Prolonged medial forebrain bundle unit responses to rewarding and aversive intracranial stimuli

Brain Research Bulletin, All rights of reproduction

Vol. 1, pp. 517-522, 1976. Copyright 0 ANKHO International in any form reserved. Printed in the U.S.A.

Inc.

Prolonged Medial Forebrain Bundle Unit Responses to Rewarding and Aversive Intracranial Stimuli’ JAMES J. KEENE’

Department of Physiology and Biophysics, School of Medicine tiniversity of Puerto Rico, San Juan, Puerto Rico 00936 (Received

14 November

197.5)

KEENE, J. J. Prolonged medial forebrain bundle unit responses ro rewarding and aversive intr~~ran~a~~~~rnul~. BRAIN RES. BULL. l(6) 517-522, 1976. - In unanesthetized cerveau isole rats, brief medial forebrain bundle (MFB) and dorsal midbrain reticular (RET) stimulus trains elicited prolonged MFB unit responses lasting up to ten or more seconds. On the MFB unit responses studied, the effects of MFB and RET stimuli were basically similar, although the stimulations were probably rewarding and aversive respectively. These data agree with the previously reported anatomical localization, in medial regions of thalamus and paIlidum, of opposite single cell responses to the behaviorally opposite inputs, and suggest that unit responses to rewarding stimuli should not be characterized as reward-related when aversive stimuh elicit similar responses in the same unit. Medial forebrain bundle

Reticular formation

Unit recording

habenula, superior colhculus, and RET itself [ 51, MFB and RET convergence was predominately excitatory, as revealed by significantly increased firing. These findings raised the questions of whether MFB units also show changes in firing rate lasting seconds after the brief stimulus trains, and whether MFB and RET stimuli differ in their effects on

THIS REPORT is one of a series [ 4-71 in which the effects of medial forebrain bundle (MFBj and dorsal midbrain reticular (RET) stimuli have been compared on single units recorded in cerveau isole rats. Previous papers in this series have emphasized that brief MFB and RET stimulus trains have elicited prolonged poststimulus changes in firing rate lasting several seconds up to a minute or more, and that these MFB and RET effects in widespread areas of the brain typically converge on greater proportions of single cells than would be expected by random distribution of the effects in the samples. The present study adds data obtained under similar experimental conditions on units located in the MFB itself. It was thought that more information on unit activity in MFB following the brief stimulus trains might assist in the interpretation of results obtained with the same methods in other brain structures. Specifically, longlasting MFB-elicited inhibition and RET-elicited excitation was anatomically localized in intralaminar thalamus [ 4,5 1. Long duration opposite responses were also found in a medial pallidalentopeduncular field [ 61 The MFB-elicited excitation and RET-elicited inhibition in medial pallidum was the inverse of the intralaminar thalamic effects. Nucleus reticularis of thalamus [ 7 1. the thalamic ventral anterior-ventral lateral complex [ 6.7 1, and sensorimotor cerebral cortex [ 51 were characterized by convergence of significantly decreased firing after MFB and RET stimuli. In internal capsule [ 61,

MFB units. METHOD To facilitate comparison of results. the same experimental procedures, criteria for definition of a unit response, and methods of data reduction were used in the present and previous studies, which contain detailed description [ 5-71 The postcollicular cerveau isoli: preparation was chosen because after brain transection under ether and anterior to the Vth nerve, local and general anesthetic and artifical respiration were not required. Blood loss was minimal since only a small slit in the skull above cerebellum was made for the section, and the basilar artery was spared. Rectal temperature was monitored and maintained as necessary with a heat pad. From a larger number of experiments, eight rats (male, 400 g, Holtzman) were selected because of accurate placement of MFB and RET stimulating electrodes, judged by study of 40 p Nissl and Weil stained brain sections, in which all ipsilateral stimulating and recording electrode tracts were identified. The twisted pairs of 250 ti double-in-

’ Partially supported by NIH General Research Support Grant RR-5419. ZReprint requests to: Dr. James Keene, Fisiologia, Escuela de Medicina, UPR, GPO Box 5067, San Juan, Puerto Rico 00936. 517

t, 1 i

518

sulated nichrome wire, exposed at the cross sections of lhc rips. were aimed 3.5 mm posterior to bregma. 1.7 111111 lateral from the midline, and X.5 mm below th? skull surface. for the MFB. and 7.0 mm, 1.7 mm, and 6.5 mm respectively. for RET, with the bregma-lambda line horiLontal. The tips of the MFB stimulating electrodes and the 07 units in the present MFB sample were all located lateral to the lateral edge of fornix, medial to internal capsule. and clearly within the diffuse field of MFB fibers visible with Weil staining. The MFB units were distributed from the bodies pas. optic chiasm anteriorly to the mamtnillary teriorly (Fig. 1).

/

i

FIG. 1. Anatomicai distribution of the 52 MI% units tested with MFB and RET stimuli, in sections 1.5, 2, 3, and 4 mm posterior to bregma where the bregma-Iambda line is horizontal. According to standard measures of skew and kurtosis, this distribution of points did not significantly differ from normal in any of the three dimensions (anterior-posterior, lateral, and dorsal-ventral), except for a slight preponderance of anterior points. None of the product-moment correlations between these anatomical dimensions and four measures of MFB unit activity (spontaneous rate and poststimulus changes after MFB, RET, or both stimuli, all in spikes per second) were statistically significant. Abbreviations: MT, mammiliotbalamic tract; Fx, fornix; CP, cerebral peduncle; OT, optic tract.

From extracellular recordings with etched tungsten microelectrodes (exposed tips of about 1 p 50-150 Kohm at 2KHz, above 20 Mohm at low test frequencies), a unit was defined by well isolated action potentials believed to originate from a nearby single cell according to conventional criteria [9]. Standard pulses triggered by such action potentials, their integrated rate, and the slow wave component (1-l 5Hz) of the microelectrode recording, were displayed on oscilloscope and polygraph. Action potentials were counted electronically over ten second periods follow-

lh, i

lng stimuli i 0.3 set, IO0 Iii. O.> nl>-t‘i 2~ltl~i1~i~ -:.ilil~k!,i. pulses at 600 MA) delivered to MFB, K t.‘f. ,~nd :n ;r!ii animal, ventrai medial thalnmus. via tran~lsior ~onir~&.*~; constant current devices isulated from grouuj. Stimuli were administered m random 8)rder and su~.c:i’l sive ten second spike counts were made :~it~bleach jtlnltlit!\ until firing rate returned, if there was an apparent rt’spon~*, to previously determined vtllucs of spont:tnri~us activity f01 each unit. Repeated trials were given to unsure that oni) consistent poststimulus effects wouiil iv %:ategorizeti .I:, significant responses, and to minimk error due to apuntaneous changes in firing rate. For each unit :~nd stimulus, 11~~ poststimulus counts were averaged and cl~nlparcd with the expected count. the tnean spontaneous rats (x:, ps. 0.01 ). Since averages were used although the ihi-squat-c test assumes only single trial count\, the pl-obal~il~ty of defining 3 significant response by chance was less than suggested by the p value. With this procedure. responses of short duration, lesser magnitude. or both wert’ not classified ;ib significant.

The mean spontaneous discharge rate of the entire MFB unit sample was 12.3 spikes/set. and most units maintained a relatively steady pattern of spontaneous firing. as illustrated by prestimulus records in Fig. 2. in contrast to the incidence of bursting and irregular patterns of activity common in other structures, such as thalamus i S], in this preparation. Following MFB. RET, and simultaneous MFB and RET stimu!i, 35%., 38X. and 38% respectively of the units tested showed significant responses ! Fig. 3). which will be referred to as excitation or inhibition with no inference intended pertaining to synaptic mechamsms. The duration of these responses ranged from several seconds to more than 10 set (Fig. 2). For each of the MFB, RET, and MFB plus RET stimulus conditions. units with poststimulus decreases in discharge slightly outnumbered those with increased rates, but there was no significant net effect of the stimuli as previous reports in this series have noted in other structures. For each stimulus treatment, comparisons between the proportions of cells showing increased and decreased rates, or between the subsamples showing excitation or inliil~itio~i, did not reveai any significant differences (Fig. 3 1according to the binomial distribution. In addition, t.he mean poststimulus rate changes for all units tested with MFB ( 0.8 spikes/set or RET C-1 .I spikes/set) stimuli or both together ( 0.5 spikes/set) did not notably differ from the corrrsponding spontaneous rates, judging by paired t tests. These data contrast the results from 15 MFB units tested with ventral medial thalamic stimuli, where no significant I’+ sponses were elicited although most of the cells ( I ? of IS: binomial p
MFB UNIT

RESPONSES

M+R

M+R

ti MFB

UNITS

I

5 SEC

4

HG. 2. Representative activity

examples of effects of brief MFB (M) and RET (R) stimulus trains (dots) given separately or together (hl + RI on the Each row illustrates a single well isolated unit with each action potential represented by a standard pulse. The to add due to polygraph pen inertia when spikes occur at high frequency. Unit 1: no rep&cable responses were elicited.

of MIPB units.

pulses appear Units 2 and 3: each stimulus is followed by prolonged increased firing. Unit 4: decreased firing occurred after the stimuli. Units 5 and 6: clearly decreased discharge follows M stimuli, while R stimuli elicit smaller decreases. Unit 7: R stimuli elicited responses similar to the ‘11effect which appear to summate after both stimuli together. Surnmation of MFB and RET effects is also apparent m Lnits 2 and 6.

increased discharge after MFB stimuli; while MFB and RET stimuli each elicited decreased activity in 21 of the 28 remaining units. This significant association (x2 , p) received converging excitation, and eight ( 17%) showed converging inhibition (Fig. 2). These frequencies of converging excitation (p
stimuli did not elicit significant responses in a greater proportion of cells than responded to the stimuli separately (Fig. 3). Since similar MFB and RET effects tended to converge on single cells, the 32 units studied under the MFB and RET as well as the MFB plus RET stimulus conditions were divided into two groups (Fig. 4). In these groups, each stimulus condition produced significant mean changes in discharge rates. compared to corresponding spontaneous rates. Summation was suggested by the magnitude of firing rate increases. hut less evident in the subsample showing predominately inhihitory responses, probably because the stimulus intensity was already well above threshold. Apparent summation of the effects occurred in indivi~~ual units, but stimulus current was not systematically varied as required to estabish summation as characteristic of these effects in all of the units studied. Lastly, further analysis checked for indirect evidence of possible presynaptic interactions of the stimuli as has been suggested in lateral globus pallidus [h]. In that case. RET-elicited excitation was typically blocked or reduced in most units when MFB stimuli were given with the RET,

60

, %OF

50

,N=52

UNITS 48

TESTED 32

60 50

40

40

30

30

20

20

1’0

10

10

10

20

20

30

30

40

40

50

50

60

60

STIMULI

MFB

RET

BOTH

FIG. 3. Prolonged responses of MFB units over ten second periods after MFB and RET stimuli and both together. Percents of units with increased or decreased poststimulus firing are shown, with percents of cells showing significantly excitatory or inhibitory responses (black bars). The three stimulus treatments elicited similar effects in the samples. even though the MFB stimuli alone evoked no postsynaptic effects observable in extracellular recordings. Considering the MFB units tested by both stimuli separately and in combination, however, there was no evidence that one of the stimuli might block or augment the effects of the other, while having little or no effect itself. Of course. other experimental designs specifically addressed to the question of presynaptic interactions are required. DISCUSSION

In summary, brief MFB and RET stimulus trains have elicited prolonged MFB unit responses sustained over much of, and often more than, the ten second periods of measurement, and several analyses of the interaction of MFB and RET effects on single cells suggest that the two stimulation sites have basically similar effects in the MFB unit responses studied. The duration of action potential waveforms in this study, with an initial segment notch discernible in many units, and the exclusion of units outside

STIMULI

% CHANGE

IN

li7SPIKES/SEC

N=18

MFB

RET

BOTH

FIG. 4. Prolonged MFB unit responses during ten second periods following MFB and RET stimuli and both simultaneously. In subsamples tested under all three stimulus conditions and divided according to the predominant direction of the effects in each cell (Excitatory Resp. and Inhibitory Resp.), significantly increased rates were elicited by MFB @<0.005) and RET @<0.02) stimuli and both @<0.03). The stimuli each elicited a significant mean decrease in firing in the other sample (p
MFB UNIT

521

RESPONSES

similarity of MFB and RET effects on these units and others reported previously. One is that independent pathways may converge on the recorded neurons or at some distant site. Another is that both the MFB and RET stimuli would be expected to directly activate some elements in both lateral hypothalamus and midbrain tegmentum, either through depolarization of nearby soma or antidromic invasion of distant soma. These alternatives are consistent with anatomical [ 1 1 ] and physiological [ 121 evidence, and are not necessarily mutually exclusive. The MFB and RET stimuli used in this study were probably rewarding and aversive respectively [ 131, but there appears to be no evidence supporting or contradicting this assumption in the cerveau isole preparation, nor any previous studies other than those already cited comparing MFB and RET effects at the unit level, except one [ 171 confirming results from this lab on hippocampal units [4,5]. There are several studies, however, in which the behavioral effects of MFB stimulation have been related to MFB unit activity [ 3. 15, 161. These studies differ from the present in the use of chronic animais, shorter poststimulus periods, and large recording probes whose sampling bias is probably different than that of the present microeiectrodes. But it may not be inappropriate to consider the possible behavioral significance of those studies and the present results, since findings from cerveau isole animals have been supported with use of behaviorally confirmed rewarding and aversive brain stimulation to test units recorded in chronically implanted, awake rats [ 11 and cats (unpublished o~s~rv~itions~ as we11 as rats tested behaviorally before acute recording [ 4.91. The frequent occurance of similar prolonged unit responses to MFB and KET stimuli in numerous forebrain and midbrain structures implies that the interpretation of unit activity in relation to the rewarding and aversive features per se of the stimuli must procede with caution. There were no substantial differences between MFB and RET effects in the present MFB unit sample. This contrasts the clear d~s~rilnination between these stimuli by single units anatomically localized in medial pallidum and intralaminar thalamus, These data suggested that the similar responses may represent some common MFB-RET factor, which was labelled on arousal dimension, and that the differential responses would be more likely than any other to pertain to an affective dimension, reflecting hoth rewarding and aversive properties of stimuli [S-7]. The

long duration of these responses iasting seconds might be partially explained by the present finding that neurons in the MFB which may contribute directly or indirectly to the responses in other structures, undergo significant changes in activity sustained over the same poststimulus periods. The present results suggest that the MFB may not be involved in the appreciation of differences between rewarding and aversive inputs in the same way as medial pallidum and intralaminar thalamus might, although this possibility is not excluded because only long duration responses were studied. It is not even clear which. if any, of the MFB elicited responses in MFB units studied thus far might be especially pertinent to the rewarding features of the stimuli, since believed aversive RET stimuli tended to elicit similar responses converging on the same units. Also, in other studies of lateral hypothalamic units. auditory cues preceding food reward have elicited conditioned reactions involving mostly increased discharge [ 10.14], hut units responsive to the presence of a food reward which elicits eating all apparently show decreased firing [ Z] These considerations raise questions regarding the speculation that MFB unit responses to MFB stimuli eliciting self-stimulation may represent the rewarding property of the stimuli [ 161, where it has not been demonstrated that these responses are significantly different from those an aversive input might evoke. Indeed, some of these unit responses might be viewed as arousal-related, an interpretation favored by the present data. First. of four substances that facilitated self-stimulatio?l, only L-norepinephrine both facilitated other behaviors nonspecifically and affected MFB neural responses to rewarding stimulation ( 15 1. contrary to what might be expected if the unit responses were primarily reward-related. Second. the demonstration of similar thresholds for MFB elicitation of self-stimulation behavior and MFB unit responses [ IhI, without evidence that the threshold for arousal phenomena was significantly different in these experiments, could support the present hypothesis because the increased self-stimulation could itself be considered as one indicator, but not the best, of increased arousal level. This discussion has focussed on techniques which might help establish how MFB unit activity may be reward-related, in view of the well known relation between this structure and reward phenomena; and the likelihood that some MFB unit activity is reward-related is not denied.

REFERENCES 1. Casey, K. L. and J. J. Keene. tJnit analysis of the effects of motivating stimuli in the awake animal: pain and self-stimulation. In: Brain Unit Activity During Behavior, edited by M. I. Phillips. Thomas, Springfield, 111: 1973, pp. 115-129. M. D. Hypothalamic unit activity and eating 2. Hamburg, behavior. ,4nz. J. Physiol. 220: 980-985, 197 1. 3. lto, M. The excitability of medial forebrain bundle neurons during self-stimulating behavior. J. Neumphysiol. 35:

652-664, 4.

1972.

Keene, J. J. Opposite rewarding and aversive

medial thalamic brain stimulation.

unit

responses

to

Expl Nertrol. 39:

19-35.1973. 5. Keene, J. J. Reward-associated excitation

lasting

seconds

inhibition and pain-associated in single intralaminar thalamic units.

Brain RES. 64: 211-224, 1973. 6. Keene, J. J. Reward-associated excitation and pain-associated inhibition lasting seconds in rat medial pallidal units. ExpZ Nefrrof. 49: 97-l

14, 1975.

7. Keene,

J. J. Prolonged unit responses in thalamic reticular. ventral, and posterior nuclei following lateral hypothalamic and midbrain reticular stimulation. J. Neurosci. Rex I: 459-469,1975. 8. Keene, J. J. and K. L. Casey. Excitatory connection from lateral hypothalamic self-stimulation sites to escape sites in medullary reticular formation. Expl Newel. 28: 155 -166, 1970. 9. Keene, J. J, and R. L. Casey. Rewarding and aversive brain stimulation: opposite effects on medial thalamic units. Pl~~ts~~l. Behav. IO: 283-287, 1973. 10. Linseman, M. A. and J. Olds. Activity changes in rat hypothalamus, preoptic area and striatum associated with Pavlovian conditioning. J. Neuroph_vsiol. 36: 103% 1050, 1973. 11. Millhouse, 0. E. A Golgi study of the descending medial forebrain bundle. Brain RCS. IS: 341-363, 1969.

12.

Murphy, J. T.. J. J. Dreifuss and P. Gluor. ‘I’opographical differences in the responses of single hypothalamic neurons to limbic stimulation. Am. J. Ph.ysio/. 214: 1443-1453. 1968. 13. Olds, J. and M. I?. Olds. Approach-avoidance analysis of rat dioncephalon. J. camp. Neural. 120: 259.--295. 1963. 14. Olds, M. E. Short-term changes in the firing pattern of hypothalamic neurons during Pavlovian conditioning. Brairr Re,s. 58: 9S- I 1A. 1973.

IS.

16. 17.

Olds. .%I. t,:. Effect ol‘ intraventrlcular Ilowpinephrmc ,J*: neuron activity in the medial forebrain bundle during wlf-blitllulation behavior. Bruin Rex 80: 461 -477. 19~4. Olds, hl. 1,:. Unit responses m the medial torcbranl hundie !(I rewarding stimulation in rht, hypotlr:rltnux Rrnili liC%. x0: 479.-495, 1974. Sinnamon. H. M. and J. S. SchaartLbaum. Ik~rcal IGppwampal unit and EXG responses tkl rewarding and aversive ~I-~III stimulation in rats. Bruit? Rc>s. 56: 183 202. 1973.