Study of the cortical representation of whisker frequency selectivity using voltage-sensitive dye optical imaging

ABSTRACT

The facial whiskers of rodents act as a high-resolution tactile apparatus that allow the animal to detect the finest details of its environment. Previously it was shown that whisker-sensitive neurons in the somatosensory cortex show frequency selectivity to small amplitude stimuli, An intravital voltage-sensitive dye optical imaging (VSDi) method in combination with the different frequency whisker stimulation was used in order to visualize neural activity in the mice somatosensory cortex in response to the stimulation of a single whisker by different frequencies. Using the intravital voltage-sensitive dye optical imaging (VSDi) method in combination with the different frequency whisker stimulation we visualized neural activity in the mice somatosensory cortex in response to the stimulation of a single whisker by different frequencies. We found that whisker stimuli with different frequencies led to different optical signals in the barrel field. Our results provide evidence that different neurons of the barrel cortex have different frequency preferences. This supports prior research that whisker deflections cause responses in cortical neurons within the barrel field according to the frequency of the stimulation. Many studies of the whisker frequency selectivity were performed using unit recording but to map spatial organization, imaging methods are essential. In the work described in the present paper, we take a serious step toward detailed functional mapping of the somatosensory cortex using VSDi. To our knowledge, this is the first demonstration of whisker frequency sensitivity and selectivity of barrel cortex neurons with optical imaging methods.

Introduction

Mice, as well as other rodents, use their whiskers to find their way in the darkness and obtain information about surrounding objects. Many scientists interested in the mechanism of the sensory perception have studied the whisker system intensively, because of the unique relationship between individual facial whisker and the corresponding barrels of the primary somatosensory cortex (S1). The barrel field has been particularly interesting as a key point for studying the relationship between cortical organization and the sensory information processes.^{5,24,28,29,36,43} The whisker system of the mice are very convenient for studying the mechanisms underlying information processing during active sensation.

Thus, recent studies have discovered that the barrel field in the somatosensory cortex contains a map of directional sensitivity, but the question about whisker frequency representation remains unclear.^22,43 Mice explore the tactile world by active whisking – whisker's periodical, approximately 50 ms per cycle in mice movements (O'Connor et al, 2010, 2013).

In rodents, the topographic whisker to barrel relationship has long been described as a main principle of the somatotopicity. But, in spite of this anatomical relationship, electrophysiological and optical imaging studies have revealed that responses to the whisker stimulation are also modulated by different property of the whisker's stimulation.²⁸

Traditional electrophysiological methods have revealed many important aspects of the neural network function but it remains unclear about the spatial organization of neural circuits. In contrast, intravital optical methods allow functional imaging of the neural activity and therefore provide a suitable technique to explore the functional representations of neuronal activity in the cerebral cortex.

VSDi offers the possibility to visualize activity of neuronal populations with high spatial resolution up to few tens of microns and high temporal resolution up to the millisecond and appears to be one of the best methods to study the cortical processing at neuronal population level.^6,25,41 After the craniotomy, the voltage-sensitive dye molecules are applied to the surface of the brain and bind to the membranes of all cells transform changes in transmembrane voltage into fluorescence signals. Physically, this signal is proportional to the membrane area of all stained molecules under each pixel of the imaging area.⁶ The recorded signal combined emitted photons from neurons and glial cells, all cellular compartments, including bodies, dendrites and axons of neurons. The fluorescent signal is then recorded by the charge-coupled device (CCD-camera) and can be used for the generation of the functional maps of the brain surface.

In the present study, we used VSDi in order to investigate the functional representation of the frequency sensitivity of the barrel field. VSDi monitors change in the transmembrane voltage over large populations of cortical neurons. This technique has yielded revolutionary results when applied to the somatosensory systems since it is based on the direct monitoring of the neural activity.^27,43. Other types of the optical imaging allow us to study spatiotemporal features of the neural pools in vivo.^15,18,19,20

Our results show that the VSDi signal has a different signal strength in response to different frequencies of the single whisker stimulation. We believe this suggests that the whisker's system has frequency preferred neurons that correspond to the frequency of the whisker deflection based on inputs from the cells of the facial whisker pad (Fig. 2A). Therefore, we demonstrated that neurons in the somatosensory cortex show preference for certain frequency of the whisker's stimulation.

Figure 2.

(A) Whisker is sweeping surfaces of different roughness. (B) Four different stimuli used in experiments.

Materials and methods

Experiments were performed on six B6 mice, 25-30 g body weight, at 6-10 weeks of age using MiCAM-02 system (MiCAM-02, Brain Vision Inc.., Japan). Animals were anesthetized with urethane (i.p. 1.5 mg/g) ²⁷; the head was shaved and placed in a stereotaxic frame. After that about 4×4 mm cranial window was made over the left parietal cortex by the dental drill. The exposed dural surface was cleansed with a hemostatic sponge (Gelfoam Ltd) dipped in the body temperature artificial cerebrospinal fluid (ACSF) and washed by ACSF. After that the hemostatic sponge dipped into the voltage-sensitive dye RH-1691 (Optical Imaging Ltd, 1.0 mg/ml in ACSF) was applied to the exposed and preliminary dried dural surface for 45 – 60 min. After staining, the cortex was washed with dye-free ACSF for about 15 min to remove unbounded molecules of the RH-1691.^31,43,44 The cortical surface was covered with high-density silicone oil (Sigma Aldridge, viscosity is 60.000 cSt) and sealed with a 0.1 mm thick cover glass. During the experiment the body temperature of the animal was kept at 37ºC by a temperature-controlled heating pad. All animal handling was in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No.80-23) and a protocol approved by the University of Maryland, Baltimore Institutional Animal Care and Use Committee.

For the imaging session the objective of MiCAM-02 (Brainvison, Tokyo Ltd) system was positioned above the recording area with its optical axis perpendicular to the cortical surface (Fig. 1). At the start of each optical recording, a gray-scale image of the cortical surface was obtained and then moved down 300 μm to the region of interest (ROI).^27,43

Figure 1.

Intravital setup for voltage-sensitive dye optical imaging in vivo. The cortex is illuminated with light 630 nm wavelengths; the emitted photons with wavelength above 690 nm are acquired with CCD camera. Dichroic mirror split emission and excitation light.

Each experimental session consisted of 375 trials, 500 frames per trial, with the stimulus (whisker C2 deflection) presented at the 300th frame, one trial per stimulus. The inter-trial interval was 12 sec. Change in fluorescence was calculated as ΔF/F (%) in the ROI using Brain Vision Analyzer (Brain Vision Inc.., Japan). A specially designed electromechanical stimulator, fitted onto an XYZ manipulator was used for mechanical deflection of the whisker. The stimulator was coupled to the imaging system through the MiCAM-02 controller, so that the stimulation's pulse and the optical signals were collected simultaneously.

The algorithm of each experiment was the following: different frequency stimulus 60 ms duration was generated by the MiCAM-02 imaging system (SciMedia Inc., 2012) every 12 s. Each stimulus was presented after beginning of the trial, the trial consisting of 500 frames of 5 ms frame duration. During the session, each stimulus was presented 75 times. The sequence was: (1) each trial began with illumination shutter opening at 300 ms prior to recording and illumination of cortex with excitation light; (2) the CCD camera started acquiring frames 1500 ms prior to whisker stimulation; (3) at 1500 ms the whisker was stimulated for 60 ms by 100, 200, 333 or 500 Hz frequency (Fig. 2B); all pulses have the square shape and 1 ms duration and move for 0.1- 0.2 mm whisker along the rostro-caudal axis; (4) at 2500 ms the camera stopped collecting frames and the shutter was closed blocking the illumination light; (5) an 12 s inter-stimulus interval separated trials.

It used four different frequencies: 100, 200, 333 and 500 Hz to test the whisker system using relatively high frequency stimulation. We wanted to see the difference in responses to different frequencies, but not stimulus duration or other parameters of the stimulation, so we kept all parameters of stimulation the same, and only the frequency was changed from stimulus to stimulus. Each stimulus was a train of pulses with a different frequency, the train duration being 60 ms. However, to have consistent duration of the train – 60 ms regardless of the frequency – the train must be begin from the pulse and completed by the pulse (Fig. 2B). So we chose the frequencies which not only lay well into the range of 100 – 500 Hz, but also allowed to start and end the stimulus by pulse with 50 ms duration of stimulus.

Multiple steps of data analysis were used to study the spatial and temporal features of the evoked fluorescence in the barrel field. To determine the areas significantly activated by a particular stimulus, the image data was thresholded by the 50% of the maximum response value inside the whole recording area.^12,32,44 Thus, we generated boundaries of the brain areas activated by a particular stimulus. The integrated fluorescence level in the 7 x7 pixels area, located in the center of activity pattern has been calculated for each frame and each stimulus. Thus, we used this value to make a chart of the time courses of the fluorescence signals.

Results

The activated areas were identified by the number of activated pixels exhibiting a change in fluorescence (ΔF/F) greater than 50% of the maximum change in signal, and the pseudocolor maps of the neural activity have been obtained. The results showed that differences between responses to the 100, 200, 333 and 500 Hz mechanical stimulation frequencies within 60 ms were indicated by a change in the voltage-sensitive dye optical signal is shown in Figure 3.

Figure 3.

Voltage-sensitive dye optical images showing single-whisker stimulation fluorescence changes in response to different frequency stimulation. Time after stimulus onset is indicated at the top left corner, upper row of the images.

The integrated level of the fluorescence signal was analyzed statistically. Regardless of frequency, the signal appeared just after the frame when the stimulus was presented and reached its peak 30–40 ms after stimulus onset. After reaching their maximum value the activated areas decreased until returning to baseline 40–70 ms after the stimulus offset.

We have not observed a difference in the spatial representation of the optical signal in response to different stimuli. In all examined animals VSDi signals elicited in response to different frequencies whisker stimulation were observed in a cortical area morphologically corresponding to the contralateral barrel field. Time courses of the signals obtained in response to all stimuli in all experiments were averaged and are shown in Figure 4. Significant difference between the peak time, onset time, and decay time was not observed.

Figure 4.

Change in fluorescence (ΔF/F(%), ordinate) in response to different frequency stimuli (denoted at the top right corner of the chart). Fluorescence signal was recorded in the small red square marked on the cortical surface in the incretion (top, right). Stimulus presented at the 200th frame (abscissa).

The responses of 333 Hz were significantly stronger than other frequencies, in spite of the amplitude and speed of the whisker motion remaining the same; only frequency varied among the stimuli.

We have compared all stimulus between each other by one way ANOVA and Tukey's multiple comparison test. For this analysis the total area, under the curves, shown at the Figure 4, were calculated separately for each experiment. Results of the statistical analysis are shown in Figure 5

Figure 5.

The comparison of the intensity of the responses (vertical) in reply to different frequency stimulation (horizontal). The under curve area of each response in each experiment was calculated in relative units and analyzed.

The difference was significant between 100 and 333 Hz (P-value 3.78⁻⁶), 200 and 333 Hz (P-value 1.06⁻⁷), 500 and 333 Hz (P-value 4.45⁻⁵) and not significant between 100 and 200 Hz, 200 and 500 Hz, 100 and 500 Hz.

These observations are consistent with previous findings, obtained by electrophysiological methods in terms of the location of the representation of different frequencies. This indicates that our VSDi are relevant to population activities of cortical neurons, because the recording method directly measures neural activities.

Discussion

The understanding of the physiological perception must be based on knowledge of the properties of the neural circuits involved into the brain sensory system. As a part of somatosensory system, the whisker system is critical to many species' ability to obtain information about the surrounding world. Thus, it is a logical inference that the whisker's frequency selectivity is an essential element of barrel cortex information processing. Frequency selectivity of the neurons have been described at the cortical level of the whisker's system ^14,16 but its mechanism is not clarified yet.

The first in Intravital VSDi data in the mammalian brain was obtained more than thirty years ago in the rat cerebral cortex (Orbach et al, 1985). The VSDi technique, which detects small changes in the fluorescence caused by neuronal activities in the cortical tissue, have proven to be useful for visualizing functional map.⁴² In the last decade, the technique of Intravital VSDI has developed rapidly, thus enabling the visualization of various cortical maps. Therefore, it was very logical to apply this method for the study of whisker system's cortical neurons in the experimental goal described in this manuscript.

The voltage-sensitive dye spread along the mediolateral and dorsoventral axes of the neocortex. It was shown ²⁷ that the voltage-sensitive dye RH-1691 loading extended across all cortical layers. It is important to note that in the imaging experiments we do not have real depth resolution. The fluorescence signal in each pixel consist of the signal from whole “cylinder” of tissue below this pixel. We set the plane of focus 300 µm below the surface, but this can only partially cut off the signal from above and below the focal plane.

It was shown in different technical approaches that RH-1691 is preferentially located in the layers II–III.^13,26 Layers I–III contain 70% of the dye and most likely the fluorescent signal.²⁶ Illumination intensity decreases with depth (because of dye concentration), whereas light scattering increases; activity in layers V–VI has smaller contribution to the signal. Therefore, the voltage-sensitive dye signal in our experiments is most likely a combination of spikes and synaptic activity in layers II–IV.

When the tip of the whisker makes contact with a rough surface, its movement changes. Sensory neurons of the whisker follicula report to the trigeminal nerve that the whiskers have made contact.^10,11 Each contact generates a spike in the neurons of the whisker follicular.^1,11 therefore, neuron firing frequency is determined by several factors, among them being the velocity of the whisker motion and the roughness of the surface (Fig. 2A). The barrel field of the somatosensory cortex is one of the structures responsible for detecting an object by the whisking, any mechanical contact between the whiskers and the surface causes neural spikes in the neocortex (O'Connor et al, 2013). Consequently, the frequency of the whisker vibration can be much higher than the whisker's motion: up to hundreds of Hz, as we used in our experiment.

Encoding in the higher parts of the whisker system starting in the lower level: thus, trigeminal ganglia neurons likely encode at least four specific events: whisking, contact with object, pressure against the object, and detachment from the object.⁴⁰ Frequencies more than a hundred Hz are located out of physiological range of the whisker's movement, but rough surface contact generates high-frequency “stickslip-ring” events.³⁸ Our results correlate with the electrophysiological data, suggesting that somatosensory neurons showed bandpass tuning, reflecting the frequency of the whisker motion.³

Previously, the methods of the computational neuroscience have been successfully applied for the modeling of the frequency and the directional selectivity of the whisker system by the tuning of synaptic amplitude and latency.³⁶ Therefore, to predict how the angular selectivity is expressed during varying stimuli, the synaptic amplitude and the latency are implemented in a simple model.³⁶

It is shown that one stimulus frequency can affect the representation of another directional selectivity. In other words, the directional selectivity is modulated by the frequency of an ongoing stimulus.³⁶ They concluded that directional selectivity is based on the synaptic tuning which is sensitive to the stimulus temporal structure.

This group suggests ³⁶ that the role of directional selectivity may be prominent in behavior, primarily in the discrimination of the object location, but more likely to be negligible in other behavioral acts such as texture discrimination, which is based on the frequency selectivity. Therefore, we suppose that the directional selectivity of the barrel's neurons is dependent on the stimulus frequency.

Neural transmission of the whisker system inputs in the whisker follicle, the trigeminal ganglion, and is relayed somatotopically through the barreletes of the brainstem and barreloids of the thalamus to the barrels of the barrel field of the contralateral neocortex. It has been shown that in the somatotopic frequency map in the somatosensory cortex the part of barrel field, corresponding to the rostral part of the whisker's pad is more sensitive to the higher frequency while corresponding to the caudal part – more sensitive to the low frequency.³ It is logical to suppose that the base of this phenomenon is the same that we observe in the auditory cortex organized tonotopically: sound frequencies are spatially arranged in an orderly manner.^45,46

It has been reported that the sensory transduction and neural processing of whisker frequency information parallels the transduction and neural representation of the auditory stimuli ³⁹ in the auditory cortex, and we support this theory.

However, in contrast to the auditory system, in the whisker system the neurons most sensitive to the particular frequency are not organized into the morphological pool, therefore the frequency-selective functional map can't be observed in the neocortex.

Thus, we obtained activity patterns evoked by a particular stimulus. We found that whisker stimuli with different frequencies led to different activation patterns in the barrel field. Our results provide preliminary evidence that the different neural pools of the barrel cortex have different frequency preferences.

Looking from the anatomical point of view, the sensory information comes from the whiskers follicle to the neocortex. Thus, the second order neurons of the lemniscal pathway are located in the trigeminal nuclei and form “barrelettes” of the brainstem.⁸ Their axons go to the opposite hemisphere to the “barreloids” of the ventral posterior medial nucleus (VPMdm) of the thalamus. The axons of VPMdm neurons reach the barrel field.^7,8

The frequency of the neuron firing in the brainstem, thalamus and (S1) is determined by several factors, among them including the velocity of the whisker motion and the physical properties of the object's surface.^10,11

So, those S1 neurons may be sensitive to the frequency of the whisker stimulation. The electrophysiological evidence found: reports about responses of S1 neurons evoked by high-frequency stimulations of the whisker in a phaselocked manner for long stimulus durations has been published few years ago.⁹ The data clearly suggest that high-frequency components of stimuli can be faithfully encoded by S1 neuronal discharges reported by several groups in the previous studies.^21,30

According to the results of numerous studies of the last decade, neurons of the barrel field are sensitive to the frequency of the whisker stimulation.^2,16,37,48 We believe that frequency selective neurons in the barrel are characterized by a V-shaped histogram of the stimulation frequency/stimulation intensity similar to how the neurons of the auditory cortex have a V-shaped histogram of sound frequency/sound intensity.⁴⁵

We hypothesize that the frequency selectivity in the somatosensory cortex is organized similarly to the auditory cortex. As well known in mammalians, the auditory cortex is arranged in tonotopic maps on the basis of the frequency selectivity of individual neurons.^44,47 The tonotopic organization exists at each stage of the auditory pathway, as well as the whiskers somatotopical organization. Layer IV of the primary auditory cortex (A1) receives its inputs from the auditory thalamus, to which most of the scientists have agreed that the thalamocortical projection defines the frequency selectivity and tonotopic organization of the A1.⁴⁷ Barrels of the somatosensory cortex receives its inputs from the ventral posteromedial nucleus (VPM) of the thalamus.

Complex mechanisms probably collaborate in getting the cortical sensitive neurons and behavior. The whisker motor cortex sends signals coding for whisker movement and the neurons of the barrel field receive this input.³³ The whisker position signal from the whisker motor cortex probably amplifying neural activity related to touch during the periods of the whisker movement. Therefore the frequency of the whisker touching might be reflected by the spike rate of the cortical neurons in the barrel field.

Finally, these experiments were carried out in anesthetized animals. Urethane alter response properties in the somatosensory cortex so this issue can be examined in future studies. An even more intriguing question is how the interactions between trigeminal ganglia, barreloids (brain stem), barrelettes (thalamus) and the barrels operate in a behavioral context. Based on the rat experiments, (Hawking and Gerdjikov, 2012) it was concluded that tactile responding is sparse and shows no stimulus frequency modulation at the level of individual medium spiny projection neurons. However, same authors report that the firing rate was related to stimulus frequency, so the vibrotactile frequency was encoded during the transition from high to low.¹⁷; Hawking and Gerdjikov, 2012).

Conclusions

In our paper we clearly demonstrated that intravital VSDi method in combination with the different frequency whisker stimulation is a powerful tool that allows to visualize neural activity in the mice somatosensory cortex in response to the stimulation of a single whisker by different frequencies. We also demonstrated that whisker stimuli with different frequencies led to different optical signals in the barrel field. Our findings provide evidence that different neurons of the barrel cortex have different frequency preferences.

This supports prior research that whisker deflections cause responses in cortical neurons within the barrel field according to the frequency of the stimulation. Increase in stimulation intensity is indicated by an increase of the number of neurons which respond to the stimulation. Presence of these neurons within the barrel is the key stone of our investigation. To the best of our knowledge this is the first experimental demonstration of whisker frequency sensitivity and selectivity of barrel cortex neurons with optical imaging methods.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank B. Okura, T. Sakuraba and K. Tsubokura from SciMedia Company for their help with MiCAM-02 installation and data analysis and Drs. Hirouki Arakawa and Anna Volnova for their valuable advices for experiments. We would also like to express our gratitude to Dr. Daniel O'Connor for his advices and Dr. Reha S. Erzurumlu for the valuable discussions and help with the manuscript writing.

Funding

This work is supported by the NSF CAREER AWARD (CBET-1254743) (YC).