of 7
Supplemental Document
Acoustic-feedback wavefront-adapted
photoacoustic microscopy: supplement
Y
UECHENG
S
HEN
,
1,†
J
UN
M
A
,
2,†
C
HENGTIAN
H
OU
,
2,†
J
IAYU
Z
HAO
,
3
Y
AN
L
IU
,
4,7,
H
SUN
-C
HIA
H
SU
,
5
T
ERENCE
T. W. W
ONG
,
6
B
AI
-O
U
G
UAN
,
2
S
HIAN
Z
HANG
,
1,8
AND
L
IHONG
V. W
ANG
5,9
1
State Key Laboratory of Precision Spectroscopy, School of Physics and Electronic Science, East China
Normal University, Shanghai 200241, China
2
Guangdong Provincial Key Laboratory of Optical Fiber Sensing and Communications, College of Physics
& Optoelectronic Engineering, Jinan University, Guangzhou 510632, China
3
School of Electronics and Information Technology, Sun Yat-Sen University, Guangzhou 510275, China
4
School of Optometry, Indiana University, Bloomington, Indiana 47405, USA
5
Caltech Optical Imaging Laboratory, Andrew and Peggy Cherng Department of Medical Engineering,
Department of Electrical Engineering,California Institute of Technology, 1200 East California Boulevard,
Pasadena, California 91125, USA
6
Translational and Advanced Bioimaging Laboratory, Department of Chemical and Biological
Engineering, Hong Kong University of Science and Technology, Hong Kong SAR, China
7
yl144@iu.edu
8
sazhang@phy.ecnu.edu.cn
9
lvw@caltech.edu
These authors contributed equally to this work.
This supplement published with Optica Publishing Group on 5 February 2024 by The Authors
under the terms of the Creative Commons Attribution 4.0 License in the format provided by the
authors and unedited. Further distribution of this work must maintain attribution to the author(s)
and the published article’s title, journal citation, and DOI.
Supplement DOI: https://doi.org/10.6084/m9.figshare.24991449
Parent Article DOI: https://doi.org/10.1364/OPTICA.511359
Acoustic-feedback
wavefront-adapted
photoacoustic microscopy: supplement
Y
UECHENG
S
HEN
,
1,†
J
UN
M
A
,
2,†
C
HENGTIAN
H
OU
,
2,†
J
IAYU
Z
HAO
,
3
Y
AN
L
IU
,
4,7
H
SUN
-C
HIA
H
SU
,
5
T
ERENCE
T.
W.
W
ONG
,
6
B
AI
-
OU
G
UAN
,
2
S
HIAN
Z
HANG
,
1,8
AND
L
IHONG
V.
W
ANG
5,9
1
State Key Laboratory of Precision
Spectroscopy, School of Physics and
Electronic Science, East China
Normal University, Shanghai 200241, China
2
Guangdong Provincial Key Laboratory of Optical
Fiber Sensing and Communications, College of
Physics & Optoelectronic Engineering,
Jinan University, G
uangzhou 510632, China
3
School of Electronics and Information Technology
, Sun Yat-sen University, Guangzhou 510275, China
4
School of Optometry, Indiana Unive
rsity, Bloomington, IN 47405, USA
5
Caltech Optical Imaging Laboratory,
Andrew and Peggy Cherng Departm
ent of Medical Engineering,
Department of Electrical Engineering, Californ
ia Institute of Technology, 1200 East California
Boulevard, Pasadena, CA 91125, USA
6
Translational and Advanced Bioi
maging Laboratory, Department
of Chemical and Biological
Engineering, Hong Kong University of Sc
ience and Technology, Hong Kong, China
7
yl144@iu.edu
8
sazhang@phy.ecnu.edu.cn
9
lvw@caltech.edu.
These authors contributed equally to this work.
Supplementary Note 1: Determination of the size of the isoplanatic patch for
in
vivo
studies
Before dynamically compensating for tissue-induced
aberration for live animals, the size of the
isoplanatic patch should be determined a priori
. This parameter determines the area within
which a single set of correction parameters
is still effective in providing substantial
improvement in image quality, which depends on sample morphology, refractive index
heterogeneity, and wavefront sensing methods [1, 2]. Previous studies in fluorescence
microscopy have shown that samples with low su
rface curvature can have a patch size of up to
hundreds of microns [3, 4]. In contrast, for certain samples, it has been shown that a single
aberration correction may only work in the vicinity of the aberration measurement location,
indicating a relatively small patch size [5-7].
In this work, we determine the size of the isoplanatic patch by examining the correlation of
the compensation phase map. We first picked
a small area containing microvessels and
determined the aberration point-by-point with a very small step size. Quantitatively, we found
that for the position being examined, the correlation coefficient drops to about 70% for two
positions separated by
r
7–8
μ
m on average. This value further drops below 50% when the
relative distance is beyond 10
μ
m. Therefore, we set the size of the isoplanatic patch to be about
15
μ
m (
≈ 2
r
). Although different imaging regions may have considerably different patch
sizes, this value was set for all animal experime
nts in this study for convenience. It is worth
noting that, in general, enlarging this size can speed up the imaging process but at the cost of
compensation accuracy.
Supplementary Note 2: Optimization pe
rformance and phase maps at different
locations in the zebrafish embryo
In this section, we present the optimization
performance and correctiv
e wavefront maps at
mulitiple locations of the zebrafish embryo. Supplementary Fig. 1a reproduces Fig. 3c of the
main text, with four representative locations denoted by the white dashed squares.
Corresponding phase maps used for compensating sample-induced aberrations are provided in
Supplementary Fig. 1b, which shows spatially-varying aberrations. At location 1 (on a large
posterior cardinal vein), the Zernike coefficien
ts for defocus and oblique astigmatism are the
largest and are 2
π
and -3
π
, respectively. This results in a si
gnal enhancement of approximately
2.1 times. In contrast, at position 2 (on a dorsal artery), the Zernike coefficients for defocus and
oblique astigmatism are quite different from those at location 1, measuring
π
each. The signal
enhancement at this location is approximately
2.6 times. At location 3 (near a large piece of
pigment), the Zernike coefficients for defocus,
oblique astigmatism, and vertical coma are -
π
,
π
, and -
π
, respectively. This compensation leads to a ~1.8 times signal enhancement. At location
4 (on a small piece of pigment), the Zernike coef
ficients for defocus,
oblique astigmatism, and
horizontal coma are 2
π
, -
π
, and -
π
, respectively. For small pigment, the signal enhancement
reaches approximately 12 times.
These results affirm the pivotal role of defocus correction
throughout the field of view. Furthermore, the di
fferent Zernike coefficien
ts and phase maps at
different locations underscore the necessity of dynamic aberration compensation.
Supplementary Fig. 1. Optimization performan
ce and phase maps at different locations in
the zebrafish embryo. a
Maximum amplitude projection image acquired with AWA (acoustic-
feedback wavefront-adapted) correction.
b
Phase maps and signal enhacement at four
representative locations denoted by the white boxes in (
a
).
Supplementary Note 3: Optimization pe
rformance and phase maps at different
locations in the mouse ear
In this section, we present the optimization
performance and correctiv
e wavefront maps at
multiple locations of the mouse ear. Supplementary Fig. 2a reproduces Fig. 4c from the main
text, and it shows four representative locatio
ns denoted by the white dashed squares.
Corresponding phase maps used to compensate
for sample-induced aberrations are presented
in Supplementary Fig. 2b, whic
h shows spatially-varying aberra
tions. At location 1 (on a large
vessel), only the Zernike coeffi
cient for defocus is nonzero (-
π
), resulting in a modest signal
enhancement of 1.1 times. This outcome suggests that AWA-correction works better on smaller
vessels. In contrast, at location 2 (on a small
vessel), the Zernike coefficients for defocus,
oblique astigmatism, and vertical astigmatism are 4
π
, -2
π
, and –
π
, respectively. The signal
enhancement at this position is approximately 3.4 times. At location 3 (near another large
vessel), the Zernike coefficients for defocus,
oblique astigmatism, vertical astigmatism, and
horizontal coma are 2
π
,
π
, -
π
, and 3
π
, respectively. This compensation leads to a roughly 1.7
times signal enhancement. At location 4 (on an
other vessel), the Zernike coefficients for
defocus, oblique astigmatism, vertical
astigmatism, and vertical coma are 4
π
, -
π
, 3
π
, and 3
π
,
respectively. The signal enhancement reaches ap
proximately 2.1 times. These results reaffirm
the crucial significance of defo
cus correction across the entire
field of view. Moreover, the
different Zernike coefficients and phase maps at
different locations highlight the essential need
for dynamic aberration correction.
Supplementary Fig. 2. Optimization performan
ce and phase maps at different locations in
the mouse ear. a
Maximum amplitude projection image acquired with AWA correction.
b
Phase
maps and signal enhacement at four represeantiv
e locations denoted by the white boxes in (
a
).
Supplementary Note 4: Characterizing resolution enhancement of AWA-PAM
with scattering gels of controlled thickness
Microscopic inhomogeneity inherent to biological tissue considerably distorts optical
wavefronts and degrades image resolution. Here, we show the proposed AWA-PAM can
dynamically offset tissue-induced aberration and improve imaging resolution. To quantitatively
assess this capability, as shown in Supplementary
Fig. 3a, we placed layers of scattering gels
with controlled thicknesses on top of a piece of
gold sheet that has a sharp edge. These
scattering gels were made by mixing 20% intralipid solution, gelatin powder, and pure water
with a mass ratio of 5: 10: 85 [8]. At 532 nm wavelength, the scattering gel has similar optical
properties to soft tissue, which was quantified w
ith a scattering coefficient and a scattering
anisotropy of about 10 mm
-1
and 0.9. To compensate for optical aberration caused by scattering
gels, point-by-point dynamic compensation was implemented. We started the optimization
process from the 4
th
order to the 10
th
order. The Zernike polynomials with orders larger than 10
were found to have relatively small coefficients,
thus being considered inefficient and neglected
here.
With scattering gels covered on top of the gold sheet, one-dimensional (1D) scanning across
the edge of the gold sheet was performed to ob
tain the edge spread function. The line spread
function, which is calculated as the derivative of the edge spread function, quantifies the lateral
resolution through its FWHM [9]. As a typical
example, when the thickness of the scattering
gel was 150
μ
m, these parameters obtained with and
without AWA correction are provided in
Supplementary Figs. 3b and 1c. For other thicknesses, the obtained resolutions with and without
AWA correction are plotted in Supplementary Fig. 3d. When there is no scattering gel, for both
cases, the imaging resolution is roughly the same, about 3
μ
m. This observation indicates that
system aberration has been well compensate
d for, and AWA-PAM does not cause evident
artifacts. As the thickness
increases, the imaging resolutions for conventional PAM
considerably deterior
ate, which are 15.9
μ
m, 37.5
μ
m, 49.4
μ
m, 59.0
μ
m, and 81.6
μ
m when
the thicknesses of the scattering gels are 150
μ
m, 200
μ
m, 250
μ
m, 350
μ
m, 450
μ
m,
respectively. These results confir
m that the existence of thin s
cattering media largely affects
the quality of the optical focus. In comparison, with AWA-PAM, imaging resolutions at
corresponding depths are improved to 9.0
μ
m, 21.0
μ
m, 31.2
μ
m, 40.8
μ
m, and 63.6
μ
m,
respectively. Since AO is known to be incapable of handling increased amounts of scattered
light, AWA-PAM could not bring the lateral resolution back to the diffraction-limited value at
depths. Nonetheless, it still improves the resolution at all depths by compensating for tissue-
induced aberration. An improvement ratio, defined as the ratio of resolutions (FWHM widths
of the line spread function) obtained without and with AWA-PAM, is also provided as a
function of scattering gel thickness. For all th
e thicknesses we tested, the best improvement
ratio of about 1.8 was achieved at thicknesses ranging from 150~200
μ
m. This observation is
because the quality of foci at these depths larg
ely depends on optical ab
erration. Nonetheless,
it should be noted that improvement ratios obtained with scattering gels are typically smaller
than those obtained during
in vivo
studies, as scattering gels are mesoscopic homogeneous.
The improvement ratio reduces to 1.28 at the imaging depth of 450
μ
m. When the thickness
of the scattering gel further goes beyond 450
μ
m, we speculate that AWA-PAM loses its power
as scattered light gradually dominates. This deduction is supported by the observation that the
obtained resolutions at the thickness of 450
μ
m, regardless of employing AWA correction or
not, had reached one-third of the acoustically
determined diffraction-limited value (0.71
λ
/NA
= 185
μ
m). This observation indicates that the fo
rmed foci under this condition are largely
deteriorated by optical scattering, which cannot be handled through aberration compensation.
Therefore, AO gradually loses its power beyond this imaging depth. Wavefront shaping [10-
14], a more sophisticated but currently inefficient methodology, has been proposed to conquer
optical scattering. However, it has not yet achieved satisfactory performance with live animals
due to several technical challenges.
Supplementary Fig. 3. Characterizing the resolution enhancement of AWA-PAM with
scattering gels of controlled thickness. a
Schematic of the characterization process, in which
a gold sheet with a sharp edge was covered
by scattering gels with various thicknesses.
b
,
c
A
typical example of the raw data (black dots),
fitted edge spread functions (red lines), and
computed line spread functions
(blue lines). We measured
the edge spread function by
performing a 1D scanning of the sample and ac
quiring the PA signal with and without AWA
correction, in order to quantify
the lateral resolution from the lin
e spread function. The thickness
of the scattering gel was 150
μ
m.
d
The obtained imaging resolutions with (red) and without
(blue) AWA correction as a function of scatte
ring gel thickness. An improvement ratio that
quantifies the resolution improvement with AWA
correction is also provided, denoted by the
dark red dots.
Supplementary Note 5: Quantification of signal improvement in large vessels of
the mouse ear
Here, we quantify the signal improvement in large
vessels of the mouse ear. First, we registered
the image acquired without AWA correction to th
e image acquired with AWA correction. Then
we computed the signal improvement by taking the ratio of the PA signal acquired with and
without AWA correction (i.e. Fig.
4c and 4f). To find the signal improvement over the large
vessels of the mouse ear, we manually create
d a binary mask that corresponds to the large
vessels (Supplementary Fig. 4a). The PA signa
l improvement in large vessels is shown in
Supplementary Fig. 4b, from which we calcula
ted the average signal improvement in large
vessels and found it to be 29.8%.
Supplementary Fig. 4. Quantification of signal
improvement in large vessels of the mouse
ear. a
Binary mask corresponding to the large vessels of the mouse ear.
b
Signal improvement
in large vessels. The average signal improvement in large vessels is 29.8%.
References
1. N. Ji, "Adaptive optical fluorescence microscopy," Nature Methods
14
, 374 (2017).
2. K. M. Hampson, R. Turcotte, D. T. Miller, K. Kurokawa
, J. R. Males, N. Ji, and M. J. Booth, "Adaptive optics for
high-resolution imaging," Nature Reviews Methods Primers
1
, 68 (2021).
3. K. Wang, W. Sun, C. T. Richie, B. K. Harvey, E. Be
tzig, and N. Ji, "Direct wavefr
ont sensing for high-resolution
in vivo imaging in scattering tissue," Nature Communications
6
, 7276 (2015).
4. N. Ji, T. R. Sato, and E. Betzig, "Characterization a
nd adaptive optical correction of aberrations during in vivo
imaging in the mouse cortex," Proceedings
of the National Academy of Sciences
109
, 22-27 (2012).
5. C. Wang, R. Liu, D. E. Milkie, W.
Sun, Z. Tan, A. Kerlin, T.-W. Chen
, D. S. Kim, and N. Ji, "Multiplexed
aberration measurement for deep tissue imaging in vivo," Nature Methods
11
, 1037 (2014).
6. K. Wang, D. E. Milkie, A. Saxena, P. Engerer, T.
Misgeld, M. E. Bronner, J. Mumm, and E. Betzig, "Rapid
adaptive optical recovery of optimal reso
lution over large volumes," Nature Methods
11
, 625 (2014).
7. M. Kim, Y. Jo, J. H. Hong, S. Kim, S. Yoon, K.-D. Song,
S. Kang, B. Lee, G. H. Kim, H.-C. Park, and W. Choi,
"Label-free neuroimaging in vivo us
ing synchronous angular scanning
microscopy with single-scattering
accumulation algorithm," Nature Communications
10
, 3152 (2019).
8. P. Lai, X. Xu, and L. V. Wang, "Dependence of opti
cal scattering from Intralipid
in gelatin-gel based tissue-
mimicking phantoms on mixing temperature a
nd time," Journal of Biomedical Optics
19
, 035002 (2014).
9. L. V. Wang and H.-i. Wu,
Biomedical optics: principles and imaging
(John Wiley & Sons, 2012).
10. T. Zhao, S. Ourselin, T. Vercauteren, and W. Xi
a, "High-speed photoacoustic-guided wavefront shaping for
focusing light in scattering media," Opt. Lett.
46
, 1165-1168 (2021).
11. P. Lai, L. Wang, J. W. Tay, and L. V. Wang, "Phot
oacoustically guided wavefront
shaping for enhanced optical
focusing in scattering media," Nature Photonics
9
, 126 (2015).
12. F. Kong, R. H. Silverman, L. Liu, P. V. Chitnis, K.
K. Lee, and Y. C. Chen, "Photoacoustic-guided convergence
of light through optically diffusive media," Opt. Lett.
36
, 2053-2055 (2011).
13. M. A. Inzunza-Ibarra, E. Premillieu
, C. Grünsteidl, R. Piestun, and T.
W. Murray, "Sub-acoustic resolution
optical focusing through scattering using photoacoustic fl
uctuation guided wavefront
shaping," Opt. Express
28
,
9823-9832 (2020).
14. D. B. Conkey, A. M. Caravaca-Agui
rre, J. D. Dove, H. Ju, T. W. Murr
ay, and R. Piestun, "Super-resolution
photoacoustic imaging through a sca
ttering wall," Nature Communications
6
, 7902 (2015).