact laser-Doppler flowmetry on experimental flow models, the power spectra of the detector signal can be obtained by homodyne or by heterodyne detection. Especially with uniformly moving probe particles coherence effects are observed, leading to changes in the width of the power spectrum. Due to destructive interference effects at the detector surface, in the measured homodyne spectra the contribution of relatively high Doppler frequencies is suppressed compared with that of lower frequencies. This results in spectra which are narrower than expected theoretically. This effect allows us to investigate the importance of the relative amount of coherence areas at the surface of the detector.

The specific intensity arising from an anisotropically scattering illuminated cylinder is calculated using the equation of radiative transfer. The solution is obtained in closed analytical form. A new simple liner integral equation is derived for the values of the specific intensity at the boundary of the cylinder. The novel method used for the solution of this problem leads immediately in a straightforward and systematical way to the appropriate basic equations for the problem at hand. This method applies equally well to similar scattering problems with other geometries.

The time-dependent equation of the radiative transport is reduced to the stationary one for the case of a radiation source being modulated by a harmonic frequency. A Monte Carlo scheme is suggested to solve the resulting equation. The technique avoids tracking the time-histories of each individual photon and allows us to take the finite single-scattering transient time into account. The algorithm directly estimates the quantities being relevant to frequency-domain measurements. A single set of photon trajectories can be used to compute the modulation and the phase of the scattered radiation at different modulation frequencies. The results of the Monte Carlo simulations are compared to predictions of the rigorous radiative transport theory and the diffusion approximation. It is found that the Monte Carlo technique provides a good agreement with the transport theory whereas the accuracy of the diffusion approximation decreases with growth of the modulation frequency. In addition, the technique is used to study the effect of the finite single-scattering transient time on the resulting modulation and phase distribution of the diffusely reflected radiation. It is shown that even a transient time as short at 0.1 ps can significantly affect the reflected signal from a medium presenting optical properties similar to those of biological tissues in the near-infrared spectral range.

We investigate whether a combination of steady-state diffusion theory with a distributed set of single scatter sources may describe the diffuse reflectance of light from tissue. We have compared the predictions of the single interaction model with the numerical results obtained by an exponentially weighted scatter source model and we have found a quite good agreement of the latter with Monte Carlo simulations over a wide range of the transport albedo.

A collimated light beam propagating through a turbid medium can be strongly attenuated by both absorption and scattering. It is usually assumed that, in first approximation, the attenuation depends linearly on the thickness of the scattering medium and on the scatterers concentration. This law, based on the assumption that the scatterers are uncorrelated, is in good agreement with experimental measurements when the particles density is low. However, in a dense distribution of particles, it has been shown that corrections need to be introduced. In this paper, we relate some experimental results concerning coherent propagation in latex spheres suspensions in water. These results are obtained by the use of polarization analysis and show the deviation of the attenuation from the linear law in spheres concentration. A good agreement is obtained with the Keller propagation theory in dense media.

In this paper on the basis of analytical and numerical consideration we have taken into account the actual photons avalanche front. It is shown that image resolution is decreased slower than it follows from DA and runs not as DA predicted (root)t but as (root)t - d/v, where d is the thickness of the layer and v-speed of light in the scattering medium, t-time passes through the moment of light pulse incident on the surface of the medium.

The basic principles of light scattering and focused beam diffraction by random and organized tissues in the framework of laser diagnostics and disease monitoring technologies design are considered.

We have developed diffusely scattering solid phantoms with optical (scattering) properties amenable to theoretical calculations. Monodisperse quartz glass spheres were used as scatterers embedded in polyester resin. An infrared dye was added to simulate absorption by biological tissue. Solid phantoms were tested for their macroscopic homogeneity. Several phantoms were built with well-defined spatial variations in their transport scattering and absorption coefficients to be used for optical tomography. Scattering, transport scattering, and absorption coefficients of solid, homogeneous phantoms and of aqueous suspensions of monodisperse quartz glass spheres were derived from measurements of time-integrated collimated transmittance and time-resolved diffuse transmittance. For aqueous suspensions of monodisperse quartz glass spheres at known number density scattering and transport scattering coefficients calculated by Mie theory are in quantitative agreement with experimentally derived values. In addition, diffuse reflectance and diffuse transmittance of aqueous suspensions at various number densities were measured and found to be in excellent agreement with results of Monte Carlo calculations using theoretical values for the scattering coefficients and anisotropy parameters.

Time-resolved measurement and FEM (finite element method) calculation were carried out for solid phantoms. Two cylindrical solid phantoms were made of epoxy resin containing titanium oxide particles as scatterer. Light source was sub-picosecond light pulse in the near infrared wavelength range, and a detector was a streak scope with 10 picosecond time resolution. One phantom was homogeneous both in scattering and absorption, and the other included two inhomogeneities with higher absorption than that of the background. Time-resolved transmittances measured at various detection angles were compared with the FEM results of the three dimensional diffusion equation. The agreement between the measured and calculated results was very good qualitatively and fairly good quantitatively. Both the measured and calculated results for the inhomogeneous case showed the reasonable deviation from the homogeneous case, and it has been shown that the existence of the inhomogeneities is measurable.

Light transmitted by thin slabs of human dental enamel is scattered and Fraunhofer diffracted. At three wavelengths we used the two dimensional transmitted Fraunhofer diffraction patterns to determine the periodicity of the prism structure and the curvature and misalignment of the prisms. This was performed by fitting theoretical curves to measured intensity functions. We found for each sample a mean value of the periodicity and its standard deviation. The deviation between the mean values found at each wavelength was much smaller than the deviation of the size distribution in each sample. The measured overall mean value of the periodicity and its standard deviation was d equals 4.9 plus or minus 0.6 micrometer. The mean angle of curvature and misalignment was (Delta) (phi) equals 32 degrees plus or minus 22 degrees.

This paper presents in-vitro-studies using the scattered intensity distribution obtained by cw- transillumination to examine the condition of rheumatic disorders of interphalangeal joints. Inflammation of joints, due to rheumatic diseases, leads to changes in the synovial membrane, synovia composition and content, and anatomic geometrical variations. Measurements have shown that these rheumatic induced inflammation processes result in a variation in optical properties of joint systems. With a scanning system the interphalangeal joint is transilluminated with diode lasers (670 nm, 905 nm) perpendicular to the joint cavity. The detection of the entire distribution of the transmitted radiation intensity was performed with a CCD camera. As a function of the structure and optical properties of the transilluminated volume we achieved distributions of scattered radiation which show characteristic variations in intensity and shape. Using signal and image processing procedures we evaluated the measured scattered distributions regarding their information weight, shape and scale features. Mathematical methods were used to find classification criteria to determine variations of the joint condition.

In order to quantify the chromophore components from in vivo blood NIRS, a blood- equivalent phantom has been developed which consists of properly diluted intralipid and ICG dye. The reflection and transmission near infrared spectroscopy (NIRS) of the phantoms with different scattering backgrounds and ICG concentrations are measured and analyzed by the spectral multicomponent analysis (MCA) method to extract ICG concentration. The experimental results show that the MCA method can be used to quantify absolute ICG concentrations in scattering media if the average path lengths are known. Moreover, it was found by the experiments that both the water absorption peak at 970 nm and the ICG absorption peak at 800 nm show similar behavior during the change of the scattering background. Thus the ratio of the MCA-estimated concentration factor of ICG to water is independent of the blood-phantom scattering.

The availability of oxygen in the human vastus medialis muscle and the tympanic, skin forehead, quadriceps, and rectal temperatures has been investigated during exercise test and post-exercise with non-invasive near-infrared spectroscopy, infrared thermometer, and an array of four thermistors, respectively. During exercise time rectal temperature was not recorded, before exercise basal values were obtained, and after exercise all the data for two hours were recorded. The signals from near-infrared spectroscopy have been studied by analogy to forced vibration and viscously damped free vibration. Other models have been used to evaluate the temperatures. The time necessary for the reoxygenation signal to cross the baseline during the post exercise period was from 30 min to over 100 min. The peak of pH values was 5 min post-exercise and to arrive at basal levels needed 25 min to more than 40 min. The peak of rectal temperature starts around 20 - 30 min post-exercise remaining 25 - 40 min at the same value, starting to slip down slowly at variable intervals of several minutes requiring over two hours to arrive at basal levels. The data obtained by near-infrared spectroscopy and skin quadriceps, rectal temperatures confirm that the oxygen consumption remains after exercise in the muscle studied.

The absorption and scattering properties of whole blood suspensions are studied by simultaneous polarimetric and imaging measurement. Blood samples contained in a cylindrical flow through cuvette are illuminated with linearly polarized light (670 nm) axially and with diffuse white light medially. A high resolution line scan camera is used to obtain the transmission image through a varying pathlength chamber. Suspensions of crenated red blood cells are examined at different concentrations. Additional measurements were performed by standard blood test techniques to provide a reference for the evaluation of experimental results.

Multispectral studies of light propagation in female breast tissue have been performed. Short pulses of white light were generated by using self-phase modulation of a high-power laser pulse focused into a cuvette filled with water. The white light pulses illuminated the tissue and the scattered light was recorded with time- and wavelength dispersion by a streak camera. Measurements were performed on breast mastectomies in vitro and measurements on healthy breast tissue in vivo. The reduced scattering coefficient and the absorption coefficient of breast tissue were obtained in different wavelength regions by fitting solutions of the diffusion equation to the experimental data. Significant variations in the magnitude of the optical properties could be seen between the different individuals. No characteristic spectral discrepancy for tumor tissue was found.

Time gating for laser imaging and tomography of biological tissues is a rapidly developing and promising approach to biomedical diagnostics based on laser probing. There are good reasons to hope that in future the tomography utilizing IR radiation will become competitive with traditional x-ray tomography due to its safety and potentially high resolution. There is also growing interest in retrieving information about functional characteristics of tissues rather than just stationary imaging. In the present paper we consider the results of Monte-Carlo simulations and discuss the opportunities of time-gated laser Doppler flowmetry as a technique providing information about moving components of biotissues. Our main concern is with possible improvements in spectral and spatial resolution characteristics of the laser Doppler flowmeters operating in back-scattered and transmitted light modes.

A general approach for the investigation of the anisotropy properties of scattering and turbid mediums is presented. These properties are given by seven independent parameters characterizing the depolarization degree, amplitude and phase anisotropy of investigated mediums. The approach is based on the statement of basis presentation of anisotropy properties of medium and on theorem of decomposition. This allows us to obtain full information on anisotropy properties of the medium which cannot be obtained by any other known sensing methods and to perform a general classification system of mediums. Such a classification system enables us to find the regularities and to conduct compared analysis of anisotropy properties of different nature mediums.

Intensity distribution of light scattered by a layer of the eye lens homogenate at temperatures around the characteristic temperature for the phase separation is experimentally analyzed. The scattered light produces a speckle pattern. The measurements are performed in the diffraction field. The parameters of the scattered intensity are shown as a function of the temperature. The intensity of the specular component dependent on the temperature is shown. The experimental results suggest the existence of three kinds of scatterers in the eye lens homogenate. The correlation length of the scatterers is estimated.

The application of probing beam space-angle intensity moments for investigation of optically inhomogeneous biological objects is considered. Special attention is paid to measurements of lens-like inhomogeneities in the presence of scattering caused by small-scale fluctuations of refraction. The results can be used in ophthalmology for non-invasive early detection of cataractous changes in human eye lens and in laboratory experiments with blood fluxes.

Antigoat IgG raised in rabbit has been used to study its aggregation with goat antigen and the cluster formation is studied using laser light scattering. A (chi) ² fitting method is used to fit the experimentally determined scattered intensity with the theoretically calculated scattered intensity. Scattered intensity is theoretically computed using a radial distribution function of the r^D-3 form. The static structure factor S(q,R_g), the radius of gyration (R_g), and the correlation function ((xi) ) are determined.

We have recorded time-resolved transillumination images of solid phantoms with objects embedded differing in their scattering and absorption coefficients from those of the bulk material. The optical properties of the phantoms were chosen similar to those of human breast tissue. Several quantities such as fractional transmittance and first moments as well as Fourier amplitudes and phase shifts have been derived from measured distributions of times of flight to form two-dimensional images. We discuss contrast and effective signal to noise ratio of such images. Assuming homogeneous optical properties, from an analysis of the Fourier transforms of the measured distributions of times of flight we have obtained effective (frequency dependent) transport scattering and absorption coefficients as a function of scanning position. The coefficients deduced in this way are in qualitative agreement with the known optical properties of the bulk of the phantoms and the embedded objects.

Using photon-density-waves (PDW), (multiple) objects embedded in a highly scattering medium with optical properties similar to tissue are detected. The measurements were performed with near-infra-red laser light at 675 nm which was either kept continuous-wave (cw) or amplitude modulated (AM) at 219 MHz or 650 MHz. We find that the spatial resolution of the projection image shows only slight improvements as the frequency is increased. This improvement comes at the expense of signal strength in the modulated part of the light. That is, the PDW shows a much stronger attenuation as compared to the cw light intensity. The implications of the lower signal-to-noise ratio at high modulation frequencies is that the modulated light projections are less suitable for further data processing. For a given data acquisition system, this fact cancels the advantages of the higher raw resolution of PDW as compared to cw light. Therefore, we find no clear advantage of PDW over cw light for obtaining sharper tomographic images of the diffuse media, regardless of the inverse scattering method used.

A possible approach to optical mammography consists of a planar tandem scan of a light source and an optical detector in a transmission geometry. The main difficulty with interpreting the data collected in this sampling geometry arises from edge effects, which are due to variable breast thickness within the scanned region and to photon losses through the sides of the breast. We have developed an algorithm for frequency-domain data analysis that corrects for edge effects, thus providing significantly enhanced image contrast with respect to raw data images. The application of our algorithm to data collected in vivo strongly improves the detectability of optical inhomogeneities in the breast.

Continuous-wave ultrasonic modulation of scattered laser light has been used to image objects in tissue-simulating turbid media for the first time. We hypothesize that the ultrasound wave focused into the turbid media modulates the laser light passing through the ultrasonic focal spot. The modulated laser light collected by a photomultiplier tube reflects the local mechanical and optical properties in the focal zone. Buried objects in 5-cm thick tissue phantoms are located with millimeter resolution by scanning and detecting alterations of the ultrasound-modulated optical signal. Ultrasound-modulated optical tomography separates the conflict between signal and resolution in purely optical imaging of tissue and does not rely on ballistic or quasi-ballistic photons but on the abundant diffuse photons. The imaging resolution is determined by the focused ultrasonic wave. This technique has the potential to provide a noninvasive, nonionizing, inexpensive diagnostic tool for diseases such as breast cancer.

This paper presents an ongoing investigation of laser induced xenofluorescence using a fluorescent marker to detect objects in turbid media. The aim is the development and validation of a method for imaging vessels using near infrared fluorescence angiography techniques in strongly scattering media. The main purpose is to show the course of blood and lymph vessels in the head and neck region of tumor patients. The first step was the selection of the best method and a suitable fluorescence dye to image these vessels. Using a phantom with the optical parameters similar to skin and fat tissue different methods of fluorescence excitation and detection were investigated. The following experiments were performed: whole area two-dimensional excitation and detection (CCD camera) as well as a focused excitation with scanned detection (photo diode). Furthermore investigations of the improvement of the vessel detection using a phase demodulation technique were performed. The measurements were simultaneously accompanied by analytic calculations using diffusion approximation. The photophysical investigations of several dyes have led to the use of tsAlClPc as fluorescence dye for the technical comparison of the investigated methods. Since this phthalocyanine is phototoxic a clinical certified dye (probably Indocyanine green) will be used for future experiments. The methods with focused excitation and scanned detection have shown the best resolution, but they are slow and expensive in comparison to the CCD camera technique. Further experiments should give more information about the application dependent decision which method is best suited.

This paper presents measurement results of absorption and scattering in tissues and tissuelike phantoms by using coherent detection imaging (CDI) technique. CDI measurements are carried out in 0.8 - 1.3 micrometer wavelength region by using several continuous wave, single frequency lasers, including a tunable Ti:sapphire laser and laser diode-pumped Nd:YAG lasers. A 120 dB dynamic range achieved with less than 10 mW incident power and a detection time of 0.1 msec enables spectroscopic measurements through attenuation of as much as exp(-27). Measurement results using cw lasers are also compared to the temporal scattering profiles of coherent photon migration in highly scattering media. For this purpose, a modified CDI system using low-coherence super luminescent diodes has been developed. The results from the laser and low coherence CDI measurements are found to be consistent.

The imaging resolution in turbid media is severely degraded by light scattering. Resolution can be improved by extracting the unscattered or weakly scattered light. In this paper the state of polarization of the emerging light is used to discriminate photon pathlength, the more weakly scattered photons maintaining their original polarization state. It is experimentally demonstrated that over a wide range of scatterer concentrations, different particle sizes possess different characteristics. Three distinct regimes are described in detail along with the techniques to improve resolution within these regimes.

The purpose of this research is to try to investigate and improve the possibilities of optical coherence tomography for evaluation and imaging of melanoma. As approaches to reach this aim by photo diode detection were not successful, we suggest making an experimental setup with a slow-scan CCD-camera that is able to detect very little numbers of photons. The interferometric arrangement had to be adapted to the quite bad signal-to-noise ratio of the camera. First measurements were made on models consisting of very small glass spheres embedded in polyester resin. It was possible to prove coherent photons from a depth of 2 mm. Objects could be found up to 1 mm depth.

Locally and individually varying optical tissue parameters (mu) _a, (mu) _s, and g are responsible for non-neglectible uncertainties in the interpretation of spectroscopic data in optical biopsy techniques. The intrinsic fluorescence signal for instance doesn't depend only on the fluorophore concentration but also on the amount of other background absorbers and on alterations of scattering properties. Therefore neither a correct relative nor an absolute mapping of the lateral fluorophore concentration can be derived from the intrinsic fluorescence signal alone. Using MC-simulations it can be shown that in time-resolved LIFS the simultaneously measured backscattered signal of the excitation wavelength (UV) can be used to develop a special, linearized rescaling algorithm to take into account the most dominant of these varying tissue parameters which is (mu) _a,ex. In combination with biochemical calibration measurements we were able to perform fiberbased quantitative NADH- concentration measurements. In this paper a new rescaling method for VIS and IR light in the frequency domain is proposed. It can be applied within the validity range of the diffusion approximation and provides full (mu) _a and (mu) _s rescaling possibility in a 2- dimensional, non-contact mapping mode. The scanning device is planned to be used in combination with a standard operation microscope of ZEISS, Germany.

We look at the transillumination imaging of relatively thick tissues, for which the diffusion approximation is valid. We solve the direct problem for a sphere of arbitrary radius, scattering and absorption, imbedded in a scattering and absorbing slab. We consider several source configurations: wide beam, point and dipole sources, either cw or modulated. We discuss the results, particularly the resolution threshold, i.e., the sphere radius above which the inverse problem can be solved accurately.

We describe a simple reconstruction method for diffuse optical imaging based on modified back-projection approach for medical tomography. These modifications are based primarily on the deconvolution of the broadened image by a point spread function which is dependent on the scattering of light in tissue. The nonlinearities in the image formation are handled empirically by coordinate transformations. Although our method is an approximate image reconstruction technique, it may provide a basis for rapid, real time medical monitoring by using optical projections. We have applied this method to experimental projections taken by parallel and fan beam tomography geometries on biomedical phantoms with multiple objects. The results presented in this report indicate considerable improvement in image resolution.

Investigations on photon migration in strongly scattering media have shown that due to the practical bandlimitation of the media's transfer characteristic a complete reconstruction of size, shape, and optical parameters of embedded inhomogeneities is generally impossible. However, using a priori information can considerably reduce the solution space. In this article we present a method to determine the position of an inhomogeneity in a cylindrical scattering object by backprojecting the changes in measurement value. For the backprojection a set of trajectories is used approximating the maximum probability photon path. Reconstruction methods as known from computed tomography (CT) are used to obtain a visually improved slice image giving the opportunity to estimate size and shape of the inhomogeneities. Although in this article the procedure is described for time-integrated measurements it should be applicable to any other measurement technique. Above all the method is computationally very efficient.

Images of realistic tissue phantoms were obtained by means of time-resolved transmittance measurements interpreted with the diffusion theory. A system for time-correlated single photon counting was used to collect time-resolved transmittance curves with an instrument response less than 50 ps (FWHM). Tissue phantoms were agarose solutions containing appropriate amounts of ink (absorber) and intralipid (diffuser). The optical properties selected for the background ((mu) '_s approximately equal to 8 cm ^-1 and (mu) _a approximately equal to 0.1 cm^-1) were typical for tissues at near infrared wavelengths and the thickness (4.3 cm) was comparable with the average values in x-ray mammography. The presence of a tumor mass was simulated through an embedded agarose cylinder (6 or 11 mm in diameter). The optical parameters of the inclusion were increased by a factor of 1.5 or 2 with respect to the values in the surrounding medium. Measurements were performed over a 4 cm by 4 cm area, with an acquisition point every 2 mm. With an incident power less than 1 mW, the typical acquisition time was 1 sec/point. The experimental data were fitted with the diffusion approximation to evaluate (mu) '_s and (mu) _a. Images were constructed by plotting the optical parameters as a function of position. For comparison, time-gated images were constructed from the same experimental curves as used for the fit. The technique proved more sensitive than time-gating to changes in scattering properties. On the contrary, it was not possible to obtain good quality images based on differences in the absorption coefficient.

This paper presents first results of an in-vivo-experimental study on the detection of pathological changes of chronical polyarthritis (c.P.) performed by a near infrared cw- transillumination method in interphalangeal joints. The inflammation of a joint system when caused by c.P. leads to changes in the synovial membrane, synovia composition and content, and anatomic geometrical variations. Measurements have shown that these rheumatic induced inflammation processes result in a variation in optical properties ((mu) _a, (mu) _s, g) of the joint system. Using a scanning system the interphalangeal joint is transilluminated with diode lasers (788 nm, 831 nm) perpendicular to the articular cavity in order to use the entire distribution of the transmitted radiation intensity for diagnostic purposes. The study includes results of in-vivo-measurements on healthy and c.P. patients and the evaluation of different distribution properties for detection of chronical polyarthritis.

The application of spatial filtering techniques to frequency domain imaging through scattering media has been investigated using a diffusion model. The criterion used to evaluate the imaging performance of any given system is the trade-off between signal to noise ratio and resolution. Spatial filtering is shown to offer the greatest improvement in system performance for objects positioned near to the detector.

Laser beam propagation inside tissues with different lateral dimensions has been considered. Scattering and anisotropic properties of tissue critically determine spatial fluence distribution and predict sizes of tissue specimens when deviations of this distribution can be neglected. Along the axis of incident beam the fluence rate weakly depends on sample size whereas its relative increase (more than 20%) towards the lateral boundaries. The finite sizes were considered to be substantial only for samples with sizes comparable with the diameter of the laser beam. Interstitial irradiance patterns simulated by Monte Carlo method were compared with direct measurements in human brain specimens.

To model the photon migration in highly scattering media, we use an approximation of the Boltzmann equation, the diffusion equation. A prerequisite for handling the inverse problem consists in solving the forward problem under realistic conditions. We discuss the influence of boundary conditions on the light propagation. The boundary conditions at the walls surrounding the object highly sensitively influence the photon flux at the boundary which means that the time-resolved transmittance is affected. An algorithm for the determination of boundary parameters is introduced and demonstrated by an instructive example. We use the finite element method for the time-resolved case as a basic method in combination with a minimization strategy. The boundary conditions are determined as conditions of the third kind, i.e. the photon density is proportional to the outward photon flux at the boundary.

Light propagation in highly scattering media can be numerically simulated by solving the diffusion equation by the finite element method (FEM). Employing an iterative algorithm, the FEM solution of the forward problem is applied to the inverse imaging problem. Good test results were previously achieved when absorbers were searched in different objects. Now the reconstruction of scattering is also taken into account. Simulated measurement data are used to test and evaluate the method at various objects with tissue-like properties. Resulting problems are very ill posed. The algorithm is specially adapted to the illposedness of the problem. Improvements in reconstruction results can be achieved in two ways, first by adapting the detector arrangements and, secondly, by using a regularization strategy. The effectiveness of these methods is demonstrated by instructive examples.