Multiunit bursts in rat pallidum during grooming and sterotyped jaw movements

0361-9230/84 $3.00 + .OO

Brain Research Bulletin, Vol. 13, pp. 493-4%, 1984.B Ankho International Inc. Printed in the U.S.A.

Multiunit Bursts in Rat Pallidum During Grooming and Sterotyped Jaw Movements J. S. MCKENZIE, Department

of Physiology,

P. W. EVERETT

AND W. A. A. KUNZE

University of Melbourne, Received

Parkville,

Victoria, 3052 Australia

8 February 1984

MCKENZIE, J. S., P. W. EVERETI AND W. A. A. KUNZE. Multiunit bursts in rat pallidurn during grooming and stereoBRAIN RES BULL 13(4) 493-4%, 1984.-Extmcellular recordingsfrom the globus pallidus of awake, unrestrained rats showed a distinctive bursting activity during grooming behaviour and in periods of stereotyped jaw movements induced by amphetamine (3 or 5 mg/kg IP) or apomorphine (2 mg/kg SC). During stereotyped licking, there was one burst for each outward movement of the tongue. The bursts were shown to consist of several separate unit spikes firing so as to produce a fusiform envelope of amplitudes, suggesting an ordered recruitment of pallidal neurons related to licking.

typed jaw movements.

Pallidurn

Multiunit analysis

Bursting activity

Stereotypy

THE dopamine (DA) stimulants amphetamine and apomor-

phine induce a dose-dependent increase of locomotion in the rat, which at higher doses (5 mg/kg IP or greater) is replaced by stereotyped actions [6,7]. Intense stereotyped behaviour involves compulsive jaw movements, including licking, gnawing and biting. Prior unilateral destruction of the DAergic pathway to the neostriatum causes the hypermotility response to be expressed as circling, without abolishing the stereotyped behaviour seen at higher doses. The neural mechanisms underlying these behavioral responses involve action on DA-ergic receptors in the dorsal striatum and possibly nucleus accumbens (2, 7, 8, 91. Neostriatal efferents project principally to the globus pallidus (GP) and aubstantia nigra pars reticulata (SNr) 141.Hence it may be assumed that the latter are important in mediating these responses. This has led us to study unit activity in GP and SNr during drug-induced stereotyped and rotational behaviout. A pattern of neural discharge related to grooming and to jaw movements seen during stereotyped behaviour is reported here. METHOD Male hooded rats received unilateral lesions of the DAergic nigmstriatal pathway by local infusion of 6-hydroxydopamine (6-OHDA) aimed at rostra1 SN (8 pg in 4 ~1 or 0.9% saline with 8 pg ascorbic acid as a preservative). Steteotaxic coordinates (AP: 4.0 mm, LR: 2.0 mm, H: +2.7 mm) were taken from the atlas of Albe-Fessard ef al. [l]. Fourteen days were allowed for complete degeneration of dopaminergic neurons. The effectiveness of the lesions was tested by the circling t‘esponse to dexamphetamine sulphate (3 m&g IP) in a mtometer based on that described by others 15). Accepted animals were then implanted with eight recording elect&es (70 pm stainless steel with insulating coat of teflon: Medwire 3 16 SS 3t) under Ketamine anaesthesia (150 m&g IP with supplementary doses of 60 mg/kg/hr). Two electrodes weE aimed at SNr and two at GP. on each side of

493

Jaw movements

the brain. As reference electrode, a bare stainless steel needle, 0.2 mm diameter, was implanted deeply through one hemisphere. Extracellular unit recordings were made during behaviour induced by dexamphetamine sulphate (3 or 5 mg/kg IP) or apomorphine hydrochloride (1 or 2 mg/kg IP). Animals were placed in the rotometer and allowed a 30-minute habituation period, before recording for 15 minutes prior to drug administration. The post-iaection recording period was 75 minutes. Neural activity was amplified (1000x), using a bandpass of 0.5-10 KHz, before storage on magnetic tape. The apparatus allows rotational behaviour to be coded and stored along with neural activity [ 10). Records of bursting were analysed from the magnetic tape by photographing continuous sweeps transversely on moving film through a Grass camera. A fast sweep-speed (2 msec/div) allowed examination of the shape of the components contributing to each burst. A second grounded oscilloscope trace established the angled baseline. The exposed film was then projected onto a screen at 10x magnification. The major negative-going or positive-going components of spikes were clearly identifiable above biological noise, and were measured from baseline to peak of deflection along the vertical direction of film travel. The correct angles were established on translucent underlay. The frequencydistributions of negative and positive spikes were plotted in separate histograms (Fig. 2). From the spikes contributing to each modal value in the histograms (see Results), ten were selected at random and superimposed tracings made of them. This was done by matching peak values and tracing spike contours together with sloping baseline from the magnified film image. Thus noise-dependent fluctuations of peak amplitude for each modal value were transformed into a spread of baseline position in a band (Fig. 3), and a clear indication of the coherence of the major spike contours was obtained. The minor waveform components, superimposed on biological noise, were not further considered in the analysis.

494

MCKENZIE.

FIti.

1. A: Bursting

activity

in GP of rat associated

with grooming

behaviour.

EVERE’I‘I

At slow sweep

!?YI) Kt’NL%l-

speed

grouped spike discharges with fusiform envelope occur on the background of low amplitude biologlcai noise activity. B: During amphetamine-induced locomotion, single unit activity occurs without txmting pattern. C: Synchronised bursting in the GP on each side. D: Internal organisation of a burst: large spike occurs only once. E: Detail of units comprising a burst: several unit shapes and amplitudes are discernable. F: Simultaneous recording from GP (upper) and SNr (lower) during amphetamine-induced activity. Calibrations: Vertical: I mV (A,C,D.E.F): 0.4 mV (B). Horizontal: l(W) msec tA.C’.Ft: 40 msec (B); 20 msec (D); 10 msec (El.

On completion of recording, 10 FA of anodal current was passed through each recording electrode to earth for 10 seconds, and the brain perfused via the left ventricle with 10% potassium ferrocyanide in 10% formalin, to form a Prussian Blue spot at the tip of each electrode. Frozen sections were cut at 60 pm and stained on the slide with cresyl violet, to verify recording sites and the extent of 6-OHDA lesions. RESULTS

Bursting multiunit activity was observed in GP of each of 5 animals during both amphetamineand apomorphineinduced stereotyped behaviour. During the pre-drug recording period, animals occasisonally spent time grooming. Pallida1 electrodes detected similar bursting activity during this behaviour also (Fig. 1A). This firing pattern appeared again during both apmorphine- and amphetamine-induced stereotyped jaw movements. It was not present during any other behaviour (e.g., locomotion, rotation, quiet resting, rearing). and was seen only on electrodes later verified as being located in GP. Simultaneous recordings from neighbouring structures, and from SNr (Fig. lF), did not show unit bursts. During amphetamine-induced locomotion, in contrast to stereotyped behaviour, pallidal recordings contained either isolated unit discharges (Fig. 1B) or only bacground noise, but no bursting activity patterns. Bursting activity was recorded in the GP of each side during stereotyped behaviour. Its occurrence was synchronous between the two sides, usually with different amplitudes (Fig. 1C). Observed at a fast sweep speed, each burst appeared to be composed of 3

or more different units (Fig. lD, 1E). The bursts had a mean duration of 96.3 msec (S.D. 20.2) with from 3 to 5 distinct spikes within a burst, and a mean period of 2.1 (S.D. 2.2) seconds between bursts. Small amplitude spikes occurred at the start and end of a burst, while larger spikes were clustered around the center. The largest spike never tired more than once in a burst. was always centrally placed in the burst and was present in all but the initial and final bursts of a sequence. This internal structure produced a fusiform envelope (Fig. 1A). In one animal observed in detail, bursting was seen during intense stereotyped licking. The use of an audio monitor allowed examination of the licking while listening to the bursting activity. The two were synchronous, one burst being heard for each outward movement of the tongue licking the side of the rotometer. Separate amplitude histograms for negative-going spikes (Fig. 2A) and for positive-going spikes (Fig. 2B), were constructed with a bin width of 0.05 mV, for a sample of 251 bursts. The histograms had several discrete peaks clearly separated from each other. There were two peaks for negative spikes (Fig. 2A) corresponding to mean amplitudes of 2.25 mV (S.D. 0.07) and 2.73 mV (S.D. 0.08). The three peaks for positive spikes (Fig. 2B) correspond to mean amplitudes of2.25 mV (S.D. 0.07), 2.73 mV (S.D. 0.19) and 3.22 mV (S.D. 0.06). The absence of overlap implies that a burst is generated by the tiring of several separable units rather than by the superimposition of unresolved spike activity. Ten samples of the spikes contributing to each peak in the amplitude histogram were traced through the enlarger and superimposed (Fig. 3). The superimposed major tran-

MULTIUNIT

1.5

BURSTS IN PALLIDUM

2.0

2.5

3.0 mV

N

1.5

2.0

2.5

3.0

4.0

3.5 mV

FIG. 2. Frequency distribution for 821 visually discriminable spikes measured in records of 251 bursts. A: Negative-going spikes. B: Positive-going spikes (see Fig. 3). Abscissae: spike Amplitude (mV) measured from base-line. Ordinate: Number of instances in 0.05 mV bins.

Gents were highly congruent in each of the 5 peaks, giving further evidence that the spikes represented unitary poten-

tials and not randomly occurring transients with a continuous distribution of amplitudes. The high tiring frequency within a burst (approximately 500 Hz) resulted in the oscilloscope trace occasionally failing to return to the baseline before the next spike occurred. In these cases, the last part of the first spike and the first part of the next added together. When this happened it was always possible to fit the major components of such spike complexes to tracings of the individual spikes occurring elsewhere in the record. DISCUSSION

The possibility of the bursting activity being an artifact has been discounted because: (1) The bursting complexes were not seen in neighbouring or distant structures outside GP during the same periods of observation. Electrodes that were aimed at GP but misplaced by distances less than one millimeter did not display this distinct bursting activity, nor did simultaneous recordings from SN (Fig. 1,. It would be highly unlikely for each of the several recording electrodes outside GP to lie on the null points of distant electromyographic current sources. (2) The amplitude-frequency histogram had distinct, well-separated peaks. Although the recorded transients could theoretically represent artifacts derived from mechanical movement and muscle electrical activity, the shape and magnitude of such transients would depend on multiple factors such as the various conductive

FIG. 3. The shape and magnitude of the units contributing to the oeaks in the soike distribution of Fig. 2. Ten tracings of each spike have been superimposed to demonstrate their coherent configuration as evidence for their unitary nature. The peaks of the principal transient in each spike were aligned in order to show clearly the shape consistency of the njor waveform. This method of superimposition yielded a range of base-line positions, the limits shown by the parallel sloping lines thus representing background noise contributions to variation of peak amplitude as measured from the grounded base-line, and indicated quantitatively in the histograms of Fig. 2. Calibration: I mV, 2 msec.

pathways from distant current generators, and the strength of the many such generators. EMG recordings would not resolve the issue, because of sampling limitations and signal degradation. The amplitudes of such artifactual transients could take many values between the biological noise and the maximal spikes recorded. Hence any large sample would have a continuous range of amplitudes with a Gaussian distribution [3]. In contrast, our recordings showing 5 discrete amplitudes (Fig. 2A and 2B) would imply that only a small group of units was isolated by the electrode. Microelectrodes with high impedance would select single units better from within immediately proximate groups of cells, but could not reduce any contaminating local potentials due to volumeconduction from distant muscles. It is unlikely that many of the large transients resulted from coincident unit discharges, since this would be evident in arithmetical summation of amplitudes or, if the unit discharges were not perfectly synchronous, in the amplitudes of such compound transients falling between the peaks in Fig. 2A and 2B, so producing a more continuous distribution of amplitudes. The multimodal histogram shown in Fig. 2, and the ability to superimpose the major components traced from fast

MCKENZIE, EVERETT AND KCJNZI:

49h

sweeps with a high degree of coherence, as shown in Fig. 3, argues that a burst is formed by several discrete units firing in a complex, rather than by an increase in random background activity. The electrodes used are regularly able to resolve single-unit activity at signal-to-biological noise ratios up to 5:l (see [lo] and Fig. IE), at firing-rates up to 800 Hz, in excess of those seen here within a complex. Microelectrode recording techniques better able to resolve the firing pattern of each unit and simultaneously determine its relationship to other units would be required to examine more precisely the structure of these multiunit complexes. Histological examination showed that all electrodes recording bursting activity were located in GP, GP neurons are dispersed among striatal efferent and other fibres passing through the region, and it is unlikely that the electrodes used in this experiment were able to record only from the neurons and never from axons. Units with firing patterns related to jaw movements associated with feeding have been demonstrated in both neostriatum and GP of the monkey [I I]. By extrapolation to the rat, striatal efferent axons approaching GP cells might also fire in relation to oropharyngeal activity. Recording electrodes in neostriatum would be needed to test this possibility.

Being correlated with licking, the bursts might constiturc responses to facial or oropharyngeal sensory stimulation produced by the movements concerned. In awake cats. :I majority of GP units were found to be responsive to mechanical stimulation of the face, particularly in the peri-oral rcgion [121.The GP bursts described here were observed only during movements that involve marked jaw action, viz. grooming and licking. Since many other movement patterns of the rat must also involve sensory stimulation of vibrissar and muzzle, the bursts were probably not related to peri-oral facial stimuli but to oropharyngeal movements per se. This is not to exclude responses to afferent input from oropharyngeal receptors in joint or muscle. difficult to test for in awake behaving rats.

The fusiform envelope of the discharge seem’s 111rcprcsent a recruitment of neural elements culminating at the middle of the burst, Since afferent axons to the GP (see above) could contribute especially to smaller components of the burst, it might be speculated that the larger GP neurons fire near the central point of a burst following a brief period of temporal facilitation. This suggests an ordered discharge of GP units concerned with the stereotyped acf of licking or grooming.

REFERENCES 1.

Albe-Fessard, II., F. Stutinsky and S. Libaurban. Atlas du Diencephale du Rat Blanc. Paris: Editionsdu

Stereotaxique

Centre National de la Recherche Scientifiaue. 1966. 2. Creese, I. and S. D. Iversen. The iharmacological and anatomical substrates of the amphetamine response in the rat. Brain Res 83: 419-436, 1974. 3. Ferguson, G. R. Statistical Analysis tion, Tokyo, Sydney: McGraw-Hill,

in Psychology

und Educa-

1971, pp. 86-87. 4. Graybiel, A. M. and D. W. Ragsdale. Fibre connections of the basal ganglia. In: Progress in Brain Research, ~0151. Drvelopment and Chemical Speciticitv of Neurons, edited bv M. Cuenod, G. W. Kreutibeig and F. E. Bloom. Amsteidam; Elsevier, 1979, pp. 239-283. 5. Greenstein, S. and S. D. Glick. Improved automated apparatus for recording rotation (circling behaviour) in rats or mice. Pharmacol

Biochem Behav 3: 507-510. 1974.

6. Iversen S. D. Striatal function and stereotyped behaviour. In: Psychobiology of the Striatum, edited by A. R. Cools, A. H. M. Lohman and J. M. L. van den Berken. Amsterdam: Elsevier, 1979, pp. 99-118. 7. Iversen, S. D. and G. F. Koob. Behavioural implications of dopaminergic neurons in the mesolimbic system. In: Advance5 in Biochemical Psychopharmacology, vol 16, edited by E. Costa and G. L. Gessa. New York: Raven Press, 1977, pp. 209-214.

8. Kelly. P. H, and S, D. lversen. Selective 6OHDA-induced destruction of mesolimbic dopamine neurons: abolition of psychostimulant induced locomotor activity in rats. tZrrr .I Pharmacol 40: 45-56. 1976. 9. Kelly. P. H.. P. W. Serviour and S. D. Iversen.

Amphetamine and apomorphine response in the rat following 6-OHDAlesions of the nucleus accumbens septi and corpus striatum. Rrl~in Res 94: 507-522, 1975. 10. McKenzie, J. S., P. W. Everett and L. J. Dally. A method for simultaneously recording neural activity and rotation in the rat. Physiol Behav 30: 643-657, 1983,

Rolls, E. T., S. J. Thorpe, S. Maddison, A. Roper-Hall, A Puerto and D. Perret, Activity of neurons in the neostriatum and related structures in the alert animal. In: The Ncostriatrcm, edited by I. Divac and R. G. E. Oberg, New York: Pergamon Press, 1979, pp. 163-182. 12. Schneider, J. S., J. R. Morse and T. I. Lidsky. Somatosensory properties of globus pallidus neurons in awake cats. E.YP Brairr Rrs 46: 311-314. 1982.

II.