The dose-dependent effect of the D2R agonist quinpirole microinjected into the ventral pallidum on information flow in the limbic system

Abstract

The ventral pallidum (VP) receives its primary inputs from the nucleus accumbens (NAC) and the basolateral amygdala (BLA). We demonstrated recently that in the VP, the D2 DA receptor (D₂R) agonist quinpirole dose-dependently facilitates memory consolidation in inhibitory avoidance and spatial learning. In the VP, D₂R can be found both on NAC and BLA terminals. According to our hypothesis, quinpirole microinjected into the VP can facilitate memory consolidation via modulation of synaptic plasticity on NAC and/or BLA terminals.

The effect of intra-VP quinpirole on BLA-VP and NAC shell-VP synapses was investigated via a high frequency stimulation (HFS) protocol. Quinpirole was administered in three doses into the VP of male Sprague-Dawley rats after HFS; controls received vehicle. To examine whether an interaction between the NAC shell and the BLA at the level of the VP was involved, tetrodotoxin (TTX) was microinjected into one of the nuclei while stimulating the other nucleus.

Our results showed that quinpirole dose-dependently modulates BLA-VP and NAC shell-VP synapses, similar to those observed in inhibitory avoidance and spatial learning, respectively. The lower dose inhibits BLA inputs, while the larger doses facilitates NAC shell inputs. The experiments with TTX demonstrates that the two nuclei do not influence each others’ evoked responses in the VP.

Power spectral density analysis demonstrated that independent from the synaptic facilitation, intra-VP quinpirole increases the amplitude of gamma frequency band after NAC HFS and BLA tonically suppresses the NAC’s HFS-induced gamma facilitation. In contrast, HFS of the BLA results in a delayed, transient increase in the amplitude of the gamma frequency band correlating with the LTP of the P1 component of the VP response to BLA stimulation.

Furthermore, our results demonstrate that the BLA plays a prominent role in the generation of the delta oscillations: HFS of the BLA leads to a gradually increasing delta frequency band facilitation over time, while BLA inhibition blocks the NAC’s HFS induced strong delta facilitation.

These findings demonstrate that there is a complex interaction between the NAC shell region and the VP, as well as the BLA and the VP, and support the important role of VP D₂Rs in the regulation of limbic information flow.

1. Introduction

The mesencephalic dopaminergic (DA) neurons in the ventral tegmental area (VTA) innervate numerous cortical and subcortical limbic structures, including the prefrontal cortex (PFC), hippocampus, nucleus accumbens (NAC), basolateral and central amygdala (BLA and CeA, respectively) and the ventral pallidum (VP) [1, 2]. D2-like dopamine receptors (D₂R) are one of the two subgroups of DA receptors and can be localized throughout the limbic system [3].

There are several lines of evidence supporting the role of DA and its D₂Rs in synaptic plasticity and memory consolidation [4]. DA receptors, including the D₂ subtype, play an important role in information flow in the NAC and can regulate the balance among the PFC, hippocampal and BLA inputs onto medium spiny neurons (MSNs) [5–7]. Similarly, DA, via its D₁ and D₂ DA receptor subtypes, regulates the information flow within the BLA as well, shifting the balance from the mPFC input to the sensory cortical ones, modulating aversive Pavlovian conditioning [8–11]. In the NAC shell region, the activation of D₂Rs blocks long-term potentiation (LTP), and it is necessary for the induction of long-term depression (LTD) in PFC terminals [5]. In the case of the BLA fibers ending on the NAC shell MSNs, both the formation of LTP and LTD require D₂R activation [6]. In the BLA, the activation of the D₂Rs facilitates LTP formation [12]. It was demonstrated that blocking the D₂Rs in the NAC shell region impairs one-trial inhibitory avoidance learning [13] and spatial learning [14] as well. Furthermore, memory consolidation in inhibitory avoidance learning requires concurrent DA receptor activation in the BLA and the NAC shell region [15]. These results suggest that D₂Rs of the NAC shell region and/or the BLA are essential for avoidance and spatial learning processes, and in parallel, these receptors can modulate LTP and/or LTD, the electrophysiologycal correlates of synaptic plasticity.It was shown that the D₂R-expressing NAC neurons innervate the VP [16], and D₂Rs can be located presynaptically, most probably on the GABAergic synaptical ending on the VP neurons [17]. In addition, the BLA also innervates the VP by glutamatergic fibers, evoking both excitatory and inhibitory responses [18]. Similar to the NAC terminals, D₂Rs can be found presynaptically on the BLA glutamatergic nerve endings as well [19]. D₂Rs are also located postsynaptically on the output and interneurons of the VP [17]. It is well-known that GABA- and glutamate-evoked responses are modulated by DA and its receptors in the VP [20], however it is not clarified how they influence synaptic plasticity there.

The ventral pallidum is a basal forebrain limbic structure, the part of the loop controlling entry of information into long-term memory [21] and via its innervation of the VTA is the main regulator of DAergic population activity [22]. We have shown recently that the intra-VP microinjection of the D₂R agonist quinpirole induces place aversion and dose-dependently modulates VTA population activity, as well asinfluences DA neuron burst activity [23]. Furthermore, the D₂R antagonist sulpiride develops place preference in experimental animals [24], likely via a remote feedback control mechanism over VTA DAergic activity, which is supported by the fact that the D₂R antagonist microinjected into the VP increases DA levels in the VP on a large scale [25], even though the number of the D₂ autoreceptors in the VP is low [17]. Additionally, we have also shown that the quinpirole microinjected into the VP dose-dependently facilitates memory consolidation related to inhibitory avoidance [26] and spatial learning [27] processes. This raises the question reagarding what is the mechanism of the learning-facilitatory effect of intra-VP D₂R agonist quinpirole? One alternative explanation, as we have shown recently [23], can be that the intra-VP quinpirole enhances VTA population activity and consequently it can increase the released DA in various brain regions to promote synaptic plasticity. However, in spatial learning, the largest dose of intra-VP quinpirole was the most effective [27], but it suppressed VTA DAergic activity [23]. Therefore, a more plausible explanation could be that DA and perhaps the activation of D₂Rs can facilitate learning processes via the local modulation of the synaptic plasticity.

In the present study, we aimed to clarify how the D₂R agonist quinpirole microinjected into the VP can affect synaptic plasticity between theNAC shell region and the VP, as well as the BLA and the VP (see Figure 1) For this reason, the NAC shell region or the BLA was stimulated and the stimulation-evoked local field potential (LFP) responses in the VP were recorded. The effects of intra-VP quinpirole on NAC shell-VP and BLA-VP synapses were investigated applying a high frequency stimulation (HFS) protocol and intracerebral VP quinpirole microinjections. To isolate the effects of the NAC shell and the BLA from each other, stimulation of the NAC shell region following TTX microinjected into the BLA, or stimulation of the BLA following TTX microinjected into the NAC shell region was carried out, recording evoked responses in the VP. In addition to the effects on the evoked responses and synaptical plasticity, the spontaneous LFP activity of the VP was also recorded before every stimulation and was analyzed by means of Fast Fourier transformation.

Figure 1:

Simplified circuit diagram representing the main connections of the ventral pallidum (VP). (A): The two main input nuclei of the VP are the nucleus accumbens (NAC) and the basolateral amygdala (BLA). The NAC sends GABAergic inhibitory innervation to VP, while the BLA innervates the VP with glutamatergic fibers, which, in addition to the excitatory effects, can exert inhibitory ones as well via metabotropic receptors in the VP [19]. The D2 dopamine receptors (D₂Rs) can also be found presynaptically in the VP, on both the NAC and the BLA terminals. We have also shown that the D₂R agonist quinpirole microinjected into the VP dose-dependently facilitates memory consolidation related to inhibitory avoidance. According to our hypothesis, quinpirole can exert its learning-facilitating effects via the modulation of synaptic plasticity, so we assume that intra-VP quinpirole modifies synaptic strength of the fibers of the two main input nuclei. (B) To test this hypothesis, stimulation electrodes were implanted into the NAC shell region or into the BLA, as well as recording electrode and guide cannula for the microinjections into the VP. NAC shell region or BLA was stimulated, LFP recorded in the VP, and quinpirole (0.1μg, 1.0μg or 5.0μg, in 0.4μl physiological saline, controls received vehicle) microinjected into the VP. The high frequency stimulation (HFS) of the NAC shell region evoked LTD in NAC-VP synapses, while HFS of the BLA induced LTP in the P1 (probably inhibitory) short latency LFP component, but the N1 (probably excitatory) component was not changed. The 1.0μg or 5.0μg doses of intra-VP quinpirole shifted LTD of the NAC-VP synapses to LTP. In contrast, the 0.1μg dose of quinpirole induced LTD in the BLA stimulation evoked N1 component, furthermore it inhibited the LTP of the P1 component. This dose-dependence is consistent with our previous behavioral findings demonstrating that the large dose of intra-VP quinpirole improves spatial learning processes, whilst the small dose facilitates inhibitory avoidance learning.

2. Methods

2.1. Acute preparations

In the acute electrophysiological experiments rats were anesthetized with chloral hydrate (400 mg/kg, i.p.; Sigma) and fixed in a stereotaxic frame. Stainless steel guide cannulae (26 gauge) were implanted unilaterally dorsal to the VP (AP: −0.3mm from bregma; ML: + 2.2mm from midline; V: −5.8mm from brain surface), the NAC shell region (AP: −2.2mm from bregma; ML: + 1.0mm from midline; V: −6.0mm from brain surface), or the BLA (AP: −2.9mm from bregma; ML: + 5.0mm from midline; V: −7.6mm from skull) according to the Paxinos and Watson stereotaxic atlas [28]. For the local field potential (LFP) measurements tungsten electodes were lowered into the center of the VP (AP: −0.3mm from bregma; ML: + 2.2mm from midline; V: −7.4mm from brain surface). Concentric bipolar stimulation electrodes (SNEX 100x) were implanted into the NAC shell region (AP: −2.2mm from bregma; ML: + 1.0mm from midline; V: −7.0mm from brain surface) or into the BLA (AP: −2.9mm from bregma; ML: + 5.0mm from midline; V: −8.6mm from skull). During the acute experiments body temperature was held constant at 37°C using a thermostatically-controlled heating pad (Fintronics) and anesthesia was maintained by i.p. injection of chloral hydrate as needed to maintain suppression of the hindlimb compression reflex.

2.2. Subjects and drug administration

Experiments were performed on 95 male Sprague-Dawley rats in accordance with the NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh. The present study is reported in accordance with ARRIVE guidelines. The rats were housed in a temperature controlled room (22±1°C) under standard housing conditions with free access to food and water and a 12h light/ dark cycle.

The D2 dopamine receptor agonist (-)-Quinpirole hydrochloride (Sigma–Aldrich Co., Q102) dissolved in physiological saline was injected unilaterally into the VP. Quinpirole was administered in three different doses (0.1μg, 1.0 μg or 5.0μg, in 0.4μl physiological saline, 0.98mM, 9.77mM and 48.89mM, respectively). Tetrodotoxin dissolved in Dulbecco's phosphate-buffered saline (dPBS) was microinjected into the NAC shell region or in the BLA in 0.4μl volume (TTX, 1.0μM, Sigma–Aldrich Co.). Control animals received vehicle only. Solutions were kept in + 4°C before administration. Drugs or vehicle were microinjected through stainless steel injection tubes inserted into the guide cannulae. The microinjection pipette was attached to a 10μl Hamilton microsyringe via polyethylene tubing. All microinjections were delivered over a 60sec period. After completion of the microinjection, the pipette was left in place for an additional 60sec to allow diffusion into the surrounding tissue as well as to prevent the backflow of the solution along the insertion track.

2.3. NAC Shell or BLA HFS stimulation - VP field potential recordings - VP D₂R agonist microinjections

Anaesthetized rats were placed in a stereotaxic apparatus and concentric bipolar stimulation electrodes were lowered into the NAC shell region or into the BLA. Ipsilateral guide cannulae were implanted into the VP and tungsten electrodes for local field potential recording were inserted into and extending 1.3mm below the tips of the implanted guide cannulae. Field potential signals were recorded and amplified 2000x with an AC amplifier and band-pass filtered at 0.3–300Hz. The electrophysiological signal and the data were transferred via a Powerlab interface (AD instruments) to a computer with LabChart v.8 software (sampling frequency 40kHZ). The protocol for the experiments are illustrated in Figure 2A. At the beginning of the recording 100 single current pulses (0.2ms, 0.5Hz) were delivered to reveal the baseline NAC shell/BLA stimulation – VP response curve in the 0.2–1.0mA stimulation intensity range (increasing intensity gradually at 0.2mA intervals). Current intensity was adjusted to evoke an average 65% of the maximal response (standard deviation:13%), which was then used for subsequent experiments. Every 10 evoked responses were averaged and then the peak of the averaged short-latency responses determined and averaged. Single pulse stimulation of the NAC shell region or the BLA evoked characteristic local filed potentials in the VP (Fig. 3B and 4B, respectively). Only the components of the responses with an average peak latency longer than 4.0msec and shorter than 15.0msec or 30.0msec (so P1, as well as N1 and P1 in the case of NAC stimulation or BLA stimulation, respectively) were considered, and their peak amplitudes were measured and used for the analysis. The NAC shell stimulation evoked a three-component response (P1, N1 and P2, Figure 3B) and only the first positive component (P1) was considered and analyzed (30 rats, average peak latency: 8.7msec, standard deviation: 0.73msec). With BLA stimulation, a four-component response was evoked (N1, P1, N2 and P2, see Figure 4B) and the first two components (N1 and P1) of the VP LFP responses were analyzed: specifically, a negative (29 rats, average peak latency: 10.4msec, standard deviation 1.8msec) and a subsequent positive (29 rats, average peak latency: 26.0msec, standard deviation 1.2msec) wave. All short-latency responses were normalized by the average of the corresponding first 30 baseline responses (1 if the amplitude of the response equals the average of the first 30 responses).

Figure 2:

Schedule of the stimulation-microinjection experiments. (A): Schedule of the experiments where the NAC shell or the BLA was stimulated and the LFP was recorded in the VP. The D₂R agonist quinpirole was microinjected into the VP in 0.1μg, 1.0μg or 5.0μg doses, the control rats received vehicle. (B): Schedule of the experiments where the NAC shell or the BLA was stimulated and TTX was microinjected into the BLA or the NAC shell region, respectively. LFP was recorded in the VP.

Figure 3:

Effect of intra-VP quinpirole on the plasticity of the NAC shell region-VP synapses and the power spectral density of the VP LFP. (A): Schematic illustration of stimulation electrode placement in the NAC shell region and recording electrodes as well as cannulae in the VP as shown in coronal sections of rat brain taken from Paxinos and Watson atlas [28]. The numbers refer to anterior–posterior distance from bregma in mm. (B): Average of baseline VP responses to NAC shell region. Each stimulus pulse in the NAC shell region evoked characteristic local filed potentials in the VP. Vertical axis shows the amplitude of the response in microvolts (μV), while the horizontal axis depicts time in milliseconds (msec). Mean baseline responses of each animal are averaged and represented as continuous black line. Standard deviation is visualized as grey bands surrounding the mean (mean±SD). Arrow above the curve indicates time of the stimulation. P1, N1 and P2 are the components of the response, P1, as a short-latency component was considered and analyzed. (C): Effect of intra-VP quinpirole on average amplitudes of P1 components of the VP responses to NAC shell stimulation. Mean of 100 normalized responses (±S.E.M.) is represented in each group and trial. During the baseline stimulation and the 1st post HFS+inj SS trial there was no significant difference among the groups. After HFS stimulation and following microinjections, a slight LTD can be observed by time in the control and 0.1μg quinpirole treated group. However, 1.0μg and 5.0μg doses of quinpirole gradually shifted LTD to LTP. (D): Effect of intra-VP quinpirole on average amplitudes of P1 components of the VP responses to NAC shell stimulation. Five consecutive means of 20 normalized responses (±S.E.M.) are represented in each group and trial. (E): Effect of intra-VP quinpirole on power spectral density in the 30–49Hz frequency band. Mean of the frequencies in 30–49Hz frequency band (±S.E.M.) is represented in each group and trial. The 5.0μg dose of quinpirole increased amplitude in gamma frequency band in third post HFS+inj SS trial. In the figures * and # indicate significant differences among the groups (p<0.05).

Figure 4:

Effect of Intra-VP quinpirole on the plasticity of the BLA-VP synapses. (A): Schematic illustration of stimulation electrode placement in the BLA and recording electrodes as well as cannulae in the VP as shown in coronal sections of rat brain taken from Paxinos and Watson atlas [28]. The numbers refer to anterior–posterior distance from bregma in mm. (B): Average of baseline VP responses to BLA stimulation. Each stimulus pulse in the BLA evoked characteristic local filed potentials in the VP. Vertical axis shows the amplitude of the response in microvolts (μV), while the horizontal axis depicts time in milliseconds (msec). Mean baseline responses of each animal are averaged and represented as continuous black line. Standard deviation is visualized as grey bands surrounding the mean (mean±SD). Arrow above the curve indicates time of the stimulation. N1, P1, N2 and P2 are the components of the response, N1 and P1 were regarded in the analysis. (C) and (D): Effect of intra-VP quinpirole on average amplitudes of N1 and P1 components of the VP responses to BLA stimulation (respectively). Mean of 100 normalized responses (±S.E.M.) is represented in each group and trial. During the baseline stimulation there was no significant difference among the groups. After HFS and following microinjections, 0.1μg quinpirole significantly decreased the amplitude of the N1 component and blocked LTP formation observed in the control group. In the figure * and # indicate significant differences among the groups (p<0.05, see details in the text). (E) and (F): Effect of intra-VP quinpirole on average amplitudes of N1 and P1 components of the VP responses to BLA stimulation. Five consecutive means of 20 normalized responses (±S.E.M.) are represented in each group and trial. (G): Analysis of power spectral density in the 0–4Hz (delta) frequency band. Mean of the frequencies (±S.E.M.) is represented in each group and trial. HFS of the BLA resulted in increased amplitudes of the delta frequency band in the post HFS trials compared to the baseline. (H): Analysis of power spectral density in the 30–49Hz (gamma) frequency band. Mean of the frequencies (±S.E.M.) is represented in each group and trial. HFS of the BLA induced a delayed, transient increase in the amplitude of the gamma frequency band, in the 1st post HFS+inj SS trial. In the figures * and # indicate significant differences among the groups (p<0.05).

After determining the baseline, a high-frequency stimulation (HFS) train of 100 pulses at 50Hz (1.0mA) was applied in the NAC shell or in the BLA. Fifteen minutes after the HFS, the LFP electrode was temporarily removed and microinjection cannula was inserted into the guide tube. Quinpirole was microinjected at different doses into the VP; control animals received vehicle. After the microinjection and lowering the LFP electrode back to its original position, 1 hour, 1.5 and 2 hours after the HFS a train of 100 single current pulses (SS, 0.2ms, 0.5Hz, using the chosen intensity) was performed and the average of 100 responses was determined. At the end of the recordings the location of the stimulation electrode was marked by administering 10sec cathodal current at 200μA. Spontaneous LFPs were recorded (sampling frequency 1.0kHZ) before every single current pulse stimulation (SS), for 15min before the first, and 3 min before subsequent SSs. Spontaneous LFPs were analyzed in 3 min bins applying the Fast Fourier transformation Welch method (length of segments: 1024, overlap of segments: 10, using Python scripts) and then normalized by the sum of the frequency amplitudes. We evaluated power spectral density in the 0–49Hz frequency domain. Thereafter, the amplitudes in the frequency domain were normalized by the average frequency amplitudes of the first 15min LFP recordings to reveal changes in the frequency amplitudes.

2.4. NAC Shell or BLA HFS stimulation - BLA or NAC Shell TTX microinjection - VP field potential recordings

Using the same procedure (Figure 2B) as above, bipolar stimulation electrodes were lowered into the NAC shell region or into the BLA, while ipsilateral guide cannulae were implanted into the BLA or into the NAC shell region, respectively. Tungsten electrodes for local field potential recordings were lowered into the VP.

Current intensity was adjusted to evoke an average 68% of the maximal responses (standard deviation:12%), which was then used for subsequent experiments. Every 10 evoked responses were averaged, and then the peak of the averaged short-latency responses determined and averaged. The VP LFP response components were considered and analyzed when the maximum or minimum poststimulus delay was within the 4.0–15.0 or 30.0msec range (i.e., P1, as well as N1 and P1 in the case of NAC shell or BLA stimulation, respectively). The average peak latency of the NAC shell stimulation-evoked short-latency component was 8.9msec (16 rats, standard deviation: 0.36msec). The peak latency of BLA stimulation-evoked first component of the VP LFP response was 11.7msec (15 rats, standard deviation 1.9msec) and the peak latency of the subsequent positive wave was 26.3msec (15 rats, standard deviation 0.87msec).

After determining the baseline, TTX was microinjected into the NAC shell region or into the BLA. Ten minutes after the microinjection a train of 100 single current pulses (SS, 0.2ms, 0.5Hz, using the chosen intensity) was delivered and the average response determined. Thirty-forty minutes after the microinjection a high-frequency stimulation (HFS) train of 100 pulses at 50Hz (1.0mA) was applied in the NAC shell or in the BLA. Six to eight minutes after the HFS, a train of 100 single current pulses (SS, 0.2ms, 0.5Hz) was applied. At the end of the recordings the location of the stimulation electrode was marked by administering 10sec cathodal current at 200μA. As described above, spontaneous LFP was recorded before every SS (for 15min before the first and 3min before the later ones) and analyzed in the same way.

2.5. Histology

At the end of electrophysiological recordings a lethal dose of chloral hydrate (additional 400mg/kg i. p.) was administered to the anesthetized rats for euthanasia. The rats were decapitated and the brains removed before fixing the tissue in 8% paraformaldehyde for approximately 48h and then transferred to 25% sucrose solution for cryoprotection. The brains were frozen and sliced coronally (60μm), and stained with a combination of neutral red and cresyl violet. Microinjection sites were reconstructed according to the rat brain stereotaxic atlas [28]. Only data from animals with correctly placed cannulae were analyzed.

2.6. Statistical analyses

For statistical analysis two-way mixed ANOVA followed by Bonferroni post hoc test were applied. Data are presented as mean ± standard error of the mean (S.E.M.). Statistical significance was established at p<0.05. In two-way mixed ANOVA analysis trial was regarded as within subject factor, while treatment as between subject factor. To examine the possible correlations between the variables Pearson Correlation test was used.

3. Results

3.1. Effects of NAC Shell or BLA HFS stimulation on VP field potentials after VP D₂R agonist microinjections

Stimulation electrodes were placed into the NAC shell region or into the BLA and recordings were made from the VP. Histological examination showed that the microinjection cannulae were located precisely in the VP and stimulation electrodes in the NAC shell region or in the BLA in 59 of the 63 rats. Schematic illustration of cannula placements is shown in Figures 3A & 4A. The remaining 4 (4/63) rats were excluded from the statistical analysis due to incorrect placements.

In the NAC shell stimulation experiments (Figure 3C, D), two-way mixed ANOVA analysis indicated a significant treatment effect (F(3,30.000)=3.697, p=0.022), but not a significant trial effect (F(3,90.000)=0.535, p=0.659); however, a significant treatment * trial interaction was found (F(9,90.000)=3.487, p=0.001). Bonferroni post hoc test revealed that there was no significant difference among groups within the baseline SS and the 1st post HFS+inj SS. However, in the 2nd post HFS+inj SS the 5.0μg quinpirole dose increased significantly the normalized amplitude of the first VP LFP component compared to the control and 0.1μg quinpirole treated groups (p=0.016, p=0.036, respectively). In the 3rd post HFS+inj SS both the 1.0μg and the 5.0μg quinpirole had similar effects facilitating the short-latency responses compared to the controls (p=0.011, p=0.014, respectively) and the 0.1μg quinpirole treated group (p=0.007, p=0.008, respectively).

The analysis of the power spectral density revealed significant differences among the trials and among the groups (summary of the results of the power spectral density analysis can be seen in Table 1). In the 0–4Hz frequency range two-way mixed ANOVA analysis showed a nonsignificant treatment effect (F(3,30.000)=0.250, p=0.860), a significant trial effect (F(3,90.000)=5.119, p=0.003); and a nonsignificant treatment * trial interaction was found (F(9,90.000)=1.069, p=0.393). Post hoc test demonstrated a slight but significant increase in the amplitudes of the 1st and 2nd post HFS+inj SS trials compared to that of the baseline (p=0.028, p=0.002, respectively). In the 4–8Hz frequency range two-way mixed ANOVA analysis demonstrated a nonsignificant treatment effect (F(3,30.000)=2.425, p=0.085), a significant trial effect (F(3,30.000)=23.138, p=0.001); and a nonsignificant treatment * trial interaction was found (F(9,30.000)=2.190, p=0.052). Post hoc test demonstrated that the amplitudes of the 1st, 2nd and 3rd post HFS+inj SS trials were slightly but significantly decreased compared to that of the baseline (p=0.001 in all cases). In the 8–13Hz frequency range significant differences cannot be found (see Supplementary Matrial). In the 13–30Hz frequency range two-way mixed ANOVA analysis revealed a nonsignificant treatment effect (F(3,30.000)=0.522, p=0.671), a significant trial effect (F(3,90.000)=11.035, p=0.001); and a significant treatment * trial interaction was found (F(9,90.000)=2.092, p=0.038). Post hoc test demonstrated a significant increase in the amplitudes of the 1st, 2nd and 3rd post HFS+inj SS trials compared to that of the baseline (p=0.014, p=0.002, p=0.001, respectively). Furthermore, in the 3rd post HFS+inj SS trial 5.0μg quinpirole significantly increased amplitude compared to the controls (p=0.025). In the 30–49Hz frequency range (Figure 3E) two-way mixed ANOVA analysis showed a significant treatment effect (F(3,30.000)=3.756, p=0.021), a significant trial effect (F(3,90.000)=8.774, p=0.001); and a significant treatment * trial interaction was found (F(9,90.000)=2.922, p=0.004). Post hoc test demonstrated a significant increase in the amplitudes of the 2nd and 3rd post HFS+inj SS trials compared to the baseline (p=0.011, p=0.001, respectively), moreover, in the 3rd post HFS+inj SS trial 5.0μg quinpirole significantly increased amplitude compared to the controls and 0.1μg quinpirole treated groups (p=0.001 in both cases). The result of the Pearson correlation test, where significant correlation was not revealed, can be found in the Supplementary Material.

Table 1:

Effects of NAC shell or BLA stimulation per se, or combined with intra-BLA or NAC shell TTX microinjection (respectively) on the frequency bands of power spectral density of the VP LFP (the quinpirole’s effects are not included into the table).

		6–8 min after HFS	60 min after HFS	90 min after HFS	120 min after HFS
0–4 Hz	NAC HFS	↑↑	↑	↑	-
	BLA HFS	↑	↑	↑↑	↑↑
	BLA HFS+NAC TTX	↑	nk	nk	nk
	NAC HFS+BLA TTX	-	nk	nk	nk
4–8 Hz	NAC HFS	↑↑	↓	↓	↓
	BLA HFS	-	↓	↓	↓
	BLA HFS+NAC TTX	-	nk	nk	nk
	NAC HFS+BLA TTX	-	nk	nk	nk
8–13 Hz	NAC HFS	↑↑	-	-	-
	BLA HFS	-	-	-	↓
	BLA HFS+NAC TTX	-	nk	nk	nk
	NAC HFS+BLA TTX	↑↑	nk	nk	nk
13–30 Hz	NAC HFS	↑↑	-	-	-
	BLA HFS	-	-	-	-
	BLA HFS+NAC TTX	-	nk	nk	nk
	NAC HFS+BLA TTX	↑↑	nk	nk	nk
30–49 Hz	NAC HFS	-	-	-	-
	BLA HFS	-	↑↑	-	-
	BLA HFS+NAC TTX	-	nk	nk	nk
	NAC HFS+BLA TTX	↑↑	nk	nk	nk

In the table “nk” means “not known”; “-“ indicates nonsignificant changes; ↑/↓ indicates significant changes compared to the baseline, when it is smaller than ±40% in the control group; while ↑↑ indicates significant increase, when it is larger than +40% in the control group. In the table, it can be seen that the high frequency stimulation of the NAC strongly facilitates amplitudes of all the frequency domains except that of the 30–49 Hz (gamma frequency band), however, this facilitation is transient, it subsides over time, or even turns into a weak decrease (see 4–8 Hz). Inhibiting the BLA by TTX, a strong gamma facilitation can be observed, reflecting that the BLA suppresses the NAC-induced gamma-facilitation. Moreover, the HFS of the BLA lead to a gradually increasing facilitation of the 0–4 Hz (delta) frequency band over time. In addition to the effect on the delta frequency range, the HFS of the BLA results in a delayed, transient increase in the amplitude of the gamma frequency band as well, which is correlated with the increased amplitude of the P1 component.

In the BLA stimulation experiments (Figure 4C, D, E, F), the first component two-way mixed ANOVA analysis showed a significant treatment effect (F(3,29.000)=7.310, p=0.001), but not a significant trial effect (F(3,87.000)=2.253, p=0.880); nevertheless, a significant treatment * trial interaction was found (F(9,87.000)=4.924, p=0.001). Bonferroni post hoc test demonstrated that during the baseline SS there was no significant difference among the groups, but post HFS+inj single pulse stimulations 0.1μg quinpirole significantly decreased the amplitude of the first component compared to control, 1.0μg and 5.0μg quinpirole treated groups (in 1st post HFS+inj SS p=0.001, p=0.001, p=0.021; in 2nd post HFS+inj SS p=0.001, p=0.001, p=0.047; in 3rd post HFS+inj SS p=0.001, p=0.001, p=0.040, respectively), below the level of the baseline stimulation. In case of the second component, in addition to the significant treatment effect (F(3,29.000)=3.077, p=0.043), there was a significant trial effect (F(3,87.000)=36.534, p=0.001) as well; furthermore, there was a significant treatment * trial interaction (F(9,87.000)=3.671, p=0.001). In the baseline SS there was no significant difference among the groups; however, in the post HFS+inj SS trials Bonferroni post hoc test revealed significant differences between the control and 0.1μg quinpirole-treated group (p=0.003, p=0.024, p=0.023; in 1st, 2nd and 3rd post HFS+inj SS, respectively), and between the 0.1μg quinpirole and 5.0μg treated groups (p=0.015, 2nd post HFS+inj SS). The facilitated second component of the BLA stimulation observed in the control and 5.0μg quinpirole treated groups (compared to the baseline of the groups) was blocked by the 0.1μg quinpirole, i.e. this group’s P1 component remains at baseline levels.

The analysis of the power spectral density revealed significant differences among the trials and among the groups also in the BLA stimulation-VP quinpirole microinjection experiments (summary of the results of the power spectral density analysis can be seen in Table 1). In the 0–4Hz frequency range (Figure 4G) two-way mixed ANOVA analysis showed a significant treatment effect (F(3,29.000)=2.981, p=0.048), a significant trial effect (F(3,87.000)=11.025, p=0.001); and a nonsignificant treatment * trial interaction (F(9,87.000)=1.097, p=0.374). Post hoc test demonstrated a significant increase in the amplitudes of 1st, 2nd and 3rd post HFS+inj SS trials compared to that of the baseline (p=0.001, p=0.001 and p=0.002, respectively). In the 4–8Hz frequency range two-way mixed ANOVA analysis demonstrated a nonsignificant treatment effect (F(3,29.000)=0.015, p=0.998), a significant trial effect (F(3,87.000)=24.796, p=0.001); and a nonsignificant treatment * trial interaction was found (F(9,87.000)=0.663, p=0.740). Post hoc test demonstrated that the amplitudes of the 1st, 2nd and 3rd post HFS+inj SS trials were slightly but significantly decreased compared to that of the baseline (p=0.001 in all cases). In the 8–13Hz frequency range two-way mixed ANOVA analysis demonstrated a nonsignificant treatment effect (F(3,29.000)=2.566, p=0.074), a significant trial effect (F(3,87.000)=6.072, p=0.001); and a nonsignificant treatment * trial interaction was found (F(9,87.000)=0.716, p=0.693). Post hoc test showed that the amplitudes of the 3rd post HFS+inj SS trial were slightly but significantly decreased compared to that of the baseline (p=0.001). In the 13–30Hz frequency range significant differences cannot be found (see Supplementary Matrial). In the 30–49Hz frequency range (Figure 4H) two-way mixed ANOVA analysis showed a nonsignificant treatment effect (F(3,28.332)=2.295, p=0.099), a significant trial effect (F(3,85.773)=12.493, p=0.001); and a nonsignificant treatment * trial interaction was found (F(9,85.735)=1.560, p=0.141) (one data from the 1.0μg quinpirole treated group, from the 3rd post HFS+inj SS trial, has been excluded, since it was determined to be an outlier as determined by Grubb’s test with >99.9% confidence level). Post hoc test demonstrated that the amplitudes in the 1st post HFS+inj SS were significantly increased compared to the baseline, the 2nd post HFS+inj SS and the 3rd post HFS+inj SS trials (p=0.001 in all cases).

The result of the Pearson correlation test, where significant correlation was not revealed, can be found in the Supplementary Material. In the 1st post HFS+inj SS trial, Pearson correlation test revealed that there was a positive correlation between the amplitude of the gamma frequency range and the amplitude of the P1 component of the LFP response evoked by the BLA stimulation (R=0.621, p=0.001).

3.2. Effects of BLA or NAC Shell inactivation on NAC Shell or BLA stimulation-evoked VP field potentials

To isolate the effects of NAC shell stimulation from antidromic activation of BLA neurons that project to the VP or antidromic activation of NAC-BLA neurons by BLA stimulation, stimulation of the NAC shell region following TTX microinjected into the BLA, or stimulation of the BLA following TTX microinjected into the NAC shell region, was done. In neither case was the LFP responses recorded in the VP affected. Histological examination showed that the the stimulation electrodes were placed in the NAC shell region or in the BLA with the cannulae placed in the BLA or NAC shell region in 31 of the 32 rats (Figure 5A and 6A). The remaining 1 (1/32) rat had the cannula placed ventral to the BLA and was excluded from the analysis.

Figure 5:

Intra-BLA TTX microinjection did not impact VP LFP response to NAC shell region stimulation, but increases amplitude in low gamma frequency range of power spectral density especially after HFS of NAC shell region. (A): Schematic illustration of stimulation electrode placement in the NAC shell region, recording electrodes in the VP and injection cannulae in the BLA as shown in coronal sections of rat brain taken from Paxinos and Watson’s atlas. The numbers refer to anterior–posterior distance from bregma in mm. (B): Effect of intra-BLA TTX on average amplitudes of P1 components of the VP responses to NAC shell stimulation. Mean of 100 normalized responses (±S.E.M.) is represented in each group and trial. There was no statistical difference between groups in any trail. (C): Effect of intra-BLA TTX on average amplitudes of P1 components of the VP responses to NAC shell stimulation. Five consecutive means of 20 normalized responses (±S.E.M.) are represented in each group and trial. (D): Effect of intra-BLA TTX on power spectral density in the 0–4Hz (delta) frequency band. Mean of the frequencies (±S.E.M.) is represented in each group and trial. HFS of the NAC resulted in increased amplitudes of the delta frequency band in the post TTX+HFS SS trial compared to the baseline, but this effect was blocked by the intra-BLA TTX. (E): Effect of intra-BLA TTX on power spectral density in the 30–49Hz frequency band. Mean of the frequency amplitudes in 30–49Hz frequency band (±S.E.M.) is represented in each group and trial. The intra-BLA TTX increased amplitude in gamma frequency band mainly after HFS of the NAC shell region. In the figure * indicates significant differences among the groups (p<0.05, see details in the text).

Figure 6:

Blocking the NAC shell region with TTX did not modify the VP response to BLA stimulation. (A): Schematic illustration of stimulation electrode placement in the BLA, recording electrodes in the VP and cannulae in the NAC shell region as shown in coronal sections of rat brain taken from Paxinos and Watson atlas [28]. The numbers refer to anterior–posterior distance from bregma in mm. (B) and (C): Effect of intra-NAC shell TTX on average amplitudes of N1 and P1 components of the VP responses to BLA stimulation (respectively). Mean of 100 normalized responses (±S.E.M.) is represented in each group and trial. There was no statistical difference between groups in any trial. (D) and (E): Effect of intra-VP quinpirole on average amplitudes of N1 and P1 components of the VP responses to BLA stimulation. Five consecutive means of 20 normalized responses (±S.E.M.) are represented in each group and trial.

Stimulation of the NAC shell region following intra-BLA TTX microinjection (Figure 5B, C) did not influence significantly the VP LFP response: two-way mixed ANOVA analysis revealed a nonsignificant treatment effect (F(1,16.000)=2.230, p=0.155), but a significant trial effect (F(2,32.000)=18.780, p=0.001); moreover, the treatment * trial interaction was nonsignificant (F(2,32.000)=1.705, p=0.198). Bonferroni post hoc test revealed that the average amplitudes of the post TTX SS trial and the post TTX+HFS SS trial was increased significantly compared to the baseline (p=0.002, p=0.001, respectively), while the average peak amplitude of the post TTX+HFS SS was significantly higher than that of the post TTX SS trial (p=0.028).

In the experiments, when the NAC shell was stimulated and TTX was microinjected into the BLA region, the analysis of the power spectral density revealed significant differences in the NAC stimulation-BLA TTX experiments (summary of the results of the power spectral density analysis can be seen in Table 1). In the 0–4Hz frequency range (Figure 5D) two-way mixed ANOVA analysis showed a significant treatment effect (F(1,15.834.000)=5.425, p=0.033), a significant trial effect (F(2,31.148)=13.796, p=0.001); and a significant treatment * trial interaction was found (F(2,31.148)=3.408, p=0.046) (one data from the TTX group, from the post injection trial, has been excluded, since it was determined to be an outlier as assessed by Grubb’s test with >99.9% confidence level). Post hoc test revealed a significant difference between the control and the TTX treated group in the post TTX+HFS SS trial (p=0.002). In the 4–8Hz frequency range two-way mixed ANOVA analysis demonstrated a significant treatment effect (F(1,16.000)=5.720, p=0.029), a significant trial effect (F(2,32.000)=8.849, p=0.001); and a nonsignificant treatment * trial interaction was found (F(2,32.000)=0.908, p=0.413). Post hoc test revealed a significant difference between the control and the TTX treated group (p=0.029), furthermore, the amplitudes in the post TTX+HFS SS trial were significantly increased compared to the baseline (p=0.001). In the 8–13Hz frequency range two-way mixed ANOVA analysis demonstrated a nonsignificant treatment effect (F(1,16.000)=0.028, p=0.869), a significant trial effect (F(2,32.000)=18.166, p=0.001); and a nonsignificant treatment * trial interaction was found (F(2,32.000)=0.154, p=0.858). Post hoc test showed that the amplitudes of the post TTX SS and post TTX+HFS SS trials were significantly increased compared to that of the baseline (p=0.026, p=0.001, respectively), furthermore, amplitudes of the post TTX+HFS SS trials were higher also compared to those of the post TTX SS trial (p=0.009). In the 13–30Hz frequency range two-way mixed ANOVA analysis demonstrated a nonsignificant treatment effect (F(1,16.000)=3.066, p=0.099), a significant trial effect (F(2,32.000)=43.586, p=0.001); and a nonsignificant treatment * trial interaction was found (F(2,32.000)=0.327, p=0.723). Post hoc test revealed that the amplitudes of the post TTX SS and post TTX+HFS SS trials were significantly increased compared to that of the baseline (p=0.012, p=0.001, respectively), furthermore, amplitudes of the post TTX+HFS SS trials were higher also compared to those of the post TTX SS trial (p=0.001). In the 30–49Hz frequency range (Figure 5E) two-way mixed ANOVA analysis showed a significant treatment effect (F(1,16.000)=5.782, p=0.029), a nonsignificant trial effect (F(2,32.000)=3.284, p=0.050); and a nonsignificant treatment * trial interaction was found (F(2,32.000)=3.091, p=0.059). Post hoc test revealed a significant difference between the amplitudes of the control and the TTX treated groups (p=0.029).

The results of the Pearson correlation test, where significant correlation was not revealed, can be found in the Supplementary Material.

When BLA was stimulated and TTX was microinjected into the NAC shell region (Figure 5B, C, D, E), with respect to the first short-latency component, statistical analysis indicated a nonsignificant treatment effect (F(1,15.000)=1.404, p=0.254), a significant trial effect (F(2,30.000)=8.397, p=0.001) and a nonsignificant treatment * trial interaction (F(2,30.000)=1.213, p=0.311). In case of the second component a nonsignificant treatment effect (F(1,15.000)=2.780, p=0.116), a significant trial effect (F(2,30.000)=15.936, p=0.001) and a nonsignificant treatment * trial interaction (F(2,30.000)=2.779, p=0.078) was revealed. In both cases, post hoc test demonstrated that the average peak amplitude in the post TTX+HFS SS was significantly higher than that of the baseline and post TTX SS trial (first component: p=0.001, p=0.039, respectively, second component: p=0.001, p=0.001).

The analysis of the power spectral density revealed significant differences (summary of the results of the power spectral density analysis can be seen in Table 1). In the 0–4Hz frequency range two-way mixed ANOVA analysis showed a nonsignificant treatment effect (F(1,15.000)=0.162, p=0.693), a significant trial effect (F(2,30.000)=4.074, p=0.027); and a nonsignificant treatment * trial interaction was found (F(2,30.000)=0.488, p=0.619). Post hoc test demonstrated that the amplitudes of the post TTX+HFS SS trial were higher also compared to those of the baseline (p=0.026). In the 4–8Hz; 8–13Hz;13–30Hz and 30–49Hz frequency ranges two-way mixed ANOVA analysis showed that there were no significant differences among the groups and among the trials (see Supplementary Material).

4. Discussion

These experiments demonstrate how the two main inputs of the VP, from the NAC shell and the BLA, influence the VP LFP, and how these afferents interact and are dose-dependently modulated by VP D2 DA receptors. Stimulating the NAC shell region, the P1 component of the LFP recorded in the VP is likely driven by the direct NAC shell-VP GABAergic pathway. It was shown that optogenetic HFS of the NAC shell region induces LTP in D₁R and LTD in D₂R expressing accumbal medium spiny neurons – VP synapses [29]. In contrast, electrical HFS applied in the present study did not permit selective stimulation of the D₁R- and D₂R-expressing MSNs; as a result, the low-amplitude LTD observed may be a consequence of attenuation by the LTP component over time in the contol group. This low-amplitude LTD was shifted to LTP by intra-VP administration of 1.0 and 5.0μg of the D₂R agonist quinpirole. This is consistent with several studies showing that activation of D₂Rs can block LTD [30] and facilitates or it is necessary for LTP induction [12, 31–34] in certain brain regions. In particular, it was shown that quinpirole facilitates GABAergic synaptic transmission in the primary motor cortex [35]. In the experiment with TTX, the results of the control group apparently contradict the aforementioned findings, since the HFS of the NAC shell region evoked LTP, but not LTD. Nevertheless, these recordings were performed 6–8min after the HFS, in contrast to the previous one, when the LTD was observed between 60–120min after HFS. Since HFS can induce both LTP and LTD in different NAC shell-VP GABAergic synapses [29], we can assume that initially, LTP predominates, and then this balance is shifted toward LTD over time.

The BLA stimulation evoked a biphasic short-latency response, presumably reflecting an excitatory N1 component and an inhibitory P1 component of the LFP, which is consistent with the previous findings that stimulating the BLA evokes both excitatory and inhibitory unit activities in the VP [18]. It is not clear how the evoked LFP components and the mono- or polysynaptic unit activities correspond to each other, i.e. which is the exact point from where we can say that a component of the evoked LFP response is caused by polysynaptic effects. The N1 and P1 components differ in their temporal characteristics (are narrower) compared to the N2 and P2 components, suggesting that they are at least partly monosynaptic. Nevertheless, we cannot exclude that polysynaptic effects take part in their genesis as well. It was shown that the stimulation of the BLA evokes short-latency (≤12msec) and long-latency (>12msec) responses as well (both can be excitatory and inhibitory) [18]. The short-latency effects likely can be due to monosynaptic connections, while the long-latency ones are more likely to be polysynaptic. One of the most plausible candidates for the polysynaptic inhibitory effect is the NAC. Yim and Mogenson demonstrated that around 50% of inhibitory responses with longer latencies were attenuated by the inhibition of the NAC [36]. Nevertheless, our present experiments with TTX microinjected into the NAC revealed that the BLA-evoked VP LFP response does not strongly depend on the NAC input to the VP (only a small trend towards a reduction could be seen). Another alternative, which was not considered by Yim and Mogenson, is that the BLA induces inhibitory responses in the VP via the GABAergic fibers of the CeA [37]. This can be evaluated by subsequent experiments. Applying HFS stimulation in the BLA, we have shown that the inhibitory (P1) component can be facilitated, leading to an LTP. This LTP was inhibited by intra-VP administration of 0.1μg quinpirole. In addition, the 0.1μg dose of the agonist blocked the excitatory P1 component as well. It was demonstrated that DA attenuates BLA-evoked responses in the VP [18], likely via presynaptic inhibition of glutamate release [19]. Moreover, it was shown that activation of D₂Rs can facilitate LTD formation in several brain areas [38–40]. Here we demonstrated that the D₂R agonist dose-dependently inhibits LTP and induces LTD on the BLA terminals in the VP.

The BLA potentiates neuronal firing in the NAC [41]; furthermore, in the present experiments the response evoked by BLA stimulation was similar to that of the NAC shell, starting with the P1 component. Therefore, to examine if a direct interaction between the BLA and NAC shell region impacted the results, TTX was microinjected into the NAC, and BLA was stimulated. Our results demonstated that the VP response to the BLA stimulation was independent of the NAC shell region. Thus, by means of the TTX microinjection into the NAC shell region or BLA, the antidromic activation of BLA or NAC shell neurons projecting to the VP also could be excluded.

In addition to the analysis of the evoked responses, the spontaneous LFP before each single stimulation trial was recorded and analyzed in the frequency domain (we summarized these results in Table 1). Our results show (see Table 1) that the HFS of the NAC strongly facilitates LFP amplitudes of all the frequency bands, except that of the 30–49 Hz (gamma frequency band), but only transiently. The facilitation declined over time, or can also become a depression. Interestingly, if the BLA was inhibited by TTX, a strong gamma facilitation could be observed, reflecting that the BLA suppresses the NAC-induced gamma-facilitation. In addition, the 5.0μg intra-VP quinpirole facilitates gamma and beta frequencies (higher beta frequencies, closer to the gamma) in the VP LFP over the long term. This effect, however, was independent of the LTP-inducing effect of the agonist, since there was no correlation between these two measures. It is well-known that the enhanced gamma oscillation can control the connectivity between different brain regions, further modulating learning processes [42, 43]. This means that in addition to the LTP-inducing effect of the the D₂R agonist, we found another mechanism through which it can facilitate the learning processes.

One of the most striking phenomena was that the HFS of the BLA led to a gradually increasing facilitation of the 0–4 Hz (delta) frequency band over time. The BLA’s role in the generation of the delta frequency component is also supported by the fact that the NAC’s HFS induced strong delta and the lower theta facilitation is almost completely abolished by the BLA inhibition. Our finding is consistent with the literature, since it was shown that BLA has a delta-range intrinsic rhythm of oscillation, and it can spread to other brain regions [44]. In the present experiments, there was no correlation between the BLA’s HFS induced LTP and the delta-range facilitation, showing that the spread of the increased delta frequencies to the VP does not depend on the synaptical strength. In addition to the effect on the delta frequency range, the HFS of the BLA results in a delayed, transient increase in the amplitude of the gamma frequency band as well, which was correlated with the increased amplitude of the P1 component. It was demontrated in neural circuit models that gamma oscillation is beneficial to synaptic potentiation [45], furthermore, similar to our present result, it was revealed that LTP affects cortical gamma-band of electroencephalograms in the hippocampo-prefrontal cortex (PFC) pathway of anesthetized rats [46, 47].

Recently, we have shown that the intra-VP injection of 5.0μg of the D₂R agonist has an acute, immediate suppressing effect on VTA DAergic population and burst activity; in contrast, the 0.1μg and 1.0μg quinpirole doses have only a delayed effect occuring approximetly 1–1.5 hours after their administration [23]. It is well-known that the activation of hippocampus-NAC axis increases [48], while the stimulation of the BLA decreases [49] VTA population activity via the VP. Accordingly, one possible explanation for the delayed increase in VTA DAergic population acivity can be that the 0.1μg quinpirole attenuated the BLA input, whereas the 1.0μg quinpirole facilitated the NAC shell input and thereby disinhibited VTA DAergic neurons. With respect to the NAC shell-VP pathway, this hypothesis is strongly supported by the fact that stimulation of D₂R-expressing MSNs increases VTA population activity via the VP [50]. Interestingly, although the 5.0μg strenghtened the NAC shell-VP connection, this dose did not result in an increased VTA population activity. As discussed above, D₂Rs in the VP are localized presynaptically on the NAC and BLA terminals, as well as on the output neurons and interneurons of the VP. It is therefore plausible that the effect on the synaptic plasticity and the consequent increased VTA DAergic population activity are caused by the activation of presynaptic D₂Rs, whilst the acute effect of the agonist is a postsynaptic one which can overcome the presynaptic effects over the VTA.

Previously, we have demonstrated that the 1.0μg and the 5.0μg doses of quinpirole in the VP facilitates spatial learning [27] and now we extend these findings to show that these two larger doses facilitate NAC shell-VP synapses. This synaptic connection is the part of the pathway regulated by the hippocampus that is responsible for spatial and contextual learning [21]. Furthermore, we have already shown that inhibitory avoidance learning is enhanced only by the 0.1μg quinpirole in the VP [26], and now we demonstrated that the BLA-VP synapses are modified only by this dose. It is well-known that the BLA plays a potent role in aversive learning [51]. In this way, we propose a logical explanation for the mechanism regarding how the intra-VP quinpirole can dose-dependently modulate synaptic plasticity in NAC shell-VP and BLA-VP fibers, improving spatial learning processes and inhibitory avoidance learning, respectively.

5. Conclusion

In total, our results revealed that intra-VP quinpirole dose-dependently modulates BLA-VP and NAC shell-VP synapses: the lower dose inhibits BLA inputs, while the larger doses facilitates NAC shell inputs. By means of the experiments with TTX, we have demonstrated that the two nuclei do not influence each others’ evoked responses in the VP.

Analysis of the power spectral density revealed that independent from the effect on synaptic strength, intra-VP quinpirole increases the amplitude of gamma frequency band after NAC HFS , while BLA tonically suppresses the NAC’s HFS-induced gamma-facilitation. In addition, HFS of the BLA results in a delayed, transient increase in the amplitude of the gamma frequency band correlating with the LTP of the P1 component of the VP response. All these results show that the gamma frequency-facilitation can be synapse-specific, but not necessarly dependent on the synaptic strenght. Furthermore, our results demonstrate that the BLA plays a prominent role in the generation of the delta oscillations: HFS of the BLA leads to a gradually increasing delta frequency band-facilitation over time, while BLA inhibition blocks the NAC’s HFS induced strong delta facilitation. .

In conclusion, we have shown in the present study that the intra-VP D₂R agonist quinpirole dose-dependently regulates information flow in the limbic system at the level of the VP.

Highlights:

The intra-VP D₂R agonist quinpirole induces LTP on the NAC shell-VP synapses.

Intra-VP quinpirole inhibits VP LFP response and LTP evoked by BLA stimulation.

Quinpirole increases the amplitude of the gamma frequency band in the VP LFP.

BLA tonically suppresses the NAC’s HFS-induced gamma facilitation.

BLA plays a prominent role in the generation of the delta oscillations.

9. Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.