Transcranial direct-current stimulation modulates synaptic mechanisms involved in associative learning in behaving rabbits

Results

In a first series of experiments, we tested whether simultaneous tDCS applied to the somatosensory cortex could modify the characteristics of local field potentials (LFPs) evoked in the vibrissa S1 area of alert behaving animals by air puff stimulation of the contralateral whisker pad (Fig. 1A). Whisker stimulation evoked a short-latency (mean ± SEM, 15.4 ± 2.5 ms; n = 103) negative LFP (N1, Fig. 1B), followed by late positive and negative components (Fig. S1A) and then a smaller LFP evoked at the end of the air puff. The amplitude of air puff-evoked LFPs depended on the stimulus intensity and on the recording site across the somatosensory cortical layers (Fig. S1 A and B).

The N1 component of the LFP was amplified by the simultaneous presence of an anodal tDCS and was reduced by cathodal tDCS (Fig. 1B). Indeed, the N1 component was significantly increased in both amplitude (up to a mean maximum of 153 ± 16.2%; n = 5; P ≤ 0.05, one-way ANOVA) and area (up to a mean maximum of 247.1 ± 82.8%, n = 5, P ≤ 0.05, one-way ANOVA) in response to anodal tDCS (from 0.5 mA to 2 mA; Fig. 1C). In contrast, cathodal tDCS (from −0.5 mA to −2 mA) decreased the amplitude (by up to a mean maximum of 63.9 ± 13.1%; n = 5) and area (by 61.3 ± 16.8%; n = 5) of the N1 component compared with N1 values collected from controls and during the presentation of anodal tDCS (Fig. 1C). Thus, concurrent tDCS applied to the somatosensory cortex was capable of increasing or decreasing the amplitude and area of LFPs evoked by the simultaneous air puff stimulation of the whisker pad. Moreover, tDCS had a persistent poststimulus effect; the inhibitory effect of 1 mA of cathodal tDCS on N1 amplitude was significantly [F_(7,28,760) = 16.943; P < 0.001, repeated-measures ANOVA] evident for up to 30 min after the end of the DC stimulation (Fig. 1D). Confirming previous results reported in humans (16), these experiments show that tDCS can modify the functional properties of LFPs evoked in the somatosensory area by natural stimulation of facial mechanoreceptors, including a noticeable poststimulus effect on cathodal currents.

In a second series of experiments, we tested whether simultaneous tDCS could also modulate the acquisition of a well-known model of associative learning, the classical conditioning of eyelid responses, using stimulation of the whisker pad as a conditioned stimulus (CS). A previous study indicated that specific areas of the somatosensory cortex are activated during this type of associative learning, depending on the skin site of the CS presentation (17). As in that study, we used a train of pulses (100 ms, 200 Hz) presented to the whisker pad as CS, followed 250 ms later by an air puff presented to the ipsilateral cornea as an unconditioned stimulus (US). Two habituation sessions (H1 and H2, when the CS was presented alone) and 10 conditioning sessions (C1–C10, with paired CS–US presentation) were carried out. Each conditioning session consisted of 66 trials (6 series of 11 trials each) separated at random by intervals of 50–70 s. Of the 66 trials, 6 were test trials in which the CS was presented alone (Fig. 2A). Fig. 2B presents representative samples of classically conditioned eyelid responses collected from C2 and C8 in the control situation (conditioning series 1, traces in black) and during anodal (red trace in C2) and cathodal (blue trace in C8) stimulation. In agreement with previous reports (18), control animals reached asymptotic values (>80% of conditioned responses per session) for their learning curves by C3 or C4 (Fig. 2C). The experimental group was presented with three blocks of anodal tDCS (≤1 mA) applied over the right somatosensory cortex, contralateral to the CS–US, during conditioning series 2, 4, and 6 of C2 (Fig. 2C). The presence of concurrent anodal tDCS during C2 was associated with a significant increase in the percentage of conditioned responses [59.9 ± 11.2 for controls vs. 95.6 ± 3.5 for stimulated animals; F_(1,5,4) = 9.279; P < 0.05, one-way ANOVA] (Fig. 2C). The experimental group also was presented with three blocks of cathodal tDCS (≤−1 mA) during conditioning series 2, 4, and 6 of C8 (Fig. 2A). In this case, stimulated animals showed a significantly lower percentage of conditioned responses compared with the control group [49.3 ± 10.5 vs. 90.5 ± 1.7; F_(1,5,4) = 14.934; P < 0.05, one-way ANOVA]. Thus, by using tDCS to decrease or enhance sensory perception process, we were able to modulate the animals’ ability to acquire a typical associative learning test, such as classical eyeblink conditioning (19). An intrasession analysis of classical conditioning evolution during sessions in which anodal (C2) or cathodal (C8) tDCS was presented to the somatosensory cortex of the experimental group is illustrated in Fig. 2D. As shown, concurrent anodal tDCS increased the rate of conditioned responses across the session, whereas cathodal tDCS decreased this rate during the stimulation session as well as thereafter. This findings support the idea that tDCS modifies the associative strength of cortical circuits involved in associative learning (20, 21).

We carried out two additional series of experiments to examine the putative mechanisms underlying the short- and/or long-term effects of tDCS on the somatosensory cortex. In the first of these experiments, we performed a paired-pulse test in two animals to determine whether presynaptic processes involved in the thalamocortical sensory input could be affected by simultaneous anodal or cathodal tDCS. Fig. 3A illustrates monosynaptic (>6 ms) field excitatory postsynaptic potentials (fEPSPs) evoked at the vibrissa S1 cortex by pairs of pulses (at 20-ms intervals) applied to the ipsilateral VPM nucleus of the thalamus (22). As shown in Fig. 3B, the amplitude (in millivolts) of fEPSPs evoked in the somatosensory cortex by the first pulse increased steadily with current strength until it reached asymptotic values. In contrast, fEPSPs evoked by the second pulse increased roughly in parallel with the fEPSPs evoked by the first pulse (but with larger values) up to a certain stimulus intensity, after which the fEPSP slopes evoked by the second pulse were smaller than those evoked by the first pulse. In fact, the second/first paired-pulse ratio decreased progressively from facilitation to depression, with an inflection point at ∼2.2 mA (Fig. 3C). The paired-pulse test indicated that thalamocortical synapses present a paired-pulse facilitation at low intensities that is reversed to paired-pulse depression at higher intensities. The simultaneous presence of anodal or cathodal tDCS modulated, and even reversed, the paired-pulse relationship at either low (Fig. 3D) or high (Fig. 3E) stimulation intensities, suggesting that tDCS modifies thalamocortical synapses at presynaptic sites (23).

In the second series of additional experiments, we examined the synaptic mechanisms underlying the long-lasting depressant effects on cortical activity produced by cathodal tDCS. Based on previous studies of the molecular processes underlying LTD (24, 25), we investigated the effect of the selective A1 adenosine receptor antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) when injected at the cortical recording site. Local microinjection of DPCPX significantly diminished the LTD induced after cathodal tDCS compared with vehicle controls (n = 3; P < 0.001, one-way ANOVA) (Fig. 3F).

Discussion

This study demonstrates that associative learning processes can be modulated by tDCS applied over the somatosensory cortex, and contributes to our understanding of the neural mechanisms involved in short-term and long-lasting tDCS-induced effects. Although previous in vitro (26–28) and acute (1–3, 29, 30) animal studies have demonstrated the modulatory effects of DC on cortical excitability, until now no direct evidence of the effects of tDCS based on cortical recordings of electrical activity in alert animals has been reported. And although behaving animals have been successfully used for the study of different tDCS implications in basic (31–33) and clinical (34–36) aspects, the animal model presented here allows the combination of tDCS application in alert animals with invasive recording of cortical electrical activity.

We report here that intracortical recording of cortical sensory LFP during simultaneous tDCS shows an increase or decrease in the amplitude and area of the main field potential components, depending on the anodal or cathodal polarity of the applied current. This short-term effect corresponds to the effects reported in the motor cortex based on measurements of motor evoked potentials in humans (7, 37–40) and in anesthetized animals (29). Although the complete neurophysiological characterization of short-term effects is difficult due to the tDCS-induced artifacts in electrophysiological recordings, the behavioral changes observed during tDCS application point to the same general polarity-specific effects when applied to somatosensory (16, 41) and visual (42, 43) cortices. The changes in sensory LFPs or fEPSPs in response to whisker or VPM nucleus stimulation during tDCS reported here suggest a modification in the number of neurons recruited by the same stimulus, with consequent reinforcement or attenuation of the subjective perception of the stimulus.

To test whether sensory thalamic afferent signals can be modified by simultaneous tDCS, we used a paired-pulse test, applying stimulation at the thalamic VPM nucleus and recording the evoked LFPs in the somatosensory cortex (22). The paired-pulse test is commonly used to highlight presynaptic effects on neurotransmitter release (23). We found that the simultaneous presence of anodal or cathodal tDCS modulates the paired-pulse relationship, suggesting that tDCS modifies thalamocortical synapses at presynaptic sites. Our findings do not exclude the possibility of additional postsynaptic modifications, however. Thus, changes in the LFPs induced by whisker stimulation during tDCS seem to be the result of direct changes in the membrane potential of cortical neurons, together with changes in the neurotransmitter release of thalamocortical sensory afferents.

Only cathodal tDCS was able to induce significant poststimulus changes (i.e., decreased amplitude and area of the N1 component) in the recorded LFPs, which persisted for up to 30 min. In contrast, no changes were observed after anodal tDCS. Similar results have been reported in sensory evoked potentials after tDCS application over the human somatosensory cortex. Dieckhöfer et al. (11) reported a significant reduction in the N20 component after cathodal tDCS, but no changes after anodal tDCS. Similarly, in another study, pain perception and the amplitude of laser-evoked potentials were decreased after cathodal tDCS over the somatosensory cortex in humans, whereas no changes were observed after anodal tDCS (44).

Previous studies performed in humans have attempted to elucidate neural mechanisms underlying the long-lasting effects observed after tDCS through the systemic administration of various drugs (45). Thus, NMDA receptors have been implicated in cathodal after-effects, and NMDA and GABA_A receptors have been implicated in anodal after-effects (46, 47). Interestingly, experiments performed in rats support the results of human studies suggesting the participation of NMDA receptors in the process after anodal polarization (48). In addition, anodal DC stimulation applied over the rat sensorimotor cortex has been shown to modify adenosine-elicited accumulation of cAMP (4), inducing increases in protein kinase C and calcium levels (5, 6). Moreover, the involvement of adenosine A1 receptors in the LTD of cortical circuits has been reported recently (24, 49). Here we show that local microinjection of the adenosine A1 receptor antagonist DPCPX prevents the LTD evoked in the somatosensory cortex after cathodal tDCS. In contrast to other adenosine receptors, the adenosine A1 receptor suppresses neuronal activity by coupling with protein Gi to inhibit the adenylate cyclase–cAMP–protein kinase A signaling pathway (50). Previous reports adenosine-dependant increase in cAMP after anodal polarization (4–6), along with the evidence presented here implicating adenosine A1 receptor in cathodal after effects, point to a major role of adenosine in the modulation of cortical excitability induced by tDCS. In accordance with this, the selective pharmacologic manipulation of adenosine receptors during simultaneous tDCS may have important clinical applications.

Although the results reported here stress the importance of using animal models to study the immediate effects and after-effects of tDCS, some limitations of this experimental model must be taken into account. These limitations are associated mainly with differences between the electrodes used in rabbits and those used in clinical applications, hindering direct comparisons of current density and voltage distribution. According to the current distribution spherical head model (51), the maximum current density used in the rabbit brain was higher than that applied in humans during tDCS (Fig. S2). Nevertheless, no histopathological alterations were observed in the stimulated region (Fig. S3), confirming previous studies on the effects of tDCS in rats (31). Another limitation of this animal model is the simplicity of the rabbit cortical surface (i.e., the absence of circumvolutions) compared with the complex surface of the human cortex.

In conclusion, our results reported here support the findings of previous studies in humans regarding the effects of tDCS on the cerebral cortex, confirm the potential of this technique for modulating associative learning, and demonstrate the participation of adenosine A1 receptors in its selective actions on cortical circuits.

Methods

Supplementary Material