In Vivo Voltage-Sensitive Dye Imaging of Subcortical Brain Function

The whisker system of rodents is an excellent model to study peripherally evoked neural activity in the brain. Discrete neural modules represent each whisker in the somatosensory cortex (“barrels”), thalamus (“barreloids”), and brain stem (“barrelettes”). Stimulation of a single whisker evokes neural activity sequentially in its corresponding barrelette, barreloid, and barrel. Conventional optical imaging of functional activation in the brain is limited to surface structures such as the cerebral cortex. To access subcortical structures and image sensory-evoked neural activity, we designed a needle-based optical system using gradient-index (GRIN) rod lens. We performed voltage-sensitive dye imaging (VSDi) with GRIN rod lens to visualize neural activity evoked in the thalamic barreloids by deflection of whiskers in vivo. We stimulated several whiskers together to determine the sensitivity of our approach in differentiating between different barreloid responses. We also carried out stimulation of different whiskers at different times. Finally, we used muscimol in the barrel cortex to silence the corticothalamic inputs while imaging in the thalamus. Our results show that it is possible to obtain functional maps of the sensory periphery in deep brain structures such as the thalamic barreloids. Our approach can be broadly applicable to functional imaging of other core brain structures.

Scientific RepoRts | 5:17325 | DOI: 10.1038/srep17325 pathways. In this system, neuronal modules, representing single whiskers on the snout, form at the brainstem (barrelettes), thalamic (barreloids), and neocortical (barrels) levels 20,21 . Because of their ease of access, identification and close to surface location, the barrel fields have been the subject of numerous imaging studies with a variety of approaches [22][23][24][25] . In contrast to imaging of the neocortex, functional mapping of the subcortical structures using optical methods was technically challenging and limited 26 . Currently, miniature endoscopic probes offer a solution for deep brain imaging by overcoming the limited depth of intravital microscopy 27 . Taking advantage of this, we built a functional imaging system, which combines VSDi with a 1 mm-GRIN rod lens to monitor temporal and spatial dynamics of neural activities in the thalamus following whisker stimulation. We imaged neural activity in the ventral posteromedial (VPM) nucleus of the mouse thalamus responding to single whisker stimulation. With this approach, we were able to obtain a functional map of the thalamic barreloids.

Results
Cortical signal recording. First we carried out an experiment to record the cortical signal corresponding to single whisker stimulation using our GRIN-rod-lens VSDi system. The purpose of this experiment was to confirm the performance of the needle-based imaging system. For imaging, the distal end of the GRIN rod lens was positioned above the recording area and directed such that its optical axis was perpendicular to the cranial window. The focusing plane was set to 300 μ m below the dural surface. As illustrated in Fig. 1(a), C2 whisker was stimulated for a period of 10 ms. The signal appeared 20 ms after the stimulus onset, and reached its peak at about 40 ms after the stimulus onset. After reaching the maximum value, the activated signals decreased until returning to baseline [ Fig. 1(b)]. The changes in activation area followed a pattern corresponding to the stimulation [ Fig. 1(c)]. Blood vessels were visualized due to different light absorption of the hemoglobin and the cortical tissue (labeled in red arrow in Fig. 1(c)). The results of cortical signal recording agreed well with the conventional VSDi system without GRIN rod lens, indicating that the GRIN-rod-lens imaging system is suitable for VSDi signal recording 24 . In addition, we imaged the cortex first with the conventional system and then the same area with the GRIN-rod-lens for comparison. As illustrated in Fig. 1(d), there was no notable difference in the boundary of the activity patterns.
Signal recording in the thalamus. Mapping VPM responses to stimulation of single whisker. Fig. 2(a) depictss C2 whisker stimulation. Fig. 2(b) shows changes in fluorescence (Δ F/F) in response to C2 whisker stimulation. Ipsilateral stimulation data was acquired as control. Fig. 2(c) shows the GRIN rod lens images of changes in fluorescence signals in response to contralateral C2 whisker stimulation. The signal appeared about 20 ms after the stimulus onset, and reached its peak at 55-60 ms after the stimulus onset. An interesting observation is that after reaching its maximum value, the activated areas started to spread at 65 ms after the stimulus onset (indicated by the red arrows in Fig. 2(c)). Another peak appeared at 95 ms, which may be due to the whisker swinging back after the 10-ms air puff stimulation. After 110 ms, the activated signals disappeared gradually.
Mapping responses to stimulation of multiple whiskers. To further investigate whether the needle-imaging system could differentiate multiple whisker-evoked responses in the VPM, we performed a three-whisker stimulation experiment. C2, D2 and E2 whiskers were stimulated at the same time [ Fig. 3(a)]. Ipsilateral stimulation data was acquired as control [ Fig. 3 Fig. 3(c) shows the signal appearance and diminishing in both activation area size and signal magnitude. At 50 ms, we can clearly see the three areas, which correspond to three different whiskers. In order to see the temporal response more closely, signals from the three "hot spots" areas [28][29][30] were plotted in Fig. 3(d). There was a clear 5 ms peak delay between one area and the other two that may be due to the different whiskers lengths. Fig. 4 shows the result of such an experiment in which whiskers B2 and E2 were stimulated at nearly the same time. B2 and E2 whiskers have a larger spatial separation on the snout. In Fig. 4(c), we can clearly see two separate responses. There was also a clear peak delay of 10 ms between these two responses.
Mapping responses to stimulation of multiple whiskers at different times. To further investigate whether the imaging system could differentiate the responses in the VPM when whiskers are stimulated at different times, we performed a time-difference stimulation experiment.
B2 and D2 whiskers were used for stimulation [ Fig. 5(a)]. Stimulation for B2 was set at 70 th frame (350 ms) and stimulation for D2 was set at 120 th frame (600 ms). The first response to B2 reached its maximum at 30 ms after B2 whisker stimulation, and the second response to D2 whisker reached its maximum at 35 ms after D2 whisker stimulation as shown in Fig. 5 Exploring the effect of corticothalamic inputs to activities in thalamic barreloids. We performed imaging experiments in the thalamus of 4 mice following muscimol injection into the barrel cortex. Blocking cortical activity abolishes VSD signal in the cortex but not in the VPM. This is exciting new results which allows one to compare thalamic activation in the absence of corticothalamic inputs or when the cortex is silenced. Fig. 6(a) shows wave plot of the changes in fluorescence (Δ F/F(%), ordinate) in response to C2 whisker stimulation of the same area (5 by 5 pixels) in cortex before and after muscimol injection. We can clearly see that the responses in the cortex were inhibited by muscimol. In the following experiment, we first inserted the GRIN rod lens to the VPM based on the same procedure described above. Then the fluorescence signals in VPM before and after muscimol injection in response to contralateral C2 whisker stimulation were acquired as shown in Fig. 6(c). Compared to the signal without silencing the cortex, we observed that the activation area is much wider and scattered. Fig. 6(b) shows changes in fluorescence (Δ F/F(%), ordinate) in response to C2 whisker stimulation before and after muscimol injection. Fluorescence signal was recorded from the small blue and green squares marked on the images shown in Fig. 6(c) at 30 ms. We can see the signal magnitude is relatively low compared to the signal magnitude before muscimol injection. To quantify the effect of corticothalamic inputs to activities in the thalamic barreloids, the signal amplitude and activated area during stimulation were determined as shown in Fig. 7(a). The group with muscimol injection consisted of 3 data sets from 3 different animals. Control experiments (without muscimol injection) were performed on 4 data sets from 4 different animals. The signal amplitude in the group with muscimol injection was statistically smaller compared with the control group (P = 0.0176). And activated area in the group with muscimol injection was nearly twice wide as the control group (P = 0.0041). We conclude that our data may provide evidence indicating that feedback from the cortex plays a crucial role in shaping thalamic responses 32 . Fig. 7(b) shows the GRIN rod lens track to the VPM in brain slice. Red color indicates RH-1691 dye fluorescence spread from the track of the GRIN probe. The positions of VPL,VPM and PO are labeled. While the GRIN rod lens causes a large area of damage to the brain, the barrel field of the somatosensory cortex and thalamocortical pathway are spared. This is mainly due to an angled approach to the VPM medially from the cortex and hippocampus. We can see the dye spread to VPM clearly and the GRIN   rod lens is right above the VPM, which indicates that the stereotaxic coordinates we used allowed for precise location of the VPM.

Discussion
Brain optical imaging has developed considerably within the last few decades 2 . Optical methods can offer both high spatial and temporal resolutions and are therefore particularly promising for measuring the hemodynamic, metabolic, and neuronal activity in vivo 7,8,13,33,34 . VSDi, which utilizes a high-speed CCD camera, offers an opportunity to study the activity of neuronal ensembles in vivo with relatively high spatial (up to 20 μ m) and temporal resolution (up to few milliseconds, which is comparable to electrophysiology) 35,36 . However, since CCD cameras integrate the back-scattered light from various depths, VSDi cannot detect depth-resolved functional activation 37 . In contrast with the single photon imaging, multi-photon microscopy has been used for functional neuronal imaging, and recent developments extend the penetration depths up to 1.6 mm in a mouse 33 . On the other hand, light penetration is fundamentally limited by scattering in the tissue. To overcome this limitation, several groups explored optical endoscopes. Miniature optical endoscopes are typically based on commercially available gradient-index (GRIN) rod lenses or imaging fiber bundles. GRIN rod lenses, which are typically 350-2,000 μ m in diameter, can provide relatively high resolution, and have been used in deep brain imaging with relatively little injury 27,[38][39][40][41][42][43] .
Many studies investigated information processing along the whisker-barrel system, using electrophysiological and morphological techniques. However, imaging studies in normal laboratory rodents and transgenic mice have been limited to the barrel cortex due to accessibility issues 24,35,[44][45][46][47] . Light penetration and scattering limits the usage of the optical imaging methods in subcortical structures. Our present in vivo results show that the combination of VSDi and GRIN optical probe can be utilized in imaging deep brain structures and can be adopted for use in freely moving animals through flexible imaging fiber bundle 48 . It is important to note the limitations of VSDi imaging at multiple levels of a sensory system. A notable one is the temporal resolution. The temporal resolution of the system we used is 5 ms, which is much longer than synaptic relays from the brainstem to cortex. Thus, we were not surprised that the latency of the VSDi signal, obtained from the VPM, and the time courses were similar for the optic signals recorded from the barrel cortex. In this regard, VSDi approach is not as sensitive as standard electrophysiological recording techniques because of the limited temporal resolution. The time differences between neural responses in the VPM and the barrel field can be just few milliseconds, which is shorter than our temporal resolution.
The VSDI signal and the physical relationship to barreloids in each experiment is not feasible at the moment. However if we express GFP or other fluorescent tags in barreloid neurons and activate them we might be able to pick up individual barreloids and specific activity patterns in relation to them. At the present time we can only compare the size of the activated areas in the VPM with respect to stained brain sections after imaging. Another limitation in imaging from deep brain structures is that the barreloids in the VPM are not all in the same plane from the tip of the GRIN-lens. The axes of the barreloids rows are not orthogonally located to the focal plane, and many of them are located slightly above or below the focus. Thus, the system we used does not have depth resolution for three-dimensional (3-D) neuronal maps. Lastly, the source of the fluorescence also has its own 3D structure located at an undetermined distance from the focal plane is projected as 2D optical patterns. 3D imaging through GRIN-rod-lens needle microscope is also feasible using advanced image reconstruction algorithms 49 . Nonetheless, the compact size and functionality of GRIN optical devices in combination with VSDi are enabling imaging deep brain structures and functions in the mammalian brain in vivo. Animal preparation. We imaged voltage-sensitive dye optical signals in eight adult mice (B6 male and female, 20-30 g body weight, age 5-10 weeks). All animals were anesthetized with urethane (1.15 g/kg) with body temperature maintained at 37 °C with a heating blanket. The animal's head was placed in a stereotaxic frame. For surgical preparation, the head was shaved, and a midline cutaneous incision was made. The skin over the skull was retracted, and a cranial window, 1.5 mm in diameter, was made using a dental drill 1 . A 10 μ L Hamilton syringe, outer needle diameter 0.3 mm, was used to inject 0.3-0.5 μ L voltage-sensitive dye RH-1691 (Optical Imaging Ltd, 1.0 mg/mL in the artificial cerebrospinal fluid (ACSF)) into the VPM using stereotaxic coordinates (− 1.7 posterior, 1.6 lateral and 3.1 mm below dura mater ) 50 over 5 minutes. The cranial opening was covered by a drop of high viscosity silicone oil to prevent drying and decrease brain pulsation 35 . Imaging was started 15-20 minutes after dye loading. Figure 8 illustrates the schematic of the needle-based VSDi imaging system.

Experimental setup.
The system is equipped with a 1-mm diameter GRIN rod lens (NA 0.113; Go Foton Corporation), which provides an imaging field-of-view (FOV) of ~0.9 mm and relays the image from the mouse brain at the distal end of the needle probe back to the focal plane of the objective (Leica Objective Planapo 2.0 × , M-series). The system utilizes a 637 nm laser diode as its light source, which is coupled with a single-mode fiber to shape its light beam. A diffuser is used to make the light more uniform. The light is collimated by an objective and goes through a dichroic mirror (650 nm, single edge dichroic beam splitter; FF650-DiO1-50 × 70 mm; Andover Corporation). A shutter is applied to control the excitation light state and avoid dye photobleaching. The light is then coupled to the GRIN rod lens by the microscope objective. A custom-built motorized 3D micro-stage facilitates accurate light coupling between the objective and GRIN rod lens. The emitted fluorescent light is collected back through the GRIN rod lens, objective, dichroic mirror, an emission filter (695 nm, 695FG07-50, Andover Corporation), and finally imaged on a high-speed CCD camera (MiCAM02-HR, SciMedia, Ltd). For deep tissue imaging, after aligning the GRIN rod lens to the objective, the GRIN-rod-lens probe was gently inserted into the anesthetized mouse thalamus by moving up the animal stage slowly and using stereotaxic coordinates 50 . Stimuli and data acquisition. Each experimental session consisted of 10-30 trials, with 200 frames per trial. Data acquisition rate was 5 ms/frame (200 Hz). For single stimulus experiments, the stimulus (whisker deflection) was presented at the 100 th frame (one stimulus per trial). For dual stimulus experiment, the two stimuli were set at the 70 th frame and the 120 th frame, respectively. The pause between trials was 10 seconds. Fluorescence changes were calculated as Δ F/F (%) in the recording area using Brain Vision Analyzer (Brain Vision Inc., Tokyo, Japan). Before stimulation, all whiskers, with the exception of the designated whiskers used for experiment, were clipped close to the skin. To perform stimulation, a glass pipette (1.0 mm in diameter) fitted on an XYZ manipulator was aimed at the designated whisker. Air-puff stimulus (duration 10-20 ms) was applied through a Picospritzer pressure valve connected to the glass pipette 35 . The Picospritzer was coupled to the imaging system through the MiCAM-02 controller, so the air could be puffed onto the whiskers at precisely controlled time points. Trains of whisker stimuli were delivered and the associated changes in fluorescence signals were recorded in the contralateral thalamus. In addition, we also performed ipsilateral stimulation as control 21 . In some experiments muscimol (0.1 μ l, 10 mM in ACSF), a selective GABA A receptor agonist, was injected at a depth of 300 μ m below the cortical surface 51 . The injection was performed with a glass pipette (20 μ m tip diameter) attached to a Nanojet II injector (Drummond Scientific, USA) 52 Data analysis. For single stimulation experiments, the final ten pre-stimulus frames (i.e., 90-99 th frame) were averaged as the baseline image. The baseline image was then subtracted from each subsequent frame to obtain changes in fluorescence signals. For dual stimuli experiments, the final ten pre-stimulus frames, i.e., 60-69 th frame and 110-119 th frame were averaged as the baseline for the 1 st and 2 nd stimulus, respectively. Pixels which exhibited a change in fluorescence (Δ F/F) greater than 50% of the maximum change were identified as activated regions 24 . Subsequently, we obtained pseudocolor maps of the areas activated by whisker stimulation. In the experiment exploring the effect of corticothalamic inputs, obtained data were expressed as mean ± standard deviation. Statistical analyses were then carried out using MATLAB. Student t-test was used to compare the signal amplitude and activated area (pixel number) in VPM before and after muscimol injection. P < 0.05 was considered to indicate a statistically significant difference.

Histology.
After the experiment, the animal was euthanized by barbiturate overdose and perfused by 4% paraformaldehyde and decapitated. The brain was extracted and sliced coronally (0.3 mm thickness) using a vibratome. The slice was photographed using fluorescence microscope. The probe track was superimposed 50 .