2.1.

ECoG–LSCI System

An ECoG–LSCI platform integrating ECoG recording and LSCI technology was used to simultaneously measure neuronal activity and full-field rCBF, respectively (Fig. 1). For the ECoG signal recording, three stainless steel epidural electrodes, including two electrodes placed on the S1FL and one reference electrode, were secured on the skull to acquire resting-state ECoG (RS-ECoG) using a multichannel data acquisition system (Blackrock Microsystems, Salt Lake, Utah) prior to peripheral stimulation. The ECoG signals were recorded through a head-stage amplifier (gain of 2) and filtered by a 0.5 to 7500-Hz band-pass filter. The signals were then digitized at a sampling rate of 1 kHz using the filter settings, and a 100-Hz low-pass filter was applied to compare changes in intercortical functional connectivity over time after PTI stroke with different peripheral stimuli. rCBF was assessed using an in-house LSCI system. Coherent light provided by a continuous-wave (CW) laser module (660 nm; 100 mW; RM-CW04-100, Unice E-O Service Inc., Taoyuan, Taiwan) was used to illuminate the region of interest (ROI). The laser beam was expended with a planoconvex lens ($f=75\mathrm{mm}$, LA1608-A, Thorlabs Inc., Newton, New Jersey) to a size of $\sim 40\times 30\mathrm{mm}$ to provide even illumination of the exposed area of the cortex. The illuminated area was imaged using a 16-bit charge-coupled device (CCD) camera ($1032\times 776\text{pixels}$; pixel size: $4.65\times 4.65\mu \mathrm{m}$, DR2-08S2M/C-EX-CS, Point Grey Research Inc., Richmond, British Columbia, Canada) via an adjustable magnification lens (0.3 to $1\times$, $f/4.5$ max) with a $2\times$ extender. To satisfy the Nyquist sampling criterion and to maximize the contrast of the imaged speckle pattern, we changed the f-stop setting of the lens to modulate the speckle size to five times the pixel pitch ($4.65\mu \mathrm{m}$) of the CCD camera.²⁸ To eliminate scattering, a linear polarizer was placed before the CCD image collection lens, and the working distance was $\sim 5\mathrm{cm}$. Laser speckle images were acquired at 30 fps, and the exposure time was fixed at 10 ms for in vivo experiments. The LSCI system was controlled by a custom LabVIEW program (National Instruments, Austin, Texas). A graphics processing unit (GPU) was also introduced into our LSCI data processing framework to achieve real-time, high-resolution blood flow visualization on a PC, which is a parallel computing platform and programming model invented by NVIDIA (Santa Clara, California) that allows for dramatic increases in computing performance by harnessing the power of the GPU (GeForce GTX 650 Ti, NVIDIA, Santa Clara, California). The rCBF of the selected ROI was acquired throughout the experiment, and changes in the rCBF of the cortical arteries and cortical veins were calculated and averaged. Nevertheless, some technical limitations of the ECoG–LSCI system still exist. For instance, bilateral ECoG recordings of the S1FL cortex through the skull-secured screw electrodes were near the imaging window of the S1FL region but could not be implanted directly in the LSCI imaging field of the S1FL to prevent interference with the optical results.

Fig. 1

Sketch of the ECoG–LSCI system and experimental setup. Schematic diagram of the ECoG–LSCI system for investigating cortical functional changes in rats. The setup included (1) ECoG recordings, (2) LSCI imaging, (3) PTI stroke in the cerebral cortex, and (4) peripheral electrical stimulation. The animal was fixed in a custom-made stereotaxic frame, and its temperature was maintained by a self-regulated heating pad. ECoG signals were recorded through two bilaterally secured epidural electrodes and one electrode (at $+2\mathrm{mm}$ posterior to the lambda) on the skull and were amplified via a front-end amplifier (lower right inset). For LSCI, an open-skull cranial window was made over the S1FL and S1HL regions (lower right inset), and a collimated beam of a CW 660-nm laser was adjusted to evenly illuminate the cortical area of interest. The CCD camera was positioned over the exposed area with a slightly larger field of view than the craniotomy site, and the data were interconnected with a personal computer via a USB interface for further image analysis. For induction of PTI, a CW 532-nm laser was also integrated and focused on the selected distal MCA arteriole of the S1FL region of rose bengal-injected rats for 15 min. Contralateral peripheral electrical stimulation was delivered using a DS3 stimulator.

2.2.

Animal Preparation and Surgery

All procedures were performed in accordance with approvals from both the Institutional Animal Care and Use Committee (IACUC) of the National Health Research Institute (NHRI) and the National Yang Ming University, Taiwan. To compare the rescue effects of different peripheral stimulations on the ischemia model, a total of 45 male adult Sprague Dawley rats ($N=45$) weighing 300 to 350 g were randomly allocated to three groups, including sham control, forelimb stimulation, and hindlimb stimulation groups. All animals were housed in an environment with a 12-h dark/light cycle at constant temperature and humidity and allowed free access to food and water.

The rats were anesthetized with pentobarbital ($50\mathrm{mg}/\mathrm{kg}$) by intraperitoneal (IP) injection, and supplemental anesthesia boosters (1/3 amount of the initial dose) were given as necessary. In this study, we aimed to examine the changes in the normalized CBF in the presence of sensory stimulation; however, pentobarbital is known to partially suppress the CBF and the cerebral blood volume of rats.²⁹ The depth of anesthesia before the surgery was confirmed by checking the response of the hindlimb withdrawal reflex. After presurgical preparation (anesthesia and hair shaving), the rats were fixed on a custom-made acrylic stereotaxic holder to reduce motion artifacts during the experiment, and the skin and muscles were removed from the skull to expose the bregma and the lambda as reference points. As shown in Fig. 1, two types of experimental setups, the ECoG recording system and the LSCI setup, were used in the experiment to simultaneously evaluate the physiological changes before and after PTI stroke induction. In the ECoG recording setup, two epidural electrodes were bilaterally fixed at the S1FL ($\mathrm{AP}=+1.0\mathrm{mm}$, $\mathrm{ML}=\pm 4.5\mathrm{mm}$), and a reference electrode was fixed 2 mm posterior to the lambda on the skull (lower right inset, Fig. 1). In the LSCI setup (lower right inset, Fig. 1), a $5\times 2.5{\mathrm{mm}}^2$ cranial window was made while keeping the dura intact in the right hemisphere (diagonal coordinates of anterior-posterior [AP]: $+3.5\mathrm{mm}$ and medial-lateral [ML]: $+2\mathrm{mm}$; AP: $-1.5\mathrm{mm}$ and ML: $+4.5\mathrm{mm}$ to the bregma), which at the intersections of the MCA and the ACA were associated with the S1FL and the S1HL.¹⁵

2.3.

Photothrombotic Ischemic Stroke

In this study, a branch of the MCA within the S1FL on the right hemisphere near the coordinates of $+0.25\mathrm{mm}$ AP and $+3.5\mathrm{mm}$ ML to the bregma was targeted with focal PTI stroke. The target arteriole was selected based on its appearance and position within the anterior segment of the forelimb somatosensory map.³⁰ To induce occlusion, rose bengal photosensitizer (${\mathrm{Na}}^{+}$ salt, R3877; Sigma-Aldrich Corp., St. Louis, Missouri) diluted to $10\mathrm{mg}/\mathrm{ml}$ in HEPES-buffered saline was injected through the tail vein of the anesthetized rat at $30\mathrm{mg}/\mathrm{kg}$ body weight. Within 2 min after injection, a surface arteriole ($80\pm 5\mu \mathrm{m}$ in diameter) was illuminated for 15 min using green-light laser (beam diameter: 0.4 mm; 532 nm, 8.3-mW; LEDP1-G_MM400-0.48_1m_FC_M8; Doric Lenses Inc., Quebec, Canada) with a $2\times$ optical laser lens beam expander for occlusion until a stable clot was formed. The diameter of the illumination zone for PTI was 0.8 mm with an average calculated light intensity of $67.5\mathrm{mW}/{\mathrm{cm}}^2$.

2.4.

Peripheral Electrical Stimulation

We followed and modified the parameters of our published study.¹² In brief, each rat underwent RS-ECoG and rCBF baseline measurements prior to PTI stroke. After 15-min PTI stroke, each rat sequentially underwent 10 ECoG–LSCI recording sessions with 15-min intervals. Peripheral electrical stimulation was followed by 210 s of ECoG–LSCI recording for each session, except for the 10th session. Representative time courses of the peripheral electrical stimulation during the 10 consecutive ECoG–LSCI recording sessions are shown in Fig. 2.

Fig. 2

The timeline of peripheral stimulation and ECoG–LSCI recording. A schematic diagram of the experimental protocol indicating the timing of PTI induction (pink box), ECoG recordings and LSCI imaging for 210 s (blue boxes) and sensory electrical stimulation (yellow boxes) (upper). The ECoG and LSCI signals were first recorded before PTI stroke as basal levels and further recorded every 15 min after PTI stroke (from poststroke 15 to 150 min). The sensory electrical stimulation paradigm, as shown in the middle panel, consists of a stimulation train with an amplitude of 3 mA, a pulse width of $200\mu \mathrm{s}$, a frequency of 5 Hz, a duration of 40 s, and a duration interval of 40 s (lower). Estimation of the infarct volume using TTC staining was performed 24 h after PTI stroke induction.

Two needle electrodes were subcutaneously inserted into a contralateral forepaw or hindpaw. The limbs were stimulated by applying trains of rectangular pulses of $200\mu \mathrm{s}$ duration at a repetition rate of 5 Hz supplied with a DS3 isolated current stimulator (Digitimer Ltd., Welwyn Garden City, Hertfordshire, United Kingdom). The maximum current amplitude was 3 mA in all stimulations. The stimulation paradigm of a single session consisted of five blocks. Each block employed 40 s of stimulation, which contained 200 stimulation pulses and between 40-s rest periods.

2.5.

Data Analysis for LSCI

The ROIs for speckle flow index (SFI) images were chosen for the areas of PTI with diameters of 0.8 mm and positions of illumination centered at AP: $+0.25\mathrm{mm}$ and ML: $+3.5\mathrm{mm}$ to the bregma, S1FL (diagonal coordinates of AP: $+0.5\mathrm{mm}$ and ML: $+4.25\mathrm{mm}$; AP: $-1.5\mathrm{mm}$ and ML: $+4.5\mathrm{mm}$) and S1HL (diagonal coordinates of AP: $+0.25\mathrm{mm}$ and ML: $+2.75\mathrm{mm}$; AP: $-1.5\mathrm{mm}$ and ML: $+2.25\mathrm{mm}$). The speckle images were computed with the algorithm based on the spatial statistics of the imaged speckle pattern running in a LabVIEW environment.^31,32 The spatial laser speckle contrast, $K$, is given by

To quantify the rCBF changes at different time points of post-PTI stroke in both the ACA and MCA regions, a normalized ratio of rCBF (${\mathrm{rCBF}}_{\mathrm{N}}$) from the SFI images was defined as

The linear correlation of resting SFI fluctuations between the ACA and MCA regions at each frequency was evaluated by a magnitude-squared coherence function given by

In this study, the two SFI signals were coherent at the frequency band $(f)$ between 0.05 and 0.15 Hz where the coherence values were larger than 0.5.^38,39 Therefore, the phase of $P_{\mathrm{AM}}(f)$ at frequencies between 0.05 and 0.15 Hz was used to indicate the relative lag between the coherent components, which was calculated to describe the temporal relationship of SFI between the ACA and the MCA regions and is defined by

2.6.

Resting State Functional Connectivity Based on Bilateral ECoG Recording

The coherence was previously utilized to analyze dynamic changes between the ECoG time series.^41,42 This approach was also used in this study to investigate the correlated relationship between time courses in bilateral S1FL cortical regions. The RS-ECoG signals were measured from the time points 15 min before PTI stroke to 150 min after stroke. The signals were filtered using a Butterworth band-pass filter with low- and high-cut off frequencies of 0.5 and 50 Hz. The temporal fluctuations of the ECoG signals from the left and right cortices were calculated for the delta (1 to 5 Hz), theta (5 to 8 Hz), alpha (8 to 15 Hz), and beta (15 to 30 Hz) frequency bands. MATLAB^® was used to analyze the ECoG recordings offline.

In this study, the temporal changes in intercortical functional connectivity after PTI stroke were confirmed by serial RS-ECoG recordings. In line with previous findings,^12,43 ECoG signals in the low-frequency delta band (1 to 5 Hz) were significantly attenuated compared with before PTI stroke. Therefore, the averaged magnitude-squared coherence was used to measure the intercortical functional connectivity,⁴⁴ which was calculated using a 210-s ECoG signal over the delta band across bilateral S1FL cortical regions. Here, the magnitude-squared coherence was calculated by

A normalized ratio of ${\mathrm{Coh}}_{\mathrm{RL}}^2(f)$ is further defined by

2.7.

Quantification of Infarct Volume

Twenty-four hours after PTI, the rats were deeply anesthetized by 10% chloral hydrate ($100\mathrm{mg}/\mathrm{kg}$, IP injection) and perfused with cold normal saline, and their brains were rapidly removed. We followed the experimental protocol from our published study to measure the ischemic infarct.¹² Each brain was rinsed with normal saline, frozen at $-20°\mathrm{C}$ for 10 min, sliced into $500\text{-}\mu \mathrm{m}$ coronal sections using a rat brain slicer (Zivic Instruments, Pittsburgh, Pennsylvania), and further stained with 2% 2,3,5-triphenyltetrazolium chloride (TTC; T8877; Sigma-Aldrich Corp., St. Louis, Missouri) solution at 37°C for 20 min. The slices were then washed twice with normal saline and fixed with 10% formalin. After fixation, the images of the stained brain slices were examined using ImageJ (National Institutes of Health, Bethesda, Maryland), and the corrected infarct volumes in the slices were qualified by the absence of staining. The corrected infarct volume = {[total lesion volume − (nonischemic hemisphere volume—ischemic hemisphere volume)] / ischemic hemisphere volume} × 100%.^45–47

2.8.

Statistical Analysis

All data were averaged and expressed as the $\text{mean}\pm \text{standard error}$ (SEM). Statistical analysis was performed using PASW Statistics, version 18 (SPSS, Chicago, Illinois). To compare time series changes in the intercortical functional connectivity (${\mathrm{CR}}_{\mathrm{inter}}$) and ACA–MCA interarterial regulation via rCBF coherence and its corresponding phase delay time, the nonparametric Wilcoxon signed-rank test was conducted to test for treatment effects with time. $p<0.05$ was considered statistically significant.

Following different peripheral electrical stimuli, differences in infract volume across groups were compared using the nonparametric Kruskal–Wallis test. A probability value of $p<0.05$ was used as the criterion to determine statistical significance.

3.1.

Comparison of Normalized rCBF Ratio Changes Before and After Focal PTI Stroke in the Presence of Peripheral Stimulation

The cranial window was formed while maintaining the dura intact to study the spatial distribution of the SFI in the rat cortex, as shown in Fig. 3(a). In this study, LSCI provided two-dimensional mappings of overall blood perfusion within the brain [Fig. 3(b)]. The time course of changes in cerebral vascular morphologies was obtained before and after PTI stroke. The MCA and the ACA as well as their branches were clearly detected. Baseline SFI images were taken precisely 15 min before PTI stroke induction. Following PTI induction, blood flow decreased most prominently in the core region at the distal MCA arteriole, as shown in Fig. 3(b). Therefore, to compare the rescue effects on focal PTI stroke with different peripheral stimuli, cortical regions and subtleties of dynamic SFI responses were evaluated before, during, and after PTI. We found that the ${\mathrm{rCBF}}_{\mathrm{N}}$ in the sham control group was significantly decreased after PTI in a time-dependent manner compared to the basal ${\mathrm{rCBF}}_{\mathrm{N}}$ level before PTI onset due to clot formation [**$p<0.01$, Wilcoxon signed-rank test, $N=15$, Figs. 3(b) and 4(a)]. Interestingly, increases in rCBF of the penumbra areas, such as the S1FL area and the S1HL area, were shown in the first measurement (Tb15) after PTI induction but were absent in the PTI area, i.e., the ischemic core. This result may arise from the collateral circulation mediated by ischemia per se.¹⁶ Nevertheless, the ${\mathrm{rCBF}}_{\mathrm{N}}$ of the PTI area of forelimb-stimulated rats markedly recovered to $0.67\pm 0.13$ compared to the other two groups from 90 min after stroke (*$p<0.05$, Wilcoxon signed-rank test, $N=15$), as shown in Figs. 3(b) and 4(a). However, the ${\mathrm{rCBF}}_{\mathrm{N}}$ values of the PTI induction area of the sham control and hindlimb-stimulated groups exhibited continual time-dependent decreases, with ${\mathrm{rCBF}}_{\mathrm{N}}$ values of $0.35\pm 0.08$ and $0.43\pm 0.07$, respectively, from 90 min after stroke [Figs. 3(b) and 4(a)]. As shown in Figs. 4(b) and 4(c), the ${\mathrm{rCBF}}_{\mathrm{N}}$ values of the S1FL and S1HL areas in the forelimb-stimulated group were significantly increased to $1.06\pm 0.25$ and $0.99\pm 0.14$, respectively, compared to those of the sham control group of $0.73\pm 0.08$ and $0.70\pm 0.07$, respectively, from 75 min after PTI stroke (**$p<0.01$, Wilcoxon signed-rank test, $N=15$). However, there was no difference in ${\mathrm{rCBF}}_{\mathrm{N}}$ between the sham control and hindlimb-stimulated groups. These results indicated that only forelimb stimulation restored rCBF in the PTI induction area, S1FL area, and S1HL area after PTI stroke.

Fig. 3

The speckle contrast images for rCBF in the presence of peripheral stimuli. (a) Open-skull photograph of the surface of the rat brain taken prior to PTI stroke. Dorsal view of a white light image indicates the three selected ROIs, including the PTI area, the S1FL, and the S1HL, over the primary somatosensory (S1) cortex. The stroke target for the S1FL cortex of PTI stroke (1.5 mm) and location of green laser illumination (centered at AP: $+0.25\mathrm{mm}$ and ML: $+3.5\mathrm{mm}$ to bregma) is shown as a dashed red circle. The dotted green box shows an ROI of the S1FL without illumination. The dashed blue box represents an ROI of the S1HL. The forelimb and hindlimb receptive fields in the S1 cortex are represented as translucent green and yellow maps, respectively.⁴⁸ The ACA–MCA collaterals are located at the interface of the distal branches of these two major supply vessels.⁴⁹ The ACA–MCA anastomosis is denoted as the white dashed line. $\text{Scale bar}=1\mathrm{mm}$. (b) Spatiotemporal SFI evolution of the in vivo blood flow of the rat cortex showing the contralateral rCBF changes in the sham control group and upon electrical stimulation in the forelimb- and hindlimb-stimulated groups at serial time points (every 15 min). $\text{Scale bar}=5\mathrm{mm}$. Tb indicates the time before PTI stroke.

Fig. 4

The rCBF was enhanced by forelimb stimulation-induced collateral circulation. The time-series ${\mathrm{rCBF}}_{\mathrm{N}}$ changes in (a) the PTI area, (b) the S1FL area, and (c) the S1HL area upon different peripheral stimuli following PTI induction. The symbol “*” indicates significant differences ($p<0.05$, Wilcoxon signed-rank test) compared with controls at each time point. Data are represented as the $\text{mean}\pm \mathrm{SEM}$. $N=15$ for each group. Tb indicates the time before PTI stroke.

3.2.

Comparison of ACA–MCA Interarterial Anastomotic Regulation After Focal PTI in the Presence of Peripheral Stimulation

To characterize the interarterial anastomotic regulation of the ACA and MCA territories of the ischemic penumbra before and after the administration of peripheral stimuli, we investigated the coherence [${\mathrm{Coh}}_{\mathrm{AM}}^2(f)$] and phase delay [$\mathrm{\Delta }t(f)$] of oscillations in rCBF at very low frequencies ($f=0.05$ to 0.15 Hz) between the ACA and MCA territories. As shown in Fig. 5, no differences in ACA–MCA ${\mathrm{Coh}}_{\mathrm{AM}}^2(f)$ and $\mathrm{\Delta }t(f)$ over time were observed between the sham control and hindlimb-stimulated groups before and after PTI stroke. However, the ACA–MCA ${\mathrm{Coh}}_{\mathrm{AM}}^2(f)$ was significantly higher in the forelimb-stimulated group than in the other groups at 75 to 150 min after PTI stroke (**$p<0.01$, Wilcoxon signed-rank test, $N=15$). This finding revealed that the ${\mathrm{Coh}}_{\mathrm{AM}}^2(f)$ was significantly enhanced in the presence of forelimb stimulation, as it was not enhanced in the sham control or hindlimb-stimulated groups. Furthermore, there was an obvious increase in ACA–MCA phase lead [$\mathrm{\Delta }t(f)>0$] in the forelimb-stimulated group compared to the other two groups at 75 to 150 min after PTI stroke (##$p<0.01$, Wilcoxon signed-rank test, $N=5$), indicating that rCBF oscillations of the ACA territory led to oscillations of the MCA territory. Taken together, the results suggested that only the forelimb stimulation enhanced the hemodynamic coherence of the ACA and MCA areas, and the CBF oscillations of the ACA preceded the CBF oscillations of the MCA.

Fig. 5

Time course analysis of ${\mathrm{Coh}}_{\mathrm{AM}}^2$ and the time of phase delay between the ACA and MCA territories with different peripheral stimuli. The changes in SFI-squared coherence (bar charts) and phase delay (red solid line) functions for the sham control and the forelimb and hindlimb stimulation groups between the ACA and the MCA are plotted in 15 min intervals from the baseline recording (before PTI induction) to 150 min after PTI stroke. The red dashed line indicates the SFI time series with zero phase delay between the ACA and MCA areas. A positive phase delay indicated that SFI oscillations in the ACA led those in the MCA. The symbols “**” and “##” indicate significantly different means ($p<0.01$, Wilcoxon signed-rank test) compared with controls at each time point. Data are represented as the $\text{mean}\pm \mathrm{SEM}$. $N=15$ for each group. Tb indicates the time before PTI stroke.

3.3.

Comparison of Intercortical Functional Connectivity After PTI Stroke in the Presence of Peripheral Stimulation

To elucidate the intercortical connectivity in the presence of peripheral electrical stimulation, we recorded ECoG and analyzed the normalized intercortical delta band coherence (${\mathrm{CR}}_{\mathrm{inter}}$). As shown in Fig. 6, a significant decrease in ${\mathrm{CR}}_{\mathrm{inter}}$ was observed at 15 min after PTI stroke compared with before PTI stroke [**$p<0.01$ vs. ${\mathrm{CR}}_{\mathrm{inter}}(T_b)$, Wilcoxon signed-rank test, $N=15$]. Similarly, the ${\mathrm{CR}}_{\mathrm{inter}}$ values in the hindlimb-stimulated group and the sham control group were also markedly decreased.

Fig. 6

Time course analysis of interhemispheric S1FL functional connectivity (RS-ECoG ${\mathrm{CR}}_{\mathrm{inter}}$) before and after PTI stroke. Comparison of changes in the normalized delta band coherence between bilateral S1FL regions with different peripheral stimuli following PTI induction. The results indicated that coherence in the delta bands between the bilateral S1FL regions significantly recovered upon timely administration of forelimb electrical stimulation. The symbols “*” and “**” indicate the significantly different means corresponding to $p<0.05$ and $p<0.01$ (Wilcoxon signed-rank test), respectively, compared with controls at each time point. Data are presented as the $\text{mean}\pm \mathrm{SEM}$. $N=15$ for each group. Tb indicates the time before PTI stroke.

In the forelimb-stimulated group, ${\mathrm{CR}}_{\mathrm{inter}}$ slowly recovered from 30 to 60 min post-PTI stroke followed by a significant improvement at 75 min post-PTI stroke compared with the ${\mathrm{CR}}_{\mathrm{inter}}$ values of the other groups at 75 min post-PTI stroke (*$p<0.05$, Wilcoxon signed-rank test, $N=15$). ${\mathrm{CR}}_{\mathrm{inter}}$ remained significantly increased up to 150 min post-PTI stroke, approaching $\sim 1.0$, compared with the other groups (**$p<0.01$, Wilcoxon signed-rank test, $N=15$). However, no statistically significant increases in ${\mathrm{CR}}_{\mathrm{inter}}$ were observed in the sham control and hindlimb-stimulated groups. Taken together, our data indicated that forelimb stimulation significantly and functionally enhanced the intercortical connectivity of the S1FL region.

3.4.

Comparison of Infarct Volume in the Presence of Peripheral Stimulation

The histology of the infract volume after PTI stroke was further evaluated by TTC staining. Representative TTC-stained brain coronal sections from each group are shown in Fig. 7(a). In the sham control rats, there was apparent damage in the S1FL area of the right hemisphere, and the infarction area extended from $\sim 0.25\mathrm{mm}$ anterior to 0.36 mm posterior to the bregma to a depth of 3 mm from the cortex surface, which was comparable to the hindlimb-stimulated rats. In contrast, the rats that immediately received forelimb stimulation after PTI stroke exhibited less brain damage with reduced infarction $\sim 0.25\mathrm{mm}$ anterior to the bregma with nearly absent damage in other slices.

Fig. 7

Analysis and quantification of infarct volume by TTC staining of the S1FL areas according to a rat brain atlas.⁵⁰ (a) Coronal sections of ischemic brain from the untreated control group (first row) and from groups that received forelimb (second row) or hindlimb (third row) stimulation after PTI stroke (the arrow indicates the infarct area). (b) Statistical analysis of infarct volume after PTI stroke. The symbol “*” indicates significantly different means ($p<0.05$, Kruskal–Wallis test) compared with the forelimb stimulation group. Data are presented as the $\text{mean}\pm \mathrm{SEM}$. $N=15$ for each group. The results showed that forelimb stimulation largely diminished the PTI stroke-induced infarct volume.

The statistical results revealed that PTI stroke animals that received forelimb stimulation exhibited the most significant reduction in infarct volumes by $\sim 44\%$ ($7.35\pm 0.23{\mathrm{mm}}^3$, *$p<0.05$, Kruskal-Wallis test, $N=15$) compared to the sham control rats ($17.05\pm 72.71{\mathrm{mm}}^3$) and the hindlimb-stimulated rats ($16.41\pm 2.21{\mathrm{mm}}^3$) [Fig. 7(b)].

4.1.

Contralateral Forelimb Stimulation Restores rCBF in the S1FL Region of PTI Stroke Rats

LSCI is a relatively inexpensive, fast, and noninvasive optical technique used to measure vascular blood flow by utilizing laser speckle to achieve higher spatial and temporal resolutions,^13,24,51 and it has been commonly applied to estimate CBF during functional stimulation and physiological fluctuations in the cortex of rodents.^24,52–54 Furthermore, LSCI has also been used to monitor cerebral hemodynamics at the scale of a single blood vessel with high spatial resolution via a cranial window⁵⁵ and to detect the development of anastomosis of collateral circulation in acute stroke rats.¹⁶ Accordingly, we measured rCBF of specific cortical areas, such as the S1FL and the S1HL, by LSCI and found that rCBF of the forelimb-stimulated group was significantly higher than rCBF of the control and hindlimb-stimulated groups after PTI stroke. These results suggest that early peripheral stimulation of target brain regions after ischemic stroke might preserve CBF of the infarct area via stimulated collateral circulation. These data are also consistent with a previous study that reported that functional stimulation of the sensory cortex largely reduced ischemic insult, including CBF.⁵⁶

4.2.

Blood Flow Pattern by LSCI Between the S1FL and S1HL Brain Regions

Figures 4(a)–4(c) show that rCBF of the S1FL and S1HL areas immediately increased during PTI stroke onset in all groups but dropped soon after PTI stroke in the control and hindlimb-stimulated groups. The rCBF of the S1FL and S1HL areas remained higher than basal levels only in the forelimb-stimulated group after PTI stroke until the end of the experiments (poststroke 150 min). These results demonstrated that forelimb stimulation not only increased rCBF in the PTI stroke area but also in the S1FL and S1HL areas, which might be because early forelimb stimulation facilitated collateral blood flow from neighboring areas, such as the ACA and the MCA, to the PTI stroke area. Therefore, rCBF of the S1FL and S1HL areas was higher than basal levels, and therefore, rCBF increases in these three areas in the forelimb-stimulated group were higher than in the control and hindlimb-stimulated groups.

4.3.

Peripheral Stimulation-Induced Interarterial Regulation of the ACA and MCA Areas

To investigate the mechanism of peripheral stimulation-induced collateral circulation, we monitored the coherence of blood perfusion oscillations and the phase delay time between the ACA and MCA watersheds, which might be involved in the blood perfusion change upon forelimb electrical stimulation. A previous study showed that interarterial anastomotic circulation was observed in normal rats at 29 ACA–MCA junctions per hemisphere and provides ACA collateral supply to the MCA tissue field.⁴⁹ Toriumi et al. further demonstrated that the T-junctions of interarterial anastomosis of the ACA and the MCA were dynamically relocated under physiological conditions, and retrograde blood flow appeared in the MCA branch after MCAo ischemia of rodents.⁵⁷ In addition, our previous study also indicated that early peripheral sensory forelimb stimulation improves the outcomes of ischemia by recruiting collateral circulation. Thus, we postulated here that the forelimb stimulation-enhanced collateral circulation was associated with interarterial anastomotic regulation to reduce the ischemic damage, and the collateral perfusion may be from the ACA to the MCA. As shown in Fig. 5, the results of this study strongly suggest that only the forelimb stimulation increased the coherence of ACA and MCA hemodynamics, as the phenomenon was absent in the sham control and hindlimb stimulation groups. Moreover, the delay time of rCBF signals at the low frequency range (0.05 to 0.15 Hz) of the ischemic penumbra between the ACA and MCA areas did not change in the presence of PTI at various time points in the sham control or hindlimb stimulation groups. However, it significantly increased in the forelimb-stimulated group from 75 to 105 min after PTI stroke, indicating that collateral perfusion of the ACA occurs prior to that of the MCA. Furthermore, the high ${\mathrm{Coh}}_{\mathrm{AM}}^2$ of the ACA and MCA areas in the forelimb-stimulated group also suggests a reliable phase relationship from the estimation. A previous study also indicated that collateral perfusion provides an alternative route for blood flow to reach ischemic tissue during stroke,⁵⁸ and therefore, forelimb stimulation-enhanced collateral circulation through interarterial anastomotic regulation might play an important role in rescuing brain tissue from ischemic insults. Taken together, these results suggest that forelimb stimulation-induced collateral circulation might be highly associated with ACA–MCA interarterial anastomotic regulation, which progressed from ACA to MCA territory, after PTI stroke.

4.4.

Peripheral Stimulation Enhances Intercortical Coherence and Rescues Brain Damage After PTI Stroke

A previous study showed that ECoG coherence represents the functional connectivity between brain regions and is interrupted after brain injury.⁵⁹ As shown in Fig. 6, the intercortical coherence of the S1FL region was largely diminished in all groups after PTI stroke, but it was enhanced only in the forelimb-stimulated group. Therefore, these results demonstrated that the functional recovery of brain damage of the S1FL region is achieved only by forelimb-sensory stimulation but not by hindlimb stimulation or in sham controls and that it might be associated with collateral circulation enhancement. The results of infarct volume after PTI stroke, as shown in Fig. 7, revealed that forelimb stimulation decreased infract volume compared to the control groups after PTI stroke. These results further corroborate the CBF and intercortical coherence data that only forelimb stimulation, not simulation of neighboring areas, significantly reduces the infarct volume of the S1FL region when delivered during the early stage of PTI in stroke rats and is consistent with previous studies that early peripheral stimulation exerts neuroprotective effects in stroke rats.^10–12

In contrast, several studies have shown that hypoperfusion of ischemia also induces repetitive spreading depression-like depolarization, called peri-infarct depolarization (PID), in the area surrounding the ischemic core.^60–62 Anoxic PID is triggered by the release of potassium and excitatory amino acids, such as glutamate, from the infarct core and is propagated at a rate of $\sim 2$ to $3\mathrm{mm}/\min$.⁶⁰ Nevertheless, von Bornstadt et al. revealed that mild tactile stimulation induces a supply-demand mismatch in the peri-infarct hot zone, known as the ischemic penumbra, and further causes stimulation-mediated PID, which worsens ischemic tissue outcome.⁶³ However, the effects of mild tactile stimulation-triggered PIDs cannot be directly deduced to other stroke models or species. Here, based on our data, we proposed that receptive field-specific electrical stimulation enhanced interarterial anastomotic collateral circulation and rescued region-specific brain activity and tissue damage. The favorable outcome of peripheral stimulation was likely related to enhanced glucose levels by stimulation-mediated collateral perfusion blood glucose support and glycogenolysis of glial cells.^64,65 However, the effects of sensory stimulation remain unclear and need to be further examined and confirmed in the future.

In summary, we measured CBF, blood flow phase changes, and intercortical coherence by using our superresolution LSCI and ECoG recording system and examined infarct volume by TTC staining of PTI stroke rats. Early-stage electrical forelimb stimulation facilitated CBF recovery, induced changes in the blood flow phase, and enhanced collateral circulation of the infarct area and the intercortical coherence of the S1FL, which were associated with functional recovery after PTI stroke. The infarct volume was also significantly decreased in the early forelimb-stimulated group after PTI stroke. Thus, electrical stimulation administered during the early stage after PTI stroke in rats improved functional recovery and reduced infarct volume in the brain and, thus, may be a potential therapeutic strategy for clinical ischemic stroke treatment.