Hybrid Monte Carlo– diffusion method for light

propagation in tissue with a low-scattering region

Toshiyuki Hayashi, Yoshihiko Kashio, and Eiji Okada

The heterogeneity of the tissues in a head, especially the low-scattering cerebrospinal fluid 共CSF兲
layer surrounding the brain has previously been shown to strongly affect light propagation in the
brain. The radiosity-diffusion method, in which the light propagation in the CSF layer is assumed
to obey the radiosity theory, has been employed to predict the light propagation in head models.
Although the CSF layer is assumed to be a nonscattering region in the radiosity-diffusion method, fine
arachnoid trabeculae cause faint scattering in the CSF layer in real heads. A novel approach, the
hybrid Monte Carlo– diffusion method, is proposed to calculate the head models, including the low-
scattering region in which the light propagation does not obey neither the diffusion approximation nor
the radiosity theory. The light propagation in the high-scattering region is calculated by means of
the diffusion approximation solved by the finite-element method and that in the low-scattering region
is predicted by the Monte Carlo method. The intensity and mean time of flight of the detected light
for the head model with a low-scattering CSF layer calculated by the hybrid method agreed well with
those by the Monte Carlo method, whereas the results calculated by means of the diffusion approx-
imation included considerable error caused by the effect of the CSF layer. In the hybrid method, the
time-consuming Monte Carlo calculation is employed only for the thin CSF layer, and hence, the
computation time of the hybrid method is dramatically shorter than that of the Monte Carlo method.
© 2003 Optical Society of America
OCIS codes: 170.0170, 170.3660, 170.3890.

1. Introduction however, it has the disadvantage of requiring enor-

The development of an adequate model to calculate
light propagation in biological tissues is essential for
near-infrared spectroscopy1–3 and imaging.4,5 De-
spite numerical analyses based on a diffusion approx-
imation having been widely applied to predict the
light propagation in biological tissues, the diffusion
approximation does not hold in low-scattering re-
gions. One of the typical low-scattering regions in
biological tissue is the cerebrospinal fluid 共CSF兲 layer
surrounding the brain, and previous results have
shown that the light propagation in the head is
strongly affected by its presence.6 – 8 The Monte
Carlo 共MC兲 method can accurately predict the light
propagation in a model with low-scattering regions,
The authors are with the Department of Electronics and Elec-
trical Engineering, Keio University 3-14-1, Hiyoshi Kohoku-ku,
Yokohama 223-8522, Japan. E. Okada’s e-mail address is
okada@elec.keio.ac.jp.
Received 1 September 2002; revised manuscript received 11 De-
cember 2002.
0003-6935兾03兾162888-09$15.00兾0
© 2003 Optical Society of America
mous computation time.7 An alternative approach,
the radiosity-diffusion method, has been proposed to
calculate the light propagation in the scattering me-
dia, including nonscattering 共void兲 regions.9,10 In
the radiosity-diffusion method, the light propagation
in the nonscattering regions is assumed to obey the
radiosity theory, whereas the diffusion approxima-
tion is applied to the diffusive regions.
In most of the previous reports,6 –10 the CSF layer is
assumed to be a nonscattering region because the
CSF itself is a nonscattering and clear fluid. In the
real head, however, fine arachnoid trabeculae cause
faint scattering in the CSF layer. The mean optical
path length predicted for the model with a low-
scattering CSF layer with a transport scattering co-
efficient of approximately 0.2 mm⫺1 agrees closer
with the experimental results than that for the model
with a nonscattering CSF layer.11 Both the diffu-
sion approximation and the radiosity theory cannot
accurately predict light propagation in the low-
scattering CSF layer.
In this study, we propose a novel approach, the
hybrid MC– diffusion method, which can be em-
ployed to calculate the models, including low-

2888 APPLIED OPTICS 兾 Vol. 42, No. 16 兾 1 June 2003


Fig. 1. Schematic diagram of the adult head model with a low-
scattering CSF layer.
scattering regions in which the light propagation
obeys neither the diffusion approximation nor the
radiosity theory. The concept of the hybrid MC–
diffusion method is basically the same as that of the
hybrid radiosity-diffusion method. The light prop-
agation in the high-scattering region is calculated
by means of the diffusion approximation solved by
the finite-element method 共FEM兲 and that in the
low-scattering region is predicted by the MC
method. The light propagation in the adult head
model with the low-scattering CSF layer was pre-
dicted with the proposed hybrid MC– diffusion
method and FEM to calculate the intensity and
mean time of flight of the detected light. The re-
sults calculated with each method were validated
by the results of the MC method.
2. Numerical Modeling of Light Propagation
A. Adult Head Model
The adult head model considered in this study is
shown in Fig. 1. Since the structure of the low-
scattering CSF layer sandwiched between the high-
scattering tissues strongly affects the light
propagation in the adult head,6 – 8 the head consists of
three different homogeneous layers. These include
the 10-mm-thick superficial layer, the 2-mm-thick in-
termediate layer, and the 24-mm-thick bottom layer
that mimic the scalp and skull, the CSF, and the
brain, respectively. The optical properties of each
layer are shown in Table 1. The optical properties of
the superficial and brain layers were based on the
reported data of the dermis,12 skull,13 and brain.14
The transport scattering coefficient of the CSF layer
was varied from 0 to 1.0 mm⫺1. The light propaga-
Table 1. Optical Properties of the Adult Head Model
Transport
Scattering Absorption
Coefficient ␮s⬘ Coefficient
Tissue Type 共mm⫺1兲 ␮a 共mm⫺1兲
Superficial layer 1.75 0.017
CSF 0, 0.1, 0.2, 1.0 0.002
Brain 2.2 0.036
tion in the head model was analyzed by means of
diffusion theory, MC method, and the hybrid MC–
diffusion method. The diffusion equation was nu-
merically solved by means of the FEM.
B. Finite-Element Method
Numerical modeling of the light propagation in bio-
logical tissues has been frequently based on the dif-
fusion approximation of the light transport equation.
The photon density ⌽共r兲 at position r caused by a
continuous-wave 共CW兲 light source can be given by
the time-independent diffusion equation15 as shown
in Eq. 共1兲:
⫺ⵜ 䡠 ␬共r兲ⵜ⌽共r兲 ⫹ ␮ a共r兲⌽共r兲 ⫽ q 0共r兲, (1)
where q0共r兲 is the isotropic light source, ␮a共r兲 is ab-
sorption coefficient, and ␬共r兲 is the diffusion coeffi-
cient, defined as
␬共r兲 ⫽ 1兾兵3关␮ a共r兲 ⫹ ␮ s⬘共r兲兴其, (2)
where ␮s⬘ 共r兲 is the transport scattering coefficient.
Theoretical and experimental results have shown the
validity of the diffusion equation under appropriate
conditions where ␮s⬘ 共r兲 ⬎⬎ ␮a共r兲 and 1兾␮s⬘共r兲 is much
smaller than the distance between the source and the
observation points, i.e., ␮s⬘共r兲 is sufficiently high.
The FEM was employed to solve the diffusion equa-
tion.15,16 The head model was divided into 9600 tri-
angular elements. The Robin boundary condition
was implemented to take into account the internal
reflection caused by the refractive-index mismatch on
the surface of the model.17 The refractive index of
the head model was assumed to be 1.4. An isotro-
pic point source located at the distance 1兾␮s⬘ below
the surface of the head model was used to approxi-
mate the collimated light source.
The outgoing flux, ⌫0共␰兲, which represents the in-
tensity of the detected light, at position ␰ on the
boundary can be obtained as follows:
⌫ 0共␰兲 ⫽ ⫺␬共␰兲n̂ 䡠 ⵜ⌽共␰兲, (3)
where n̂ is the outward normal at position ␰. The
mean time of flight was also obtained from the time-
independent diffusion equation by the direct calcula-
tion proposed by Arridge and Schweiger.18
C. Monte Carlo Method
The MC method has the advantages of being concep-
tually simple and permitting the direct handling of
complex geometries and optical heterogeneity, includ-
ing the nonscattering regions. The MC19 –21 method
based on a variance reduction technique was employed
to predict the light propagation in the head model.
The propagation of a photon in the head model was
determined with the transport scattering coefficient
and random numbers. Isotropic scattering was as-
sumed in all the layers. The scattering length was
corrected when the photon crossed the boundary be-
tween the different types of tissue. The reflection due
to the refractive-index mismatch was considered only
on the boundary between the scalp and the air. When
1 June 2003 兾 Vol. 42, No. 16 兾 APPLIED OPTICS 2889

Fig. 2. Finite-element mesh of the adult head model for the hy-
brid MC– diffusion method.
the photon reached the detector, the ultimate survival
weight of photon was calculated from the absorption
coefficient and partial optical path length in each layer.
The survival weight and optical path length were re-
corded to calculate the intensity and mean time of
flight of the detected light. Since the MC method can
be employed in the model, including the nonscattering
regions, the results from the FEM and hybrid MC–
diffusion method were compared with the results of the
MC method. The MC method has the disadvantage
of requiring enormous computation time to keep sta-
tistical errors within reasonable limit.
D. Hybrid Monte Carlo– diffusion Method
The concept of the hybrid MC– diffusion method is
basically the same as the hybrid radiosity-diffusion
method.9 The light propagation in the high-
scattering tissue, i.e., the superficial and brain layers,
is calculated by the FEM and that in the low scatter-
ing CSF layer is predicted by the MC method. A
two-dimensional finite-element mesh for the head
model is shown in Fig. 2. The superficial and brain
layers in the model were divided into 5366 and 8724
triangular elements, respectively.
The diffusion theory to analyze the light propaga-
tion in the scattering region was combined with the
MC method to analyze the light propagation in the
low-scattering region by the following five-step pro-
cess. In the first step, the photon density in the
superficial layer arising from a light source on the
head surface was calculated with the FEM to obtain
the reflectance of the head surface and the outgoing
flux, ⌫1共p兲, at any position p on the inner surface of
the superficial layer. In the second step, we calcu-
lated the incoming flux, ⌫2共q兲, on an region dA2 at
position q on the surface of the brain layer. The
incoming flux, ⌫2共q兲, can be calculated by integration
of the contribution from over the total inner surface,
As-in, of the superficial layer across the CSF layer as
shown in Fig. 3,
兰
⌫ 2共q兲 ⫽ ⌫ 1共p兲 F共p, q兲dA 1, (4)
As-in
2890 APPLIED OPTICS 兾 Vol. 42, No. 16 兾 1 June 2003
Fig. 3. Geometry for calculation of the transmittance and reflec-
tance ratios by the MC method.
where F共p, q兲 is the transmittance ratio that gives
the contribution of the outgoing flux on dA1 to the
incoming flux on dA2. Since the diffusion equation
is not valid in the CSF layer, the light propagation
was predicted with the MC method to obtain the
transmittance ratio. Similarly, the incoming flux
⌫3共r兲 on a region dA3 at position r on the inner surface
of the superficial layer was obtained with the reflec-
tance ratio that gives the contribution of the outgoing
flux on dA1 to the incoming flux on dA3. The reflec-
tance ratio was also calculated by the MC method.
In the third step, the photon density in the brain
layer arising from the incoming flux ⌫2共q兲 was calcu-
lated with the FEM to obtain the outgoing flux ⌫4共s兲
from the brain surface. In the fourth step, the light
propagation in the CSF layer caused by the flux,
⌫4共s兲, was predicted with the MC method to obtain
the flux, ⌫5共r兲, coming back onto the inner surface of
the superficial layer. Since the CSF layer in the
head model was a simple slab, the flux can be calcu-
lated from the transmittance ratio as well as from the
incoming flux, ⌫2共q兲, in the second step. The total
flux coming back onto the inner surface of the super-
ficial layer, ⌫5共r兲⬘, is the sum of ⌫5共r兲 and ⌫3共r兲 cal-
culated in the second step. In the fifth step, the
outgoing flux from the head surface and from the inner
surface of the superficial layer arising from flux ⌫5共r兲⬘
was calculated with the FEM. The processes from
the second to the fifth step were iterated until changes
in the total flux become negligible. Thus the intensity
of the detected light on the head surface could be ob-
tained.
The mean time of flight of the detected light was
calculated by a similar process. The mean time of
flight 具t12共p兲典 of the flux ⌫1共p兲 on the inner surface of
the superficial layer was calculated with the FEM.
The time taken for the light to propagate from p to q
through the CSF is given by
t pq共q兲 ⫽ 具l典兾c, (5)
where 具l典 is the mean optical path length of the light
propagated from p to q predicted by the MC method,
and c is the speed of light in the CSF layer. Sum-
ming all the contributions from the inner surface of
the superficial layer, As-in, yields the mean time of
Fig. 4. Angular distribution 共a兲 of transmitted and 共b兲 of reflected
light from the superficial layer.
flight of flux ⌫2共q兲 at a point q on the brain surface,
and is shown by
具t 13共q兲典 ⫽ 再兺
p
关具t 12共p兲典 ⫹ t pq共q兲兴⌫ pq共q兲 冎 冒兺
p
where ⌫pq共q兲 is the flux from p to q. The mean time
of flight, 具t14共s兲典, of the flux, ⌫4共s兲, from the brain
surface was obtained from this mean time distribu-
tion and the mean time of flight, 具tqs共s兲典, of light that
traveled in the brain layer from q to s calculated with
the FEM:
再兺 冎 冒兺
具t 14共s兲典 ⫽ 关具t13共q兲典⫹具tqs共s兲典兴⌫qs共s兲
q q
where ⌫qs共s兲 is the flux at s from q. The mean time
of flight of flux, ⌫5共r兲, can be calculated with Eq. 共6兲,
and the mean time of flight in the superficial layer
was calculated with the FEM. These processes were
iterated until changes in the mean time of flight of
the detected light become negligible.
Since the light propagating in the CSF layer is
hardly scattered, the angular dependence of the light
radiated to the CSF layer significantly affects the
Fig. 5. Detected intensity for homogeneous models as a function
of source– detector spacing. The transport scattering coefficient
of the model ␮⬘s: 共a兲 0.1 mm⫺1 and 共b兲 1.0 mm⫺1.
⌫ pq共q兲 (6) distribution of the incoming flux from the CSF. This
problem was previously investigated by Firbank et
al.9 Figure 4 shows the angular distribution of the
light exiting from the superficial layer predicted by
the MC method. The refractive-index mismatch on
the superficial–CSF-layer boundaries was ignored
and a hundred million photons were injected to ob-
tain the result. The distribution of the surface ra-
diance is proportional to the cosine of the exiting
zenithal angle ␪ for the exiting light that is isotropi-
⌫ qs共s兲, (7) cally distributed, whereas the distribution of the ra-
diance caused by the photon density inside a
scattering body can be considered proportional to
cos ␪ 共1 ⫹ cos ␪兾0.71兲. The angular distribution of
the transmitted light shown in Fig. 4共a兲 indicates
that we should consider the 共cos ␪兾0.71兲 term for the
first calculation of the light propagation in the CSF
layer.9 The angular distribution of the light re-
flected from the high-scattering medium is almost
proportional to cos ␪ as shown in Fig. 4共b兲. Conse-
quently, the term of 共cos ␪兾0.71兲 for the light source
1 June 2003 兾 Vol. 42, No. 16 兾 APPLIED OPTICS 2891
Fig. 6. Mean time of flight of detected light for homogeneous
models as a function of source– detector spacing. The transport
scattering coefficient of the model ␮⬘s: 共a兲 0.1 mm⫺1 and 共b兲 1.0
mm⫺1.
distribution in the MC method was only considered in
the first iteration of the hybrid calculation.
3. Results and Discussion
A. Validity of Diffusion Equation for Low-Scattering
Medium
Figures 5 and 6 shows the intensity and mean time of
flight of the detected light for a homogeneous model
calculated by the FEM and MC method as a function
of source– detector spacing, respectively. The trans-
port scattering coefficient of the homogeneous model
was 0.1 and 1.0 mm⫺1, and the absorption coefficient
of both models was 0.002 mm⫺1. Since the diffusion
equation is not valid in a low-scattering medium, the
intensity and mean time of flight for the low-
scattering model calculated with the FEM are signif-
icantly underestimated. The results of the FEM for
the model with a scattering coefficient is 1.0 mm⫺1
agreed well with that of the MC method. These re-
2892 APPLIED OPTICS 兾 Vol. 42, No. 16 兾 1 June 2003
Fig. 7. CSF layer 共a兲 reflectance and 共b兲 transmittance caused by
an isotropic point source on the surface of the CSF layer.
sults suggest that the diffusion approximation is not
valid in the low-scattering CSF layer of the head
model and that it is not appropriate to employ the
FEM to calculate the light propagation in the head
model with the CSF.
B. Transmittance and Reflectance Ratios
In the hybrid MC– diffusion method, the incoming flux
from the CSF layer such as ⌫2共q兲, ⌫3共r兲 and ⌫5共r兲 can be
obtained from the transmittance and reflectance ra-
tios. We do not have to recalculate the light propa-
gation in the CSF by the MC method in each iterative
process once the transmittance and reflectance ratios
are obtained. The transmittance and reflectance ra-
tios as a function of the detected position are shown in
Fig. 7. The angular distribution of the point source on
the surface of the CSF layer is assumed to be propor-
tional to cos ␪. These ratios can be used to calculate
the incoming flux from the CSF layer except for the
first calculation in the iterative processes. In case of
Fig. 8. Detected intensity for the adult head models as a function of the source– detector spacing. The transport scattering coefficient
of the CSF layer ␮s⬘: 共a兲 0 mm⫺1, 共b兲 0.1 mm⫺1, 共c兲 0.2 mm⫺1, 共d兲 1.0 mm⫺1.
the nonscattering CSF layer, the transmittance ratio is
only shown because no reflection can occur in the non-
scattering CSF layer. The total intensity of the trans-
mittance decreases with the increasing scattering
coefficient of the CSF layer, whereas the total intensity
of the reflectance increases with the increasing scat-
tering coefficient. The distribution of both the trans-
mittance and the reflectance of light propagated in the
CSF layer caused by a point source on the boundary
converges with the increasing scattering coefficient of
the CSF layer. Since the geometry of the CSF layer in
the head model was a simple slab, the transmittance
and reflectance ratios at any position were the same.
The MC method calculation for 50 million incident
photons to obtain the transmittance and reflectance
ratios for each model took approximately 10 min in
computation time on a dual 450 MHz Pentium III PC.
C. Validity of Hybrid Monte Carlo–Diffusion Method
The intensity of detected light on the surface of the
head model as a function of the source– detector spac-
ing is shown in Fig. 8. The results calculated by the
hybrid MC– diffusion method and FEM are compared
with those by the MC method to discuss the validity
of the methods. Figure 8共a兲 shows the detected in-
tensity for the model with the nonscattering CSF
layer. The result of the hybrid radiosity-diffusion
method is also shown in Fig. 8共a兲 since that method
can be applied only to the model with the nonscatter-
ing CSF. The intensity calculated by the hybrid ra-
diosity and hybrid MC methods and FEM agree with
that by the MC method up to a source– detector spac-
ing of 15 mm. Once the source– detector spacing
exceeds 15 mm, the result of the FEM does not match
the MC method results, whereas the results by the
hybrid methods are in good agreement with that by
the MC method. Since the diffusion equation is not
valid in the nonscattering region, the influence of the
CSF layer should be considered for calculation of the
propagation of light detected at more than 15 mm
from the source. The detected intensity for the
1 June 2003 兾 Vol. 42, No. 16 兾 APPLIED OPTICS 2893
Fig. 9. Mean time of flight of detected light for the adult head models as a function of the source– detector spacing. The transport
scattering coefficient of the CSF layer ␮⬘s: 共a兲 0 mm⫺1, 共b兲 0.1 mm⫺1, 共c兲 0.2 mm⫺1, 共d兲 1.0 mm⫺1.
model with the low-scattering CSF layer calculated
with the FEM is slightly overestimated compared
with the MC method results beyond the source–
detector spacing of 20 mm, whereas the results from
the hybrid MC method agree well with the MC
method results, as shown in Figs. 8共b兲 and 8共c兲. The
error in the FEM results for the head model decreases
with the increasing scattering coefficient of the CSF
layer. The diffusion equation holds in the scattering
medium with a transport scattering coefficient of 1.0
mm⫺1 as shown in Fig. 5共b兲. The detected intensity
calculated with the FEM agrees with the MC method
results as well as the hybrid MC method in the case
in which the transport scattering coefficient of the
CSF layer is 1.0 mm⫺1.
The mean time of flight for the head models as a
function of the source– detector spacing calculated
by the MC method, the FEM, and the hybrid MC
method is shown in Fig. 9. The result of the hybrid
radiosity method is also shown in Fig. 9共a兲. The
2894 APPLIED OPTICS 兾 Vol. 42, No. 16 兾 1 June 2003
mean time of flight for the model with the nonscat-
tering CSF layer predicted by the MC method and
the hybrid methods increases slowly for source–
detector spacing greater than 30 mm and levels off,
whereas that by the FEM remains approximately
constant for the source– detector spacing greater
than 20 mm as shown in Fig. 9共a兲. The FEM un-
derestimates the mean time of flight for the light
detected at more than 20 mm from the source be-
cause of the collapse of the diffusion approximation
in the CSF layer. The slope of the mean time of
flight for the head models with the low-scattering
CSF layer with 0.1 and 0.2 mm⫺1 transport scat-
tering coefficients predicted by the MC method
slightly decreases beyond the source– detector spac-
ing of 30 mm. This tendency can be observed in
the experimentally measured mean time of flight
for adult foreheads.11 The mean time of flight cal-
culated by the hybrid MC method agrees well with
the MC method results, whereas that by the FEM
was underestimated for the source– detector spac-
ing greater than 25 mm. The error in the mean
time of flight decreases with the increasing scatter-
ing coefficient of the CSF layer. The mean time of
flight for the model with the CSF layer having a 1.0
mm⫺1 transport scattering coefficient predicted by
both the FEM and the hybrid MC method agrees
well with that by the MC method because the dif-
fusion approximation holds in the tissue with a 1.0
mm⫺1 transport scattering coefficient.
The calculation by the MC method for 50 million
photons took approximately 30 h in computation
time, whereas the hybrid MC method took approxi-
mately 200 s except for the precalculation of the
transmittance and reflectance ratios. The computa-
tion time of the hybrid MC method is longer than that
of the FEM, which takes approximately 2 s, however,
it is significantly reduced compared with that of the
MC method.
4. Conclusion
In this study, the hybrid MC– diffusion method is
proposed for calculation of the light propagation in
a scattering medium with a low-scattering layer.
The light propagation in the high-scattering region
is calculated by means of the diffusion approxima-
tion that is solved with the FEM and that in the
low-scattering region is predicted by the MC
method. The results for the head model with the
nonscattering or low-scattering CSF layer calcu-
lated by means of the hybrid method agree well
with those by the MC method, whereas the results
calculated by the diffusion approximation include
considerable error caused by the effect of the CSF
layer. In the hybrid method, the MC method was
employed only for the thin CSF layer, and hence,
the computation time of the hybrid method is dra-
matically shorter than that of the MC method. We
calculated only the light propagation in the two-
dimensional model, but it is not difficult to extend
the model to the three-dimensional geometry that
allows us to calculate the light propagation in the
more realistic model of the head.
The authors thank S. R. Arridge, H. Dehghani, and
M. Schweiger for their helpful discussions. This re-
search was partly supported by the Ministry of Edu-
cation, Science and Culture of Japan under a Grants-
in-Aid for Developmental Scientific Research
13558116 and 13750397, The Research Grant of The
Okawa Foundation for Information and Telecommu-
nications, and The Research Grant of Casio Science
Promotion Foundation.
References
1. S. J. Matcher, C. E. Elwell, C. E. Cooper, M. Cope, and D. T.
Delpy, “Performance comparison of several published tissue
near-infrared spectroscopy algorithm,” Anal. Biochem. 227,
54 – 68 共1995兲.
2. G. Gratton, M. Fabiani, D. Friedman, M. A. Franceschini, S.
Fantini, P. Corballis, and E. Gratton, “Rapid changes of optical
parameters in the human brain during a tapping task,” J.
Cogn. Neurosci. 7, 446 – 456 共1995兲.
3. S. Kohri, Y. Hoshi, M. Tamura, C. Kato, Y. Kuge, and N.
Tamaki, “Quantitative evaluation of the relative contribution
ratio of cerebral tissue to near-infrared signals in the adult
human head: a preliminary study,” Physiol. Meas. 23, 301–
312 共2002兲.
4. J. C. Hebden, S. R. Arridge, and D. T. Delpy, “Optical imaging
in medicine: I. Experimental techniques,” Phys. Med. Biol.
42, 825– 840 共1997兲.
5. B. Chance, E. Anday, S. Nioka, S. Zhou, L. Hong, K. Worden,
C. Li, T. Murray, Y. Ovetsky, D. Pidikiti, and R. Thomas, “A
novel method for fast imaging of brain function, non-
invasively, with light,” Opt. Express 2, 411– 423 共1998兲, http:兾兾
www.opticsexpress.org.
6. M. Firbank, M. Schweiger, and D. T. Delpy, “Investigation of
‘light piping’ through clear regions of scattering objects,” in
Optical Tomography, Photon Migration, and Spectroscopy of
Tissue and Model Media: Theory, Human Studies, and In-
strumentation, B. Chance and R. R. Alfano, eds., Proc. SPIE
2389, 167–173 共1995兲.
7. E. Okada, M. Firbank, M. Schweiger, S. R. Arridge, M. Cope,
and D. T. Delpy, “Theoretical and experimental investigation
of near-infrared light propagation in a model of the adult
head,” Appl. Opt. 36, 21–31 共1997兲.
8. M. Firbank, E. Okada, and D. T. Delpy, “A theoretical study of
the signal contribution of regions of the adult head to near-
infrared spectroscopy studies of visual evoked responses,” Neu-
roimage 8, 69 –78 共1998兲.
9. M. Firbank, S. R. Arridge, M. Schweiger, and D. T. Delpy,
“An investigation of light transport through scattering bod-
ies with non-scattering regions,” Phys. Med. Biol. 41, 767–
783 共1996兲.
10. S. R. Arridge, H. Dehghani, M. Schweiger, and E. Okada,
“The finite element model for the propagation of light in
scattering media: a direct method for domains with non-
scattering regions,” Med. Phys. 27, 252–264 共2000兲.
11. E. Okada and D. T. Delpy, “Near infrared light propagation in
an adult head model. Part 1: Modeling of low-level scattering
in cerebrospinal fluid layer,” Appl. Opt. 42, 2906 –2914 共2003兲.
12. C. R. Simpson, M. Kohl, M. Essenpreis, and M. Cope, “Near-
infrared optical properties of ex vivo human skin and sub-
cutaneous tissues measured using the Monte Carlo inversion
technique,” Phys. Med. Biol. 43, 2465–2478 共1998兲.
13. M. Firbank, M. Hiraoka, M. Essenpreis, and D. T. Delpy,
“Measurement of the optical properties of the skull in the
wavelength range 650 –950 nm,” Phys. Med. Biol. 38, 503–510
共1993兲.
14. P. van der Zee, M. Essenpreis, and D. T. Delpy, “Optical prop-
erties of brain tissue,” in Photon Migration and Imaging in
Random Media and Tissues, B. Chance and R. R. Alfano, eds.,
Proc. SPIE 1888, 454 – 465 共1993兲.
15. M. Schweiger, S. R. Arridge, M. Hiraoka, and D. T. Delpy,
“Application of the finite element method for the forward
model in infra-red absorption imaging,” in Mathematical
Methods in Medical Imaging, D. C. Wilson and J. N. Wilson,
eds., Proc. SPIE 1768, 97–108 共1992兲.
16. S. R. Arridge, M. Schweiger, M. Hiraoka, and D. T. Delpy, “A
finite element approach for modelling photon transport in tis-
sue,” Med. Phys. 20, 299 –309 共1993兲.
17. M. Schweiger, S. R. Arridge, M. Hiraoka, and D. T. Delpy, “The
finite element method for the propagation of light in scattering
media: boundary and source conditions,” Med. Phys. 22,
1779 –1792 共1995兲.
1 June 2003 兾 Vol. 42, No. 16 兾 APPLIED OPTICS 2895
18. S. R. Arridge and M. Schweiger, “Direct calculation of the mo-
ments of the distribution of photon time of flight in tissue with
a finite-element method,” Appl. Opt. 34, 2683–2687 共1995兲.
19. P. van der Zee and D. T. Delpy, “Simulation of the point spread
function for light in tissue by a Monte Carlo technique,” Adv.
Exp. Med. Biol. 215, 179 –191 共1987兲.
20. M. Hiraoka, M. Firbank, M. Essenpreis, M. Cope, S. R. Ar-
ridge, P. van der Zee, and D. T. Delpy, “A Monte Carlo inves-
2896 APPLIED OPTICS 兾 Vol. 42, No. 16 兾 1 June 2003
tigation of optical pathlength in inhomogeneous tissue and its
application to near-infrared spectroscopy,” Phys. Med. Biol. 38,
1859 –1876 共1993兲.
21. E. Okada, M. Schweiger, S. R. Arridge, M. Firbank, and D. T.
Delpy, “Experimental validation of Monte Carlo and finite-
element methods for the estimation of the optical path
length in inhomogeneous tissue,” Appl. Opt. 35, 3362–3371
共1996兲.