Rats generate sweeping whisker movements in order to explore their environments and identify objects. In somatosensory pathways, neuronal activity is modulated by the frequency of whisker vibration. However, the potential role of rhythmic neuronal activity in the cerebral processing of sensory signals and its mechanism remain unclear. Here, we showed that rhythmic vibrissal stimulation with short duration in anesthetized rats resulted in an increase or decrease in the amplitude of somatosensory-evoked potentials (SEPs) in the contralateral barrel cortex. The plastic change of the SEPs was frequency dependent and long lasting. The long-lasting enhancement of the vibrissa-to-cortex evoked response was side- but not barrel-specific. Local application of dl-2-amino-5-phosphonopentanoic acid into the barrel cortex revealed that this vibrissa-to-cortex long-term plasticity in adult rats was N-methyl-d-aspartate receptor-dependent. Most interestingly, whisker trimming through postnatal day (P)1–7 but not P29–35 impaired the long-term plasticity induced by 100 Hz vibrissal stimulation. The short period of rhythmic vibrissal stimulation did not induce long-lasting plasticity of field potentials in the thalamus. In conclusion, our results suggest that natural rhythmic whisker activity modifies sensory information processing in cerebral cortex, providing further insight into sensory perception.
The vibrissal sensory system of rats is functional for tactile discrimination with high resolution or object location, and has been regarded as one of the key model systems to study the mammalian sensory processing (Welker 1964; Zucker and Welker 1969; Woolsey and Van der Loos 1970; Kleinfeld et al. 2006; Kleinfeld and Deschenes 2011). The vibrissal input is conducted from the trigeminal ganglion, and is relayed somatotopically to the brainstem and thalamus, and then to the barrel cortex (Bosman et al. 2011).
Rats and mice actively move their facial vibrissae over objects and their surfaces in rhythmic sweeps at frequencies between 5 and 15 Hz when exploring the environment (Carvell and Simons 1990; Berg and Kleinfeld 2003). Rats show high acuity, and can use their whiskers to discriminate gratings between 1 and 1.06 mm (Carvell and Simons 1990, 1995). Resonant vibrations occur while whisking against textures (Hartmann et al. 2003; Wolfe et al. 2008). The speeds of typical ‘whisking’ over textures generate 100–1000 Hz microvibrations of the vibrissa, and the driving frequency between the 2 textures differs by only 5% (Neimark et al. 2003; Andermann et al. 2004). During whisking, neuronal activity in barrel cortex varies rhythmically and is synchronized with whisker movement (Crochet and Petersen 2006; von Heimendahl et al. 2007).
Recent evidence suggests that a brief period of rhythmic auditory or visual stimulation induces synaptic plasticity such as long-term potentiation (LTP) of cortical responses induced by sensory inputs (Clapp et al. 2005, 2006; Teyler et al. 2005). Previous work has shown that the fast oscillatory component in single-unit spike trains recorded from the rat barrel cortex is correlated with whisker position in the whisk cycle (Fee et al. 1997). We hypothesize that neurons in somatosensory pathway have frequency-tuning plasticity because of the resonance properties of their somatotopically associated vibrissa. Here, we determined whether a period of passive rhythmic whisker stimulation at different frequencies induced long-lasting changes of evoked sensory responses in the barrel cortex and investigated the mechanisms underlying these effects. We showed that a brief period of rhythmic vibrissal stimulation in anesthetized rats resulted in a frequency-dependent and long-lasting increase or decrease in the amplitude of somatosensory-evoked potentials (SEPs) in the contralateral barrel cortex. This suggests that natural rhythmic whisker activity modifies sensory information processing in the cerebral cortex, providing further insight into sensory perception.
Materials and Methods
Male or female Sprague-Dawley rats weighing 220–300 g (Experimental Animal Center, Zhejiang University, Certiﬁcate No. 22–9601018) were housed under standard laboratory conditions of 12 h light/dark (lights on from 7:00 to 19:00), 22–26 °C, and 40–70% humidity. They were provided food and water ad libitum. All studies were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All efforts were made to minimize suffering and to keep the numbers of animals used to a minimum.
Rats were anesthetized with urethane (1.2 g/kg i.p. initial dose). Supplemental doses (10% of initial dose) were administered to maintain areflexic hind paw-pinch responses and a respiration rate of 90–120 breaths per min. Body temperature was monitored and maintained at 37 °C with a controlled heating blanket and rectal thermometer (FHC, Brunswick, ME, USA). The rat was held in a stereotaxic frame on a floating table to minimize external vibrations. A craniotomy and durotomy were performed. The following coordinates of barrel cortex with respect to lambda and bregma in the same horizontal plane were used: posterior to bregma, −1.9 mm; lateral to midline, +5.5 mm; ventral from the dura, −0.9 mm, according to the Paxinos and Watson atlas. The coordinates of the ventral posteromedial thalamic nucleus (VPM) were posterior to bregma, −3.2 mm; lateral to midline, +2.6 mm; ventral from the dura, −5.6 mm. The surface of the brain was kept moist with artificial cerebrospinal fluid (ACSF). At the end of each experiment, the rat was euthanized by urethane overdose (3 g/kg, i.p.) and in a subset of cases perfused for identification of location of recording or injection.
Along with the recording electrode, a syringe for microinjection (Hamilton, Reno, NV, USA) was inserted into layer IV of the barrel cortex. dl-2-Amino-5-phosphonopentanoic acid (APV, 200 µM; Sigma, St Louis, MO, USA) was used as selective antagonist of NMDA receptor. APV was dissolved in saline solution. The microinjection volume of APV or vehicle was 1 µL. The injection was finished within 3 min, and the injection pipette was kept in place after the infusion.
A picospritzer (Parker Hannifin Instrumentation, Cleveland, OH, USA) controlled by a PG4000A digital stimulator (Cygnus Technology, Southport, NC, USA) was used for whisker stimulation. The picospritzer pipette was mounted at a position allowing air-puff pulse stimulation precisely vertical to the long axis of the vibrissa at about 10 mm from the vibrissa pad. The principal vibrissa (PV) was determined as the whisker that produced the largest amplitude and shortest latency response.
A vibrissa was stimulated at 30 s intervals with 5 ms duration. Following exposure of barrel cortex, normal ACSF was used to keep the brain surface moist. Recording probes were placed and baseline signals of both spontaneous and evoked potentials were recorded at least 30 min after surgery. After 30 min stable recording from the barrel, 6 different stimulation protocols were applied to the contralateral PV to determine plastic changes of the activity in the cortex: 100 Hz, 1 s; 50 Hz, 10 s; 50 Hz, 2 s; 20 Hz, 30 s; 20 Hz, 5 s; and 5 Hz, 10 min.
The air-puff pressure controlled by the picospritzer ranged from 1 to 30 pounds per square inch (psi). A video camera was horizontally mounted near the pipette and vibrissa, so the whisker deflection was captured when the air-puff stimulus was delivered. The whisker deflection angle was measured with a protractor. In most experiments, 20 psi was used to deliver puffs to the whisker.
Parylene-coated tungsten microelectrodes (FHC, Brunswick, ME, USA) with 1–2 MΩ impedance at 1 kHz were used to record field potentials at a single depth corresponding to the middle layers of the barrel cortex. Recordings from the barrel cortex were made contralateral to vibrissal stimulation (1–4 mm posterior and 4–7 mm lateral to bregma), 700–950 µm deep to the cortical surface, targeting layer IV as judged by micromanipulator readings and negative polarity of the local field potential. These depths were confirmed by histological reconstruction of electrolytic lesions in 2 rats. Signals were amplified 1000× by Model 1700 4-channel amplifiers (A-M Systems, Inc., Sequim, WA, USA) with a filter frequency ranging from 0.1 Hz to 5 kHz for both spontaneous and evoked potentials and stored with stimulus markers. Signals were then sampled at 20 kHz with an ML795 PowerLab/4SP data acquisition system (AD Instruments, Bella Vista, NSW, Australia) and stored on a hard disk for off-line analysis. The amplitude of SEP was calculated from the layer IV recordings to evaluate changes in synaptic transmission. The sampling data of spontaneous field potentials were filtered by 10th-order high-pass and low-pass digital Butterworth filters with cutoff frequencies of 5 and 400 Hz, respectively. Data are expressed as mean ± standard error. Student's t-test was used for statistical analysis.
Cytochrome Oxide Reaction
After recording, some rats were anesthetized with urethane and sacrificed by transcardial perfusion with cold normal saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer solution (PBS). The brain was postfixed for 6 h and then cryoprotected in 30% sucrose for 48 h. The cortex was separated and stored in 30% sucrose, flattened, and then tangentially sectioned from the lateral surface of the barrel cortex. Sections obtained from the cutting plane were 40 μm thick and were stored in 0.1 M PBS as briefly as possible to preserve the best obtainable cytochrome oxide activity. The sections were collected in 0.1 M PBS and then transferred to a staining solution consisting of 0.05% 3,3′-diaminobenzidine tetrahydrochloride, 0.04% cytochrome C from horse heart, and 3% sucrose (all from Sigma) in 0.1 M PBS. Incubation was performed free-floating with gentle agitation at 37 °C under visual control. The staining reaction was stopped by washing 3 times in 0.1 M PBS (5 min each). The sections were transferred onto glass slides for drying, and then immersed twice in 50%, 70%, 95%, and 100% ethanol, and twice in dimethylbenzene in sequence for 10 min each. After that, the sections on glass slides mounted with 1 or 2 drops of resinene were examined under a microscope.
Three groups of rats were used in this part of study. At 12–24 h after birth, infant-trimmed rats (n = 6) were gently restrained and all of their mystacial vibrissae on both sides of the face were clipped to within 2–3 mm of the skin surface; trimming continued daily to 7 days of age. All whiskers were subsequently allowed to regrow for 60–70 days before electrophysiological recording began. Under brief anesthesia with ether, one group of rats (n = 6) had their whiskers clipped daily to within 2–3 mm of the skin surface for 7 days during postnatal day (P) 29–35; all vibrissae were allowed to regrow for 30–40 days before electrophysiological recording began. Normally reared animals (n = 8) were electrophysiologically recorded with no previous whisker clipping.
Air Pressure-Dependent SEPs in Barrel Cortex Induced by Different Vibrissal Stimulus Frequencies
To investigate the relationship between stimulation frequency and whisker deflection, we applied 20 psi air-puffs at 5, 20, 50, and 100 Hz, and measured the whisker deflection angles with a protractor. With air-puffs of 20 psi, the deflection angle was 11–12° (Fig. 1A,B). At this fixed stimulus intensity (20 psi), changing the frequency did not affect the deflection angle (Fig. 1B). But with increasing stimulus intensity (1, 5, 10, 15, 20, 25, and 30 psi, single stimulus), the deflection angles increased (n = 5, Fig. 1C). Higher pressures caused larger deflection angles and SEP amplitudes (Fig. 1D).
After obtaining a stable baseline, vibrissal stimulation at 100 Hz for 1 s at 5 or 20 psi both caused a long-lasting increase in the amplitude of the SEPs (Fig. 1E). The normalized amplitude of the SEP at 5 psi was 100 ± 1.01% at baseline, 103.50 ± 2.37% after stimulation for 0–30 min (P > 0.05), and 115.33 ± 6.10% after stimulation for 30–60 min (P < 0.05, n = 5, Fig. 1E,F). The normalized amplitude of the SEP at 20 psi was 100 ± 3.40% at baseline, 134.80 ± 5.93% after stimulation for 0–30 min (P < 0.01), and 144.34 ± 11.60% after stimulation for 30–60 min (P < 0.01, n = 8, Figs 1E,F and 2B,F).
Because 20 psi reliably induces the long-term plasticity, this was used to deliver puffs to the whisker in the subsequent experiments.
Frequency-Dependent Long-Term Plasticity Induced by Rhythmic Vibrissal Stimulation in Barrel Cortex
To compare the SEP amplitudes induced by 20 psi vibrissal stimulation, we recorded SEPs every 100 µm at known depths within the cortex (Fig. 2A). The SEP waveform had a major component lasting for 90–100 ms from the start of the signal with a latency of 13–15 ms, and the amplitude varied from 0.1 to 1.0 mV. The SEP had a maximal amplitude at intermediate depths (around 400 µm electrode advancement). The amplitude progressively decreased with further electrode advancement (Fig. 2A). This pattern of SEP depth profile was very consistent among animals.
To determine whether a period of passive rhythmic whisker stimulation at different frequencies induced long-lasting changes of evoked sensory responses in the barrel cortex, we obtained a stable baseline response by stimulating the PV at a 30-s interval for a minimum of 30 min and recording SEPs from the cortex. Every 6 consecutive responses were averaged, so that a 30-min baseline was composed of 10 data points. After obtaining a stable baseline, vibrissal stimulation at different frequencies (see Materials and Methods) was applied to the vibrissa. Stimulation at 100 Hz for 1 s with 20 psi caused a significant long-lasting increase in the amplitude of the SEPs. The normalized amplitude of the SEP was 100 ± 3.40% at baseline, and 144.34 ± 11.60% after stimulation for 30–60 min, P < 0.01, n = 8 (Fig. 2B,F). Stimulation at 50 Hz applied to the vibrissa for 10 or 2 s with 20 psi caused a significant long-lasting increase in the amplitude of the SEPs. The effect of stimulation at 50 Hz was less remarkable than that at 100 Hz: 50 Hz, 10 s stimulation, 100 ± 3.84%, 119.03 ± 9.57%, P < 0.05, n = 6; 50 Hz, 2 s stimulation, 100 ± 3.16%, 135.67 ± 8.17%, P < 0.01, n = 6 (Fig. 2C,F). Rhythmic stimulation (20 Hz) of the PV for 30 or 5 s with 20 psi did not significantly increase the amplitude of the vibrissa-tocortex-evoked response: 20 Hz, 30 s stimulation, 100 ± 3.65%, 115.06 ± 11.13%, P > 0.05, n = 6; 20 Hz, 5 s stimulation, 100 ± 2.89%, 102.96 ± 6.36%, P > 0.05, n = 6 (Fig. 2D,F). Furthermore, stimulation at 5 Hz for 10 min with 20 psi caused a significant long-lasting decrease in the amplitude of the SEPs. The normalized amplitude of the SEP was 100 ± 1.30% at baseline, and 74.65 ± 9.93% after stimulation during 30–60 min, P < 0.05, n = 7 (Fig. 2E,F). The changes of SEPs caused by high- or low-frequency rhythmic vibrissal stimulation lasted longer than 1–2 h.
The Absence of Long-Term Plasticity in the VPM
To determine the derivation of plasticity evoked by short-term rhythmic vibrissal stimulation, we recorded from the VPM in the contralateral thalamus. We applied a period of passive rhythmic whisker stimulation at 20 psi and 100 Hz for 1 s to determine whether this could induce LTP of evoked sensory responses in the VPM as in the barrel cortex. We obtained a stable baseline response by stimulating the PV at 30 s intervals for 30 min and recording the field potential in the VPM. Every 6 consecutive responses were averaged, so that a 30-min baseline was composed of 10 data points. After obtaining a stable baseline, we found that vibrissal stimulation at 20 psi and 100 Hz for 1 s did not cause a significant long-lasting increase in the amplitude of the SEP (Fig. 3A,B). The normalized amplitude of the SEP was 100 ± 4.63% at baseline, 108.73 ± 2.53% after stimulation for 0–30 min (P > 0.05), and 107.84 ± 4.09% after stimulation for 30–60 min (P > 0.05, n = 4, Fig. 3A,B).
The Induced Long-Term Plasticity Is Input- but Not Barrel-Specific
To demonstrate whether the 100 Hz whisker stimulation-induced potentiation of SEPs was specific for the sensory input, we recorded SEPs in both sides of barrel cortex in response to deflection of whiskers on contralateral snouts before and after unilateral 100 Hz stimulation (n = 6). Ten baseline SEPs were recorded at 3 min intervals. The PV in right snout was then rhythmically deflected at a frequency of 100 Hz for 1 s, after which SEPs were again recorded every 3 min for 60 min. In left snout, the PV according to the SEPs recorded in right barrel cortex was not stimulated by 100 Hz deflection (Fig. 3C). To make sure there was no influence of the SEP from the barrel cortex of one hemisphere to the other, we calculated the ratio of amplitudes of the 2 SEPs with different interstimulus intervals when stimulating the PV of both sides of the snout with random sequence. We measured the ratio of the amplitudes of the evoked responses when the left side or the right side of the snout was stimulated first. There was no influence of the SEPs recorded from the 2 hemispheres (Fig. 3D).
Stimulation at 100 Hz induced a significant long-lasting increase in the peak amplitude of responses recorded at left barrel cortex to whisker deflection on the right side of the snout (Fig. 3E–G). The normalized amplitude of the SEP was 100 ± 3.85% at baseline, and 157.70 ± 12.13% after rhythmic stimulation during 30–60 min, P < 0.01, n = 6. In contrast, at the same time, the responses to whisker deflection on the unstimulated (left) side of the snout (recorded in the right barrel cortex) were not potentiated. The normalized amplitude of the SEP was 100 ± 4.26% at baseline, and 104.81 ± 6.33% after rhythmic stimulation during 30–60 min, P > 0.05, n = 6 (Fig. 3E–G). This result demonstrated that 100 Hz stimulation induced an input-specific potentiation of sensory responses.
To investigate whether the long-lasting enhancement of the vibrissa-to-cortex-evoked response caused by 100 Hz rhythmic stimulation is barrel-specific, we recorded the evoked responses from 2 electrodes targeting 2 barrels within one side of barrel cortex. The PVs of the 2 recording sites are called whisker A and B separately (Fig. 4A). An example showing the recording sites of both electrodes in a slice stained with cytochrome oxidase is shown in Figure 4B. To ensure there was no overlap of the responses recorded by the 2 electrodes, we calculated the ratio of the 2 SEPs with different interstimulus intervals when stimulating the whisker A or B in the same side of the snout with random sequence (Fig. 4C).
Rhythmic vibrissal stimulation at 100 Hz applied to whisker A for 1 s caused a long-lasting increase in the amplitudes of vibrissa-to-cortex-evoked responses (Fig. 4D). The normalized amplitude of the SEP was 100 ± 1.85% at baseline, and 148.58 ± 8.83% after 100 Hz vibrissal stimulation during 30–60 min, P < 0.01, n = 6 (Fig. 4E,F). This long-lasting enhancement of the SEPs was also observed in the other recording site according to whisker B which was not stimulated by 100 Hz. The normalized amplitude of the SEP was 100 ± 1.71% at baseline, and 152.02 ± 4.94% after the end of baseline during 30–60 min, P < 0.001, n = 6 (Fig. 4E,F). These results demonstrate that a short period of passive rhythmic somatosensory stimulation induces nonspecific, long-lasting plasticity of the cerebral processing of somatosensory input within one side of barrel cortex.
NMDA Receptor Is Involved in Long-Term Plasticity After Short-Term Rhythmic Vibrissal Stimulation
To investigate whether the NMDA receptor is required for this plasticity, the competitive NMDA receptor antagonist APV was microinjected into the recorded barrel cortex. Several minutes later, the normalized amplitude of the SEP decreased from 100 ± 2.07% to 56.37 ± 4.08%, P < 0.001. After we obtained a stable baseline by stimulating the PV and recording the SEP, short-term rhythmic vibrissal stimulation at 100 Hz for 1 s did not increase the amplitude of the vibrissa-to-cortex-evoked response compared with control (n = 6, P > 0.05). The normalized amplitude of the SEP in the APV group was 107.52 ± 7.47% and 133.68 ± 7.94% in the control group after rhythmic stimulation during 30–60 min (P < 0.05, Fig. 5A,C). Microinjection of the same dose of APV into the barrel cortex also abolished the long-lasting depression of the vibrissa-to-cortex-evoked response caused by stimulation of the PV at 5 Hz for 10 min (n = 6). The normalized amplitude of the SEP in the APV group was 108.60 ± 9.11% and 85.27 ± 2.19% in the control group after rhythmic stimulation during 30–60 min (P < 0.05, Fig. 5B,C). These results show that cortical NMDA receptors are involved in the long-lasting plasticity induced by short-term rhythmic vibrissal stimulation. In other words, the vibrissa-to-cortex long-term plasticity in adult rats in vivo is NMDA receptor-dependent.
The normalized amplitude of the SEP decreased after APV injection. But, it shows there is no correlation between the amplitudes of SEPs evoked by vibrissal stimulation and the long-term plasticity after short-term rhythmic vibrissal stimulation (Fig. 5D). So, the decreased amplitudes of SEPs after APV injection will not affect the efficacy of the long-term plasticity after short-term 100 Hz vibrissal stimulation.
Critical Period and the Long-lasting Plasticity Caused by Rhythmic Vibrissal Stimulation
To investigate the involvement of the critical period of the development in this long-lasting plasticity, 3 groups of rats were used in this part of study. At 12–24 h after birth, rats with whisker trimmed at P1–7 were gently restrained and all of their mystacial vibrissae on both sides of the snout were clipped to within 2–3 mm of the skin surface; trimming continued daily to 7 days of age. All whiskers were subsequently allowed to regrow for 60–70 days before electrophysiological recording began. Rhythmic vibrissal stimulation at 100 Hz applied to the vibrissa for 1 s could not reliably cause an obvious long-lasting increase in the amplitudes of vibrissa-to-cortex-evoked responses (Fig. 6A). The normalized amplitude of the SEP was 100 ± 2.42% at baseline, and 113.84 ± 4.33% after 100 Hz vibrissal stimulation during 30–60 min (P < 0.05, compared with control group, n = 6, Fig. 6A,C).
Under brief anesthesia with ether, one group of rats had their whiskers clipped daily to 2–3 mm of the skin surface for 7 days during P29–35; all vibrissae were allowed to regrow for 30–40 days before electrophysiological recording began. Rhythmic vibrissal stimulation at 100 Hz applied to the vibrissa for 1 s caused a long-lasting increase in the amplitudes of vibrissa-to-cortex-evoked responses (Fig. 6B). The normalized amplitude of the SEP was 100 ± 4.40% at baseline, and 134.94 ± 8.81% after 100 Hz vibrissal stimulation during 30–60 min (P > 0.05, compared with control group, n = 6, Fig. 6B,C).
Previous studies have described frequency-specific neurons in barrel cortex and these neurons show vibrissal frequency tuning (Andermann et al. 2004). A variety of studies indicated that the neurons in rat barrel cortex are sensitive to velocity (Armstrong-James and Fox 1987; Pinto et al. 2000), and the response of these neurons show velocity sensitivity to sinusoidal frequency-varying stimuli (Arabzadeh et al. 2003). Neurons specifically encode the product of frequency and amplitude, which is proportional to the mean speed of the whisker vibration (Arabzadeh et al. 2003, 2005). In our study, with increased speed of whisker vibration, we recorded increased SEP amplitude (Fig. 1D), in agreement with previous studies (Arabzadeh et al. 2003, 2005).
In the hippocampus and neocortex, LTP is usually elicited by brief stimulation (e.g., 0.5 or 1 s) at high frequency (e.g., 50 or 100 Hz) (Bliss and Collingridge 1993; Huang et al. 1996). But long-term depression (LTD) is usually electrically induced by low-frequency stimulation, typically in the range of 1–5 Hz (Popkirov and Manahan-Vaughan 2011; Goh and Manahan-Vaughan 2013). In this regard, the long-lasting change of SEPs after rhythmic vibrissal stimulation in vivo could also qualify as LTP or LTD.
In previous studies, transient vibrissal deflections generated robust high-frequency oscillations in rat barrel cortex, which could interact with the high-frequency input driven by vibrissal resonance (Jones et al. 2000; Crochet and Petersen 2006). The present study revealed that stimulation of a vibrissa at different frequencies affected the efficacy of synapses in neocortex. This could well serve as a way to regulate the efficacy of the vibrissa–barrel cortex pathway in vivo during behavior.
Although 5 and 20 Hz both fall in the range of natural whisking, 20 Hz whisker stimulation was ineffective in long-term plasticity induction in our study and 5 Hz induced LTD. A recent study also showed that a brief period of rhythmic whisker stimulation (8 or 2 Hz) in mice resulted in a frequency-specific long-lasting increase in the amplitude of SEPs in the contralateral barrel cortex, but 20 Hz stimulation failed to induce a significant increase (Megevand et al. 2009). We used higher frequencies (20–100 Hz) to stimulate whiskers, and our study differed from the previous study in the species used, the direction of vibrissal stimulation, and the stimulation strategy (Megevand et al. 2009). Vibrissal microvibrations are generated by different textures at much higher frequencies (100–1000 Hz) than the whisking movements themselves (8–12 Hz). We investigated the effects of the higher-frequency texture-induced whisker microvibrations on processing in the barrel cortex by applying continuous 20–100 Hz passive whisker stimuli to anesthetized rats for various durations. In the previous study, multichannel epicranial recording of SEPs was used. Although this is a minimally invasive method, the spatial resolution is limited to ∼1 mm and allows for recording from multiple cortical areas simultaneously; thus, it may include signals from outside the barrel cortex of mice (Troncoso et al. 2000). The frequency-specific plasticity in the rat barrel cortex suggests that persistent synaptic information-processing in the rat barrel cortex depends on the convergence of a very specific pattern of afferent activity. Functionally, this may play an important role in encoding similar but distinct sets of information.
Our study showed that the long-lasting enhancement of the vibrissa-to-cortex-evoked response was hemisphere-specific, but not barrel-specific within one hemisphere. Anatomically, injections within barrels show radiating horizontal projecting fibers extending furthest in the extragranular layers (Bernardo et al. 1990). Other experiments in rats have also shown that the septum contains a pattern of linked horizontal pathways within layer IV (Kim and Ebner 1999). Lesions of the septal region between the neighboring barrels are able to prevent horizontal information transmission between the barrels (Fox 1994). This horizontal transmission occurs between neighboring barrels and pathways linking septal regions to neighboring barrels can potentiate this transmission. Previous work also showed that potentiation can be induced in a neighboring barrel on the same side but not the contralateral side (Fox 2002). Similar to our results, experience-dependent plasticity in the barrel cortex includes an increase in transmission between columns corresponding to spared whiskers and those corresponding to deprived whiskers. This change of horizontal transmission is also probably due to a change in the horizontal connections between columns (Glazewski et al. 1998; Fox 2002).
Since one property of the long-lasting plasticity caused by brief rhythmic vibrissal stimulation is the increased strength of horizontal transmission, it is possible that intracortical pathways between neighboring barrels are capable of potentiating under the control of this rhythmic sensory input. So in our study, the long-lasting enhancement of the vibrissa-to-cortex-evoked response diffused within one barrel field due to the horizontal fiber connections within the field.
Our results showed no difference in the amplitude of LTP between the 2 barrel areas corresponding to whiskers A and B. A previous study showed that sites that are able to induce similar LTP in layers II/III by stimulating layer IV are >500 μm distant (Fox 2002). Moreover, if there were fine differences in the LTP amplitude between the 2 barrel areas, they may have been detectable by single-unit recordings.
It is known that tactile learning can transfer from a trained finger to the corresponding finger on the other hand, or from a whisker to that on the other side of the body (Harris and Diamond 2000). But in our study, the duration of conditional stimuli was only 1 s, not long enough to be regarded as sufficient to induce tactile learning. So, the long-lasting enhancement of the unilateral vibrissa-to-cortex-evoked response did not diffuse to the contralateral hemisphere during recording.
It has been reported that electrically induced potentiation in the cortex has a number of features in common with LTP in the hippocampus, that is, they are both dependent on the activation of NMDA receptors (Kirkwood et al. 1993). NMDA receptors are critically involved in the generation of LTP or LTD in the neocortex and hippocampus (Malenka and Bear 2004; Morris 2013; Nabavi et al. 2013). These receptors are also critical for experience-dependent plasticity in the barrel, visual, and auditory systems (Cui et al. 2009, 2010; Huang et al. 2012). Our study showed for the first time that the form of plasticity caused by rhythmic vibrissal stimulation was also dependent on NMDA receptors like the electrically induced or experience-dependent plasticity in the hippocampus and neocortex.
In our study, application of an NMDA receptor antagonist to the barrel cortex directly decreased the vibrissa–barrel cortex-evoked responses. The correlation between the amplitudes of the evoked SEPs and the changes of the LTP amplitudes induced by 100 Hz rhythmic stimulation was very weak (Fig. 5D). So, the influence of APV on the amplitude of the synaptic response did not affect the plasticity.
APV greatly suppressed the long-term plasticity in the barrel cortex, which implies that the induction of LTP occurs only in the neocortex. Our finding that a short period of rhythmic vibrissal stimulation did not induce long-lasting plasticity of the field potential in the thalamus confirmed this speculation. Previous studies showing that experience-induced plasticity always occurs in the thalamocortical projection in the barrel cortex is in agreement with our finding (Rema et al. 2006; Oberlaender et al. 2012). This result is reasonable because the sensory projections from brainstem to thalamus is a relatively simple relay station; however, more active information processing occurs in the thalamocortical projection.
Primary sensory cortex contains topographical representations of the corresponding sensory inputs and these representations are refined during development in response to sensory experience. Barrels form soon after birth (Schlaggar and O'Leary 1994) and are resistant to sensory manipulation after an early critical period (P5) (Henderson et al. 1992; Fox et al. 1996). The experience-related plasticity occurs most markedly during a postnatal time window—the critical period (Fox 1992; Hensch 2004; Itami and Kimura 2012). Sensory deprivation by whisker trimming continues to modify receptive fields (Maravall et al. 2004). Whisker trimming in young, but not mature rats leads to deficits in whisker-based sensory discrimination (Carvell and Simons 1996). In layer IV, LTP and LTD are expressed in thalamocortical synapses throughout the critical period (approximately the first postnatal week in rodents) (Crair and Malenka 1995; Feldman et al. 1998).
But the mechanisms underlying this critical period for plasticity and its closure are unknown. It has been reported that during the critical period, dysregulation of the maturation of glutamatergic signaling affects the timing of the critical period for synaptic plasticity in layer IV (Harlow et al. 2010). There is a hypothesis that a change in subunit composition of NMDA receptors reduces the ability to induce synaptic plasticity that may contribute to the end of the critical period in visual cortex and barrel cortex (Barth and Malenka 2001; Daw, Scott, et al. 2007). These findings indicate that the NMDA receptor subunit NR2B is required for thalamocortical LTP and reduction of the expression of this subunit may terminate the synaptic plasticity in layer IV of barrel cortex at the end of the first postnatal week. However, a switch in excitatory/inhibitory balance may also play an important role in the closure of the barrel cortex critical period, as has been suggested in the visual cortex (Daw, Ashby, et al. 2007).
Although cognitive impairment and loss of recall are important features of general anesthesia, neurons still code in the anesthetized state, and plasticity experiments remain relevant. Mechanisms related to LTP and LTD studied in brain slices or anesthetized animals are still assumed to be important steps in learning and memory formation (Simon et al. 2001). However, further studies need to be carried out in freely moving animals.
Altogether, we investigated neural responses in the cortex and its plasticity to high- or low-frequency low-amplitude vibrissal stimuli. We showed that a brief period of rhythmic vibrissal stimulation in anesthetized rats resulted in a frequency-dependent long-lasting increase or decrease in the amplitude of SEPs in the contralateral barrel cortex. This vibrissa-to-cortex long-term plasticity was side- but not barrel-specific, NMDA receptor-dependent, and impaired by the development of the barrel system. Our results suggest that natural rhythmic whisker activity modifies sensory information processing in the cerebral cortex.
All authors contributed to the design of the experiments. Y.H., M.D.H., and M.L.S. performed experiments. All authors contributed to the analysis and interpretation of data. Y.Q.Y. wrote the manuscript. Y.Q.Y. and S.D. supervised the work.
We thank Professor I.C. Bruce and Dr H. Wang for critical reading of this manuscript, and the anonymous reviewers for critical and insightful suggestions. Conflict of Interest: None declared.