Author`s personal copy

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
NeuroImage 45 (2009) 68–74
Contents lists available at ScienceDirect
NeuroImage
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / y n i m g
Technical Note
Simultaneous, live imaging of cortical spreading depression and associated cerebral
blood flow changes, by combining voltage-sensitive dye and laser speckle
contrast methods
Tihomir P. Obrenovitch a,⁎, Shangbin Chen b, Eszter Farkas a,c
a
b
c
Division of Pharmacology, School of Life Sciences, University of Bradford, Bradford, BD7 1DP, UK
Britton Chance Center for Biomedical Photonics – Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, Wuhan, China
Department of Physiology, School of Medicine, University of Szeged, Szeged, Hungary
a r t i c l e
i n f o
Article history:
Received 17 September 2008
Revised 5 November 2008
Accepted 13 November 2008
Available online 6 December 2008
a b s t r a c t
Cortical spreading depression (i.e. waves of cellular depolarization, CSD) causes the aura symptoms in
classical migraine, and may contribute to delayed cellular damage after an ischemic or traumatic insult to the
brain. In the latter cases, secondary neuronal injury may be worsened by some of the cerebral blood flow
(CBF) changes that are associated with CSD. Here, we describe a new method for the simultaneous, live
imaging of local cellular depolarization and CBF changes (i.e. two variables with well-defined and important
biological significance), through a closed cranial window prepared in anesthetized rats. This novel
experimental strategy was validated by imaging the changes associated with CSD elicited by application of
high K+ medium on the cortical surface. CSD was visualized directly by using a fluorescent voltage-sensitive
(VS) dye, whereas laser speckle contrast (LSC) imaging allowed the visualization of the corresponding CBF
changes. In addition to the high temporal and spatial resolution of VS dye and LSC imaging, their novel
combination allows to determine how CBF changes relate to a heterogeneous and evolving pattern of cellular
depolarization, in any area of interest of the cortical region under study. This methodological development is
especially pertinent and timely for investigations into the peri-lesion depolarizations that occur in models of
focal brain injury, situations where their site of spontaneous elicitation and propagation pattern cannot be
predicted. It should also help advance our knowledge in epilepsy, CBF pharmacology, and neurovascular
coupling under normal and pathophysiological conditions.
© 2008 Elsevier Inc. All rights reserved.
Introduction
Cortical spreading depression (CSD) is a transient disruption of
cellular homeostasis (i.e. depolarization of both neurons and glia) that
self-propagates across the cerebral cortex at a rate of 2–5 mm/min (for
review, see Martins-Ferreira et al., 2000). As normal brain function
requires physiological transmembrane ionic gradients, transient
suppression of electrical activity (i.e. EEG silence) is a feature of CSD.
This neurological abnormality is attracting increasing attention. Firstly
because it is finally established that CSD causes the aura symptoms
(e.g. visual illusions) in classical migraine (Lauritzen, 1994; Goadsby,
2007). Secondly, because spontaneous, transient depolarization-like
events, some very similar to CSD, were demonstrated repeatedly in
animal models of focal ischemia (for review, see Hossmann, 1996). In
addition, subdural electrocorticography has recently provided clear
evidence of slow potential changes in the acutely injured human
brain, with spreading depression of electrocorticographic activity
typical of CSD (Strong et al., 2002; Dreier et al., 2006; Fabricius et al.,
⁎ Corresponding author. Fax: 44 1274 233363.
E-mail address: t.obrenovitch@bradford.ac.uk (T.P. Obrenovitch).
1053-8119/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.neuroimage.2008.11.025
2006; Dohmen et al., 2008). Convincing experimental evidence
suggested that such events did not merely reflect a progression of
the initial lesion, but actually favoured its expansion (Busch et al.,
1996; Dijkhuizen et al., 1999), and emerging clinical data tend to
support this notion. In patients with aneurysmal subarachnoid
hemorrhage, the combination of electrocorticography with serial
neuroimaging demonstrated that the occurrence of delayed ischemic
stroke was time-locked to a cluster of recurrent prolonged CSD in
every single case (Dreier et al., 2006), suggesting that clusters of
prolonged CSD indicate ischemic lesion progression. Patients with
malignant hemispheric stroke showed also a high incidence of
prolonged CSD clusters (Dohmen et al., 2008).
Recently, we have developed and validated a method for direct, live
imaging of CSD in animal models, using a voltage-sensitive (VS) dye
(Farkas et al., 2008). We have demonstrated that the CSD-associated
changes of VS dye fluorescence with time are equivalent to the
electrical signature of CSD (i.e. extracellular direct current or DC
potential shift). Here, we describe the coupling of the VS dye method
to laser speckle contrast (LSC) imaging, a well validated method to
image local cerebral blood flow (CBF) changes with a high temporal
and spatial resolution (Ayata et al., 2004; Briers, 2007; Paul et al.,
Author's personal copy
T.P. Obrenovitch et al. / NeuroImage 45 (2009) 68–74
2006; Strong et al., 2006). This is a pertinent and timely methodological development for the aforementioned fields of study, because
the deleterious effects of CSD may be secondary to the blood flow
changes that are associated with this event. Indeed, in the otherwise
normal brain, CSD produces complex CBF changes that may lead to the
migraine headache (Ayata et al., 2004; Bolay et al., 2002; Fabricius et
al., 1995). In the injured brain, a long-lasting hypoperfusion following
the electrical signal of CSD may be detrimental to the survival of
vulnerable regions by lowering their compromised blood supply
further (Dreier et al., 1998; Strong et al., 2007). There is also evidence
that, in contrast to the CSD-associated hyperemia that occurs
consistently in otherwise physiological conditions (Fabricius et al.,
1995), cortical depolarization can be associated with a rapid reduction
of local CBF (similar to a vasospasm) under some pathological
conditions, possibly resulting from increased resistance to flow at
microvascular level (Dreier et al., 2004; Shin et al., 2006; Strong et al.,
2007).
Materials and methods
Surgical preparation
Animal procedures were authorized by the Ethical Review Panel of
the University of Bradford, and carried out in accordance with the
British Home Office Animals (Scientific Procedures) Act 1986. Surgical
procedures were similar to those reported earlier (Farkas et al., 2008).
Briefly, adult male Sprague–Dawley rats (310–340 g; Harlan UK Ltd.,
Bicester, UK) were anaesthetized with 1.5–2.0% halothane in N2O:O2
(2:1) during the entire surgery and data acquisition period. The animal
body temperature was kept between 37.1 and 37.4 °C using a heating
pad feedback-controlled by a rectal probe. A cannula was inserted in a
femoral vein to allow the termination of each experiment by cardiac
arrest (i.v. injection of 1 ml air). The animal was placed in a stereotactic
frame, and a cranial window (approximately 4 × 5 mm) opened by
removing the parietal bone. This was carried out by gradual thinning
of the skull with an electrical drill, followed by careful removal of the
remaining bone layer. A doughnut shape ring of acrylic dental cement
was built around the craniotomy, incorporating a perfusion inlet and
outlet, and a glass capillary connected to a syringe pump (CMA/100,
CMA/Microdialysis, Stockholm, Sweden). The latter was used to elicit
CSD by ejection of 1 μl 1 M KCl in 1.2 s directed selectively towards a
small area of the posterior cortical surface. Note that we did not
observe any detectable effect of the ejected K+ beyond the target area,
with either VS dye or LSC imaging, presumably because the high K+
medium was instantly and sufficiently diluted away from the target
site. Once prepared, the chamber was filled with artificial cerebrospinal fluid (aCSF; mM concentrations: 126.6 NaCl, 3 KCl, 1.5 CaCl2, 1.2
MgCl2, 24.5 NaHCO−3, 6.7 urea, 3.7 glucose, bubbled with 95% O2 and 5%
CO2), and the dura carefully dissected before sealing the cranial
window with a glass coverslip glued to the dental cement ring. Unless
otherwise stated, aCSF was continuously perfused through the
chamber at 25 μl/min using a peristaltic pump.
Live imaging of CSD with a fluorescent VS dye
Cortical tissue loading of the VS dye (RH-1838, Optical Imaging
Ltd., Rehovot, Israel) was as previously described (Farkas et al., 2008):
i.e. the dye solution was circulated through the cortical chamber for
2 h at 80 μl/min, with a subsequent 30–35-min aCSF-rinsing preceding
the experimental procedure. For live imaging, the cortical area under
study was illuminated in stroboscopic mode (100 ms illumination
each second) with a high-power light emitting diode (LED) (625 nm
peak wavelength; SLS-0307-A, Mightex Systems, Pleasanton, CA)
equipped with an excitation filter (620–640 nm bandpass; 3RD620640, Omega Optical Inc. Brattleboro, VT). Images of VS dye fluorescence (emission filter, bandpass 3RD 670–740; Omega Optical Inc.)
69
were captured with a monochrome CCD camera (Pantera 1M30,
DALSA, Gröbenzell, Germany) attached to a stereomicroscope
(MZ12.5, Leica Microsystems, Wetzlar, Germany) via a video tube
replacing the binocular piece (Fig. 1). The overall magnification was
×3.15 and the field of view size 3.8 × 3.8 mm. The maximum resolution
of the 1M30 Dalsa camera was 1024 × 1024 pixel (i.e. 1 × 1 binning
mode), but the VS dye images were captured using the 2 × 2 binning
mode to increase the sensitivity of the VS dye fluorescence detection
by a factor of 4, which implied a reduction of image resolution to
512 × 512 pixel. An external trigger synchronized the illumination and
camera exposure (i.e. 100 ms exposure time; 1 frame/second), and
6600 consecutive images corresponding to approximately 110 min of
recording were captured in each experiment.
Laser speckle contrast (LSC) imaging
For LSC imaging, the same cortical area was illuminated with a
laser diode (Sanyo DL7140-201S; 70 mW; 785 nm emission wavelength) driven by a power supply (ITC502, Thorlabs Ltd., Cambridge,
UK) set to deliver a 100-mA current. The raw laser speckle images
were captured by a second CCD camera, identical to that used for VS
dye imaging, and attached to the same stereomicroscope by using a
1:1 binocular/video-tube beam splitter (Fig. 1). As for VS dye imaging,
an external trigger synchronized the laser illumination and camera
exposure (1 frame/second; 20 ms for both illumination and exposure
time). After the experiments, individual LSC images were computed
from the corresponding raw images, as previously described (Cheng et
al., 2003), using MATLAB (The MathWorks Inc., Natick, MA) or ImagePro-Plus (Media Cybernetics UK, Marlow, UK) software. Laser Speckle
Contrast Analysis (LASCA) is the mathematical transformation that
needs to be applied to each raw laser speckle image to obtain the
corresponding CBF image. The first step of LASCA requires the
computation of two separate images, generated from the raw speckle
image, by calculating the mean and standard deviation of the grey
levels within adjacent small regions (i.e. 5 × 5 pixel areas in our study).
As such image processing implies a loss of data resolution, the camera
used for LSC imaging was set to 1 × 1 binning mode to achieve maximal
initial resolution (i.e. 1024 × 1024). This high resolution was also
necessary for appropriate speckle detection on the raw images (i.e.
single speckle size was slightly larger than individual camera pixels).
Note that the image resolution was still 1024 × 1024 after LASCA
analysis, because the computations on 5 × 5 pixel regions were carried
out by ‘sliding’ this window size across the two dimensions of the
image by 1 pixel increments. Ultimately, CBF maps were calculated
from the LSC images as 1 / (LSC-image)2 (Cheng and Duong, 2007).
These final images (1024 × 1024) were converted to 512 × 512 pixel
format to allow direct comparisons with the VS dye images, and
application of the same data extraction method to the two sets of
images (i.e. VS dye and LSC images) (see the Kinetics of changes in VS
dye fluorescence and LSC perfusion measured in selected AOI section).
Equipment and strategy for dual VS dye/LSC imaging
Using two separate cameras mounted to the same stereomicroscope as shown in Fig. 1 allowed their sensitivity and related spatial
resolution to be optimized for the two variables, by using different
binning modes (see above). As the cameras were identical, and
attached to a high quality beam splitter, a very good match was
achieved in term of field of view, magnification/image size and
resolution, between the VS dye and LSC images. Each camera was
connected to a dedicated computer, and controlled by Image-Pro-Plus
via a Camera Link board (Phoenix, PHX-D24CL; Active Silicon Ltd.,
Uxbridge, UK). External triggering was carried out by a third computer
equipped with a data converter board (Metrabyte DAS-20, controlled
by ASYST Macmillan software; Keithley Instruments Inc., Reading, UK)
supplying 2 digital/analog (D/A) outputs. Two TTL signals were
Author's personal copy
70
T.P. Obrenovitch et al. / NeuroImage 45 (2009) 68–74
generated, synchronized at one pulse per second, but out of phase in
such a way that the LSC image capture immediately followed that of
the VS dye image. The pulse duration of the VS dye camera/LED was
100 ms, and that for the LSC camera/laser diode 20 ms.
Protocol for CSD elicitation and data analysis
After 10 min of image acquisition that was used as baseline, 4
consecutive CSD were elicited, each followed by 20 min of recovery. In
our preparation, this interval was much larger than the refractory
period, and it allowed a sufficient recovery of the resting cellular
membrane potential from the preceding CSD. However, note that the
recovery was too short for CBF to return to its initial baseline. No
spontaneous CSD occurred with this protocol. Experiments were
terminated by cardiac arrest induced by injection of 1-ml air embolus
through the venous line.
To determine the evolution of local changes in VS dye fluorescence
and CBF with time, selected areas of interest (AOI) of 3 × 3 pixel size
were positioned manually on the images. They were carefully located
to avoid any overlap with superficial blood vessels. In each experiment, any given AOI was used for the extraction of both VS dye and
CBF kinetics. ‘Scanning’ all the images of a sequence allowed the
extraction of all the averaged grey levels from the corresponding AOI
sequence, ultimately used to plot the changes in this variable with
time. For VS dye CSD imaging, each kinetic of changes in fluorescence
was corrected for dye bleaching, by subtraction of a computed drift
(i.e. 5th-degree polynomial curve fitting, calculated for each AOI, by
using as input the sequence of average grey levels from first frame of
data capture to the last frame before cardiac arrest, after removal of
the CSD ‘peaks’ from the data sequence).
Only the first 2 CSD (CSD1 and CSD2) elicited by high K+ were
analysed in details, because the pattern of CSD1 was found
consistently different from that of the subsequent CSD, whereas all
the latter had a quite similar pattern. For both CSD1 and CSD2, the
mean kinetics of changes in VS dye fluorescence from several
experiments (n = 5 or 6) were calculated after conversion of the
original grey values to percentage, taking the baseline before CSD
elicitation as 0%, and the CSD peak amplitude as 100%. For LSC
imaging, conversion of grey values to percentage was as follows: The
average of the signal during the last minute of the recording (after
cardiac arrest) was taken as 0%, whereas the average of the first
minute of the baseline (prior to CSD elicitation) was taken as 100%. All
corresponding data are given as mean ± standard deviation.
Results and discussion
Methodological considerations and alternative strategies
Experimental design and equipment
Our primary goal was to achieve dual image acquisition of CSDassociated changes in VS dye fluorescence (for live imaging of cellular
depolarization) and LSC imaging of CBF, with identical temporal and
spatial resolution. Obviously, a custom designed optical system such
as that described by Ratzlaff and Grinvald (1911) would be optimal,
but financial considerations prompted us to assemble our system
around a versatile research stereomicroscope (Figs. 1 and 2). Several
elements allowed us to design a simple and efficient dual illumination system for fluorescence and LSC imaging, by relying on direct,
stroboscopic illumination of the cortical window: (i) Availability of
powerful LED, with emitting peak wavelength corresponding closely
to that of the VS dye excitation, and which could be driven within
20 μs by a computer-controlled power supply; (ii) Availability of a
laser diode, emitting in the near-infrared to favour good tissue
penetration and large individual speckle size (Briers, 2007), which
could be driven also in strobe mode; and (iii) the well-separated
excitation and emission wavelength of the VS dye (RH-1838), together
with extremely efficient band-pass filters (i.e. very sharp cut-on and
cut-off transparency for both excitation and emission filters), which
allowed the complete block of any reflected excitation light (Fig. 1).
This simple setup was found to deliver an even illumination of the
preparation, which could be easily synchronized with camera
exposure without any electro-mechanical component (i.e. shutter or
filter-wheel).
Acquisition of VS dye and LSC images — single-camera versus
two-camera system
Providing the laser diode emission wavelength falls within the
transmission range of the band-pass (or long-pass) filter used for VS
dye imaging, a single camera may be used to capture, alternatively,
both the VS dye fluorescence and LSC images (Fig. 2). In our case, this
would have implied using an emission filter with a high-cut-off
beyond 785 nm (i.e. the emission wavelength of the laser diode that
was used), or keeping the same band-pass filter (690–740 nm) but
using a diode emitting a laser wavelength within the filter transmission window (e.g. 735 nm; QLD-735-10S, QPhotonics, Ann Harbour,
MI). The latter combination was tested in our laboratory and worked
satisfactorily (data not shown). The 2-camera system offers the
following advantages: (i) Separate optimization of the camera settings
(gain, binning, exposure) — 2 × 2 binning is especially appropriate for
VS dye imaging as it increases the camera sensitivity by a factor of 4
(see next subsection on VS dye bleaching); (ii) Synchronization and
control of illumination and camera exposure are easier and more
convenient, as the external trigger may be used only for the initiation
of illumination and camera exposure (i.e. their duration remains
‘programmable’) – with the single-camera system, the camera
specifications must allow the exposure to be initiated and terminated
by the trigger pulses rising- and falling-edge, respectively (Fig. 2); and
(iii) VS dye and LSC images are taken virtually simultaneously – with
the single-camera system, enough time must be allowed for the
camera-to-computer data transfer, before initiating the next exposure
(note the different trigger signals in Figs. 1 and 2). Using a singlecamera system provides different advantages: (i) Perfect spatial match
between the VS dye and LSC images — with the 2-camera system, a
slight mismatch is inherent to the mechanical tolerance of the sensor
position within the camera and/or to the video tubes alignment
relative to the optical path (Fig. 1); (ii) the stereomicroscope/video
tube assembly may be replaced by a macro-lens, which is likely to
provide the system with a larger, overall numerical aperture; and (iii)
lower equipment costs.
Bleaching of VS dye fluorescence and longer experiments
In this study, an experimental procedure of around 2 h was ample
for imaging multiple K+-induced CSD. Over this period, with our
illumination parameters and 1 image/second frame rate, the basal
level of the VS dye fluorescence fell from ∼ 1500 to 1000 grey levels;
i.e. only ∼ 67% of the initial VS dye fluorescence remained after 2 h of
imaging. Complementary experiments with different LED illumination
patterns and intensities showed that this signal loss was due to the
dye fluorescence bleaching rather than to its dissociation from the
cellular membranes. Clearly, one pertinent application of the
methodology described herein would be the study of peri-lesion
depolarizations in animal models of acute brain injury, for which
much longer experimental procedures would be required. As dye
bleaching is directly related to the intensity and duration of the
excitation light, VS dye bleaching may be reduced by the following: (i)
Reduction of the image capture rate — since CSD is a relatively slow
phenomenon, the acquisition of one paired-image set each 4 s would
be still adequate; (ii) optimization of the optics to increase the
numerical aperture (i.e. amount of light reaching the camera sensor);
and (iii) high camera sensitivity (e.g. 2 × 2 binning of adjacent pixels;
high camera gain). Minimizing the LED illumination used for VS dye
imaging would also reduce the risk of phototoxicity, which may occur
Author's personal copy
T.P. Obrenovitch et al. / NeuroImage 45 (2009) 68–74
71
DC potential signatures of propagating CSD at different cortical depth,
but the corresponding barrier for vertical CSD propagation was
between 800 and 1200 μm deep, i.e. further below the layers explored
by both LSC and VS dye.
As briefly mentioned above, we used a 785 nm laser diode, firstly
because this is the most commonly used wavelength for LSC imaging
of CBF (presumably because a thicker cortical layer is explored with
this near-infrared illumination), and secondly because we wanted to
maximize the speckle size. Indeed, LSC imaging with a digital camera
requires the speckle size to be larger than the camera individual
pixels, and the DALSA 1M30 has large size pixels (12 μm) to optimize
its sensitivity and dynamic range. However, as the speckle size
depends also on the optic numerical aperture, one could use the 2camera system (Fig. 1) with a 630 nm laser diode for LSC imaging to
ensure a perfect match between VS dye and LSC imaging.
Live imaging of cellular depolarization waves and associated local blood
flow changes
Fig. 1. Diagram of the strategy and equipment assembly that were used for dual imaging
of cellular depolarization and associated local blood flow changes, using a voltagesensitive (VS) dye and the laser speckle contrast (LSC) imaging technique, respectively.
The upper and central part of the figure shows the optics and illumination systems. The
central piece of equipment was a versatile, modular stereomicroscope that allowed the
same field of view to be projected on the sensors of two identical CCD cameras, by using
a 1:1 beam splitter fitted with appropriate video tubes. The lower part and sides of the
figure depict the external triggering strategy used to synchronize the exposure of the
cameras to the respective illumination.
even in the absence of chemically-induced photosensitization (Jou
et al., 2004), especially given the illumination power that can be
achieved with collimated LED.
Exposure of the cortical surface — a necessary feature of the method
Although LSC imaging of CBF can be carried out through the skull
(Ayata et al., 2004; Li et al., 2006), exposure of the cortex is required
for VS dye tissue loading. This feature implies the following: (i) The
temperature of the exposed cortical surface (i.e. the region under
study) is slightly lower than when the skull is intact — in our
experiments, measurements of the temperature of the perfusion
medium in the chamber with a miniature thermocouple gave
consistent readings of 35 °C (without any external temperature
control, besides the heating pad used to maintain the animal's body
temperature at 37 °C); and (ii) the exposed cortex is continuously
bathed by an oxygenated, ‘physiological’ medium — in some models,
this may help the tissue to cope with, or recover from, transient
insults. Data from a separate series of experiments suggested that,
with the experimental conditions used for the present study, the
cortex was slightly more susceptible to CSD elicitation than an
equivalent, close skull preparation.
Potential mismatch between the depth of cortex contributing to the
VS dye and LSC signals
The depth of penetration of the light used to illuminate the cortex
decreases with its wavelength. According to Nielsen et al. (2000), the
recording of CBF changes with laser light is effective down to an
approximate 500 μm depth from the cortical surface with a 780 nm
laser, but only down to half this depth with a 543 nm illumination. In
our study, a 785 nm laser diode was used for LSC, and a 620–640 nm
band-pass filter for VS dye excitation; therefore, a slight mismatch
between the cortical layers contributing to the LSC and VS dye signals
cannot be ruled out. This could become relevant if the depolarization
waves under study propagate differently among the different cortical
layers explored. Richter and Lehmenkühler (1993) recorded different
Direct, live imaging of CSD with a VS dye was recently developed
and validated in our laboratory (Farkas et al., 2008). We demonstrated
that it provided the imaging equivalent of electrical extracellular DC
potential recording, with waves of cellular depolarization translating
directly to fluorescence increase. The main advantage of the VS dye
signal is its direct coupling to the primary biological change of CSD (i.e.
cellular depolarization), in contrast to intrinsic optical signal (IOS)
imaging which relies on changes in variables that are altered as a
consequence of CSD (e.g. local vasodilation, tissue oxygen availability;
Ba et al., 2002). However, multiwavelength IOS imaging may provide
useful, complementary information (Brennan et al., 2007).
This new study, for which some parameters were altered (e.g.
images taken at a rate 1, instead of 2 per second; 3.15× overall
magnification, instead of 2×) provided image sequences of CSD of the
same quality, and with the same effectiveness (Fig. 3, images Ax). On
these pictures, the waves of cellular depolarization appeared as a
propagating increase in tissue fluorescence, with maximal change
detected for CSD1 by the CCD camera as a grey level shift of 142.9± 34.0
(n = 6); i.e. 8.10 ± 1.8% of baseline fluorescence measured prior to CSD1,
Fig. 2. Alternative system, using a single camera for the capture of both voltagesensitive (VS) dye and laser speckle contrast (LSC) images. The upper, central part of the
figure shows the optics and illumination systems, whereas the lower part and sides of
the figure depict the external triggering of the LED, laser diode and camera. Note that,
with this system, 3 separate TTL signals are required, and that the square pulses for VS
dye and LSC image capture control must be out-of-phase sufficiently to allow time for
the transfer and storage of the captured image to the computer to which the camera is
linked.
Author's personal copy
72
T.P. Obrenovitch et al. / NeuroImage 45 (2009) 68–74
and a change that corresponds to around 3.5% of the 12 bit grey level
camera range (0–4095).
LSC imaging was previously used to examine the CBF changes
associated with CSD or CSD-like events, but it was either used on its
own (Ayata et al., 2004; Strong et al., 2006) or combined with the
imaging of inherent optical signals (Dunn et al., 2003; Wang et al.,
2007). As the latter relied on the characteristic optical absorption of
oxy- and deoxy-hemoglobin, they were essentially indicative of tissue
oxygenation and, therefore, not directly linked to the initial phenomenon (i.e. cellular depolarization).
The most important methodological development emerging from
the present study is that each VS dye image can be associated with a
paired, LSC image of tissue perfusion, thus allowing the correlation
between local cellular depolarization and its corresponding effect on
CBF, in any AOI of the cortical region under study. The spatial
resolution and signal-to-noise ratio of the LSC images that we
obtained was satisfactory and equivalent to those reported previously
(Ayata et al., 2004; Wang et al., 2007). The LSC images might be
improved by using a higher frame rate of the camera, and temporal
LSC (tLSC) instead of conventional (spatial) LSC imaging. With tLSC,
the statistics for LSC image analysis are carried out primarily in the
temporal direction, using a sliding set of images captured sequentially,
whereas the conventional analysis is carried out within each image of
the sequence, on adjacent unit areas (5 × 5 pixels in our case). We did
not test the more advanced, temporal approach (Cheng et al., 2003).
Kinetics of changes in VS dye fluorescence and LSC perfusion measured in
selected AOI
Depolarization waves and associated changes in local CBF were
investigated together in several studies, under both normal and
pathophysiological conditions. An early attempt to compare the onset
of CSD and that of the associated CBF response in the otherwise
normal cortex was made by Hansen et al. (1980); in this case,
extracellular K+ was monitored using ion-selective microelectrodes,
and autoradiographic mapping of blood flow changes obtained with
brain sections prepared after rapid freezing of the brain in situ.
Subsequently, extracellular DC (or K+ concentration) and laser Doppler
flowmetry (LDF) signals were recorded simultaneously, either at
different sites of the cortex (Gold et al., 1998) or within the same area
(Obrenovitch et al., 2004). With regard to models of cerebrovascular
diseases, Dreier et al. (2000) used a surface electrode and LDF to
examine the inverse neurovascular coupling during CSD in a
preparation simulating the situation after subarachnoid hemorrhage,
whereas Shin et al. (2006) combined intracortical microelectrode and
LSC in a stroke model. None of these studies allowed investigators to
correlate cellular depolarization and associated local CBF changes
within several cortical areas, and at multiple time points.
The novelty and power of the technology presented here lies in the
fact that changes in cellular membrane potential and corresponding
vascular responses can be monitored together with a high time and
Fig. 3. Dual imaging of cortical spreading depression (CSD)-related changes in membrane potential and cerebral blood flow (CBF). Representative image sequence of changes in
voltage sensitive (VS) dye fluorescence intensity (A1–A7), and corresponding changes in flow velocity as assessed with laser speckle contrast (LSC) imaging (B1–B7). For the VS dye
images, subtraction of the background (average of the 3 first frames of the sequence) preceded contrast enhancement. The same image treatment was applied for the flow velocity
images, except that the “static flickering” was smoothed by running average on 15 consecutive frames. Note that the images displayed were not selected at regular intervals. (A),
contrasted picture of the cerebral cortex fluorescence and (B), contrasted average of 15 flow velocity maps, taken shortly after VS dye loading. In (A), the dotted hemi-circle delineates
the site of CSD elicitation, and the arrow indicates the direction of CSD propagation. Note the delayed flow velocity response (indicating hyperemia) relative to the maximum
brightness of the VS dye signal, and the diffuse flow velocity signal in contrast to the sharp CSD wave front shown by the VS dye.
Author's personal copy
T.P. Obrenovitch et al. / NeuroImage 45 (2009) 68–74
73
Fig. 4. Kinetics of local cellular depolarization and blood flow changes associated with cortical spreading depression (CSD), recorded by voltage-sensitive (VS) dye fluorescence (A)
and laser speckle contrast (LSC) imaging (B), respectively. Each trace is an average (mean ± stdev) of signals obtained from separate experiments (n = 6 for CSD1, and n = 5 for CSD2). In
each experiment, the VS dye and LSC imaging data were extracted from the same selected area of interest (AOI), and the VS dye signal corrected for fluorescence bleaching (i.e. 5thdegree polynomial curve fitting). All the kinetics are expressed as relative values: For the VS dye, as % of the maximal peak amplitude; for the flow velocity, as % of the initial baseline
after subtraction of the 0 flow level recorded after cardiac arrest. The traces from separate experiments were synchronized by aligning the points of maximal rate of depolarization of
the VS dye signal, with the flow velocity then remaining ‘locked’ to its respective VS dye signal. The vertical, dotted lines fitted to the peak of each VS dye signal help appreciate the
temporal relationship between the two variables. The short time bar between (A) and (B) is the time window corresponding to the pictures presented in Fig. 3.
2D-spatial resolution. In practice, we found that relevant sets of
quantitative data can be assembled conveniently as follows: (i) Visual
scanning of the image sequence assembled as a short video; (ii)
through this examination of the overall data, selection of one or
multiple AOI; (iii) use of these AOI as probes to extract the kinetics of
grey level changes within these selected regions. This is how the data
assembled in Fig. 4 were obtained. This capability (i.e. post-selection
of AOI for the extraction of quantitative data) will be especially useful
for the application of this dual imaging technique to animal models of
neurological conditions where the elicitation site and propagation
pattern of depolarization waves cannot be predicted (e.g. peri-lesion
depolarization in models of focal ischemia and traumatic brain injury).
The multiphasic pattern of CBF changes associated with CSD (Fig.
4B) was essentially the same as those reported previously. For CSD1,
this included: (i) Small, transient decrease immediately preceding the
hyperemia; (ii) marked hyperemia well synchronized with the
depolarization phase of the VS dye signal; and (iii) a sustained,
secondary hypoperfusion. Only the magnitude of the initial, brief CBF
decrease appears to be different (i.e. smaller) than in previous
experiments carried out with close skull preparation and LDF
(Obrenovitch et al., 2004). With CSD2, as observed previously, there
was no longer any further decrease of CBF relative to the level prior to
CSD, either before or after the hyperemia (Fig. 4B). It is important to
note that, in comparison to the dual DC potential/LDF monitoring, the
VS dye/LSC imaging combination provides a perfect match for the two
variables under study, in both space and time — again, this is a
pertinent feature for studies with brain ischemia models, where the
most relevant cortical locations cannot be pre-determined.
The combination of VS dye and LSC imaging is especially suitable
for the analysis of the temporal relationship between depolarization
and corresponding CBF response, at any selected cortical site. For
example, in this study, the delay between onset of CSD1 (determined
from the VS dye signal) and the associated CBF response was
measured at two areas of interest (AOI1 and AOI2) that were located
2.0 and 2.5 mm from the K+ trigger, respectively. The delay of onset at
these two sites was 21.3 ± 5.9 and 13.5 ± 5.5 s (n = 6; p b 0.04), at AOI1
and AOI2, respectively, which suggests that the delay between
depolarization and corresponding CBF response may not be consistent
throughout the cortex, and possibly decrease with increasing distance
from the elicitation site.
Conclusions and potential applications of dual VS dye/LSC imaging
Simultaneous VS dye fluorescence and LSC imaging of the cortex
in animal models allows the 2-dimensional recording of information
on local cellular depolarization and tissue perfusion; i.e. two variables
with well-defined and important biological significance. These
variables are important, individually: (i) Cellular depolarization,
because depolarization waves contribute to the pathophysiology of
common neurological disorders (see Introduction), and also because
normal brain function involves partial depolarization of brain cells
and their processes; (ii) Local blood supply, because the brain is
entirely dependent on its continuous oxygen supply for normal
function, and for the survival of regions that have been subjected
to ischemic or traumatic insults. The capability to image these
variables together is novel, timely, and particularly suitable for
further investigation into the important field of neurovascular
coupling (Lauritzen, 2001), currently the fundamental basis of neuroimaging in humans (Lauritzen and Gold, 2003).
The potential applications of this new technology are multiple and
important. In addition to classical migraine and the genesis of cellular
damage after an acute insult to the brain, it should help advance our
knowledge in epilepsy, CBF pharmacology, and neurovascular coupling. With regard to the latter, this novel imaging strategy will be
Author's personal copy
74
T.P. Obrenovitch et al. / NeuroImage 45 (2009) 68–74
especially appropriate for further investigation into ‘inverse coupling’
between cortical depolarization and CBF; i.e. the atypical vascular
response that was observed when depolarization events were superimposed to inhibition of nitrergic vascular relaxation (Dreier et al.,
2004), or associated with focal ischemia (Shin et al., 2006; Strong
et al., 2007).
Finally, applying the principle of dual imaging with a single camera
(Fig. 2) to the 2-camera system described in Fig. 1 would make it
possible to monitor a third variable, in addition to cellular depolarization (VS dye) and tissue perfusion (LSC imaging). Pertinent, complementary variables would be tissue pH (e.g. with neutral red; Chen
et al., 1999), and its degree of oxygenation and vascular volume (using
the optical properties of hemoglobin; Chen et al., 2006; Shin et al.,
2007).
Acknowledgments
This study was supported by grants from the Wellcome Trust
(079430/A/06/Z) and The Royal Society (International Incoming Short
Visits, 2007/R3). This work would not have been possible without the
contribution from D. Brown, G. Pearson and A. Kershaw (Workshop,
Bradford School of Life Sciences).
References
Ayata, C., Shin, H.K., Salomone, S., Ozdemir-Gursoy, Y., Boas, D.A., Dunn, A.K., Moskowitz,
M.A., 2004. Pronounced hypoperfusion during spreading depression in mouse
cortex. J. Cereb. Blood Flow Metab. 24, 1172–1182.
Ba, A.M., Guiou, M., Pouratian, N., Muthialu, A., Rex, D.E., Cannestra, A.F., Chen, J.W.,
Toga, A.W., 2002. Multiwavelength optical intrinsic signal imaging of cortical
spreading depression. J. Neurophysiol. 88, 2726–2735.
Bolay, H., Reuter, U., Dunn, A.K., Huang, Z., Boas, D.A., Moskowitz, M.A., 2002. Intrinsic
brain activity triggers trigeminal meningeal afferents in a migraine model.
Nat. Med. 8, 136–142.
Brennan, K.C., Beltrán-Parrazal, L., López-Valdés, H.E., Theriot, J., Toga, A.W., Charles,
A.C., 2007. Distinct vascular conduction with cortical spreading depression.
J. Neurophysiol. 97, 4143–4151.
Briers, J.D., 2007. Laser speckle contrast imaging for measuring blood flow. Opt. Appl. 37,
139–152.
Busch, E., Gyngell, M.L., Eis, M., Hoehn-Berlage, M., Hossmann, K.A., 1996. Potassiuminduced cortical spreading depressions during focal cerebral ischemia in rats:
contribution to lesion growth assessed by diffusion-weighted NMR and biochemical imaging. J. Cereb. Blood Flow Metab. 16, 1090–1099.
Chen, G., Hanson, C.L., Dunbar, R.L., Ebner, T.J., 1999. Novel form of spreading
acidification and depression in the cerebellar cortex demonstrated by neutral red
optical imaging. J. Neurophysiol. 81, 1992–1998.
Chen, S., Li, P., Luo, W., Gong, H., Zeng, S., Luo, Q., 2006. Time-varying spreading
depression waves in rat cortex revealed by optical intrinsic signal imaging.
Neurosci. Lett. 396, 132–136.
Cheng, H., Duong, T.Q., 2007. Simplified laser-speckle-imaging analysis method and its
application to retinal blood flow imaging. Opt. Lett. 32, 2188–2190.
Cheng, H., Luo, Q., Zeng, S., Chen, S., Cen, J., Gong, H., 2003. Modified laser speckle
imaging method with improved spatial resolution. J. Biomed. Opt. 8, 559–564.
Dijkhuizen, R.M., Beekwilder, J.P., van der Worp, H.B., Berkelbach van der Sprenkel, J.W.,
Tulleken, K.A., Nicolay, K., 1999. Correlation between tissue depolarizations and
damage in focal ischemic rat brain. Brain Res. 840, 194–205.
Dohmen, C., Sakowitz, O.W., Fabricius, M., Bosche, B., Reithmeier, T., Ernestus, R.I.,
Brinker, G., Dreier, J.P., Woitzik, J., Strong, A.J., Graf, R., Co-Operative Study of Brain
Injury Depolarisations (COSBID), 2008. Spreading depolarizations occur in human
ischemic stroke with high incidence. Ann. Neurol. 63, 720–728.
Dreier, J.P., Körner, K., Ebert, N., Görner, A., Rubin, I., Back, T., Lindauer, U., Wolf, T.,
Villringer, A., Einhäupl, K.M., Lauritzen, M., Dirnagl, U., 1998. Nitric oxide scavenging
by hemoglobin or nitric oxide synthase inhibition by N-nitro-L-arginine induces
cortical spreading ischemia when K+ is increased in the subarachnoid space.
J. Cereb. Blood Flow Metab. 18, 978–990.
Dreier, J.P., Ebert, N., Priller, J., Megow, D., Lindauer, U., Klee, R., Reuter, U., Imai, Y.,
Einhäupl, K.M., Victorov, I., Dirnagl, U., 2000. Products of hemolysis in the
subarachnoid space inducing spreading ischemia in the cortex and focal necrosis
in rats: a model for delayed ischemic neurological deficits after subarachnoid
hemorrhage? J. Neurosurg. 93, 658–866.
Dreier, J.P., Tille, K., Dirnagl, U., 2004. Partial antagonistic effect of adenosine on inverse
coupling between spreading neuronal activation and cerebral blood flow in rats.
Neurocrit. Care 1, 85–94.
Dreier, J.P., Woitzik, J., Fabricius, M., Bhatia, R., Major, S., Drenckhahn, C., Lehmann, T.N.,
Sarrafzadeh, A., Willumsen, L., Hartings, J.A., Sakowitz, O.W., Seemann, J.H., Thieme,
A., Lauritzen, M., Strong, A.J., 2006. Delayed ischaemic neurological deficits after
subarachnoid haemorrhage are associated with clusters of spreading depolarizations. Brain 129, 3224–3237.
Dunn, A.K., Devor, A., Bolay, H., Andermann, M.L., Moskowitz, M.A., Dale, A.M., Boas,
D.A., 2003. Simultaneous imaging of total cerebral hemoglobin concentration,
oxygenation, and blood flow during functional activation. Opt. Lett. 28, 28–30.
Fabricius, M., Akgoren, N., Lauritzen, M., 1995. Arginine–nitric oxide pathway and
cerebrovascular regulation in cortical spreading depression. Am. J. Physiol. 269,
H23–H29.
Fabricius, M., Fuhr, S., Bhatia, R., Boutelle, M., Hashemi, P., Strong, A.J., Lauritzen, M.,
2006. Cortical spreading depression and peri-infarct depolarization in acutely
injured human cerebral cortex. Brain 129, 778–790.
Farkas, E., Pratt, R., Sengpiel, F., Obrenovitch, T.P., 2008. Direct, live imaging of cortical
spreading depression and anoxic depolarization using a fluorescent, voltagesensitive dye. J. Cereb. Blood Flow Metab. 28, 251–262.
Goadsby, P.J., 2007. Recent advances in understanding migraine mechanisms, molecules
and therapeutics. Trends Mol. Med. 13, 39–44.
Gold, L., Back, T., Arnold, G., Dreier, J.P., Einhäupl, K.M., Reuter, U., Dirnagl, U., 1998.
Cortical spreading depression-associated hyperemia in rats: involvement of
serotonin. Brain Res. 783, 188–193.
Hansen, A.J., Quistorff, B., Gjedde, A., 1980. Relationship between local changes in
cortical blood flow and extracellular K+ during spreading depression. Acta Physiol.
Scand. 109, 1–6.
Hossmann, K.A.,1996. Periinfarct depolarizations. Cerebrovasc. Brain Metab. Rev. 8,195–208.
Jou, M.J., Jou, S.B., Guo, M.J., Wu, H.Y., Peng, T.I., 2004. Mitochondrial reactive oxygen
species generation and calcium increase induced by visible light in astrocytes. Ann.
N. Y. Acad. Sci. 1011, 45–56.
Lauritzen, M., 1994. Pathophysiology of the migraine aura. The spreading depression
theory. Brain 117, 199–210.
Lauritzen, M., 2001. Relationship of spikes, synaptic activity, and local changes of
cerebral blood flow. J. Cereb. Blood Flow Metab. 21, 1367–1383.
Lauritzen, M., Gold, L., 2003. Brain function and neurophysiological correlates of signals
used in functional neuroimaging. J. Neurosci. 23, 3972–3980.
Li, P., Ni, S., Zhang, L., Zeng, S., Luo, Q., 2006. Imaging cerebral blood flow through the
intact rat skull with temporal laser speckle imaging. Opt. Lett. 31, 1824–1826.
Martins-Ferreira, H., Nedergaard, M., Nicholson, C., 2000. Perspectives on spreading
depression. Brain Res. Rev. 32, 215–234.
Nielsen, A.N., Fabricius, M., Lauritzen, M., 2000. Scanning laser-Doppler flowmetry of rat
cerebral circulation during cortical spreading depression. J. Vasc. Res. 37, 513–522.
Obrenovitch, T.P., Wang, M., Urenjak, J., Butler, M.J., Dreier, J.P., 2004. The role(s) or nitric
oxide during cortical spreading depression. In: Krieglstein, J., Klumpp, S. (Eds.),
Pharmacology of Cerebral Ischemia. Medpharm, Stuttgart, pp. 157–166.
Paul, J.S., Luft, A.R., Yew, E., Sheu, F.S., 2006. Imaging the development of an ischemic
core following photochemically induced cortical infarction in rats using Laser
Speckle Contrast Analysis (LASCA). NeuroImage 29, 38–45.
Ratzlaff, E.H., Grinvald, A., 1911. A tandem-lens epifluorescence macroscope: hundredfold brightness advantage for wide-field imaging. J. Neurosci. Methods 36, 127–137.
Richter, F., Lehmenkühler, A., 1993. Spreading depression can be restricted to distinct
depths of the rat cerebral cortex. Neurosci. Lett. 152, 65–68.
Shin, H.K., Dunn, A.K., Jones, P.B., Boas, D.A., Moskowitz, M.A., Ayata, C., 2006.
Vasoconstrictive neurovascular coupling during focal ischemic depolarizations.
J. Cereb. Blood Flow Metab. 26, 1018–1030.
Shin, H.K., Dunn, A.K., Jones, P.B., Boas, D.A., Lo, E.H., Moskowitz, M.A., Ayata, C., 2007.
Normobaric hyperoxia improves cerebral blood flow and oxygenation, and inhibits
peri-infarct depolarizations in experimental focal ischaemia. Brain 130, 1631–1642.
Strong, A.J., Fabricius, M., Boutelle, M.G., Hibbins, S.J., Hopwood, S.E., Jones, R., Parkin,
M.C., Lauritzen, M., 2002. Spreading and synchronous depressions of cortical activity
in acutely injured human brain. Stroke 33, 2738–2743.
Strong, A.J., Bezzina, E.L., Anderson, P.J., Boutelle, M.G., Hopwood, S.E., Dunn, A.K., 2006.
Evaluation of laser speckle flowmetry for imaging cortical perfusion in experimental stroke studies: quantitation of perfusion and detection of peri-infarct
depolarisations. J. Cereb. Blood Flow Metab. 26, 645–653.
Strong, A.J., Anderson, P.J., Watts, H.R., Virley, D.J., Lloyd, A., Irving, E.A., Nagafuji, T.,
Ninomiya, M., Nakamura, H., Dunn, A.K., Graf, R., 2007. Peri-infarct depolarizations
lead to loss of perfusion in ischaemic gyrencephalic cerebral cortex. Brain 130 (Pt 4),
995–1008.
Wang, Z., Li, P., Luo, W., Chen, S., Luo, Q., 2007. Peri-infarct temporal changes in intrinsic
optical signal during spreading depression in focal ischemic rat cortex. Neurosci.
Lett. 424, 133–138.