Biomedical Engineering Reference
In-Depth Information
[43]. The measured dependence of reduced scattering coefficient on wavelength and corresponding
theoretical fits is plotted in Figure 8.12. The experimental trends are well approximated by the Mie
equation, and, as expected, the reduced scattering coefficient decreases with increased glycerol con-
centrations. In the case of Intralipid, a very high correlation of 0.97 was found between the reduced
scattering coefficient and the OCA refractive index, clearly showing that refractive index matching is
the mechanism of optical clearing in the absence of diffusion through tissue and molecular interac-
tions. Consequently, scattering properties of tissue phantoms can be precisely controlled to simulate
biological tissues.
8.9 Retention of SHG Polarization Signatures through
optical clearing
The polarization response exhibited by SHG is one of the exclusive benefits of this optical modality. It
is governed by the local geometrical distribution of SHG-producing dipoles that result in a respective
tensor of second-order nonlinear susceptibilities [44,45]. It has been shown that dependence of SHG
intensity on the polarization of the fundamental wave is related to the pitch angle of the protein helix [46],
whereas anisotropy of SHG signal indirectly measures regularity of dipole assembly [47]. Thus, polar-
ization measurement is yet another way SHG imaging can probe submicron structures with resolution
beyond any other optical method. For example, whenever protein assembly is altered in disease state,
the SHG polarization response will change accordingly. However, the polarization responses in tissues
become scrambled due to multiple scattering since each photon collision introduces depolarization, both
in fundamental and SHG waves [48]. In fact, at a depth of just a few mean-free paths, additional informa-
tion related to polarization can be lost completely, as shown in Figure 8.13, which plots the dependence
of forward SHG intensity on the angle between collagen fibril and excitation polarization. The data are
taken from murine tail tendon at two depths of 5 and 45 μm, the latter showing a complete loss of the typ-
ical collagen response [46], since the fundamental wave gets highly depolarized by the time the photons
reach the focal volume. The second panel shows the polarization profiles from depths of 10 and 100 μm
measured in the tendon cleared with 50% glycerol. Owing to reduction of scattering coefficient by two
orders of magnitude, the fundamental retains its polarization and the profile measured deep in the tissue
essentially overlaps with the one from the top. However, minor deviations from uncleared sample are
evident here as the minimum becomes sharper, whereas both maxima appear wider. This might partially
support the argument for alteration of higher-order collagen structures by glycerol.
Similar trends occur in SHG from striated muscle, where we observed that the input polarization is
again completely randomized at the depth of 100 μm, although the exact depth threshold has not been
experimentally verified. The treatment with 50% glycerol retains the polarization response over the first
100 μm of depth, although considerable depolarization takes place at 180 μm, where amplitude of the
curve decreases by about 40% (data not shown).
One way the effects of optical clearing on SHG polarization state can be assessed is by measurement
of the SHG signal anisotropy defined by
β = (
I
par
−
I
orth
)/(
I
par
+ 2
I
orth
)
(8.5)
where
I
par
and
I
orth
are components of SHG signal parallel and orthogonal to the polarization of the
excitation laser. These are successively measured with the respectively oriented Glan polarizer in the
detection path. The signal anisotropy β varies from 1 (all dipoles are aligned with the laser polariza-
tion) through 0 (isotropic state, circular polarization) to −0.5 (completely out-of-phase polarization
response). When β is measured from samples of any thickness beyond a single isolated fiber, there will
be some depolarization present due to scattering. In fact, forward SHG signal that originated at the
top of 60 μm tendon is almost depolarized with β < 0.2, and is completely randomized deeper in the
sample (Figure 8.14, squares). Both primary and secondary filter effects are responsible for the loss of
