- Email: [email protected]
Stimulation of the pedunculopontine tegmental nucleus in the rat produces burst firing in A9 dopaminergic neurons
Pergamon PII:
Neuroscience Vol. 92, No. 1, pp. 245–254, 1999 Copyright 䉷 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00 S0306-4522(98)00748-9
STIMULATION OF THE PEDUNCULOPONTINE TEGMENTAL NUCLEUS IN THE RAT PRODUCES BURST FIRING IN A9 DOPAMINERGIC NEURONS S. J. A. LOKWAN,*† P. G. OVERTON,*‡ M. S. BERRY† and D. CLARK* *Neuropsychopharmacology Laboratory, Department of Psychology, †School of Biological Sciences, University of Wales, Swansea SA2 8PP, U.K.
Abstract—Stimulation of the medial prefrontal cortex in the rat produces events in midbrain dopaminergic neurons which resemble natural bursts, and which are closely time-locked to the stimulation, albeit with a very long latency. As a consequence, we have previously argued that such bursts are polysynaptically generated via more proximal excitatory amino acidergic afferents, arising, for example, from the pedunculopontine tegmental nucleus. In the present study, single-pulse electrical stimulation applied to this nucleus (and other sites in the rostral pons) was found to elicit responses in the majority of substantia nigra (A9) dopaminergic neurons. Responses usually consisted of long-latency, long-duration excitations or inhibition–excitations. Thirty-seven percent of responses (currents combined) elicited by stimulation of the pedunculopontine tegmental nucleus contained time-locked bursts, the bursts being embedded in the long-duration excitatory phases of excitation and inhibition–excitation responses. Stimulation sites located within 0.5 mm of the pedunculopontine tegmental nucleus were also effective at eliciting timelocked bursts (although less so than sites located in the nucleus itself), whereas more distal sites were virtually ineffective. For responses containing time-locked bursts, a higher percentage of stimulations produced a burst when the response was elicited from within the pedunculopontine tegmental nucleus than when it was elicited from outside: the bursts themselves having a very long latency (median of 96.2 ms; shorter than that of medial prefrontal cortex-induced bursts). Finally, although there was no difference in the distribution within the substantia nigra pars compacta of cells which exhibited time-locked bursting and those which did not, stimulation-induced bursts were elicited more frequently in dopaminergic neurons which were classified as “bursting” on the basis of their basal activity. The pedunculopontine tegmental nucleus appears to be a critical locus in the rostral pons for the elicitation of time-locked bursts in A9 dopaminergic neurons. Since time-locked bursts were more often elicited from cells which exhibited bursting under basal conditions, this suggests that rostral pontine sites, in particular the pedunculopontine tegmental nucleus, may play a role in the natural burst activity of dopaminergic neurons. Given that bursts in dopaminergic neurons are generated in response to primary and secondary reinforcers, the projection from the pedunculopontine tegmental nucleus could be one means by which motivationally relevant information (arising, for example, from the medial prefrontal cortex) reaches these cells. 䉷 1999 IBRO. Published by Elsevier Science Ltd. Key words: electrical stimulation, excitatory amino acids, substantia nigra pars compacta, A9 cell group, prefrontal cortex.
Dopaminergic (DA) neurons in the substantia nigra pars compacta (SNPc; A9 cell group) and the ventral tegmental area (VTA; A10 cell group) of the anaesthetized rat are known to discharge action potentials in a pattern which consists of regular or irregular single spikes, or bursts of spikes with short interspike intervals, decreasing spike amplitude and ‡To whom correspondence should be addressed. Abbreviations: AP-5, (^)-2-amino-5-phosphonopentanoic acid; CPP, 3-((^)-2 carboxy-piperazin-4-yl)propyl-1-phosphonic acid; DA, dopaminergic; E, excitation (response); EAA, excitatory amino acid; EPSP, excitatory postsynaptic potential; IE, inhibition–excitation (response); mPFC, medial prefrontal cortex; NMDA, N-methyl-d-aspartate; PPTg, pedunculopontine tegmental nucleus; PSTH, peri-stimulus time interval histogram; SNPc, substantia nigra pars compacta; STN, subthalamic nucleus; VTA, ventral tegmental area.
increasing spike duration (see Ref. 23 for a review). Burst firing seems to be correlated with the occurrence of “salient” stimuli, i.e. those to which responses have adaptive consequences. For example, in the monkey, DA neurons produce a burst of spikes when the animal is presented with a stimulus previously associated with a food reward. 31 This observation suggests that burst activity is under the control of afferent inputs. It is likely that the afferent in question utilizes an excitatory amino acid (EAA), since iontophoretic application of the competitive Nmethyl-d-aspartate (NMDA) antagonists CPP 22 [3((^)-2 carboxy-piperazin-4-yl)propyl-1-phosphonic acid] or AP-5 6 [( ^ )-2-amino-5-phosphonopentanoic acid] reduces the level of bursting in DA neurons. One source of EAAergic afferents to midbrain DA
245
246
S. J. A. Lokwan et al.
neurons is the frontal cortex. This area of the cortex has been shown to send a glutamatergic projection to the substantia nigra 4,18 and an aspartatergic projection to the VTA, 8 both of which synapse directly with tyrosine hydroxylase-positive (presumably DA) dendrites. 21,32 Our previous work has highlighted the medial prefrontal area of the frontal cortex (mPFC) as one EAAergic source which plays a role in the burst activity of DA neurons. 36 We found that single-pulse electrical stimulation of the mPFC produced two main patterns of activity in the majority of responsive A9 and A10 neurons: (i) responses characterized by an initial excitation (E responses); and (ii) responses characterized by excitation following an initial inhibition (IE responses). Analysis of the excitatory phases of E and IE responses revealed that approximately onethird contained events which resembled natural bursts in DA neurons, and which were closely time-locked to the stimulus. As with natural bursts in DA neurons, these time-locked bursts were blocked by CPP. 37 The most parsimonious explanation for the timelocked bursts induced in DA neurons by stimulation of the mPFC is that they are produced by the excitatory action of the monosynaptic input to DA neurons from the mPFC, which has been demonstrated anatomically. 21,32 The principal problem with this hypothesis is the latency of the burst events. The bursts themselves had a median latency of around 150 ms, which is well in excess of that normally expected for monosynaptic events. Cortico-nigral/VTA fibres are very thin, 21 and therefore conduction velocity is low. However, although Thierry et al. 35 found velocities as low as 0.7 m/s for fibres projecting from the mPFC to the substantia nigra, with an approximate interelectrode distance in our study of 10 mm, a median burst latency of 150 ms would give a conduction velocity of 0.07 m/s: 10 times slower than the slowest fibres reported by Thierry et al. 35 Given this, and the high degree of temporal variation in the latency of each burst, we hypothesized that the bursts were generated polysynaptically. 36 In addition to the mPFC, areas of the ventral midbrain which contain DA neurons receive major EAAergic projections from the pedunculopontine tegmental nucleus (PPTg) and the subthalamic nucleus (STN). The PPTg projects to both the SNPc and the VTA, 15 and the STN directly to the SNPc 17 and indirectly to the VTA via a projection to the PPTg. 15 Stimulation of either nucleus produces excitation in a high proportion of responsive DA neurons. 28,30 Both the PPTg 10 and the STN 1 contain neurons which appear to be glutamatergic, and the excitatory action of the PPTg on the activity of DA neurons is blocked 30 or attenuated 12 by EAA antagonists. The most straightforward scenario to explain the polysynaptic mediation of mPFC-induced bursts in DA neurons is that stimulation of the mPFC
affects pathways which ultimately result in the activation of EAAergic afferents to the DA neurons, originating (for example) from the STN or PPTg. In this regard, it is interesting that ibotenic acid lesions of the STN reduce the number of cells in the substantia nigra pars reticulata which show excitatory responses to cortical stimulation. 11 Studies have already implicated the STN in the burst activity of DA neurons. Hence, lesions of the STN and local injections of a GABAA agonist into the STN reduce bursting in A9 DA neurons. 34 Conversely, local injections of a GABAA antagonist into the STN increase bursting in A9 DA neurons, 34 and this increase is blocked by the NMDA antagonist AP-5 7 (as with natural bursts; see Ref. 6). Finally, electrical stimulation of the STN has been shown to produce long-latency burst-like events in 35% of A9 DA neurons. 34 Similar information with respect to natural bursting or stimulation-induced bursting in DA neurons is not available for the PPTg. Indeed, only a small number of studies have examined the effects of PPTg stimulation on the activity of these cells. Scarnati et al. 29,30 reported short- to moderatelatency, single-spike excitations in DA neurons following PPTg stimulation; bursting was not mentioned in these studies. This is possibly due to the fact that ketamine was used as an anaesthetic. mPFC-induced bursts (and natural bursts) in DA neurons are mediated by NMDA receptor activation (see above), and ketamine is an NMDA receptor antagonist. 2 However, Kelland et al. 16 reported short- to moderate-latency excitations following PPTg stimulation in locally anaesthetized animals, which by their durations (⬍20 ms) probably consisted of only single spikes rather than bursts. It is therefore possible that PPTg stimulation does not lead to bursting in DA neurons. However, it is also possible that Kelland et al. 16 did not consider very-long-latency phenomena. Based on the available anatomical evidence (e.g., Ref. 15), these investigators would have been expecting monosynaptic events, which would normally have a much shorter latency than the bursts so far reported in DA neurons following stimulation of the mPFC 36 and the STN. 34 Therefore, the aim of the present study was to reexamine the effects of PPTg stimulation, using rats anaesthetized with chloral hydrate. Previous experience suggests that this anaesthetic agent does not block the production of stimulation-induced bursts in DA neurons (e.g., Ref. 36). Since stimulationelicited bursts in DA neurons can occur with a very long latency, consideration was given not only to short-latency events, but also to events which occurred with some delay after the stimulation. As with our previous work with the mPFC, an analysis procedure was applied to the patterns of activity elicited by stimulation of the PPTg in order to assess whether any elicited events resembled natural bursts in DA neurons.
Rostral pontine-induced bursts in dopaminergic neurons EXPERIMENTAL PROCEDURES
Recording and identification of dopaminergic neurons Cells were recorded from 17 male Sprague–Dawley rats (Harlan Olac, Bicester, U.K.) weighing 400–550 g at the time of the study. Animals were anaesthetized with chloral hydrate (400 mg/kg, i.p.) and mounted in a stereotaxic frame (David Kopf Instruments, Tuajanga, CA), with the skull level. Body temperature was maintained at 37⬚C with a heating pad and supplementary chloral hydrate was administered via a lateral tail vein. Extracellular single-unit recordings were made with glass microelectrodes pulled using a vertical electrode puller (Narashige Laboratory Instruments Ltd, Tokyo, Japan), and broken back against a fire-polished glass rod to a tip diameter of approximately 1 mm. Electrodes were filled with 1.5 M NaCl, saturated with Pontamine Sky Blue (BDH Chemicals Ltd, Poole, U.K.). The impedance of these electrodes ranged from 3 to 8 MV, measured at 135 Hz in 1.5 M NaCl. After manufacture, the electrode was lowered into the vicinity of the SNPc (coordinates: 3.0–3.2 mm anterior to the interaural line, 1.6–2.9 mm lateral to the midline and 6.6–8.4 mm below the cortical surface) with a hydraulic microdrive (David Kopf Instruments). Spike-related potentials were passed through a highimpedance amplification system (Fintronics Inc., Orange, CT), displayed on an analogue oscilloscope and fed to an audio monitor. The amplified signal was also sent to a Fintronics window discriminator to isolate spike activity. The digital output of the window discriminator, set to produce one pulse per spike, was connected to an intelligent interface (Cambridge Electronic Design, Cambridge, U.K.), linked to a 486 DX33 PC running CED data capture software (Spike2). This software package allowed spike-related data to be stored so that detailed analyses could be performed off-line using the data analysis section of the same application. A concentric bipolar stimulating electrode (Rhodes Medical Instruments, Woodland Hills, CA; for animals 1–4, type NE-100, with a 0.5-mm-diameter centre contact; thereafter, type SNE, with a 0.25-mm-diameter centre contact) was implanted at various positions within the rostral pons, ipsilateral to the recording electrode (coordinates: 0.0– 0.8 mm anterior to the interaural line, 1.0–1.9 mm lateral to the midline and 6.0–6.8 mm below the cortical surface). Single pulses of 0.5 ms duration, 0.25 or 0.5 mA amplitude, were delivered at 0.5 Hz using a stimulus isolation unit (model PS106; Grass Instruments Co., Quincy, MA), driven by a Grass stimulator (model S48). DA neurons were identified using previously established electrophysiological criteria. 13 Between one and seven (median two) cells were recorded from each animal, after which the final recording site was marked by passing a 30mA cathodal current through the electrode for 25–30 min to eject dye. Between one and 19 (median eight) stimulation sites were used in each animal, with each site separated by 0.2 mm from its predecessor. The final stimulation site was marked by passing anodal current (30 s at 10–100 mA) through the electrode to produce a small lesion. Brains were then removed and fixed in formal saline (10% formalin in 0.9% saline). Serial 50-mm sections were cut through the midbrain and pons using a cryostat (⫺25⬚C). Recording sites were identified in Cresyl Violet-stained sections, and stimulation sites were identified in sections stained using a silver impregnation technique. 24 Data analysis Burst analysis was performed on the activity patterns induced by stimulation of the rostral pons. The activity of each cell was recorded for a period of at least 2 min before stimulation commenced. This control period, and the subsequent stimulation periods (each current intensity was
247
applied for around 3 min at each site, with at least 80 stimulations being delivered in each case), were analysed for the occurrence of bursts using the criteria of Grace and Bunney. 13 Burst onset was defined as an interspike interval of 80 ms or less, and burst termination as an interspike interval of 160 ms or greater. These criteria have been found previously to be optimal for discriminating burst from non-burst events in A10 and A9 DA neurons. 9,14 All analyses of the temporal characteristics of the bursts (i.e. their relationship to the stimulus pulse) were conducted using the first spike of each burst as a marker of burst onset. An analytical method was applied to the bursts which allowed stimulus-induced bursts to be discriminated from bursts which were unrelated to the stimulus. This method was developed during a previous study which examined the effects of mPFC stimulation on the activity of DA neurons, 36 and it was derived as follows. Initially, the period following stimulation when stimulation-induced bursts were likely to occur was determined. For this purpose, a peri-stimulus time interval histogram (PSTH) was compiled (using bins of 50 ms) which combined the burst onset data (derived as detailed above) for all cells in the study. This initial analysis suggested that only those bursts occurring within 500 ms of the stimulation were likely to be related to current application. For a significant proportion of cells which showed evidence of stimulusbound bursting, the burst initiations occurring within the first 500 ms were grouped within a period of 150 ms, i.e. three consecutive 50-ms epochs (as in the cases shown in Fig. 1 of the present study). A measure of the degree to which bursts showed stimulus coupling could therefore be derived by expressing the greatest number of bursts occurring within three consecutive bins as a proportion of the total number of bursts occurring within 500 ms. Given this, the following criterion was adopted for a cell to be classified as showing time-locked bursting: the number of burst initiations occurring within any three consecutive 50-ms epochs had to be equal to or greater than 60% of the total number of bursts occurring within the 500 ms following stimulation. If burst activity was not influenced by the stimulation, this value would be close to 30%. In the present study, cells meeting this 60% criterion (which exhibited more than 10 bursts in any three consecutive 50-ms bins, to avoid false positives) were classified as showing timelocked burst firing in response to PPTg stimulation. Those bursts which occurred within the 150-ms epoch were classified as time-locked bursts. Statistics Statistical analyses were performed using t-tests (for paired scores) and chi-squared (x 2) tests. Probabilities ⬍0.05 were considered to be significant (two-tailed). RESULTS
Overall, the activity of 46 A9 DA neurons was recorded during the stimulation of various sites in the rostral pons. Cells were located mainly in, or adjacent to, the SNPc, although a small number of cells were located in the substantia nigra pars reticulata, where DA neurons have been reported previously (e.g., see Ref. 27). The majority of cells (37/46) were subjected to stimulation at a number of sites (median 3, range 2–6), and the majority of sites (122/134) were stimulated using both 0.25 and 0.5 mA. All stimulation sites elicited responses in DA neurons at 0.25 mA and only one site (0.8%) was ineffective at 0.5 mA. When burst analysis was performed on the single-spike firing patterns
Fig. 1.
Rostral pontine-induced bursts in dopaminergic neurons
249
Table 1. The number of stimulation sites in the rostral pons which elicited time-locked bursts (burst-positive sites) or which did not elicit time-locked bursts (burst-negative sites) at each of the two current intensities (0.25 and 0.5 mA) Current (mA)
Burst-positive sites
Burst-negative sites
Burst success
0.25 mA
0.5 mA
0.25 mA
0.5 mA
(0.25 and 0.5 mA)
PPTg ⬍ 0.5 mm ⱖ 0.5 mm
14 (50.0%) 12 (42.9%) 2 (7.1%)
16 (55.2%) 12 (41.4%) 1 (3.5%)
29 (27.6%) 46 (43.8%) 30 (28.6%)
22 (22.5%) 47 (48.0%) 29 (30.0%)
30/81 (37.0%) 24/117 (20.5%) 3/62 (4.8%)
Totals
28 (100%)
29 (100%)
105 (100%)
98 (100%)
57/260 (21.9%)
Sites are divided into those within the PPTg, those ⬍ 0.5 mm away from the PPTg and those ⱖ 0.5 mm away from the PPTg. As well as displaying the number of burst-positive and burst-negative sites for each location and current, the table also expresses these values as a percentage of the total number of sites for each current (in each burst condition). A “burst success” score has also been derived, which expresses the number of burst-positive “responses” (currents combined; the term “site” is not applicable here, since data for the two currents have been combined, and the majority of stimulation sites have contributed more than one data point) as a percentage of the total number of responses for each location.
elicited by the stimulation (see above), several sites were found to elicit events which resembled natural bursts in DA neurons, and which fulfilled our criterion for time-locking. It can be seen (for example) that the majority of burst initiations which followed stimulation in the cells shown in Fig. 1 occurred with a latency which ranged across a 150-ms time period (three 50-ms epochs). In the following sections of the Results, stimulation sites will be considered separately depending on whether bursts were elicited or not. In some cases, the same site may have elicited both bursts (a burst-positive response) and no bursts (a burst-negative response), depending on the current applied. Stimulation sites which elicited time-locked bursts The vast majority of sites which elicited timelocked bursts (Fig. 2; Table 1) at 0.25 mA were either in the PPTg, as delineated by Paxinos and Watson 26 (50.0%), or ⬍0.5 mm away from the PPTg (“peri-PPTg” sites; 42.9%). A small number of sites (7.1%) located more distal (ⱖ0.5 mm) also elicited time-locked bursts at 0.25 mA. Sites which elicited time-locked bursts at 0.5 mA were mainly in the PPTg (55.2%) or ⬍ 0.5 mm away from the PPTg (41.4%). Only 3.5% of sites located more distal elicited time-locked bursts at 0.5 mA. Cells in which time-locked bursts were elicited (n 19) were found in all parts of the SNPc, and had a similar distribution to burst-negative cells. The majority
(79.0%) of cells which exhibited time-locked bursts were “bursting” cells, as defined by a standard criterion (cells which exhibited two or more threespike bursts in 500 spikes of baseline activity; see Ref. 14). The bursts elicited by stimulation of the rostral pons (data for 0.25 and 0.5 mA combined) occurred with a very long latency (96.2 ms, range 21.3– 367.8 ms, median and range, respectively, of the mean values for each stimulation-elicited “response”; since data for the two currents have been combined, and several sites have contributed more than one data point to the calculation, we will refer to “responses” rather than “sites” whenever current-combined data are discussed), and exhibited a substantial degree of latency variability within each response, the median standard deviation (S.D.) of the latencies being 26.2 ms (range 4.1– 51.8 ms). Not all stimulation trials produced a time-locked burst in burst-positive responses; indeed, a median of only 26.7% of stimulations were effective in this regard (range 12.4–79.6%). The percentage of stimulations which led to a time-locked burst was higher for responses elicited from the PPTg (median 32.9%, range 12.8–79.6%) than for responses elicited from outside the PPTg (median 24.1%, range 12.4–76.0%). In all cases, each stimulus pulse produced only a single timelocked burst. The bursts themselves consisted of a median of 2.1 spikes (range 2.0–3.7), followed by a post-burst quiescent period (median duration
Fig. 1. Examples of an excitation (E) response (left) and an inhibition–excitation (IE) response (right) recorded in A9 DA neurons upon stimulation of the rostral pons. (A) The upper part of the figure shows raster displays of the single-spike firing patterns of the cells during stimulation at 0.25 mA (each dot represents an action potential; stimulation sites: E response, 0.05 mm ventral to the PPTg; IE response, within the PPTg). Successive stimulations (n 106 and 123 for the E and IE responses, respectively) are displayed on the horizontal axis, whilst a period covering 0.3 s before and 1.0 s after the stimulation (delivered at time zero) is displayed on the vertical axis. The lower part of the figure shows PSTHs (5-ms bins) derived from the data displayed in the corresponding rasters. The point at which stimulation was applied (time zero) is indicated by a dotted line. (B) The results of burst analysis performed on the single-spike firing patterns. This time the two rasters show the position of burst onsets during the stimulation. Likewise, the two PSTHs (50-ms bins) are derived from the burst onset data displayed in the corresponding rasters.
250
S. J. A. Lokwan et al.
Fig. 2.
Rostral pontine-induced bursts in dopaminergic neurons
251
Fig. 3. Example of the short- to moderate-latency (single-spike) excitation elicited in some DA neurons by stimulation of the rostral pons. In this case, the excitation was elicited by stimulation of the PPTg. The effects of PPTg stimulation are displayed as a PSTH (2-ms bins; 90 stimulations). A period covering 0.05 s before the stimulation (delivered at time 0; indicated by a dotted line) and 0.25 s after the stimulation is displayed.
273.0 ms, range 23.6–1509.0 ms), as in natural bursts in DA neurons (see Ref. 9). For 19 sites, both 0.25 and 0.5 mA currents elicited time-locked bursts. In these cells, there was no evidence that current intensity affected either burst latency (mean ^ S.E.M., here and below, 102.2 ^ 16.7 vs 114.1 ^ 20.0 ms at 0.25 and 0.5 mA, respectively; t18 0.8, P ⬎ 0.05), latency variability as assessed by S.D. (27.6 ^ 2.2 vs 24.3 ^ 1.7 ms at 0.25 and 0.5 mA, respectively; t18 1.9, P ⬎ 0.05) or the percentage of stimulations which produced a burst (34.5 ^ 4.4 vs 39.7 ^ 4.8 at 0.25 and 0.5 mA, respectively; t18 1.2, P ⬎ 0.05). When the single-spike firing patterns elicited by stimulation of burst-positive sites were examined, the bursts were found to be embedded in the excitatory phases of two response patterns: responses characterized by an initial, long-latency, long-duration excitation (E responses; 17.5% of burst-positive responses; currents combined) and responses characterized by a long-latency, long-duration excitation following an initial inhibition (IE responses; 82.5% of burst-positive responses). Examples of the two response types—which we have characterized extensively elsewhere 36 —are shown in Fig. 1. In a small number of cases (6.4% of IE responses; all stimulation sites were located in the PPTg), the broad, long-latency excitation was preceded by a
shorter-latency (6.0–25.0 ms) excitation (Fig. 3), which usually consisted on any one trial of a single spike, closely associated with the stimulus in time. The latency variability of these events (which could vary by as much as 18.0 ms) suggested that they were not antidromic spikes. Stimulation sites which did not elicit time-locked bursts Some of the sites which did not elicit time-locked bursts (Fig. 2, Table 1) at 0.25 mA were in the PPTg (27.6%) or ⬍ 0.5 mm away from the PPTg (43.8%). However, a number of these sites were located more distal ( ⱖ 0.5 mm) to the PPTg (28.6%). Likewise, some of the sites which did not elicit time-locked bursts at 0.5 mA were in the PPTg (22.5%) or ⬍0.5 mm away from the PPTg (48.0%). However, a number of these sites were also located distal ( ⱖ 0.5 mm) to the PPTg (29.6%). The majority (81.5%) of cells which did not exhibit time-locked bursts were “non-bursting” cells, as defined by a standard criterion (see above). When the single-spike firing patterns elicited by stimulation of burst-negative sites were examined, it was found that the majority of responses were either of the E type (18.7% of responses; currents combined) or IE type (72.2% of responses). The
Fig. 2. Coronal sections (taken from the atlas of Paxinos and Watson 26) through the rostral pons, showing stimulation sites (black dots) which elicited time-locked bursts (top two rows) and those which did not (lower two rows). Data for the two current intensities used in the study (0.25 and 0.5 mA) have been combined. The anterior–posterior position of each section relative to the interaural line is indicated (in mm) adjacent to each section. Sections extend from 2.28 mm in front of the interaural line to 0.68 mm behind the interaural line. CB, cerebellum; IC, inferior colliculus; PPTg, pedunculo pontine tegmental nucleus; SC, superior colliculus; SCP, superior cerebellar peduncle; xscp, decussation of the superior cerebellar peduncle.
252
S. J. A. Lokwan et al.
remaining responses (7.9%) consisted of inhibition alone. Again, in a small number of E and IE responses (7.9% and 8.8%, respectively; associated with both PPTg sitmulation sites and sites ⬍0.5 mm away from the PPTg), the broad, long-latency excitation was preceded by a shorter-latency (4.0– 40.0 ms) single-spike excitation. Comparative distribution of burst-positive and burst-negative sites When the total complement of stimulation sites (both burst positive and burst negative; currents combined) is considered (Table 1), 37.0% of responses elicited from the PPTg were burst-positive rather than burst-negative. The percentage of burstpositive responses (relative to burst-negative responses) elicited from the PPTg was higher than for responses elicited from sites located ⬍0.5 mm away from the PPTg (20.5%) and for responses elicited from sites located ⱖ0.5 mm away (4.8%). Indeed, the frequency of burst-positive responses (vs burst-negative responses) was significantly higher for sites located in the PPTg than for sites outside the PPTg (x 2 8.4, P ⬍ 0.01). DISCUSSION
Stimulation of sites within the rostral pons elicited events in 41.3% (19/46) of A9 DA neurons which resembled natural bursts, and which were closely time-locked to the stimulation. A time-locked burst response was more frequent in DA neurons which were classified as “bursting” under basal conditions, whilst a burst-negative response was more frequent in DA neurons which were classified as “nonbursting” cells, which raises the possibility that rostral pontine sites may be involved in the natural burst activity of DA neurons. The distribution of burst-positive and burst-negative cells in the SNPc was similar, both types being found throughout this subnucleus. Such a distribution fits with what is known about the anatomy of the PPTg–SNPc projection. Tracer injections into the PPTg result in terminal label throughout the whole rostrocaudal and mediolateral extents of the SNPc. 15 Our results show, that of the sites in the rostral pons, those associated with the PPTg are particularly effective at producing time-locked bursting in A9 DA neurons. A higher percentage of burst-positive responses was obtained from sites located in the PPTg than from sites located ⬍0.5 mm away from the PPTg. Furthermore, the proportion of burstpositive responses declined dramatically as sites became more distal to the PPTg, with only a small number of responses elicited from sites 0.5 mm or more away from the PPTg containing time-locked bursts. Finally, the number of stimulations which led to a time-locked burst in burst-positive responses was higher for responses elicited from the PPTg
than for responses elicited from outside the PPTg. Given the above, the time-locked bursts elicited by stimulation of the rostral pons in the present study will be referred to as PPTg-induced bursts for the remainder of the Discussion. The simplest explanation for the production of time-locked bursts by stimulation of the PPTg is that they were produced by the excitatory action of the monosynaptic input to DA neurons from the PPTg, which has been demonstrated anatomically. 3,5 Indeed, there was a one-to-one relationship between the stimulus pulses and the time-locked bursts (i.e. each effective pulse elicited only a single burst), which is often the case for monosynaptic events. Furthermore, there was no evidence that current intensity affected either the latency of the timelocked bursts or latency variability of the timelocked bursts, as is often the case for polysynaptic events. Perhaps the biggest problem for the hypothesis of monosynaptic mediation is that the latency of the bursts (96.2 ms median) is well in excess of that normally associated with monosynaptic events. However, we have recently argued 23 that the depolarizing envelope which underlies burst events in DA neurons consists in part of a long-duration, NMDA-mediated excitatory postsynaptic potential (EPSP). Such a potential has been demonstrated in A9 and A10 neurons following stimulation of EAAergic afferents in vitro by Mercuri et al. 20 This EPSP, which was presumably monosynaptic since it was elicited by stimulation of the SNPc or VTA in close proximity to the recording electrode, had a protracted “on” time constant (the time to peak could be several hundred milliseconds). In this regard, it is similar to the burst events reported in the present study, which therefore may also be monosynaptically mediated, in spite of their long latency. The burst events reported here resemble those which are elicited by stimulation of the mPFC. We have argued elsewhere 36 that time-locked bursts elicited in DA neurons by stimulation of the mPFC are induced indirectly via the activation of more proximal EAAergic sources, such as the PPTg. The present results offer some support for this hypothesis, in that the median burst latency (96.2 ms) for PPTg-induced bursts was shorter than the median for mPFC-induced bursts in both E (122.0 ms) and IE (187.0 ms) responses (see Ref. 36). However, if a long latency does not necessarily mean that PPTg-induced bursts are polysynaptically mediated, this raises the possibility that some bursts elicited by stimulation of the mPFC (it is likely to be only a proportion, due to the sparseness of the projection; see Ref. 33) are also generated monosynaptically. Indeed, the fact that neither the latency nor the latency variability of time-locked bursts produced by mPFC stimulation are affected by current intensity supports the idea of
Rostral pontine-induced bursts in dopaminergic neurons
monosynaptic mediation. 36 Unfortunately, the high degree of latency variability present in mPFCinduced (and PPTg-induced) bursts would not normally be expected for monosynaptic events. However, bursts in DA neurons appear to be multi-component phenomena, only part of which consists of the initial EPSP. 23 Hence, the latency of the first spike within the burst may be a complex function of multiple factors, making it particularly susceptible to variability. Furthermore, the possibility cannot be excluded that some natural bursts are intermixed with the time-locked bursts. Since natural bursts are stimulation independent, they will occur randomly within the 150-ms epoch designated as containing time-locked bursts (see Experimental Procedures), increasing the temporal variability of the full complement of burst events in this period. Finally, given that the degree of temporal variability in mPFC (and PPTg-induced) bursts differed between responses, it is possible that responses which exhibit low variability are produced monosynaptically, whereas responses which exhibit high variability are produced polysynaptically, with an intermediate level of variability indicating a mixture of the two types of mediation. The bursts we report here following stimulation of the PPTg in the rat have not been reported before. Scarnati et al. 29,30 reported short-latency (2.9 ^ 1.6 ms) and longer-latency (5.2 ^ 1.8, 6.6 ^ 3.3 ms) single-spike excitations in A9 DA neurons. The latencies of these excitations resemble those of the single-spike excitations which we saw in a small number of DA neurons following PPTg stimulation. Kelland et al. 16 also reported excitations in a number of A9 and A10 DA neurons following stimulation of the PPTg (short, medium and long latency, up to a mean of 27.6 ms). Since the duration of these excitations was ⬍20 ms, these were also probably single spike, again with a similar latency to the single-spike excitations we report here. As we mentioned in the Introduction, the lack of bursting in the studies by Scarnati et al. 29,30 may be explained by the choice of ketamine, an NMDA antagonist, as the
253
anaesthetic. The failure of Kelland et al. 16 to report bursting is more puzzling, since some of the animals in this study were only locally anaesthetized. The stimulation frequency (0.5 Hz), intensity (0.3 mA) and duration (300 ms) used by these investigators were similar to our own, as was the stimulating electrode (type SNE 100). We can only conclude (as we mentioned in the Introduction) that verylong-latency events were not considered by the authors. CONCLUSION
The present study demonstrates that stimulation of the PPTg produces events in A9 DA neurons which resemble natural bursts, thereby implicating the PPTg in the generation of the natural phenomenon. Natural bursts in DA neurons are generated in response to (amongst other things) primary and secondary reinforcers. 23 The PPTg receives extensive inputs from the limbic system, 38 and hence the projection from the PPTg to DA neurons may be one means by which motivationally relevant information—including that from the mPFC, which projects directly to the PPTg in the rat 32 — reaches DA neurons. This signal is conveyed, via bursting, to forebrain structures involved in response generation, such as the neostriatum. Since the neostriatum is considered to represent the main input nucleus of the basal ganglia, 25 a group of structures which form part of the extrapyramidal motor system, 19 whilst the PPTg is considered to lie on the output side of the basal ganglia, 38 the projection from the PPTg to DA neurons could therefore represent a means by which inputs to and outputs from the basal ganglia are co-ordinated in the sequential organization of behaviour, in which the basal ganglia seem to play a role. 19 Acknowledgements—This research was supported by the Wellcome Trust (D.C.; University Award 036936) and the Medical Research Council (P.G.O., D.C. and M.S.B.; project grant G9629658N). We would like to thank Dr C. D. Richards for helpful comments on an earlier version of the manuscript.
REFERENCES
1. Albin R. L., Aldridge J. W., Young A. B. and Gilman S. (1989) Feline subthalamic nucleus neurons contain glutamate-like but not GABA-like or glycine-like immunoreactivity. Brain Res. 491, 185–188. 2. Anis N. A., Berry S. C., Burton N. R. and Lodge D. (1983) The dissociative anaesthetics, ketamine and phencyclidine, selectively reduce excitation of central mammalian neurons by N-methyl-aspartate. Br. J. Pharmac. 79, 565–575. 3. Bolam J. P., Francis C. M. and Henderson Z. (1991) Cholinergic input to dopaminergic neurons of the substantia nigra: a double immunocytochemical study. Neuroscience 41, 483–494. 4. Carter C. J. (1982) Topographical distribution of possible glutamatergic pathways from the frontal cortex to the striatum and substantia nigra in rats. Neuropharmacology 21, 379–383. 5. Charara A., Smith Y. and Parent A. (1996) Glutamatergic inputs from the pedunculopontine nucleus to midbrain dopaminergic neurons in primates: Phaseolus vulgaris-leucoagglutinin anterograde labelling combined with postembedding glutamate and GABA immunohistochemistry. J. comp. Neurol. 364, 254–266. 6. Chergui K., Charlety P. J., Akaola H., Saunier C. F., Brunet J. L., Buda M., Svensson T. H. and Chouvet G. (1993) Tonic activation of NMDA receptors causes spontaneous burst discharge of rat midbrain dopamine neurons in vivo. Eur. J. Neurosci. 5, 137–144. 7. Chergui K., Akaoka H., Charlety P. J., Saunier C. F., Buda M. and Chouvet G. (1994) Subthalamic nucleus modulates burst firing of nigral dopamine neurones via NMDA receptors. NeuroReport 5, 1185–1188.
254
S. J. A. Lokwan et al.
8. Christie M. J., Bridge S., James L. B. and Beart P. M. (1985) Excitotoxin lesions suggest an aspartatergic projection from the medial prefrontal cortex to the ventral tegmental area. Brain Res. 333, 169–172. 9. Clark D. and Chiodo L. A. (1988) Electrophysiological and pharmacological characterisation of identified nigrostriatal and mesoaccumbens dopamine neurons in the rat. Synapse 2, 474–485. 10. Clements J. R. and Grant S. (1990) Glutamate-like immunoreactivity in neurons of the laterodorsal tegmental and pedunculopontine nucleus in the rat. Neurosci. Lett. 120, 70–73. 11. Fujimoto K. and Kita H. (1993) Responses of rat substantia nigra pars reticulata units to cortical stimulation. Neurosci. Lett. 142, 105–109. 12. Futami T., Takakusaki K. and Kitai S. T. (1995) Glutamatergic and cholinergic inputs from the pedunculopontine tegmental nucleus to dopamine neurons in the substantia nigra pars compacta. Neurosci. Res. 21, 331–342. 13. Grace A. A. and Bunney B. S. (1983) Intracellular and extracellular electrophysiology of nigral dopaminergic neurons—1. Identification and characterisation. Neuroscience 10, 301–315. 14. Grace A. A. and Bunney B. S. (1984) The control of firing pattern in nigral dopamine neurons: burst firing. J. Neurosci. 4, 2877–2890. 15. Jackson A. and Crossman A. R. (1983) Nucleus tegmenti pedunculopontinus: efferent connections with special reference to the basal ganglia, studied in the rat by anterograde and retrograde transport of horseradish peroxidase. Neuroscience 10, 725–765. 16. Kelland M. D., Freeman A. S., Rubin J. and Chiodo L. A. (1993) Ascending afferent regulation of rat midbrain dopamine neurons. Brain Res. Bull. 31, 539–546. 17. Kita H. and Kitai S. T. (1987) Efferent projections of the subthalamic nucleus in the rat. Light and electron microscopic analysis with the PHA-L method. J. comp. Neurol. 260, 435–452. 18. Kornhuber J., Kim J.-S., Kornhuber M. E. and Kornhuber H. H. (1984) The cortico-nigral projection: reduced glutamate content in the substantia nigra following frontal cortex ablation in the rat. Brain Res. 322, 124–126. 19. Marsden C. D. (1982) The mysterious motor function of the basal ganglia: the Robert Wartenberg lecture. Neurology 32, 514–539. 20. Mercuri N. B., Grillner P. and Bernardi G. (1996) N-Methyl-d-aspartate receptors mediate a slow excitatory postsynaptic potential in the rat midbrain dopaminergic neurons. Neuroscience 74, 785–792. 21. Naito A. and Kita H. (1994) The cortico-nigral projection in the rat: an anterograde tracing study with biotinylated dextran amine. Brain Res. 637, 317–322. 22. Overton P. and Clark D. (1992) Iontophoretically administered drugs acting at the N-methyl-d-aspartate receptor modulate burst firing in A9 dopamine neurons in the rat. Synapse 10, 131–140. 23. Overton P. G. and Clark D. (1997) Burst firing in midbrain dopaminergic neurons. Brain Res. Rev. 25, 312–334. 24. Palmgren A. (1948) A rapid method for selective silver staining of nerve fibres and nerve endings in mounted paraffin sections. Acta zool. 29, 377–392. 25. Parent A. (1990) Extrinsic connections of the basal ganglia. Trends Neurosci. 13, 254–258. 26. Paxinos G. and Watson C. (1992) The Rat Brain in Stereotaxic Coordinates, 2nd edn. Academic, Sydney. 27. Richards C. D., Shiroyama T. and Kitai S. T. (1997) Electrophysiological and immunocytochemical characterization of GABA and dopamine neurons in the substantia nigra of the rat. Neuroscience 80, 545–557. 28. Robledo P. and Feger J. (1990) Excitatory influences of the rat subthalamic nucleus to substantia nigra pars reticulata and the pallidal complex: electrophysiological data. Brain Res. 518, 47–54. 29. Scarnati E., Campana E. and Pacitti C. (1984) Pedunculopontine-evoked excitation of substantia nigra neurons in the rat. Brain Res. 304, 351–361. 30. Scarnati E., Proia E., Campana E. and Pacitti C. (1986) A microiontophoretic study on the nature of the putative neurotransmitter involved in the pedunculopontine–substantia nigra pars compacta excitatory pathway in the rat. Expl Brain Res. 62, 470–478. 31. Schultz W. and Romo R. (1990) Dopamine neurons of the monkey midbrain: contingencies of responses to stimuli eliciting immediate behavioural reactions. J. Neurophysiol. 63, 607–624. 32. Sesack S. R., Deutch A. Y., Roth R. H. and Bunney B. S. (1989) Topographical organization of the efferent projections of the medial prefrontal cortex in the rat: an anterograde tract-tracing study with Phaseolus vulgaris leucoagglutinin. J. comp. Neurol. 290, 213–242. 33. Sesack S. R. and Pickel V. M. (1992) Prefrontal cortical efferents in the rat synapse on unlabelled neuronal targets of catecholamine terminals in the nucleus accumbens septi and on dopamine neurons in the ventral tegmental area. J. comp. Neurol. 320, 145–160. 34. Smith I. D. and Grace A. A. (1992) Role of subthalamic nucleus in the regulation of nigral dopamine neuron activity. Synapse 12, 287–303. 35. Thierry A. M., Chevalier G., Ferron A. and Glowinski J. (1983) Diencephalic and mesencephalic efferents of the medial prefrontal cortex in the rat: electrophysiological evidence for the existence of branched axons. Expl Brain Res. 50, 275–282. 36. Tong Z.-Y., Overton P. G. and Clark D. (1996) Stimulation of the prefrontal cortex in the rat induces patterns of activity in midbrain dopaminergic neurons which resemble natural burst events. Synapse 22, 195–208. 37. Tong Z.-Y., Overton P. G. and Clark D. (1996) Antagonism of NMDA receptors but not AMPA/kainate receptors blocks bursting in dopaminergic neurons induced by electrical stimulation of the prefrontal cortex. J. neural Transm. 103, 889–904. 38. Winn P., Brown V. J. and Inglis W. L. (1997) On the relationship between the striatum and the pedunculopontine tegmental nucleus. Crit. Rev. Neurobiol. 11, 241–261. (Accepted 14 December 1998)
Neuroscience Vol. 92, No. 1, pp. 245–254, 1999 Copyright 䉷 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00 S0306-4522(98)00748-9
STIMULATION OF THE PEDUNCULOPONTINE TEGMENTAL NUCLEUS IN THE RAT PRODUCES BURST FIRING IN A9 DOPAMINERGIC NEURONS S. J. A. LOKWAN,*† P. G. OVERTON,*‡ M. S. BERRY† and D. CLARK* *Neuropsychopharmacology Laboratory, Department of Psychology, †School of Biological Sciences, University of Wales, Swansea SA2 8PP, U.K.
Abstract—Stimulation of the medial prefrontal cortex in the rat produces events in midbrain dopaminergic neurons which resemble natural bursts, and which are closely time-locked to the stimulation, albeit with a very long latency. As a consequence, we have previously argued that such bursts are polysynaptically generated via more proximal excitatory amino acidergic afferents, arising, for example, from the pedunculopontine tegmental nucleus. In the present study, single-pulse electrical stimulation applied to this nucleus (and other sites in the rostral pons) was found to elicit responses in the majority of substantia nigra (A9) dopaminergic neurons. Responses usually consisted of long-latency, long-duration excitations or inhibition–excitations. Thirty-seven percent of responses (currents combined) elicited by stimulation of the pedunculopontine tegmental nucleus contained time-locked bursts, the bursts being embedded in the long-duration excitatory phases of excitation and inhibition–excitation responses. Stimulation sites located within 0.5 mm of the pedunculopontine tegmental nucleus were also effective at eliciting timelocked bursts (although less so than sites located in the nucleus itself), whereas more distal sites were virtually ineffective. For responses containing time-locked bursts, a higher percentage of stimulations produced a burst when the response was elicited from within the pedunculopontine tegmental nucleus than when it was elicited from outside: the bursts themselves having a very long latency (median of 96.2 ms; shorter than that of medial prefrontal cortex-induced bursts). Finally, although there was no difference in the distribution within the substantia nigra pars compacta of cells which exhibited time-locked bursting and those which did not, stimulation-induced bursts were elicited more frequently in dopaminergic neurons which were classified as “bursting” on the basis of their basal activity. The pedunculopontine tegmental nucleus appears to be a critical locus in the rostral pons for the elicitation of time-locked bursts in A9 dopaminergic neurons. Since time-locked bursts were more often elicited from cells which exhibited bursting under basal conditions, this suggests that rostral pontine sites, in particular the pedunculopontine tegmental nucleus, may play a role in the natural burst activity of dopaminergic neurons. Given that bursts in dopaminergic neurons are generated in response to primary and secondary reinforcers, the projection from the pedunculopontine tegmental nucleus could be one means by which motivationally relevant information (arising, for example, from the medial prefrontal cortex) reaches these cells. 䉷 1999 IBRO. Published by Elsevier Science Ltd. Key words: electrical stimulation, excitatory amino acids, substantia nigra pars compacta, A9 cell group, prefrontal cortex.
Dopaminergic (DA) neurons in the substantia nigra pars compacta (SNPc; A9 cell group) and the ventral tegmental area (VTA; A10 cell group) of the anaesthetized rat are known to discharge action potentials in a pattern which consists of regular or irregular single spikes, or bursts of spikes with short interspike intervals, decreasing spike amplitude and ‡To whom correspondence should be addressed. Abbreviations: AP-5, (^)-2-amino-5-phosphonopentanoic acid; CPP, 3-((^)-2 carboxy-piperazin-4-yl)propyl-1-phosphonic acid; DA, dopaminergic; E, excitation (response); EAA, excitatory amino acid; EPSP, excitatory postsynaptic potential; IE, inhibition–excitation (response); mPFC, medial prefrontal cortex; NMDA, N-methyl-d-aspartate; PPTg, pedunculopontine tegmental nucleus; PSTH, peri-stimulus time interval histogram; SNPc, substantia nigra pars compacta; STN, subthalamic nucleus; VTA, ventral tegmental area.
increasing spike duration (see Ref. 23 for a review). Burst firing seems to be correlated with the occurrence of “salient” stimuli, i.e. those to which responses have adaptive consequences. For example, in the monkey, DA neurons produce a burst of spikes when the animal is presented with a stimulus previously associated with a food reward. 31 This observation suggests that burst activity is under the control of afferent inputs. It is likely that the afferent in question utilizes an excitatory amino acid (EAA), since iontophoretic application of the competitive Nmethyl-d-aspartate (NMDA) antagonists CPP 22 [3((^)-2 carboxy-piperazin-4-yl)propyl-1-phosphonic acid] or AP-5 6 [( ^ )-2-amino-5-phosphonopentanoic acid] reduces the level of bursting in DA neurons. One source of EAAergic afferents to midbrain DA
245
246
S. J. A. Lokwan et al.
neurons is the frontal cortex. This area of the cortex has been shown to send a glutamatergic projection to the substantia nigra 4,18 and an aspartatergic projection to the VTA, 8 both of which synapse directly with tyrosine hydroxylase-positive (presumably DA) dendrites. 21,32 Our previous work has highlighted the medial prefrontal area of the frontal cortex (mPFC) as one EAAergic source which plays a role in the burst activity of DA neurons. 36 We found that single-pulse electrical stimulation of the mPFC produced two main patterns of activity in the majority of responsive A9 and A10 neurons: (i) responses characterized by an initial excitation (E responses); and (ii) responses characterized by excitation following an initial inhibition (IE responses). Analysis of the excitatory phases of E and IE responses revealed that approximately onethird contained events which resembled natural bursts in DA neurons, and which were closely time-locked to the stimulus. As with natural bursts in DA neurons, these time-locked bursts were blocked by CPP. 37 The most parsimonious explanation for the timelocked bursts induced in DA neurons by stimulation of the mPFC is that they are produced by the excitatory action of the monosynaptic input to DA neurons from the mPFC, which has been demonstrated anatomically. 21,32 The principal problem with this hypothesis is the latency of the burst events. The bursts themselves had a median latency of around 150 ms, which is well in excess of that normally expected for monosynaptic events. Cortico-nigral/VTA fibres are very thin, 21 and therefore conduction velocity is low. However, although Thierry et al. 35 found velocities as low as 0.7 m/s for fibres projecting from the mPFC to the substantia nigra, with an approximate interelectrode distance in our study of 10 mm, a median burst latency of 150 ms would give a conduction velocity of 0.07 m/s: 10 times slower than the slowest fibres reported by Thierry et al. 35 Given this, and the high degree of temporal variation in the latency of each burst, we hypothesized that the bursts were generated polysynaptically. 36 In addition to the mPFC, areas of the ventral midbrain which contain DA neurons receive major EAAergic projections from the pedunculopontine tegmental nucleus (PPTg) and the subthalamic nucleus (STN). The PPTg projects to both the SNPc and the VTA, 15 and the STN directly to the SNPc 17 and indirectly to the VTA via a projection to the PPTg. 15 Stimulation of either nucleus produces excitation in a high proportion of responsive DA neurons. 28,30 Both the PPTg 10 and the STN 1 contain neurons which appear to be glutamatergic, and the excitatory action of the PPTg on the activity of DA neurons is blocked 30 or attenuated 12 by EAA antagonists. The most straightforward scenario to explain the polysynaptic mediation of mPFC-induced bursts in DA neurons is that stimulation of the mPFC
affects pathways which ultimately result in the activation of EAAergic afferents to the DA neurons, originating (for example) from the STN or PPTg. In this regard, it is interesting that ibotenic acid lesions of the STN reduce the number of cells in the substantia nigra pars reticulata which show excitatory responses to cortical stimulation. 11 Studies have already implicated the STN in the burst activity of DA neurons. Hence, lesions of the STN and local injections of a GABAA agonist into the STN reduce bursting in A9 DA neurons. 34 Conversely, local injections of a GABAA antagonist into the STN increase bursting in A9 DA neurons, 34 and this increase is blocked by the NMDA antagonist AP-5 7 (as with natural bursts; see Ref. 6). Finally, electrical stimulation of the STN has been shown to produce long-latency burst-like events in 35% of A9 DA neurons. 34 Similar information with respect to natural bursting or stimulation-induced bursting in DA neurons is not available for the PPTg. Indeed, only a small number of studies have examined the effects of PPTg stimulation on the activity of these cells. Scarnati et al. 29,30 reported short- to moderatelatency, single-spike excitations in DA neurons following PPTg stimulation; bursting was not mentioned in these studies. This is possibly due to the fact that ketamine was used as an anaesthetic. mPFC-induced bursts (and natural bursts) in DA neurons are mediated by NMDA receptor activation (see above), and ketamine is an NMDA receptor antagonist. 2 However, Kelland et al. 16 reported short- to moderate-latency excitations following PPTg stimulation in locally anaesthetized animals, which by their durations (⬍20 ms) probably consisted of only single spikes rather than bursts. It is therefore possible that PPTg stimulation does not lead to bursting in DA neurons. However, it is also possible that Kelland et al. 16 did not consider very-long-latency phenomena. Based on the available anatomical evidence (e.g., Ref. 15), these investigators would have been expecting monosynaptic events, which would normally have a much shorter latency than the bursts so far reported in DA neurons following stimulation of the mPFC 36 and the STN. 34 Therefore, the aim of the present study was to reexamine the effects of PPTg stimulation, using rats anaesthetized with chloral hydrate. Previous experience suggests that this anaesthetic agent does not block the production of stimulation-induced bursts in DA neurons (e.g., Ref. 36). Since stimulationelicited bursts in DA neurons can occur with a very long latency, consideration was given not only to short-latency events, but also to events which occurred with some delay after the stimulation. As with our previous work with the mPFC, an analysis procedure was applied to the patterns of activity elicited by stimulation of the PPTg in order to assess whether any elicited events resembled natural bursts in DA neurons.
Rostral pontine-induced bursts in dopaminergic neurons EXPERIMENTAL PROCEDURES
Recording and identification of dopaminergic neurons Cells were recorded from 17 male Sprague–Dawley rats (Harlan Olac, Bicester, U.K.) weighing 400–550 g at the time of the study. Animals were anaesthetized with chloral hydrate (400 mg/kg, i.p.) and mounted in a stereotaxic frame (David Kopf Instruments, Tuajanga, CA), with the skull level. Body temperature was maintained at 37⬚C with a heating pad and supplementary chloral hydrate was administered via a lateral tail vein. Extracellular single-unit recordings were made with glass microelectrodes pulled using a vertical electrode puller (Narashige Laboratory Instruments Ltd, Tokyo, Japan), and broken back against a fire-polished glass rod to a tip diameter of approximately 1 mm. Electrodes were filled with 1.5 M NaCl, saturated with Pontamine Sky Blue (BDH Chemicals Ltd, Poole, U.K.). The impedance of these electrodes ranged from 3 to 8 MV, measured at 135 Hz in 1.5 M NaCl. After manufacture, the electrode was lowered into the vicinity of the SNPc (coordinates: 3.0–3.2 mm anterior to the interaural line, 1.6–2.9 mm lateral to the midline and 6.6–8.4 mm below the cortical surface) with a hydraulic microdrive (David Kopf Instruments). Spike-related potentials were passed through a highimpedance amplification system (Fintronics Inc., Orange, CT), displayed on an analogue oscilloscope and fed to an audio monitor. The amplified signal was also sent to a Fintronics window discriminator to isolate spike activity. The digital output of the window discriminator, set to produce one pulse per spike, was connected to an intelligent interface (Cambridge Electronic Design, Cambridge, U.K.), linked to a 486 DX33 PC running CED data capture software (Spike2). This software package allowed spike-related data to be stored so that detailed analyses could be performed off-line using the data analysis section of the same application. A concentric bipolar stimulating electrode (Rhodes Medical Instruments, Woodland Hills, CA; for animals 1–4, type NE-100, with a 0.5-mm-diameter centre contact; thereafter, type SNE, with a 0.25-mm-diameter centre contact) was implanted at various positions within the rostral pons, ipsilateral to the recording electrode (coordinates: 0.0– 0.8 mm anterior to the interaural line, 1.0–1.9 mm lateral to the midline and 6.0–6.8 mm below the cortical surface). Single pulses of 0.5 ms duration, 0.25 or 0.5 mA amplitude, were delivered at 0.5 Hz using a stimulus isolation unit (model PS106; Grass Instruments Co., Quincy, MA), driven by a Grass stimulator (model S48). DA neurons were identified using previously established electrophysiological criteria. 13 Between one and seven (median two) cells were recorded from each animal, after which the final recording site was marked by passing a 30mA cathodal current through the electrode for 25–30 min to eject dye. Between one and 19 (median eight) stimulation sites were used in each animal, with each site separated by 0.2 mm from its predecessor. The final stimulation site was marked by passing anodal current (30 s at 10–100 mA) through the electrode to produce a small lesion. Brains were then removed and fixed in formal saline (10% formalin in 0.9% saline). Serial 50-mm sections were cut through the midbrain and pons using a cryostat (⫺25⬚C). Recording sites were identified in Cresyl Violet-stained sections, and stimulation sites were identified in sections stained using a silver impregnation technique. 24 Data analysis Burst analysis was performed on the activity patterns induced by stimulation of the rostral pons. The activity of each cell was recorded for a period of at least 2 min before stimulation commenced. This control period, and the subsequent stimulation periods (each current intensity was
247
applied for around 3 min at each site, with at least 80 stimulations being delivered in each case), were analysed for the occurrence of bursts using the criteria of Grace and Bunney. 13 Burst onset was defined as an interspike interval of 80 ms or less, and burst termination as an interspike interval of 160 ms or greater. These criteria have been found previously to be optimal for discriminating burst from non-burst events in A10 and A9 DA neurons. 9,14 All analyses of the temporal characteristics of the bursts (i.e. their relationship to the stimulus pulse) were conducted using the first spike of each burst as a marker of burst onset. An analytical method was applied to the bursts which allowed stimulus-induced bursts to be discriminated from bursts which were unrelated to the stimulus. This method was developed during a previous study which examined the effects of mPFC stimulation on the activity of DA neurons, 36 and it was derived as follows. Initially, the period following stimulation when stimulation-induced bursts were likely to occur was determined. For this purpose, a peri-stimulus time interval histogram (PSTH) was compiled (using bins of 50 ms) which combined the burst onset data (derived as detailed above) for all cells in the study. This initial analysis suggested that only those bursts occurring within 500 ms of the stimulation were likely to be related to current application. For a significant proportion of cells which showed evidence of stimulusbound bursting, the burst initiations occurring within the first 500 ms were grouped within a period of 150 ms, i.e. three consecutive 50-ms epochs (as in the cases shown in Fig. 1 of the present study). A measure of the degree to which bursts showed stimulus coupling could therefore be derived by expressing the greatest number of bursts occurring within three consecutive bins as a proportion of the total number of bursts occurring within 500 ms. Given this, the following criterion was adopted for a cell to be classified as showing time-locked bursting: the number of burst initiations occurring within any three consecutive 50-ms epochs had to be equal to or greater than 60% of the total number of bursts occurring within the 500 ms following stimulation. If burst activity was not influenced by the stimulation, this value would be close to 30%. In the present study, cells meeting this 60% criterion (which exhibited more than 10 bursts in any three consecutive 50-ms bins, to avoid false positives) were classified as showing timelocked burst firing in response to PPTg stimulation. Those bursts which occurred within the 150-ms epoch were classified as time-locked bursts. Statistics Statistical analyses were performed using t-tests (for paired scores) and chi-squared (x 2) tests. Probabilities ⬍0.05 were considered to be significant (two-tailed). RESULTS
Overall, the activity of 46 A9 DA neurons was recorded during the stimulation of various sites in the rostral pons. Cells were located mainly in, or adjacent to, the SNPc, although a small number of cells were located in the substantia nigra pars reticulata, where DA neurons have been reported previously (e.g., see Ref. 27). The majority of cells (37/46) were subjected to stimulation at a number of sites (median 3, range 2–6), and the majority of sites (122/134) were stimulated using both 0.25 and 0.5 mA. All stimulation sites elicited responses in DA neurons at 0.25 mA and only one site (0.8%) was ineffective at 0.5 mA. When burst analysis was performed on the single-spike firing patterns
Fig. 1.
Rostral pontine-induced bursts in dopaminergic neurons
249
Table 1. The number of stimulation sites in the rostral pons which elicited time-locked bursts (burst-positive sites) or which did not elicit time-locked bursts (burst-negative sites) at each of the two current intensities (0.25 and 0.5 mA) Current (mA)
Burst-positive sites
Burst-negative sites
Burst success
0.25 mA
0.5 mA
0.25 mA
0.5 mA
(0.25 and 0.5 mA)
PPTg ⬍ 0.5 mm ⱖ 0.5 mm
14 (50.0%) 12 (42.9%) 2 (7.1%)
16 (55.2%) 12 (41.4%) 1 (3.5%)
29 (27.6%) 46 (43.8%) 30 (28.6%)
22 (22.5%) 47 (48.0%) 29 (30.0%)
30/81 (37.0%) 24/117 (20.5%) 3/62 (4.8%)
Totals
28 (100%)
29 (100%)
105 (100%)
98 (100%)
57/260 (21.9%)
Sites are divided into those within the PPTg, those ⬍ 0.5 mm away from the PPTg and those ⱖ 0.5 mm away from the PPTg. As well as displaying the number of burst-positive and burst-negative sites for each location and current, the table also expresses these values as a percentage of the total number of sites for each current (in each burst condition). A “burst success” score has also been derived, which expresses the number of burst-positive “responses” (currents combined; the term “site” is not applicable here, since data for the two currents have been combined, and the majority of stimulation sites have contributed more than one data point) as a percentage of the total number of responses for each location.
elicited by the stimulation (see above), several sites were found to elicit events which resembled natural bursts in DA neurons, and which fulfilled our criterion for time-locking. It can be seen (for example) that the majority of burst initiations which followed stimulation in the cells shown in Fig. 1 occurred with a latency which ranged across a 150-ms time period (three 50-ms epochs). In the following sections of the Results, stimulation sites will be considered separately depending on whether bursts were elicited or not. In some cases, the same site may have elicited both bursts (a burst-positive response) and no bursts (a burst-negative response), depending on the current applied. Stimulation sites which elicited time-locked bursts The vast majority of sites which elicited timelocked bursts (Fig. 2; Table 1) at 0.25 mA were either in the PPTg, as delineated by Paxinos and Watson 26 (50.0%), or ⬍0.5 mm away from the PPTg (“peri-PPTg” sites; 42.9%). A small number of sites (7.1%) located more distal (ⱖ0.5 mm) also elicited time-locked bursts at 0.25 mA. Sites which elicited time-locked bursts at 0.5 mA were mainly in the PPTg (55.2%) or ⬍ 0.5 mm away from the PPTg (41.4%). Only 3.5% of sites located more distal elicited time-locked bursts at 0.5 mA. Cells in which time-locked bursts were elicited (n 19) were found in all parts of the SNPc, and had a similar distribution to burst-negative cells. The majority
(79.0%) of cells which exhibited time-locked bursts were “bursting” cells, as defined by a standard criterion (cells which exhibited two or more threespike bursts in 500 spikes of baseline activity; see Ref. 14). The bursts elicited by stimulation of the rostral pons (data for 0.25 and 0.5 mA combined) occurred with a very long latency (96.2 ms, range 21.3– 367.8 ms, median and range, respectively, of the mean values for each stimulation-elicited “response”; since data for the two currents have been combined, and several sites have contributed more than one data point to the calculation, we will refer to “responses” rather than “sites” whenever current-combined data are discussed), and exhibited a substantial degree of latency variability within each response, the median standard deviation (S.D.) of the latencies being 26.2 ms (range 4.1– 51.8 ms). Not all stimulation trials produced a time-locked burst in burst-positive responses; indeed, a median of only 26.7% of stimulations were effective in this regard (range 12.4–79.6%). The percentage of stimulations which led to a time-locked burst was higher for responses elicited from the PPTg (median 32.9%, range 12.8–79.6%) than for responses elicited from outside the PPTg (median 24.1%, range 12.4–76.0%). In all cases, each stimulus pulse produced only a single timelocked burst. The bursts themselves consisted of a median of 2.1 spikes (range 2.0–3.7), followed by a post-burst quiescent period (median duration
Fig. 1. Examples of an excitation (E) response (left) and an inhibition–excitation (IE) response (right) recorded in A9 DA neurons upon stimulation of the rostral pons. (A) The upper part of the figure shows raster displays of the single-spike firing patterns of the cells during stimulation at 0.25 mA (each dot represents an action potential; stimulation sites: E response, 0.05 mm ventral to the PPTg; IE response, within the PPTg). Successive stimulations (n 106 and 123 for the E and IE responses, respectively) are displayed on the horizontal axis, whilst a period covering 0.3 s before and 1.0 s after the stimulation (delivered at time zero) is displayed on the vertical axis. The lower part of the figure shows PSTHs (5-ms bins) derived from the data displayed in the corresponding rasters. The point at which stimulation was applied (time zero) is indicated by a dotted line. (B) The results of burst analysis performed on the single-spike firing patterns. This time the two rasters show the position of burst onsets during the stimulation. Likewise, the two PSTHs (50-ms bins) are derived from the burst onset data displayed in the corresponding rasters.
250
S. J. A. Lokwan et al.
Fig. 2.
Rostral pontine-induced bursts in dopaminergic neurons
251
Fig. 3. Example of the short- to moderate-latency (single-spike) excitation elicited in some DA neurons by stimulation of the rostral pons. In this case, the excitation was elicited by stimulation of the PPTg. The effects of PPTg stimulation are displayed as a PSTH (2-ms bins; 90 stimulations). A period covering 0.05 s before the stimulation (delivered at time 0; indicated by a dotted line) and 0.25 s after the stimulation is displayed.
273.0 ms, range 23.6–1509.0 ms), as in natural bursts in DA neurons (see Ref. 9). For 19 sites, both 0.25 and 0.5 mA currents elicited time-locked bursts. In these cells, there was no evidence that current intensity affected either burst latency (mean ^ S.E.M., here and below, 102.2 ^ 16.7 vs 114.1 ^ 20.0 ms at 0.25 and 0.5 mA, respectively; t18 0.8, P ⬎ 0.05), latency variability as assessed by S.D. (27.6 ^ 2.2 vs 24.3 ^ 1.7 ms at 0.25 and 0.5 mA, respectively; t18 1.9, P ⬎ 0.05) or the percentage of stimulations which produced a burst (34.5 ^ 4.4 vs 39.7 ^ 4.8 at 0.25 and 0.5 mA, respectively; t18 1.2, P ⬎ 0.05). When the single-spike firing patterns elicited by stimulation of burst-positive sites were examined, the bursts were found to be embedded in the excitatory phases of two response patterns: responses characterized by an initial, long-latency, long-duration excitation (E responses; 17.5% of burst-positive responses; currents combined) and responses characterized by a long-latency, long-duration excitation following an initial inhibition (IE responses; 82.5% of burst-positive responses). Examples of the two response types—which we have characterized extensively elsewhere 36 —are shown in Fig. 1. In a small number of cases (6.4% of IE responses; all stimulation sites were located in the PPTg), the broad, long-latency excitation was preceded by a
shorter-latency (6.0–25.0 ms) excitation (Fig. 3), which usually consisted on any one trial of a single spike, closely associated with the stimulus in time. The latency variability of these events (which could vary by as much as 18.0 ms) suggested that they were not antidromic spikes. Stimulation sites which did not elicit time-locked bursts Some of the sites which did not elicit time-locked bursts (Fig. 2, Table 1) at 0.25 mA were in the PPTg (27.6%) or ⬍ 0.5 mm away from the PPTg (43.8%). However, a number of these sites were located more distal ( ⱖ 0.5 mm) to the PPTg (28.6%). Likewise, some of the sites which did not elicit time-locked bursts at 0.5 mA were in the PPTg (22.5%) or ⬍0.5 mm away from the PPTg (48.0%). However, a number of these sites were also located distal ( ⱖ 0.5 mm) to the PPTg (29.6%). The majority (81.5%) of cells which did not exhibit time-locked bursts were “non-bursting” cells, as defined by a standard criterion (see above). When the single-spike firing patterns elicited by stimulation of burst-negative sites were examined, it was found that the majority of responses were either of the E type (18.7% of responses; currents combined) or IE type (72.2% of responses). The
Fig. 2. Coronal sections (taken from the atlas of Paxinos and Watson 26) through the rostral pons, showing stimulation sites (black dots) which elicited time-locked bursts (top two rows) and those which did not (lower two rows). Data for the two current intensities used in the study (0.25 and 0.5 mA) have been combined. The anterior–posterior position of each section relative to the interaural line is indicated (in mm) adjacent to each section. Sections extend from 2.28 mm in front of the interaural line to 0.68 mm behind the interaural line. CB, cerebellum; IC, inferior colliculus; PPTg, pedunculo pontine tegmental nucleus; SC, superior colliculus; SCP, superior cerebellar peduncle; xscp, decussation of the superior cerebellar peduncle.
252
S. J. A. Lokwan et al.
remaining responses (7.9%) consisted of inhibition alone. Again, in a small number of E and IE responses (7.9% and 8.8%, respectively; associated with both PPTg sitmulation sites and sites ⬍0.5 mm away from the PPTg), the broad, long-latency excitation was preceded by a shorter-latency (4.0– 40.0 ms) single-spike excitation. Comparative distribution of burst-positive and burst-negative sites When the total complement of stimulation sites (both burst positive and burst negative; currents combined) is considered (Table 1), 37.0% of responses elicited from the PPTg were burst-positive rather than burst-negative. The percentage of burstpositive responses (relative to burst-negative responses) elicited from the PPTg was higher than for responses elicited from sites located ⬍0.5 mm away from the PPTg (20.5%) and for responses elicited from sites located ⱖ0.5 mm away (4.8%). Indeed, the frequency of burst-positive responses (vs burst-negative responses) was significantly higher for sites located in the PPTg than for sites outside the PPTg (x 2 8.4, P ⬍ 0.01). DISCUSSION
Stimulation of sites within the rostral pons elicited events in 41.3% (19/46) of A9 DA neurons which resembled natural bursts, and which were closely time-locked to the stimulation. A time-locked burst response was more frequent in DA neurons which were classified as “bursting” under basal conditions, whilst a burst-negative response was more frequent in DA neurons which were classified as “nonbursting” cells, which raises the possibility that rostral pontine sites may be involved in the natural burst activity of DA neurons. The distribution of burst-positive and burst-negative cells in the SNPc was similar, both types being found throughout this subnucleus. Such a distribution fits with what is known about the anatomy of the PPTg–SNPc projection. Tracer injections into the PPTg result in terminal label throughout the whole rostrocaudal and mediolateral extents of the SNPc. 15 Our results show, that of the sites in the rostral pons, those associated with the PPTg are particularly effective at producing time-locked bursting in A9 DA neurons. A higher percentage of burst-positive responses was obtained from sites located in the PPTg than from sites located ⬍0.5 mm away from the PPTg. Furthermore, the proportion of burstpositive responses declined dramatically as sites became more distal to the PPTg, with only a small number of responses elicited from sites 0.5 mm or more away from the PPTg containing time-locked bursts. Finally, the number of stimulations which led to a time-locked burst in burst-positive responses was higher for responses elicited from the PPTg
than for responses elicited from outside the PPTg. Given the above, the time-locked bursts elicited by stimulation of the rostral pons in the present study will be referred to as PPTg-induced bursts for the remainder of the Discussion. The simplest explanation for the production of time-locked bursts by stimulation of the PPTg is that they were produced by the excitatory action of the monosynaptic input to DA neurons from the PPTg, which has been demonstrated anatomically. 3,5 Indeed, there was a one-to-one relationship between the stimulus pulses and the time-locked bursts (i.e. each effective pulse elicited only a single burst), which is often the case for monosynaptic events. Furthermore, there was no evidence that current intensity affected either the latency of the timelocked bursts or latency variability of the timelocked bursts, as is often the case for polysynaptic events. Perhaps the biggest problem for the hypothesis of monosynaptic mediation is that the latency of the bursts (96.2 ms median) is well in excess of that normally associated with monosynaptic events. However, we have recently argued 23 that the depolarizing envelope which underlies burst events in DA neurons consists in part of a long-duration, NMDA-mediated excitatory postsynaptic potential (EPSP). Such a potential has been demonstrated in A9 and A10 neurons following stimulation of EAAergic afferents in vitro by Mercuri et al. 20 This EPSP, which was presumably monosynaptic since it was elicited by stimulation of the SNPc or VTA in close proximity to the recording electrode, had a protracted “on” time constant (the time to peak could be several hundred milliseconds). In this regard, it is similar to the burst events reported in the present study, which therefore may also be monosynaptically mediated, in spite of their long latency. The burst events reported here resemble those which are elicited by stimulation of the mPFC. We have argued elsewhere 36 that time-locked bursts elicited in DA neurons by stimulation of the mPFC are induced indirectly via the activation of more proximal EAAergic sources, such as the PPTg. The present results offer some support for this hypothesis, in that the median burst latency (96.2 ms) for PPTg-induced bursts was shorter than the median for mPFC-induced bursts in both E (122.0 ms) and IE (187.0 ms) responses (see Ref. 36). However, if a long latency does not necessarily mean that PPTg-induced bursts are polysynaptically mediated, this raises the possibility that some bursts elicited by stimulation of the mPFC (it is likely to be only a proportion, due to the sparseness of the projection; see Ref. 33) are also generated monosynaptically. Indeed, the fact that neither the latency nor the latency variability of time-locked bursts produced by mPFC stimulation are affected by current intensity supports the idea of
Rostral pontine-induced bursts in dopaminergic neurons
monosynaptic mediation. 36 Unfortunately, the high degree of latency variability present in mPFCinduced (and PPTg-induced) bursts would not normally be expected for monosynaptic events. However, bursts in DA neurons appear to be multi-component phenomena, only part of which consists of the initial EPSP. 23 Hence, the latency of the first spike within the burst may be a complex function of multiple factors, making it particularly susceptible to variability. Furthermore, the possibility cannot be excluded that some natural bursts are intermixed with the time-locked bursts. Since natural bursts are stimulation independent, they will occur randomly within the 150-ms epoch designated as containing time-locked bursts (see Experimental Procedures), increasing the temporal variability of the full complement of burst events in this period. Finally, given that the degree of temporal variability in mPFC (and PPTg-induced) bursts differed between responses, it is possible that responses which exhibit low variability are produced monosynaptically, whereas responses which exhibit high variability are produced polysynaptically, with an intermediate level of variability indicating a mixture of the two types of mediation. The bursts we report here following stimulation of the PPTg in the rat have not been reported before. Scarnati et al. 29,30 reported short-latency (2.9 ^ 1.6 ms) and longer-latency (5.2 ^ 1.8, 6.6 ^ 3.3 ms) single-spike excitations in A9 DA neurons. The latencies of these excitations resemble those of the single-spike excitations which we saw in a small number of DA neurons following PPTg stimulation. Kelland et al. 16 also reported excitations in a number of A9 and A10 DA neurons following stimulation of the PPTg (short, medium and long latency, up to a mean of 27.6 ms). Since the duration of these excitations was ⬍20 ms, these were also probably single spike, again with a similar latency to the single-spike excitations we report here. As we mentioned in the Introduction, the lack of bursting in the studies by Scarnati et al. 29,30 may be explained by the choice of ketamine, an NMDA antagonist, as the
253
anaesthetic. The failure of Kelland et al. 16 to report bursting is more puzzling, since some of the animals in this study were only locally anaesthetized. The stimulation frequency (0.5 Hz), intensity (0.3 mA) and duration (300 ms) used by these investigators were similar to our own, as was the stimulating electrode (type SNE 100). We can only conclude (as we mentioned in the Introduction) that verylong-latency events were not considered by the authors. CONCLUSION
The present study demonstrates that stimulation of the PPTg produces events in A9 DA neurons which resemble natural bursts, thereby implicating the PPTg in the generation of the natural phenomenon. Natural bursts in DA neurons are generated in response to (amongst other things) primary and secondary reinforcers. 23 The PPTg receives extensive inputs from the limbic system, 38 and hence the projection from the PPTg to DA neurons may be one means by which motivationally relevant information—including that from the mPFC, which projects directly to the PPTg in the rat 32 — reaches DA neurons. This signal is conveyed, via bursting, to forebrain structures involved in response generation, such as the neostriatum. Since the neostriatum is considered to represent the main input nucleus of the basal ganglia, 25 a group of structures which form part of the extrapyramidal motor system, 19 whilst the PPTg is considered to lie on the output side of the basal ganglia, 38 the projection from the PPTg to DA neurons could therefore represent a means by which inputs to and outputs from the basal ganglia are co-ordinated in the sequential organization of behaviour, in which the basal ganglia seem to play a role. 19 Acknowledgements—This research was supported by the Wellcome Trust (D.C.; University Award 036936) and the Medical Research Council (P.G.O., D.C. and M.S.B.; project grant G9629658N). We would like to thank Dr C. D. Richards for helpful comments on an earlier version of the manuscript.
REFERENCES
1. Albin R. L., Aldridge J. W., Young A. B. and Gilman S. (1989) Feline subthalamic nucleus neurons contain glutamate-like but not GABA-like or glycine-like immunoreactivity. Brain Res. 491, 185–188. 2. Anis N. A., Berry S. C., Burton N. R. and Lodge D. (1983) The dissociative anaesthetics, ketamine and phencyclidine, selectively reduce excitation of central mammalian neurons by N-methyl-aspartate. Br. J. Pharmac. 79, 565–575. 3. Bolam J. P., Francis C. M. and Henderson Z. (1991) Cholinergic input to dopaminergic neurons of the substantia nigra: a double immunocytochemical study. Neuroscience 41, 483–494. 4. Carter C. J. (1982) Topographical distribution of possible glutamatergic pathways from the frontal cortex to the striatum and substantia nigra in rats. Neuropharmacology 21, 379–383. 5. Charara A., Smith Y. and Parent A. (1996) Glutamatergic inputs from the pedunculopontine nucleus to midbrain dopaminergic neurons in primates: Phaseolus vulgaris-leucoagglutinin anterograde labelling combined with postembedding glutamate and GABA immunohistochemistry. J. comp. Neurol. 364, 254–266. 6. Chergui K., Charlety P. J., Akaola H., Saunier C. F., Brunet J. L., Buda M., Svensson T. H. and Chouvet G. (1993) Tonic activation of NMDA receptors causes spontaneous burst discharge of rat midbrain dopamine neurons in vivo. Eur. J. Neurosci. 5, 137–144. 7. Chergui K., Akaoka H., Charlety P. J., Saunier C. F., Buda M. and Chouvet G. (1994) Subthalamic nucleus modulates burst firing of nigral dopamine neurones via NMDA receptors. NeuroReport 5, 1185–1188.
254
S. J. A. Lokwan et al.
8. Christie M. J., Bridge S., James L. B. and Beart P. M. (1985) Excitotoxin lesions suggest an aspartatergic projection from the medial prefrontal cortex to the ventral tegmental area. Brain Res. 333, 169–172. 9. Clark D. and Chiodo L. A. (1988) Electrophysiological and pharmacological characterisation of identified nigrostriatal and mesoaccumbens dopamine neurons in the rat. Synapse 2, 474–485. 10. Clements J. R. and Grant S. (1990) Glutamate-like immunoreactivity in neurons of the laterodorsal tegmental and pedunculopontine nucleus in the rat. Neurosci. Lett. 120, 70–73. 11. Fujimoto K. and Kita H. (1993) Responses of rat substantia nigra pars reticulata units to cortical stimulation. Neurosci. Lett. 142, 105–109. 12. Futami T., Takakusaki K. and Kitai S. T. (1995) Glutamatergic and cholinergic inputs from the pedunculopontine tegmental nucleus to dopamine neurons in the substantia nigra pars compacta. Neurosci. Res. 21, 331–342. 13. Grace A. A. and Bunney B. S. (1983) Intracellular and extracellular electrophysiology of nigral dopaminergic neurons—1. Identification and characterisation. Neuroscience 10, 301–315. 14. Grace A. A. and Bunney B. S. (1984) The control of firing pattern in nigral dopamine neurons: burst firing. J. Neurosci. 4, 2877–2890. 15. Jackson A. and Crossman A. R. (1983) Nucleus tegmenti pedunculopontinus: efferent connections with special reference to the basal ganglia, studied in the rat by anterograde and retrograde transport of horseradish peroxidase. Neuroscience 10, 725–765. 16. Kelland M. D., Freeman A. S., Rubin J. and Chiodo L. A. (1993) Ascending afferent regulation of rat midbrain dopamine neurons. Brain Res. Bull. 31, 539–546. 17. Kita H. and Kitai S. T. (1987) Efferent projections of the subthalamic nucleus in the rat. Light and electron microscopic analysis with the PHA-L method. J. comp. Neurol. 260, 435–452. 18. Kornhuber J., Kim J.-S., Kornhuber M. E. and Kornhuber H. H. (1984) The cortico-nigral projection: reduced glutamate content in the substantia nigra following frontal cortex ablation in the rat. Brain Res. 322, 124–126. 19. Marsden C. D. (1982) The mysterious motor function of the basal ganglia: the Robert Wartenberg lecture. Neurology 32, 514–539. 20. Mercuri N. B., Grillner P. and Bernardi G. (1996) N-Methyl-d-aspartate receptors mediate a slow excitatory postsynaptic potential in the rat midbrain dopaminergic neurons. Neuroscience 74, 785–792. 21. Naito A. and Kita H. (1994) The cortico-nigral projection in the rat: an anterograde tracing study with biotinylated dextran amine. Brain Res. 637, 317–322. 22. Overton P. and Clark D. (1992) Iontophoretically administered drugs acting at the N-methyl-d-aspartate receptor modulate burst firing in A9 dopamine neurons in the rat. Synapse 10, 131–140. 23. Overton P. G. and Clark D. (1997) Burst firing in midbrain dopaminergic neurons. Brain Res. Rev. 25, 312–334. 24. Palmgren A. (1948) A rapid method for selective silver staining of nerve fibres and nerve endings in mounted paraffin sections. Acta zool. 29, 377–392. 25. Parent A. (1990) Extrinsic connections of the basal ganglia. Trends Neurosci. 13, 254–258. 26. Paxinos G. and Watson C. (1992) The Rat Brain in Stereotaxic Coordinates, 2nd edn. Academic, Sydney. 27. Richards C. D., Shiroyama T. and Kitai S. T. (1997) Electrophysiological and immunocytochemical characterization of GABA and dopamine neurons in the substantia nigra of the rat. Neuroscience 80, 545–557. 28. Robledo P. and Feger J. (1990) Excitatory influences of the rat subthalamic nucleus to substantia nigra pars reticulata and the pallidal complex: electrophysiological data. Brain Res. 518, 47–54. 29. Scarnati E., Campana E. and Pacitti C. (1984) Pedunculopontine-evoked excitation of substantia nigra neurons in the rat. Brain Res. 304, 351–361. 30. Scarnati E., Proia E., Campana E. and Pacitti C. (1986) A microiontophoretic study on the nature of the putative neurotransmitter involved in the pedunculopontine–substantia nigra pars compacta excitatory pathway in the rat. Expl Brain Res. 62, 470–478. 31. Schultz W. and Romo R. (1990) Dopamine neurons of the monkey midbrain: contingencies of responses to stimuli eliciting immediate behavioural reactions. J. Neurophysiol. 63, 607–624. 32. Sesack S. R., Deutch A. Y., Roth R. H. and Bunney B. S. (1989) Topographical organization of the efferent projections of the medial prefrontal cortex in the rat: an anterograde tract-tracing study with Phaseolus vulgaris leucoagglutinin. J. comp. Neurol. 290, 213–242. 33. Sesack S. R. and Pickel V. M. (1992) Prefrontal cortical efferents in the rat synapse on unlabelled neuronal targets of catecholamine terminals in the nucleus accumbens septi and on dopamine neurons in the ventral tegmental area. J. comp. Neurol. 320, 145–160. 34. Smith I. D. and Grace A. A. (1992) Role of subthalamic nucleus in the regulation of nigral dopamine neuron activity. Synapse 12, 287–303. 35. Thierry A. M., Chevalier G., Ferron A. and Glowinski J. (1983) Diencephalic and mesencephalic efferents of the medial prefrontal cortex in the rat: electrophysiological evidence for the existence of branched axons. Expl Brain Res. 50, 275–282. 36. Tong Z.-Y., Overton P. G. and Clark D. (1996) Stimulation of the prefrontal cortex in the rat induces patterns of activity in midbrain dopaminergic neurons which resemble natural burst events. Synapse 22, 195–208. 37. Tong Z.-Y., Overton P. G. and Clark D. (1996) Antagonism of NMDA receptors but not AMPA/kainate receptors blocks bursting in dopaminergic neurons induced by electrical stimulation of the prefrontal cortex. J. neural Transm. 103, 889–904. 38. Winn P., Brown V. J. and Inglis W. L. (1997) On the relationship between the striatum and the pedunculopontine tegmental nucleus. Crit. Rev. Neurobiol. 11, 241–261. (Accepted 14 December 1998)