Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jun 13.
Published in final edited form as: Brain Res. 2007 Apr 25;1156:46–58. doi: 10.1016/j.brainres.2007.04.053

BRAIN STIMULATION REWARD IS INTEGRATED BY A NETWORK OF ELECTRICALLY-COUPLED GABA NEURONS

Matthew B Lassen a, J Elliott Brown a, Sarah H Stobbs a, Seth H Gunderson a, Levi Maes b, C Fernando Valenzuela b, Andrew P Ray c, Steven J Henriksen c, Scott C Steffensen a
PMCID: PMC4056590  NIHMSID: NIHMS26673  PMID: 17524371

Abstract

The neural substrate of brain stimulation reward (BSR) has eluded identification since its discovery more than a half-century ago. Notwithstanding the difficulties in identifying the neuronal integrator of BSR, the mesocorticolimbic dopamine (DA) system originating in the ventral tegmental area (VTA) of the midbrain has been implicated. We have previously demonstrated that the firing rate of a subpopulation of γ–aminobutyric acid (GABA) neurons in the VTA increases in anticipation of BSR. We show here that GABA neurons in the VTA, midbrain, hypothalamus and thalamus of rats express connexin-36 (Cx36) gap junctions (GJs) and couple electrically upon DA application or by stimulation of the internal capsule (IC), which also supports self-stimulation. The threshold for responding for IC self-stimulation was the threshold for electrical coupling between GABA neurons, the degree of responding for IC self-stimulation was proportional to the magnitude of electrical coupling between GABA neurons, and GJ blockers increased the threshold for IC self-stimulation without affecting performance. Thus, a network of electrically-coupled GABA neurons in the ventral brain may form the elusive neural integrator of BSR.

Keywords: self-stimulation, electrical coupling, gap junctions, GABA, ventral tegmental area, dopamine

1. Introduction

Since the landmark report by Olds and Milner (Olds and Milner, 1954) the phenomenon of brain stimulation reward (BSR) has undergone considerable scrutiny. The electrical stimuli to the brain that are typically used to support BSR have well-known and robust hedonic impact, and an understanding of the neural underpinnings of BSR would elucidate those neural substrates responsible for drug reward and natural rewarding behaviors. One of the primary objectives of research on BSR is the identification of the neural substrate that is directly activated by electrical self-stimulation. However, uncovering the neural substrate or transducer of BSR has proven to be problematical as electrical stimulation of multiple, often diverse, brain structures, as well as pathways with complex fiber systems, support BSR (Miguelez and Bielajew, 2004). Notwithstanding the difficulties in identifying the primary transducer of BSR, pharmacological (Wise and Rompre, 1989), neurochemical (Hernandez et al., 2006), and lesion studies (Gallistel et al., 1996) have implicated the mesocorticolimbic dopamine (DA) system originating in the midbrain ventral tegmental area (VTA) and projecting to the nucleus accumbens (NAcc) via the medial forebrain bundle (MFB). However, studies have determined that the primary transducer of BSR could not be mesencephalic DA neurons as the reward-related fibers directly activated by the self-stimulating electrode descend through the MFB while DA fibers ascend through the MFB (Lindvall and Bjorklund, 1974), and have conduction velocities too low and refractory periods too high to account for the physiological evidence (Bielajew and Shizgal, 1986; German et al., 1980; Guyenet and Aghajanian, 1978; Maeda and Mogenson, 1980; Yim and Mogenson, 1980). Collectively, these conclusions have been known as the ‘descending path hypothesis’ (Gallistel, 1983; Shizgal et al., 1982; Wise, 1980; Wise, 1996)(67,196,246). Perhaps the greatest paradox is that very large lesions of the MFB often have very little effect on the psychophysically-measured magnitude of BSR (Gallistel et al., 1996; Simmons et al., 1998). Thus, it has been suggested that the primary neural substrate of BSR is a diffuse net-like connection between forebrain nuclei and the VTA (Simmons et al., 1998).

Electrical synaptic transmission between neurons occurs via membrane-to-membrane appositions called gap junctions (GJs; (Connors and Long, 2004)). In particular, networks of γ–aminobutyric acid (GABA) neurons may regulate oscillatory patterns in the brain through mechanisms often dependent on GJs (Galarreta and Hestrin, 2001). Studies on network oscillatory behavior have mostly involved the hippocampus, thalamus, limbic system, and neocortex; however, neurons of the midbrain may also discharge in a rhythmic oscillatory mode (Kitai et al., 1999). Connexin-36 (Cx36) GJs that mediate electrical coupling between neurons are readily detected in subsets of GABA interneurons in many regions of the central nervous system (Degen et al., 2004; Liu and Jones, 2003). We have previously characterized a homogeneous population of midbrain GABA neurons (Steffensen et al., 1998a) that express Cx36 GJs (Allison et al., 2006), and that their coupling is enhanced by DA or by stimulation of the internal capsule (IC; (Stobbs et al., 2004)). By virtue of their proximity to midbrain DA neurons, wide dynamic range, widespread axonal distribution, sensitivity to DA (Stobbs et al., 2004), their presumed synaptic connectivity to DA neurons, and their electrical coupling (Allison et al., 2006), midbrain GABA neurons may be critical integrators of mesocorticolimbic DA neurotransmission implicated in natural and drug reward. Given the dearth of evidence implicating the location of the neuronal substrate of BSR (Miguelez and Bielajew, 2004; Simmons et al., 1998), we hypothesized that electrical coupling between VTA GABA neurons and other reticular GABA neurons could be the elusive “diffuse” integrator of BSR. We present anatomical and immunohistochemical evidence for the distribution of Cx36 GJs in GABA neurons in the VTA and other areas along the ventral brain neuraxis, and correlate this distribution with physiological evidence for electrical coupling. Using single-cell recording electrophysiological techniques, we show physiological evidence demonstrating the input/output characteristics of ventral brain GABA neuron coupling via IC stimulation. Finally, using the IC self-stimulation behavioral paradigm, we present evidence that IC stimulation supports BSR, that the threshold for IC-induced coupling between these GABA neurons is the threshold for BSR, that the degree of coupling between these GABA neurons correlates with the degree of BSR, and that GJ blockers raise the threshold for BSR.

2. Results

Expression of Cx36 in GABA neurons in the VTA and other ventral brain structures

We have recently demonstrated that dorsal VTA GABA neurons express Cx36 transcripts and protein (Allison et al., 2006). To determine the extent and distribution of Cx36 expressing GABA neurons in the VTA (Fig. 1A-C) and surrounding structures, we evaluated mRNA expression of glutamic acid decarboxylase (GAD) 65/67, a marker of GABA neurons, and Cx36 with fluorescent in-situ hybridization (FISH). Low-power and high-power images of horizontal brain slices obtained at the level of the VTA from 4 rats revealed high-density co-expression of GAD65/67 in neurons in the dorsal VTA (Fig. 1D; 27% of total cells were Cx36+/GAD65/67+, 39% were Cx36+ only and 34% were Cx36-/GAD65/67-; all cells that were GAD65/67+ were also Cx36+; n=4). While the dorsal VTA was characterized by high-density co-expression of Cx36 and GAD65/67, the ventral VTA was characterized by only low-density co-expression of these transcripts. To confirm that the mRNA transcripts resulted in expression of protein we evaluated the expression of GAD65/67 and Cx36 protein in the dorsal VTA with immunohistochemistry. Low-power and high-power images of horizontal dorsal brain slices revealed co-expression of GAD65/67 and Cx36 protein in the dorsal VTA (Fig. 1E), albeit not at the levels of the co-expression of transcripts seen in the FISH images. In order to further evaluate the anatomical distribution of neurons co-expressing Cx36 and GAD65/67 we evaluated FISH images along the ventral brain rostrocaudal neuraxis (Fig. 2). While low-density co-expression of Cx36 and GAD65/67 was found in the ventral VTA (Fig. 2A), we found high-density co-expression of Cx36 and GAD65/67 transcripts in the substantia nigra reticulata (SNr), lateral hypothalamus (LH), anterior hypothalamus (AH), medial preoptic area (MPA), medial preoptic nucleus (MPO), ventral hippocampus (HPC) and reticular thalamic nuclues (RTN; Fig. 2B-G).

Figure 1. Co-expression of Cx36 and GAD65/67 in the dorsal VTA.

Figure 1

Open in a new tab

(A-C) These are 20X magnification images of Cx36, GAD65/67 and superimposed Cx36/GAD65/67 fluorescent in-situ hybridization (FISH) in the dorsal VTA of mature rats. Cx36 and GAD65/67 were detected with CY3-labeled (red) and FITC-labeled (green) riboprobes, respectively. All images illustrate DAPI nuclear staining in blue. The insets show enlarged views of the area indicated by the dashed square on the 20X images and reveal the co-expression of Cx36/GAD65/67 in dorsal VTA neurons. Scale bars in B corresponds to all images. The scale bar is equal to 50μm. (D) This graph summarizes the FISH analysis of Cx36+, GAD65/67+, Cx36+ and GAD65/67+, and Cx36- and GAD65/67-neurons in horizontal slices from 4 rat brains. Of the total number of neurons in the dorsal VTA, 27% were Cx36+/GAD65/67+, 39% were Cx36+ only, and 34% were Cx36-/GAD65/67-. All neurons that were GAD65/67+ were Cx36+. (E) This image shows a 20X magnification image of the distribution of Cx36 (red) and GAD65/67 (green) immunoreactivity (IR) in the dorsal VTA. The arrows refer to points of overlap of Cx36- and GAD-IR, seen as yellow pixels, indicative of colocalization of the Cx36 and GAD antigens. Their presence as punctae suggests colocalization is in terminals or neuropil rather than in neuronal somata. These data show that the mRNA transcripts result in expression of Cx36 and GAD65/67 protein in the dorsal VTA.

Figure 2. Co-expression of Cx36 and GAD65/67 in structures along the rostrocaudal neuraxis from the ventral tegmental area (VTA) to the reticular thalamic nucleus (RTN).

Figure 2

Open in a new tab

(A-G) These are 4X and 20X (A’-G’) magnification images of Cx36/GAD65/67 fluorescent in-situ hybridization (FISH) in the ventral VTA, dorsal VTA, substantia nigra (SN), lateral hypothalamus (LH)/anterior hypothalamus (AH), medial preoptic area (MPA)/medial preoptic nucleus (MPO), ventral hippocampus (HPC), and reticular thalamic nucleus (RTN). Cx36 and GAD65/67 were detected with CY3-labeled (red) and FITC-labeled (green) riboprobes, respectively. All images illustrate DAPI nuclear staining in blue. The insets show enlarged views of the area indicated by the dashed square on the 20X images and reveal the co-expression of Cx36/GAD65/67 in each respective area. Scale bars in A,A’ correspond to all images and is equal to 50μm.

Dopamine-sensitive, electrically-coupled, Cx36-expressing GABA neurons are localized to the VTA and other ventral brain structures

We wanted to determine if electrical coupling between GABA neurons might occur in the VTA and regions of the rostrocaudal ventral brain neuraxis containing high-density co-expression of GAD65/67 and Cx36. In order to accomplish these studies we first recorded from VTA GABA neurons and then injected the anterograde tracer biotinylated dextran amine (BDA) to label the recording sites. The mean firing rate of VTA GABA neurons recorded was 21 ± 1.3 Hz (n=85). VTA GABA neurons were characterized by their response to DA iontophoretic current (Stobbs et al., 2004). Dopamine current coupled VTA GABA neurons spikes (Fig. 3A) and enhanced their firing rate (Fig. 3B). Brief, high-frequency stimulation of the IC evoked VTA GABA neurons post-stimulus spike discharges (ICPSDs) that persisted for 300-600 msec after termination of the stimulus train (Fig. 3C). The mean number of VTA GABA neuron ICPSDs at 50% maximum stimulus level (200 Hz and 10 pulses) was 52.3 ± 14.7 (n=96), more than 4 times the number of stimulus pulses in the train (i.e., 10). Following evaluation of the effects of DA application and IC stimulation on putative GABA neurons in the VTA, we labeled recording sites that evinced DA and IC-induced coupling of these neurons with BDA (Fig. 3D). Immunohistochemical analysis of BDA-immunoreactive (IR) sites revealed labeled neuronal clusters in the dorsal VTA that overlapped with the co-expression of GAD65/67 and Cx36 (Fig. 3E). The distribution of DA-sensitive, ICPSD-producing putative GABA neurons was isolated to the dorsal VTA (n=8).

Figure 3. Dopamine-sensitive electrically-coupled GABA neurons are localized to the VTA and nearby structures and correlate with areas showing high expression of GAD 65,67 and Cx36.

Figure 3

Open in a new tab

(A) These superimposed traces show unfiltered recordings of a representative VTA GABA neuron spike before (heavy line) and after (fine line) in situ microelectrophoretic application of DA (+20 nA). Note that VTA GABA neuron spike waveforms are characterized by an initial negative-going deflection followed by a small positive-going potential. The duration of the negative-going component of the spike waveform is approximately 100 μsec. Microelectrophoretic application of DA elicited a trailing spike couplet to the waveform. (B) This ratemeter record shows that DA markedly enhances the firing rate of VTA GABA neurons without any diminution in their activation by repetitive DA current. Despite a marked increase in firing rate with microelectrophoretic pulsing of DA current the leading spike appeared to be unaffected by DA. The coupled spike followed the leading spike faithfully even at DA-evoked firing rates approaching 200 Hz. (C) This trace shows the effects of brief, high-frequency (200 Hz, 10 pulses) stimulation of the internal capsule (IC) on the discharge activity of a VTA GABA neuron. VTA GABA neuron spikes not only accompany each IC stimulation pulse, but are elicited for hundreds of msec after the stimulus train has ended, hence the term “IC-evoked post-stimulus spike discharges” or ICPSDs. (D) Recording sites associated with spikes that were coupled by DA and elicited ICPSDs were subsequently labeled by microelectrophoretic application of the anterograde tracer biotinylated dextran amine (BDA), which labels multiple neurons near the pipette. Calibration bar is 50 μm. (E) This Paxinos and Watson coronal plate at 5.8 mm posterior to bregma shows BDA-labeling sites where GABA neurons were sensitive to DA and ICPSDs were elicited (filled circles). Ventral brain areas showing high levels of Cx36 and GAD65/67 co-expression are depicted in the shaded areas. Note the overlap of BDA-labeled sites and co-labeling for Cx36 and GAD65/67 in the dorsal VTA and the substantia nigra pars reticulata (SNr). Abbreviations: IC – internal capsule; ml - medial lemniscus; PBP – parabrachial pigmented nucleus; RN – red nucleus; SNr - Substantia Nigra pars reticulata; VTA – ventral tegmental area.

A similar strategy was pursued in brain structures other than the VTA that showed high-density co-expression of GAD65/67 and Cx36. We found that neurons in these regions had similar discharge profiles (i.e., high firing rate, phasic activity) and spike waveform characteristics (i.e., short-duration spikes; Fig. 4). Most importantly, they produced ICPSDs similar to that produced by IC stimulation of VTA GABA neurons (Fig. 4). However, none of the neurons outside of the dorsal VTA responded to DA current. Neurons in those areas that produced ICPSDs were labeled with microelectrophoretic application of BDA. Immunohistochemical analysis of BDA-IR sites wherein ICPSDs could be elicited revealed labeled neuronal clusters in the dorsal VTA, SNr, LH, and RTN (n=30; Fig. 4), and correlated with areas containing high-density co-expression of Cx36 and GAD65/67 transcripts. Despite the fact that the AH, MPA/MPO, and ventral HPC showed high-density co-expression of Cx36 and GAD65/67 transcripts (Fig. 2) we were unable to elicit ICPSDs in neurons in these areas (n=5 each).

Figure 4. Coupling of neurons along the rostrocaudal ventral neuraxis at locations demonstrating co-expression of Cx36 and GAD65/67 transcripts.

Figure 4

Open in a new tab

Spike waveforms and IC-evoked post-stimulus spike discharges (ICPSDs) are shown above Paxinos and Watson coronal plates at locations posterior to bregma (−2.8 to −5.3 mm) in 0.5 mm serial sections. Brief, high-frequency (10 pulses, 200 Hz) stimulation elicited ICPSDs at locations along the rostrocaudal neuraxis from the thalamus to the VTA. The traces above show spike waveforms at stereotaxic locations where ICPSDs were elicited. Biotinylated dextran amine (BDA) was iontophoretically applied from the recording pipette at sites producing ICPSDs. Filled circles indicate locations where BDA-labeled cells were detected (see BDA-labeled neurons in Fig. 3D). Shaded areas indicate high density of co-expression of Cx36 and GAD65/67 (see Fig. 2). BDA-labeled cells corresponding to locations where ICPSDs could be elicited were found in areas that also showed high-density co-expression of Cx36 and GAD65/67. The overlap extended from the midbrain through the lateral hypothalamus to the midbrain. Abbreviations: IC - internal capsule; LH - lateral hypothalamus, ml - medial lemniscus; reticular thalamic nucleus (RTN); SNr - Substantia Nigra pars reticulata; VTA – ventral tegmental area; ZI – zona incerta.

Input/output characteristics of ICPSDs

It has been suggested that the primary neural substrate of BSR is a diffuse net-like connection between forebrain nuclei and the VTA (Simmons et al., 1998). Given our anatomical and physiological evidence for electrical coupling between GABA in the VTA, we sought to determine if a GABA network might be the “diffuse net-like: integrator of BSR. Before initiating BSR studies we needed to evaluate the input/output characteristics of ICPSDs in order to adjust our stimulation protocols and to make correlations between ICPSDs and IC self-stimulation responding. We evaluated the input/output properties of VTA GABA neuron excitability and electrical coupling by varying the stimulus parameters of IC stimulation. Since the input/output functions of IC stimulation for neurons outside the VTA were similar to that of the VTA, we focused our analysis on 56 VTA GABA neurons recorded in 15 rats. This was accomplished by constructing peri-stimulus interval spike histograms (PSHs) of VTA GABA neuron ICPSDs (Fig. 5A). We have previously demonstrated that the optimal frequency for generation of VTA GABA neuron ICPSDs is 200 Hz (Steffensen et al., 1998a), likely due to the fact that this frequency falls within the supernormal period for antidromically-activated spikes (Steffensen et al., 1998a). While holding frequency constant (i.e., 200 Hz), the number of ICPSDs monotonically increased as pulse number increased (Fig. 5B,C; n=21). The threshold for activation of VTA GABA neuron ICPSDs was 4 pulses, the minimum current necessary at 4 pulses for activation of VTA GABA neuron ICPSDs was 0.3 mA, and the mean number of ICPSDs/pulse was 4, as assessed by linear regression of the curves (Fig. 5D; n=42). To determine the laterality of ICPSDs, 10 VTA GABA neurons that produced ICPSDs ipsilaterally were evaluated for response to contralateral IC stimulation. None of these 10 VTA GABA neurons produced ICPSDs, suggesting that the IC input to VTA GABA neurons is restricted ipsilaterally.

Figure 5. Input/output functions of IC-evoked post-stimulus spike discharges (ICPSDs): Monotonic function of pulse number and stimulus intensity.

Figure 5

Open in a new tab

(A) The inset shows a representative extracellular recording of a VTA GABA neuron during high-frequency stimulation (10 pulses, 200 Hz) of the internal capsule (IC). VTA GABA neuron spikes not only accompany each IC stimulation pulse, but are elicited for hundreds of msec after the stimulus train has ended. The peri-stimulus histogram shows the average of 12 IC stimulations producing ICPSDs. (B) These representative peri-stimulus spike histograms demonstrate that VTA GABA neuron ICPSDs increase with increasing pulse number. (C) This graph summarizes the effects of pulse number on ICPSDs. Fitting these data with linear regression analysis reveals a slope of approximately 4, indicating that 4 discharges occur for every pulse. Moreover, a threshold of 4-5 pulses at half-maximum current intensity is needed to induce discharges. (D) This graph summarizes the effects of pulse number and current on VTA GABA neuron ICPSDs. The goodness of linear fit is shown next to each of the pulse # isobars.

IC self-stimulation responding as a function of pulse number and ICPSDs

We have previously demonstrated that VTA GABA neuron firing rate accelerates prior to MFB self-stimulation, but is inhibited immediately thereafter (Steffensen et al., 2001). We postulated that electrical coupling of VTA GABA neurons might underlie BSR, at least IC self-stimulation. To evaluate IC self-stimulation responding, a rate-pulse study was performed in 10 rats using an FR-1 schedule of reinforcement (Fig. 6). Rats were trained on a FR-1 schedule of reinforcement at 20 IC pulses, 0.5mA, and 200Hz. The criterion for responding was 1000 pokes in a 30 minute session, with little or no responding on the inactive nose poke hole, which was monitored during each session. In each session, the first 15 min of IC self-stimulation served as control for the latter 15 min of the session when the pulse number was adjusted (Fig. 6A). The mean responding rate at 20 pulses was 0.61 ± 1.1 Hz (n=10). The threshold for IC self-stimulation responding was 4 pulses. None of the 10 rats responded for IC self-stimulation at 3 pulses. The degree of responding for IC self-stimulation was characterized by an inverted U-shaped function of pulse number vs IC self-stimulation responding (Fig. 6B; n=10). While IC self-stimulation responding decreased monotonically with increasing pulse number, when this data was expressed as total current density (in microCoulombs) vs pulse number a sigmoidal response function was obtained (Fig. 7A). A linear relationship between IC self-stimulation and VTA GABA neuron coupling was obtained when IC self-stimulation responding was expressed as total current density delivered vs ICPSDs (Fig. 7B).

Figure 6. IC self-stimulation responding as a function of pulse number.

Figure 6

Open in a new tab

Once rats reached criterion for IC self-stimulation responding (1000 nosepokes/30 min session; 0.56 Hz) at the active nosepoke responding typically stabilized to throughout the 30 min session and between sessions, with little or no responding at the inactive nosepoke hole. (A) These histograms show IC self-stimulation responding for a typical rat with average responding during 3 separate sessions (top, middle, bottom) wherein the IC pulses were changed from 20 pulses to the value indicated in each histogram. At 15 min into the 30 min session the number of stimulus pulses delivered to the IC was switched from 20 to 4 pulses (top; black to grey on the graph), 20 to 20 pulses (middle), and 20 to 40 pulses (bottom). The rate of responding for the 20 to 4 pulse transition transiently decreased and then subsided markedly, for the 20 to 20 pulse transition produced a mild increase in responding, and for the 20 to 40 pulse transition moderately decreased responding (~ 50%). (B) This graph shows the rate of IC self-stimulation responding as a function of pulse number. Values are expressed as the grand average across 10 rats of the ratio of the average rate of responding after the transition from 20 pulses to the number on the abscissa (second half of the session) vs the average rate of responding with 20 pulses (first half of the session). Compared to the response rate at 20 pulse IC self-stimulation, responding was characterized by an inverted U-shaped curve with threshold at 4-5 pulses, variable responding from 5-10 pulses, and progressive decreases in responding from 20 to 80 pulses.

Figure 7. IC self-stimulation responding as a function of coupling between VTA GABA neurons.

Figure 7

Open in a new tab

(A) The data in Fig. 6B for IC self-stimulation responding as a function of pulse number is expressed here as a function of total current density (in microCoulombs) delivered in a self-stimulation session. Total current density was the product of IC self-stimulation responding times the pulse duration times the current. Note that the threshold remains at 4 pulses, but responding saturates at pulse numbers above 40. (B) This graph shows the relationship between IC self-stimulation current density and the number of VTA GABA neuron ICPSDs produced by a given pulse number. The goodness of linear fit is shown next to the plot, and the equation describing the line is superimposed on the data. There is a strong correlation between VTA GABA neurons ICPSDs and the total current current density delivered during IC self-stimulation.

Gap junction blockers raise the threshold for IC self-stimulation responding

We have recently demonstrated that dorsal VTA GABA neurons dye-couple, and that ICPSDs are suppressed by GJ antagonists including mefloquine (Allison et al., 2006), a potent blocker of Cx36 GJ communication (Cruikshank et al., 2004). In order to evaluate the role of VTA GABA neuron electrical coupling via GJs, we evaluated the effects of the fast-acting, reversible, but non-specific and non-selective GJ blocker quinidine, and the Cx36 GJ blocker mefloquine on IC self-stimulation response rate and threshold for responding (Fig. 8). Compared to saline control, 20 mg/kg ip quinidine, a dose that reduces VTA GABA neuron ICPSDs by 75 % in 10 min for 30 min (Stobbs et al., 2004), significantly decreased overall IC self-stimulation responding (i.e., 10 pulses; 200 Hz) when administered 10 min before the session (P<0.02, t(2,5)=3.6; mean saline rate = 0.52 ± 0.11 Hz vs mean quinidine rate = 0.11 ± 0.04 Hz; n=6 each) and significantly increased the threshold for IC self-stimulation (P<0.002, t(2,5)=4.54; mean saline threshold pulses = 4.5 ± 0.62; n=6 each; Fig. 8A). Compared to vehicle control, 30 mg/kg ip mefloquine, a dose that reduces VTA GABA neuron ICPSDs by 50% at 24 hr for 3 days (Allison et al., 2006), did not significantly affect overall IC self-stimulation responding (P>0.05, t(2,3)=2.2; mean vehicle rate = 0.69 ± 0.15 Hz vs mean mefloquine rate = 0.28 ± 0.11 Hz; n=4 each), but significantly increased the threshold for IC self-stimulation when administered 24 hr before the session (P<0.01, t(2,5)=3.53; mean saline threshold pulses = 3.5 ± 0.62; n=6 each; Fig. 8B). In order to determine if decreases in IC self-stimulation thresholds might be due to deficits in motor performance we evaluated response rates during the first 200 sec of control and drug sessions. There was no difference in response rates between quinidine vs saline control (P>0.05, t(2,3)=0.03; mean saline control rate = 1.08 ± 0.4 Hz vs quinidine rate = 1.1 ± 0.3 Hz; n=4) or mefloquine vs vehicle control (P>0.05, t(2,6)=0.25; mean vehicle rate = 1.63 ± 0.7 Hz vs mean mefloquine rate = 1.91 ± 0.8 Hz; n=4), suggesting that a performance deficit was not responsible for the decreased responding for IC self-stimulation produced by these GJ antagonists.

Figure 8. Gap junction blockers raise the threshold for IC self-stimulation responding.

Figure 8

Open in a new tab

Six rats were trained to perform IC self-stimulation to criterion on 20 IC pulses. To study the effects of the gap junction (GJ) blockers quinidine and mefloquine on threshold for responding we systematically decreased the pulse number within a session for a set number of responses at each pulse level. (A) This graph shows the effects of intraperitoneal administration of the GJ blocker quinidine vs saline on the pulse threshold for IC self-stimulation. Quinidine (20 mg/kg) was administered 10 min before the session. Quinidine significantly decreased overall responding without affecting initial performance for 20 IC pulses at the onset of the session (data not shown--see text). Note that the IC pulse threshold for IC self-stimulation responding was raised by quinidine. (B) This graph shows the effects of intraperitoneal administration of the Cx36 GJ blocker mefloquine vs DMSO vehicle on the pulse threshold for IC self-stimulation. Mefloquine (30 mg/kg) was administered 24 hr before the session. Mefloquine decreased overall responding without affecting initial performance for 20 IC pulses at the onset of the session (data not shown-see text). Note that the IC pulse threshold for IC self-stimulation responding was raised by mefloquine.

3. Discussion

Using FISH analysis, we quantified the expression of GAD65,67 and Cx36 in the VTA and surrounding structures, and correlated the expression with physiological and pharmacological evidence for electrical coupling between VTA GABA neurons. The subpopulation of midbrain GABA neurons studied here can be distinguished from midbrain DA neurons by their electrophysiological (Allison et al., 2006; Steffensen et al., 1998b; Stobbs et al., 2004) and pharmacological (Stobbs et al., 2004) properties including relatively short duration action potentials, high firing rates, activation of firing rate by DA, and spike coupling by DA and IC stimulation (i.e., ICPSDs). High density co-expression of GAD65/67 and Cx36 was found in the dorsal VTA where ICPSDs and DA spike coupling were elicited, supporting our previous studies using single-cell quantitative RT-PCR and immunohistochemical techniques demonstrating that VTA GABA neurons express Cx36 transcripts and protein (Allison et al., 2006). High density co-expression of GAD65/67 and Cx36 was also found in the SNr, LH, AH, MPA, MPO, ventral HPC and RTN; however, evidence for electrical coupling could only be found in the SNr, LH, and RTN, suggesting that IC stimulation selectively activates coupling in specific structures along the ventral neuraxis and that GABA neurons in these areas, similar to the VTA, are electrically coupled via Cx36 GJs. Moreover, while iontophoretic application of DA activated and coupled VTA GABA neuron spikes it had no effect on SNr, LH or RTN neurons suggesting that DA activation and modulation of coupling is restricted to GABA neurons in the VTA.

Pharmacological studies have shown that midbrain DA neurons are involved in BSR, but may constitute second or third-order neurons in the mesocorticolimbic reward circuit (Wise and Rompre, 1989). A role for DA in BSR has been called into question by reports demonstrating that there is no DA release (Kruk et al., 1998), or decreased DA release (Garris et al., 1999), during each operant response for MFB self-stimulation, and that GABA inhibitory projections from the VTA to the NAcc play an unappreciated role in BSR (Cheer et al., 2005). Moreover, although DA release is tonically elevated during BSR (Hernandez et al., 2006), there is no phasic release of NAcc DA in association with each self-stimulation response (Garris et al., 1999). It has been suggested that the primary neural substrate of BSR is a diffuse net-like connection between forebrain nuclei and the VTA (Simmons et al., 1998). Gap junction-mediated electrical coupling between GABA neurons in the VTA and surrounding structures may fit this description of a “diffuse” neuronal network. Indeed, the refractory period of this population of midbrain GABA neurons (i.e., 0.6 msec; (Steffensen et al., 1998a)) is within the range of proposed regulator for BSR (i.e., 0.4-1.2 msec (Yeomans, 1975)). We have previously demonstrated that VTA GABA neuron firing rate accelerates prior to MFB self-stimulation, but is inhibited immediately thereafter (Steffensen et al., 2001). The fact that VTA GABA neurons anticipate MFB self-stimulation is somewhat counterintuitive, given the assumption that VTA GABA neurons inhibit reward-related DA neurons. Indeed, working under this prevailing model, we initially assumed that IC stimulation in freely-behaving rats would be aversive, given the prevailing dogma that VTA GABA neurons inhibit DA neurons and that the pronounced enhancement of their activity and electrical coupling by IC stimulation would inhibit DA neural firing. Indeed, our initial intent was to use IC stimulation as an aversive stimulus to determine if the acceleration of VTA GABA neuron firing rate before a rewarding stimulus might be similar to that during the rat’s approach to the termination of an aversive stimulus, which might suggest that VTA GABA neurons were more involved in selective attention than reward per se. However, it became surprisingly evident that rats would perform IC self-stimulation, and they typically would perform at similar, if not higher, rates than that of MFB self-stimulation. Moreover, if we stimulated the MFB with a monopolar electrode, similar to that of other self-stimulation MFB studies, we elicited VTA GABA neuron PSDs in a manner identical to that produced by bipolar stimulation electrodes in the IC, suggesting that VTA GABA neuron coupling, or ventral brain GABA neuron coupling, may be responsible for the BSR obtained in conventional studies utilizing MFB self-stimulation.

In order to study correlations between electrical coupling (i.e., ICPSDs) and IC self-stimulation, we performed an input/output parametric analysis of VTA GABA neuron ICPSDs. We found that the threshold for elicitation of ICPSDs was 4 pulses and approximately 4 spike discharges accompanied each stimulus pulse of the train. The number of ICPSDs was monotonically related to current and pulse number, characterized by a sigmoidal relationship that approached saturation beyond 30 pulses. While pulse number and current intensity were varied, frequency was held constant at 200 Hz, as we have previously determined that this frequency sharply optimizes ICPSDs due to the supernormal period (i.e., 5 msec) of antidromically-driven VTA GABA neuron axons (Steffensen et al., 1998b). As a result of this input/output parametric analysis we chose a pulse number of 20 and a current intensity of 0.5 mA as the stimulus parameters for IC self-stimulation training, and average response rates approaching 1 Hz were obtained at this level of stimulation. By varying the pulse number midway through each session and expressing the ratio of responding in the latter half of the session to the responding in the first half of the session we were able to obtain a U-shaped response curve of pulse number vs IC self-stimulation responding, characteristic of cocaine self-administration responding. The rats may have responded proportionally less at longer pulse trains because the rewarding impact of longer pulse trains is greater. The high variance on the ascending limb of the response curve (i.e., for pulse numbers less than 10) was probably related to varying absolute thresholds across the population. However, the descending limb of the response curve was fairly linear. This obtained despite the fact that VTA GABA neuron ICPSDs reached asymptotic levels at pulse numbers around 30. Therefore, in order to understand this apparent discrepancy between pulse number, ICPSDs and response rate we expressed the data in terms of total IC self-stimulation current delivered in microCoulombs, and subsequently obtained a sigmoidal relationship with saturation at around 80 pulses. Most importantly, when the number of VTA GABA neuron ICPSDs were plotted vs total IC self-stimulation current delivered we obtained a linear relationship, providing strong evidence that the degree of electrical coupling between VTA GABA neurons is related directly to the degree of BSR.

Although we did not evaluate the input/output parameters of other putative GABA neurons along the ventral neuraxis, clearly they are driven by IC stimulation in a manner similar to VTA GABA neurons, and they co-express GAD65,57 and Cx36 transcripts, suggesting that this subpopulation of VTA GABA neurons may be part of a larger network of GABA neurons that operate independently of the mesolimbic DA system. Their electrical coupling might endow them with the ability to synchronize GABA inhibitory networks that project to the corticolimbic system. Indeed, their antidromic activation by IC stimulation (Allison et al., 2006; Steffensen et al., 1998a), and their correlation with electrocortical activity (Lee et al., 2001) suggests that they might be involved in cortical activation, perhaps the attention to rewarding stimuli, which might require fast synchronous activity between GABA neurons.

To provide pharmacological evidence relating VTA GABA neuron coupling to BSR we evaluated the effects of GJ blockers on IC self-stimulation. While the non-specific and non-selective GJ blocker quinidine and the selective Cx36 GJ blocker mefloquine did not significantly affect the ability of rats to perform IC self-stimulation, they significantly increased the threshold for responding. The rewarding impact of the stimulation may be lessened under the influence of these GJ antagonists because stronger stimulation is required to produce a given threshold level for responding and performance. Thus, the distribution of Cx36 and GAD65/67 co-labeled neurons in the VTA and along the ventral brain neuraxis provides anatomical evidence that supports the physiological (i.e., ICPSDs) and pharmacological (i.e., block by GJ antagonists) evidence demonstrating that VTA GABA neurons are connected electrically. Interestingly, electrically-coupled RTN GABA neurons involved in the generation of cortical spindles (Fuentealba and Steriade, 2005) were coupled by IC stimulation and exhibited high-density co-expression of Cx36 and GAD65/67, suggesting that GABA neurons in the VTA, and other areas along the ventral brain neuraxis, may form a GABA electrical network with RTN neurons. This web of GABA neurons may explain the diffuse nature of BSR and the difficulty in interpreting lesion studies, in that self-stimulation could be effectively elicited from multiple structures in the ventral brain. This is consistent with the results of careful experiments that have failed to block BSR by lesioning the MFB and presumed reward-relevant pathways (Simmons et al., 1998).

Shizgal has proposed a functional logistic model for BSR (Sonnenschein et al., 2003) that incorporates sigmoidal reward-growth functions based on the matching law experiments of Gallistel and Herrnstein (Gallistel, 1978; Gallistel and Leon, 1991; Herrnstein, 1971). Interestingly, for values greater than the threshold for eliciting VTA GABA neuron ICPSDs (i.e., 4 pulses), we can effectively predict the number of ICPSDs produced as a function of stimulation current and pulse number according to Shizgal’s logistical model. In conclusion, our findings that neurons in the midbrain, thalamus, and hypothalamus co-express GAD65,67 and Cx36 and couple electrically, that IC self-stimulation is a direct function of VTA GABA neuron electrical coupling, and that GJ blockers disrupt IC self-stimulation provide compelling evidence that GABA neurons along the ventral brain reticulum/tegmentum form a “diffuse net-like” substrate that forms the elusive integrator of BSR.

4. Experimental Procedure

Animal subjects

Seventy-two male Wistar rats (Charles River Laboratory, Hollister, CA) weighing 250 - 450 g. were used in the experiments in this study. They were housed individually with ad libitum access to food and water, and maintained on a reverse 12 hr light/dark cycle (off 10:00, on 22:00) this enables us to study the animals during their most active period. Animal care, maintenance and experimental procedures were in accordance with the Brigham Young University Animal Research Committee. Adequate measures were taken to minimize pain or discomfort to the animals during the surgical procedures and behavioral testing.

Single-unit activity

Extracellular potentials in anesthetized rats were recorded by a single 3.0 M KCl-filled micropipette (5-10 MΩ; 1-2 μm inside diameter), cemented 20-40 μm distal to a 4-barrel micropipette (30-80 MΩ resistance), and amplified with an Axoprobe-1A microelectrode amplifier/headstage (Axon Instruments, Union City, CA). Microelectrode assemblies were oriented into the VTA [from bregma: 5.6-6.5 posterior (P), 0.5-1.0 lateral (L), 7.0-8.5 ventral (V)] with a piezoelectric microdrive (Burleigh, Fishers, NY). Single-unit activity was filtered at 1-3 kHz (−3dB) for “filtered” recordings and 0.1-10.0 kHz for “unfiltered” recordings and displayed on analog and digital oscilloscopes. Square-wave constant current pulses (50-1000 μA; 0.15 msec duration; average frequency, 0.1 Hz) were generated by an IsoFlex isolation unit controlled by a MASTER-8 Pulse Generator (AMPI, Israel) or by computer. The internal capsule (IC) (−2.5 AP, 2.5-3.0 ML, 5.0-6.5 V) was stimulated with insulated, bipolar stainless steel electrodes. We evaluated spikes that had greater than 5:1 signal-to-noise ratio. Extracellularly recorded action potentials were discriminated with a Mentor N-750 (Minneapolis, MN), Fintronics WDR-420 (Orange, CT), or WPI −121 (Sarasota, Fl) spike analyzer and converted to computer-level pulses. Single-unit potentials, discriminated spikes, and stimulation events were captured by National Instruments NB-MIO-16 digital I/O and counter/timer data acquisition boards (Austin, TX) in Macintosh-type computers. Potentials were digitized at 20kHz and 12-bit voltage resolution.

Characterization of VTA GABA neurons

All neurons classified as VTA GABA neurons in this study were located in the VTA, met the criteria established in previous studies for spike waveform characteristics and response to IC stimulation (Allison et al., 2006; Steffensen et al., 1998a; Stobbs et al., 2004), and were activated and spike-coupled by microelectrophoretic DA ((Stobbs et al., 2004)). Presumed VTA GABA neurons were characterized by short-duration (< 200 μsec; measured at half-peak amplitude of the spike), initially negative-going, non-bursting spikes, and were identified by the following IC stimulation criteria (Steffensen et al., 1998a): short latency (i.e., 2-5 ms) antidromic or orthodromic activation via single stimulation of the IC, and multiple spiking following high-frequency (10 pulses, 200 Hz) stimulation of the IC (Allison et al., 2006; Steffensen et al., 1998a; Stobbs et al., 2004).

Self-stimulation behavior

Rats were given at least one week to recover following cranial implant surgery, and were habituated by daily handling. All self-stimulation sessions took place in sound-attenuated chambers measuring 40.5 cm wide by 24 cm deep by 38 cm high. A commutator was suspended above the operant box and the headstage cable assembly was connected to the commutator. Two small holes drilled in opposite walls of the long-side of the operant box were equipped with infrared photocell beams. Each break of the beam in the active hole by a nosepoke delivered a train of stimuli to the IC (150 μsec pulse width, 3-80 pulses, 200 Hz, +0.1-1.0 mA constant current; AMPI Master-8 (Jerusalem, Israel) pulse generator and constant current isolation unit). No stimulation accompanied nosepoke to the inactive hole; however, all responses were monitored. Rats were trained on a FR-1 schedule of reinforcement at 20 pulses, 0.35mA, and 200Hz. The criterion for responding was 1000 minimum pokes in a 30 minute session, with little or no responding on the inactive nose poke hole, which was monitored during each session. The first 15 min of the session served as control responding and the pulse number was adjusted during the latter 15 min of the session. The threshold for IC self-stimulation responding was determined by reducing the number of pulses received in the train by one, every 100 times the rat poked. Once rats reached criterion for IC self-stimulation responding (i.e., 1000 nosepokes/30 min session with little or no responding at the inactive nosepoke hole), the number of IC pulses was varied during the latter 15 min of each 30 min session from 20 to either 4,8,10,20,40,60, or 80 IC pulses. The rate of responding in the latter 15 min of the session was then expressed as a ratio of responding in the first 15 min of the session.

Histology—Biotinylated dextran amine labeling

At the termination of acute recordings, to mark DA-sensitive sites wherein VTA GABA neuron ICPSDs were elicited, and to provide morphological information on recorded neurons, 5 % biotinylated dextran amine (BDA; 10,000 MW; Molecular Probes) in 1 M KCl was injected iontophoretically through one of 4 micropipette barrels whose interior was coated with Sigmacote™ lubricant to prevent sticking. Current was applied through a Midgard or AMPI Iso-flex (Jerusalem, Israel) constant-current isolation source at 1-5 μA using a 7 sec ON/7 sec OFF cycle for 10 minutes. The pipette was left in situ for a minimum of 10 min. The rat was given an overdose of halothane (5 %) and transcardially perfused with heparinized saline, followed by 4 % paraformaldehyde/0.2 % glutaraldehyde in 0.1M phosphate buffer (PB, pH 7.4). The brains were removed and equilibrated in 30 % sucrose in PB, quick-frozen, and then cut coronally into 0.1M PB saline (PBS, pH 7.4) on a freezing microtome. BDA was visualized by incubating for a minimum of 3 hours (or as long as overnight) in avidin-biotin peroxidase complex (ABC kit, Vector Labs) as per kit protocol, then rinsed 3 times in PBS. Visualization was accomplished by rinsing tissue in 0.1M imidazole-acetate buffer (IAc, pH 7.3), then developing with 0.05 % diaminobenzidine, 2 % nickel ammonium sulfate, 0.005 % H2O2 in IAc. After processing for BDA, sections were mounted on slides, counterstained with neutral red, dehydrated, and coverslipped with Permount™ (Fisher) under glass coverslips. At the termination of the chronic recordings, electrolytic lesions were made by passing current (± 3 mA, 5 sec duration) with a constant current stimulus isolation unit (Grass Instruments, Quincy, Mass) through the recording electrodes during deep anesthesia. The animals were subsequently administered a lethal dose of halothane or pentobarbital and the brains were removed and preserved as described above. The brains were frozen and sectioned in a cryostat for inspection of the lesion site.

Histology—fluorescent in-situ hybridization

Horizontal brain slices 250 μm were prepared with a vibratome. Slices were covered with OCT, flash-frozen in isopentane and sectioned (20 μm) with a cryostat. Sections were processed for FISH as described previously (Guzowski et al., 1999). Images were acquired with a Nikon TE2000U epifluorescence microscope equipped with a spinning disk confocal imaging system as described previously (Guzowski et al., 2006). Antisense digoxigenin-labeled Cx36 and fluorescein-labeled GAD 65/67 cRNA riboprobes were hybridized overnight at 56°C and detected using a commercial biotin tyramide amplification kit with streptavidin CY3 and a FITC tyramide amplification kit (DirectFISH; PerkinElmer Life Sciences), respectively. Sections were then counterstained with DAPI (Molecular Probes, Invitrogen, Carlsbad, CA). The GAD 65/67 riboprobes were generously provided by Dr. John Guzowski (Center for the Neurobiology of Learning and Memory, University of California, Irvine). Plasmid DNA for preparation of the Cx36 riboprobe was kindly provided by Dr. Viviana Berthoud (Department of Pediatrics, University of Chicago). The plasmid, a 653 bp DNA fragment containing part of the coding sequence for mouse Cx36 inserted in pGEM-T Easy (Promega, Madison, WI), was linearized with NcoI and the FISH probe was generated using SP6 RNA polymerase as described previously (Guzowski and Worley, 2001).

Immunohistochemistry

To demonstrate protein colocalization, 30 μm thick horizontal rat brain sections were processed immunohistochemically for 1) GAD67-IR (Chemicon mouse monoclonal, 1:800 dilution) and 2) Cx36-IR (Zymed rabbit polyclonal, 1:500). Tissue was preincubated in 3% H2O2 and 1% Triton X-100 (TX) for 15 min, then rinsed for 5 min 3 times in PB. Tissue was then incubated in primary antibody solution (5% normal horse serum, 0.3% TX, in 0.1M PB). After overnight incubation, tissue was rinsed as before, then placed in HRP-conjugated goat anti-rabbit (Jackson Immuno, 1:600) and biotinylated goat anti-mouse (Jackson Immuno, 1:600) secondary antibodies (5% normal serum, 0.1% TX in 0.1M TBS). After incubating 90 min, tissue was again rinsed 3 times in PB. It was then reacted with AlexaFluor 488-conjugated tyramide and 0.003% H2O2 in PB for 15 min, rinsed once in PB with 3% H2O2 to quench excess HRP, then 3 times in PB, and finally placed in AlexaFluor 594-conjugated streptavidin (Molecular Probes, 1:300) for 60 min. Tissue was rinsed thoroughly, then mounted on chrome-alum dipped glass slides. All slides were air-dried, dehydrated through successive ethanols, cleared in xylene, and then coverslipped in DEPEX.

Image Analysis

The quantification of images acquired from the dorsal VTA was done using MetaMorph software (Universal Imaging Corporation, West Chester, PA). Images of the dVTA were taken lateral to the caudal linear raphe nu (CLi) in 4 horizontal sections representing three different animals as previously described previously (Allison et al., 2006). Manual cell counts were done on putative neurons from the median planes of the Z-stack images (representing 20% of the stack thickness). Neurons were outlined and identified as either negative, Cx36 positive, GAD65/67 positive or as positive for both. Putative glial cells were identified and excluded based on their small size and bright, uniform nuclear counterstaining with DAPI as described previously (Vazdarjanova et al., 2002). The results were expressed as the percentage of total positive cells for each riboprobe. Image capture of Cx36/GAD67 immunolabeled cells was performed with a Nikon laser scanning confocal microscope. Images of relevant areas were captured as 12 μm thick z-stacks made up of 1 μm optical thickness individual images, and merged using Volocity software.

Drug delivery

For systemic administration of drugs in vivo, mefloquine was solubilized in 1% Tween80 in distilled water and delivered intraperitoneally. Quinidine was solubilized in saline and delivered intravenously. Systemic drug responses were compared to vehicle (1% Tween80) and saline injections. Mefloquine was a gift from the Walter Reed Army Institute of Research (Silver Spring, MD).

Analysis of responses

Waveforms, discriminated spikes and stimulation events were processed with National Instruments LabVIEW and IGOR Pro software (Wavemetrics, Lake Oswego, OR). Spike durations were measured by orienting cursors on the waveforms at half-maximum peak amplitude of the negative-going spike. Ratemeter records were analyzed by orienting cursors on the ratemeter records to integrate the average firing rate of VTA GABA neurons over 5 min epochs before drug, during drug, and following recovery. Peri-stimulus spike histograms (PSHs) were constructed for determinations of the number of VTA GABA neuron ICPSDs. The histograms were normalized to number of internal capsule stimulations before and after drug treatment (12 stimulation trains at 10 sec intervals, 1 sec epoch, 2 ms bin width). The number of driven spikes following internal capsule stimulation was determined by rectangular integration using IGOR Pro software. As the number of discharges varied across neurons within each animal and across animals, we integrated spikes on PSHs falling in bins immediately after the stimulation epoch and extending to a point on the PSH where the discharges appeared to be just above the floor of spontaneous activity (range between 250-600 ms beyond the stimulus artifact). To further reduce variability across treatment groups, we standardized ICPSDs to percent control. The results for control and drug treatment groups were derived from calculations performed on spontaneous firing rate and PSHs and expressed as means ± S.E.M. The results were compared before and after drug treatment using the paired two-sample for means t-test, and single factor ANOVA, for comparisons between groups of unequal sample size. The criterion of significance was set at p<0.05.

Acknowledgements

This work was supported by PHS grant AA13666 to SCS.

Abbreviations

BDA

Biotinylated dextran amine

BSR

brain stimulation reward

Cx

connexin

DA

dopamine

FISH

fluorescent in-situ hybridization

GAD

glutamic acid decarboxylase

GABA

gamma-aminobutyric acid

IC

internal capsule

NAcc

nucleus accumbens

GJ

gap junction

ICPSDs

internal capsule post-stimulus spike discharges

IR

immunoreactivity

LH

lateral hypothalamus

MFB

medial forebrain bundle

PSH

peri-stimulus spike histogram

RTN

reticular thalamic nucleus

SNr

substantia nigra reticulata

VTA

ventral tegmental area

Footnotes

Classification: Section 3. Neurophysiology, Neuropharmacology and other forms of Intercellular Communication

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Literature References

  1. Allison DW, et al. Connexin-36 gap junctions mediate electrical coupling between ventral tegmental area GABA neurons. Synapse. 2006;60:20–31. doi: 10.1002/syn.20272. [DOI] [PubMed] [Google Scholar]
  2. Bielajew C, Shizgal P. Evidence implicating descending fibers in self-stimulation of the medial forebrain bundle. J. Neurosci. 1986;6:919–929. doi: 10.1523/JNEUROSCI.06-04-00919.1986. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Cheer JF, et al. Simultaneous dopamine and single-unit recordings reveal accumbens GABAergic responses: implications for intracranial self-stimulation. Proc Natl Acad Sci U S A. 2005;102:19150–5. doi: 10.1073/pnas.0509607102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Connors BW, Long MA. Electrical synapses in the mammalian brain. Annu Rev Neurosci. 2004;27:393–418. doi: 10.1146/annurev.neuro.26.041002.131128. [DOI] [PubMed] [Google Scholar]
  5. Cruikshank SJ, et al. Potent block of Cx36 and Cx50 gap junction channels by mefloquine. Proc Natl Acad Sci U S A. 2004;101:12364–9. doi: 10.1073/pnas.0402044101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Degen J, et al. Expression pattern of lacZ reporter gene representing connexin36 in transgenic mice. J Comp Neurol. 2004;473:511–25. doi: 10.1002/cne.20085. [DOI] [PubMed] [Google Scholar]
  7. Fuentealba P, Steriade M. The reticular nucleus revisited: intrinsic and network properties of a thalamic pacemaker. Prog Neurobiol. 2005;75:125–41. doi: 10.1016/j.pneurobio.2005.01.002. [DOI] [PubMed] [Google Scholar]
  8. Galarreta M, Hestrin S. Electrical synapses between GABA-releasing interneurons. Nat Rev Neurosci. 2001;2:425–33. doi: 10.1038/35077566. [DOI] [PubMed] [Google Scholar]
  9. Gallistel CR. Self-stimulation in the rat: quantitative characteristics of the reward pathway. J Comp Physiol Psychol. 1978;92:977–98. doi: 10.1037/h0077513. [DOI] [PubMed] [Google Scholar]
  10. Gallistel CR. Self-stimulation. In: Deutsch JA, editor. The physiological basis of memory. Academic; New York: 1983. pp. 269–349. [Google Scholar]
  11. Gallistel CR, Leon M. Measuring the subjective magnitude of brain stimulation reward by titration with rate of reward. Behav Neurosci. 1991;105:913–25. [PubMed] [Google Scholar]
  12. Gallistel CR, et al. Destruction of the medial forebrain bundle caudal to the site of stimulation reduces rewarding efficacy but destruction rostrally does not. Behav Neurosci. 1996;110:766–90. doi: 10.1037//0735-7044.110.4.766. [DOI] [PubMed] [Google Scholar]
  13. Garris PA, et al. Dissociation of dopamine release in the nucleus accumbens from intracranial self-stimulation. Nature. 1999;398:67–9. doi: 10.1038/18019. [DOI] [PubMed] [Google Scholar]
  14. German DC, et al. Electrophysiological examination of the ventral tegmental (A10) area in the rat. Brain Res. 1980;181:191–197. doi: 10.1016/0006-8993(80)91269-x. [DOI] [PubMed] [Google Scholar]
  15. Guyenet PG, Aghajanian GK. Antidromic identification of dopaminergic and other output neurons of the rat substantia nigra. Brain Res. 1978;130:69–84. doi: 10.1016/0006-8993(78)90654-6. [DOI] [PubMed] [Google Scholar]
  16. Guzowski JF, et al. Environment-specific expression of the immediate-early gene Arc in hippocampal neuronal ensembles. Nat Neurosci. 1999;2:1120–4. doi: 10.1038/16046. [DOI] [PubMed] [Google Scholar]
  17. Guzowski JF, et al. Recent behavioral history modifies coupling between cell activity and Arc gene transcription in hippocampal CA1 neurons. Proc Natl Acad Sci U S A. 2006;103:1077–82. doi: 10.1073/pnas.0505519103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Guzowski JF, Worley PF. Current Protocols in Neuroscience. Vol. 1. Wiley; New York: 2001. [DOI] [PubMed] [Google Scholar]
  19. Hernandez G, et al. Prolonged rewarding stimulation of the rat medial forebrain bundle: neurochemical and behavioral consequences. Behav Neurosci. 2006;120:888–904. doi: 10.1037/0735-7044.120.4.888. [DOI] [PubMed] [Google Scholar]
  20. Herrnstein RJ. Quantitative hedonism. J Psychiatr Res. 1971;8:399–412. doi: 10.1016/0022-3956(71)90033-1. [DOI] [PubMed] [Google Scholar]
  21. Kitai ST, et al. Afferent modulation of dopamine neuron firing patterns. Curr Opin Neurobiol. 1999;9:690–7. doi: 10.1016/s0959-4388(99)00040-9. [DOI] [PubMed] [Google Scholar]
  22. Kruk ZL, et al. Real time measurement of stimulated dopamine release in the conscious rat using fast cyclic voltammetry: dopamine release is not observed during intracranial self stimulation. J Neurosci Methods. 1998;79:9–19. doi: 10.1016/s0165-0270(97)00156-8. [DOI] [PubMed] [Google Scholar]
  23. Lee RS, et al. Discharge profiles of ventral tegmental area GABA neurons during movement, anesthesia, and the sleep-wake cycle. J. Neurosci. 2001;21:1757–1766. doi: 10.1523/JNEUROSCI.21-05-01757.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lindvall O, Bjorklund A. The organization of the ascending catecholamine neuron system in the rat brain as revealed by the glyoxylic acid fluorescence method. Acta Physiol. Scand. Suppl. 1974;412:1–48. [PubMed] [Google Scholar]
  25. Liu XB, Jones EG. Fine structural localization of connexin-36 immunoreactivity in mouse cerebral cortex and thalamus. J Comp Neurol. 2003;466:457–67. doi: 10.1002/cne.10901. [DOI] [PubMed] [Google Scholar]
  26. Maeda H, Mogenson GJ. An electrophysiological study of inputs to neurons of the ventral tegmental area from the nucleus accumbens and medial preoptic-anterior hypothalamic areas. Brain Research. 1980;197:365–377. doi: 10.1016/0006-8993(80)91122-1. [DOI] [PubMed] [Google Scholar]
  27. Miguelez M, Bielajew C. Mapping the neural substrate underlying brain stimulation reward with the behavioral adaptation of double-pulse methods. Rev Neurosci. 2004;15:47–74. doi: 10.1515/revneuro.2004.15.1.47. [DOI] [PubMed] [Google Scholar]
  28. Olds J, Milner P. Positive reinforcement produced by electrical stimulation of septal area and other regions of rat brain. J Comp Physiol Psychol. 1954;47:419–27. doi: 10.1037/h0058775. [DOI] [PubMed] [Google Scholar]
  29. Shizgal P, et al. Psychophysical and electrophysiological studies of the substrate for brain stimulation reward. In: Hoebel BG, Novin D, editors. The neural basis of feeding and reward. Haer Institute; Brunswick: 1982. pp. 419–429. [Google Scholar]
  30. Simmons JM, et al. Medial forebrain bundle lesions fail to structurally and functionally disconnect the ventral tegmental area from many ipsilateral forebrain nuclei: implications for the neural substrate of brain stimulation reward. J Neurosci. 1998;18:8515–33. doi: 10.1523/JNEUROSCI.18-20-08515.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Sonnenschein B, et al. Growth of brain stimulation reward as a function of duration and stimulation strength. Behav Neurosci. 2003;117:978–94. doi: 10.1037/0735-7044.117.5.978. [DOI] [PubMed] [Google Scholar]
  32. Steffensen SC, et al. Responses of ventral tegmental area GABA neurons to brain stimulation reward. Brain Res. 2001;906:190–197. doi: 10.1016/s0006-8993(01)02581-1. [DOI] [PubMed] [Google Scholar]
  33. Steffensen SC, et al. Electrophysiological characterization of GABAergic neurons in the ventral tegmental area. J. Neurosci. 1998a;18:8803–8815. doi: 10.1523/JNEUROSCI.18-19-08003.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Steffensen SC, et al. Electrophysiological characterization of GABAergic neurons in the ventral tegmental area. J Neurosci. 1998b;18:8003–15. doi: 10.1523/JNEUROSCI.18-19-08003.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Stobbs SH, et al. Ethanol suppression of ventral tegmental area GABA neuron electrical transmission involves NMDA receptors. J Pharmacol Exp Ther. 2004;311:282–289. doi: 10.1124/jpet.104.071860. [DOI] [PubMed] [Google Scholar]
  36. Vazdarjanova A, et al. Experience-dependent coincident expression of the effector immediate-early genes arc and Homer 1a in hippocampal and neocortical neuronal networks. J Neurosci. 2002;22:10067–71. doi: 10.1523/JNEUROSCI.22-23-10067.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Wise RA. Action of drugs of abuse on brain reward systems. Pharmacology, Biochemistry and Behavior. 1980;13:213–223. doi: 10.1016/s0091-3057(80)80033-5. [DOI] [PubMed] [Google Scholar]
  38. Wise RA. Addictive drugs and brain stimulation reward. Annu. Rev. Neurosci. 1996;19:319–340. doi: 10.1146/annurev.ne.19.030196.001535. [DOI] [PubMed] [Google Scholar]
  39. Wise RA, Rompre P-P. Brain Dopamine and Reward. Ann. Rev. of Psychol. 1989;40:191–225. doi: 10.1146/annurev.ps.40.020189.001203. [DOI] [PubMed] [Google Scholar]
  40. Yeomans JS. Quantitative measurement of neural post-stimulation excitability with behavioral methods. Physiol. Behav. 1975;15:593–602. doi: 10.1016/s0091-6773(75)92216-6. [DOI] [PubMed] [Google Scholar]
  41. Yim CY, Mogenson GJ. Electrophysiological studies of neurons in the ventral tegmental area of Tsai. Brain Research. 1980;181:301–313. doi: 10.1016/0006-8993(80)90614-9. [DOI] [PubMed] [Google Scholar]

RESOURCES