Brain Research, 106 (1976) 257-270 © ElsevierScientificPublishing Company, Amsterdam - Printed in The Netherlands

257

BEHAVIORAL STUDIES FOLLOWING LESIONS OF THE MESOLIMBIC AND MESOSTRIATAL SEROTONERGIC PATHWAYS

MARK A. GEYER, AMADEO PUERTO, DAVID B. MENKES, DAVID S. SEGAL ANt) ARNOLD J. MANDELL

Department of Psychiatry, School of Medicine, University of California, San Diego, La Jolla, Calif. 92093 (U.S.A.) (Accepted September 15th, 1975)

SUMMARY

The behavior of rats with selective lesions of either the dorsal (B7), median (B8), or lateral (B9) raphe nuclei was compared to that of sham-lesioned controls in a variety of experimental situations. As described previously 17, the extent of damage to the midbrain raphe nuclei was determined by fluorescence histochemistry, and the tryptophan hydroxylase and tyrosine hydroxylase activities of 6 forebrain regions were measured for each rat. None of the lesions affected tyrosine hydroxylase activity. Lesions of B7, which reduced tryptophan hydroxylase in the striatum, thalamus, cortex, and hypothalamus, had no significant effect on any of the behavioral measures. Lesions of B9, although twice as large, neither reduced forebrain tryptophan hydroxylase significantly nor affected any of the behavioral variables. However, B8 lesions, which reduced hippocampal, septal, cortical, and hypothalamic tryptophan hydroxylase, had behavioral effects similar to those reported after combined raphe lesions or parachlorophenylalanine. Median raphe-lesioned rats were hyperactive when placed in a novel environment and throughout the dark phase of the light/dark cycle. With respect to locomotor activity, B8-1esioned rats were also hyper-responsive to amphetamine. When placed in a stabilimeter and subjected to repeated air puff stimuli, rats with B8 lesions exhibited larger startle responses. Furthermore, only B8lesioned animals perseverated when given two unreinforced trials in a Y-maze. All these histologic, biochemical, and behavioral variables were assessed individually for all 39 animals, and a multivariate correlational analysis incorporating the data of this and the preceding paper 17 is presented here. These experiments suggest that the mesolimbic serotonergic pathway originating in B8 subserves some of the inhibition necessary to dampen behavioral responsivlty.

258 INTRODUCTION

Relationships between central serotonergic function and behavioral responsiveness to sensory stimuli have been the focus of much research. We recently reported that intraventricular infusion of serotonin (5-hydroxtryptamine, 5-HT) decreases the responsiveness of rats to repeated air puff stimulilS; central infusions of the putative hallucinogenic congener of 5-HT, bufotenin (N,N-dimethyl-5-HT) have the opposite effect. Our findings corroborated the work of Davis and his associates, who reported that combined electrolytic lesions of the midbrain raphO 3 or the administration of mdoleamine hallucinogens12 increased responsiveness to auditory stimuli. Further evidence for an inhibitory role of 5-HT in sensory responsivity comes from other studies involving a variety of behavioral measures after less direct pharmacological manipulations of whole brain 5-HT6-S,al,aL Many of the behavioral phenomena thought to be related to alterations in brain 5-HT have been elucidated by combined lesions of the dorsal and median raphe nuclei la,29,3°,34. To explore the possibility that these behavioral effects may be specifically related to one or another of the raphe nuclei, we tested rats with discrete lesions of B7, B8, or B9 with a variety of behavioral measures that are sensitive to manipulations of brain 5-HT and are thought to reflect aspects of behavioral inhibition. In the preceding paper 17 we described the effects of discrete lesions of the raphe nuclei on regional tryptophan hydroxylase and tyrosine hydroxylase activities in the same rats. We discriminated two separate serotonergic pathways derived from the midbrain raphe nuclei: the mesolimbic, innervating the hippocampus and septum, originating in the median or B8 nucleus, and the mesostriatal, innervating the striatum and the thalamus, originating m the dorsal or B7 nucleus. During the 4 weeks between stereotaxic surgery and sacrifice we tested the animals in a variety of behavioral situations. First, we examined their spontaneous locomotor actwity upon initial presentation to the chambers and during 6-h periods of 'day' and 'night'. Then we examined the magnitude and patterning of their startle responses to repeated air puff stimuli. Third, we tested them with two unreinforced trmls in a Y-maze to see whether they would perseverate, as do rats with septal or hippocampal lesionsa, lo. This simple behavioral test has commonly been used to assess response inhibition processes and is sensltwe to serotonergic manipulations43. Although raphe-lesioned rats have not been reported to perseverate, animals with combined lesions of B7 and B8 have been found to be both generally hyperactive and hyper-responsive to auditory stimuh. All the lesions, histological verification, biochemical determinations, and behavioral observations were performed on the same groups of animals, permitting a multivariate correlational analysis based on all the data presented in this and the preceding paper 17. METHODS

The 39 male rats used m the biochemical studies 17 were also used in the behavioral experiments. Stereotaxic lesions had been made in B7, B8, or B9 as described.

259 Controls were treated identically except that no current was applied. In the preceding paper 17 we give details of our histological verification of the extent of damage to the serotonergic cell bodies in the midbrain of each animal.

Locomotor activity Seventeen to 19 days after surgery each animal was placed in one of 39 activity chambers (12 in. × 12 in. × 15 in.), in which crossovers from one quadrant to another are detected by contacts in the floor and continuously monitored with a Nova 1200 ® computer ag. Locomotor activity was recorded as crossovers for 48 h during which a 12/12 h light/dark cycle and access to food and water were maintained. We divided the first hour into five 12-min blocks to assess the initial response to this novel environment. The animals were removed briefly for weighing about 18 h later, and we analyzed the first five 12-min blocks of crossovers after they were returned to the chambers. To obtain the most representative time blocks for light and dark cycles, we discarded the crossover data for the first 2 h following the weighing and tallied crossovers in three 6-h blocks, that is, the last half of the light phase and both halves of the dark phase. After 43 h in the chambers, half the animals in each group were injected with D-amphetamine sulfate (2.5 mg/kg free base s.c., Sigma) and the remainder with isotonic saline (1.0 ml/kg s.c.). Recording continued for 4 h more, and analysis of variance (ANOVA) indicated that the injections of amphetamine had no effect on any other behavioral or biochemical variable in the study. The data relating to amphetamine-induced behavioral changes have been replicated with larger groups and will be reported in detail separately (Segal et al., manuscript in preparation).

Startle responding Three weeks after surgery each rat was given two daily sessions of 15 min each in the stabilimeter chamber; the third day he had a 5-min warm-up period in the chamber, and 30 trials of air puff stimuli on a 30-see fixed interval schedule. Deflections of the stabilimeter chamber were detected potentiometrically, amplified, and recorded on a strip-chart recorder to indicate the magnitude of the startle responses. The details of the apparatus and procedures have been described elsewhere is.

Spontaneous alternation Two days after they were tested in the stabilimeter, we subjected the animals to two unreinforced trials in a Y-maze to see whether they alternated choices (as normal rats do 10) or perseverated (as rats with hippocampal or septal lesions do9,10). The Ymaze, consisting of 3 identical wooden alleys (19 in. × 5 in. × 6 in.) and a grid floor, was set up in an otherwise empty room. Each rat was placed in the start alley and allowed to explore the maze. His first entry (all 4 feet across the line) into one of the other alleys was recorded, and he was removed from the maze. After 1 rain he was returned to the start alley for the second trial. All animals responded within the 5-rain limit. Perseveration was defined as choosing the same alley on both trials. These binary data were analyzed with a simple proportionality test.

260

Statistical analysis A one-way ANOVA was performed on the grouped data, excluding the two B8-1esioned animals with less than 10~ damage to the B8 nucleus, which were also excluded from the graphed data. To determine the statistical rehability of a specific lesion manipulation, comparisons were made using the F-test with a pooled variance estimate. Because of the inherent variability in the extent of damage produced by the lesions, an overall correlational analysis was also performed, incorporating damage estimates17 independent of the target nucleus. A few stepwise regression analyses of particular variables were used to further assess some of the relationships suggested by the correlational analysis. RESULTS

Although B7 lesions produced significant decrements in tryptophan hydroxylase activity in several forebrain regions and selective reduction in this enzyme activity in the striatum and thalamus, they did not significantly alter behavior. Lesions of the B9 nuclei had no demonstrable effects on either tryptophan hydroxylase activity or behavior. Only lesions of the B8 nucleus, which resulted in selective reductions of tryptophan hydroxylase in the hippocampus and septal nuclei as welt as in hypothalamus and cortex 17, produced significant behavioral changes; these were simdar to those reported after combined raphe lesions 13,29,a°,34, reflecting a general increase in responsivity. Lesions of B8 increased the initial exploratory activity in a novel environment, increased the general locomotor activity during the dark phase of the light cycle, increased hyperactivity in response to amphethamine injection, increased the magnitude of startle responses, and produced perseveration in the Y-maze.

Locomotor activity The mean number of crossovers per 12-rain block during the first h in the activity chambers is presented for each group in Fig. 1. Animals with lesions in B8 were significantly more active than the other 3 groups during the first 24 rain (F [1,33] 30.4, P - < 0.001 and F [1,33] = 12.1, P < 0.002). By the end of the first hour all groups exhibited comparable levels of activity, and this trend was maintained over the next 2 h of the light phase. The same phenomenon was observed the following day when the animals were removed from the chambers for weighing; after being returned to the chambers the B8-1esioned animals were hyperactive during the first 12 mm ( r [1,33] = 13.7, P < 0.005). Median raphe-lesioned animals were not significantly different from the other 3 groups during the representative 6-h light period. Their response to light offset was more dramatic than that of any other group, however, and they maintained a higher level of activity for the 12 h of dark, which is reflected in the mean crossovers per 6-h block diagrammed in Fig. 2. This relative hyperactivity was most dramatic during the 3 h immediately preceding light onset, during which the number of crossovers registered by B8-1esioned animals was roughly twice that of any other group. Within 2 h after light onset the B8 animals returned to control levels of activity (not shown).

261 60

\ 50

\

= = o---o A__4 D 0

"~ \ O3 m,,,,

CONTROL (n=/O) 87 (n=lO) BS(n=8) B.9(n;9)

40

\

,,,

O t.#') t.,O

".

'

0 o

20

" x".

\..

~0 0

\.

.... I

I

I

2

'\ " ° 3

l

~ 4

5

t2 MINUTE BLOCKS

Fig. 1. Mean crossovers in successive 12-rain blocks during the first hour in the behavioral activity chambers for sham-, B7-, B8-, and B9-1eslonedrats. *P < 0.002; **P < 0.001. The response of half the raphe-lesioned animals (n ---- 5 in each group) was monitored after the injection of 2.5 mg/kg D-amphetamine. With respect to the locomotor hyperactivity induced by amphetamine, B8-1esioned animals were hyperresponsive (F [1,16] = 5.41, P < 0.05). During the 2 h beginning 90 min after injection the mean number of crossovers 4- S.E.M. for each group was as follows: control: B7: B8: B9:

131.4 115.2 274.4 165.8

4444-

50.8 31.7 73.5 37.6

All animals given amphetamine exhibited some of the stereotypy characteristic of this dose range (cf. ref. 39), but the small sample size precluded any analyses of group differences in the intensity of stereotypy.

Startle responding The mean magnitudes of startle responses are presented for each group in Fig. 3 as the first response and the respective means of responses 1-10, 11-20, and 21-30. Although consistent mean differences were observed for the first response, these were not statistically significant (F [1,33] = 2.86, n.s.). In a subsequent replication of this experiment with B7 and B8 lesions, the first startle response in B8-1esioned animals was significantly elevated (Geyer et aL, manuscript in preparation). As is apparent in Fig. 3, the mean of responses 1-10 was significantly greater in animals that had received B8 lesions than in any other group (F [1,33] = 7.2, P < 0.02). By the end of the test session all animals were responding comparably. All groups showed some degree of sensitization, assessed by observing the

262 525

260 240 220 200 160 O3 Q::

uJ o

140

7

I

(/3

120

o 13:: I00

T

8O

T

60 40

"i 0

C B7 B8 B9

I-6

C B7 B8 B9

7-12

13-18

6 HOUR BLOCKS

Fig. 2. Mean crossovers ~L S.E.M. during 6-h blocks of time in the behavioral activity chambers. Block 1-6 represents the last half of the light phase; blocks 7-12 and 13-18 represent the two halves of the dark phase of the light/dark cycle. Only rats with B8 lesions were significantly different than controls. *P < 0.001.

difference between the mean of responses 2-10 and the first response (Fig. 3). No lesion significantly affected this difference (F [3,33] = 0.23, n.s.), nor was it correlated with B8 damage (r = +0.007) or hippocampal tryptophan hydroxylase activity (r = --0.076). This index of sensitization was, however, significantly related to septal tryptophan hydroxylase activity (r = --0.329, P < 0.05) and to hippocampal and thalamic tyrosine hydroxylase activity (r----- +0.576, P < 0.001; r = +0.445, P < 0.01, respectively). The rate of habituation, assessed as the difference between the means of responses 21-30 and 1-10, was significantly increased by B8 lesions (F [1,33] ---- 10.8, P < 0.001). The habituation index was also correlated with t}8 damage (r = +0.474, P < 0.01) and hippocampal tryptophan hydroxylase (r = --0.317, P < 0.05), but not with any other enzymatic variable. These data suggest that low tryptophan hydroxylase in the hippocampus is related to a greater decrement in responding over trials, although further analysis indicated that rate of habituation p e r se is not generally

263 H o--o A--A [] []

44 ¸

\ 4O

\

\

I.....m

36

"'

32

CONTROL(n=/O) B?(n=/O) B8(~=8) B9(n=9)

\

\

z, o o.. f,,o ,,, ,.,.-

28

L,J " I.-ft.."

24

'-.,

F--

13

2O

I I

I I-I0

lit. 20

I 21-50

TRIALS

Fig. 3. Mean startle response magnitude on trial 1, trials 1-10, 11-20, and 21-30 for sham-, B7-, B8-, and B9-1esioned rats. Median raphe lesions significantly increased the magnitude of the startle response for trials 1-10 and produced a proportional, although not significant, increase in the first response. *P < 0.02. related to this enzymatic variable. These two behavioral variables were intended to reflect the presumably separate processes of sensitization and habituation (cf. refs. 13 and 21); they are discussed again with the multivariate correlational analysis below. Each animal was weighed just prior to the startle test, and those weights were compared with the response measures, by means of the correlational analysis, to determine whether the slightly, although not significantly, heavier weights of B8lesioned animals might have contributed to the effects. N o reliable correlations were observed (first response: r = --0.254; responses 1-10: r = --0.082; responses 11-20: r = --0.219; responses 21-30: r = --0.048; overall mean: r = --0.143).

Spontaneous alternation When given two unreinforced trials in a Y-maze, B8-1esioned animals once again responded differently than the other 3 groups did. Only 3 of the 10 controls, 3 of the 10 B7-1esioned animals, and one o f the 9 B9-1esioned animals perseverated. In contrast, 5 of the 8 animals with more than 10 ~o damage to the B8 nucleus perseverated. The response of the B8-1esioned animals was significantly different from that of the remaining animals taken as a group, according to a simple proportionality test (z score = 2.04, P < 0.01). In a replication with a T-maze, only one o f 12 controls and 3 of 10 B7-1esioned rats perseverated. N o t surprisingly, 8 of 11 B8-1esioned animals perseverated (z score = 3.46, P < 0.001).

Multivariate correlational analysis The correlation matrix for the relevant variables from this behavioral study and

264 TABLE 1 CORRELATION MATRIX

TRYPTOPHAN HYDROXYLASE Strl ~

B7 ]eslons

a~'t B8 leslons

Strlatum ~ . Thalamus ~ Z

Thal

LOCOMOTOR ACTIVITY

Hypoth Cortex H1ppo Sept

Ist 12 mln

Day l-6

STARTLE MAGNITUDE

Night I-6

N~ght 7-12

]st resp

Resp I-I0

S e n s Hablt dlff dl f f

- 813

-

549

- 454

- 5]0

+

048

-

-,259

-

+.054

+

094

+ 030

+ 133

+,082

+ (10]

-

-

144

-

- 451

-

759

- 451

+

+.164

+.35Z

+

408

+ 297

÷ 35l

+ 007

+.474

+ 561

+ 640

+ 079

+ 203

+.226

+.162

+.009

-.OJ&

- ]03

- 093

+.022

- r,qI

+ 567

+ 67]

+ 249

+ 397

+ 062

+.138

+.094

- 026

- 017

-.062

-.036

-.oI~

+ 688

+ 488

+ 420

-.oq8

- 033

- 245

- 417

- 191

- 288

-.057

- 217

+ 457

+ 432

-.227

+ 037

- 352

- 416

-.074

- 226

-.118

-.056

+ 615

- 445

- 086

-.251

- 382

-.265

-.406

- 076

-,317

- 288

-.035

- 187

-.225

+ 225

- 133

- 329

- 048

4.065

137

+ 718

580

062

386

023

J

~J

Hypothalamus

~

Cortex

~

Hlppocampus Septal nuclel

tu~l

Strlatum

+ 341

+.242

+ 437

+.364

+,048

+.213

+,132

~I

HI ppocampus

+.I04

+.318

+.041

+.057

- 485

+.129

+.576

-.116

- 021

+.4]0

+.006

+ 148

+.142

+ 103

-.051

+.147

~-'~1 Cortex

Significantcorrelations are underlined and in bold type. With an n of 39, the significancelevels are 0.315, 0.405, and 0.513 for P < 0.05, 0.01, and 0.001, respectively.

from the biochemical and histological report on the same 39 raphe-lesioned rats is presented in Table I. The matrix includes all the significant findings with respect to lesion groups B7 and B8 that have been reported in this and the preceding paper. The basic findings, which have been confirmed with ANOVA, will not be reiterated here. Because multivariate correlational analyses are vulnerable to spuriously significant results, some of the present findings must be interpreted with caution. When possible we have confirmed the correlational results with ANOVA. The correlations not amenable to ANOVA will require replication before definitive conclusions can be drawn. It will be particularly important to test the relationships in untreated rats. With some notable exceptions, the matrix shows a high degree of correlation between tryptophan hydroxylase levels in various forebrain regions. However, as one would predict from the separate origins of the mesolimbic and mesostriatal pathways, neither hippocampal nor septal tryptophan hydroxylase correlates significantly with striatal tryptophan hydroxylase. Furthermore, the tryptophan hydroxylase activities in the forebrain regions were generally not well correlated with their respective tyrosine hydroxylase activities, nor were the tyrosine hydroxylase activities reliably correlated between regions (not shown). The correlations between locomotor activity and startle responding were also generally low (not shown); apparently these measures reflect different and perhaps independently controlled behaviors, as suggested by the work of Horlington2L The first 12-min block of crossovers after the introduction of the animals to the chambers was, as mentioned, significantly correlated with the extent of B8 damage and negatively correlated with hippoeampal tryptophan hydroxylase. This measure of

265 exploratory activity also correlated positively with striatal tyrosine hydroxylase. The 6-h block of crossovers representative of daytime behavioral activity did not correlate reliably with any of the tryptophan hydroxylase activities; it was, however, marginally related to both hippocampal and cortical tyrosine hydroxylase levels. The night-time behavioral activity was of course significantly related to the extent of B8 damage. Particularly during the second 6 h of the dark phase, hippocampal, cortical, and hypothalamic tryptophan hydroxylase activities were all negatively correlated with the level of behavioral activity. Furthermore, striatal tyrosine hydroxylase correlated positively with both 6-h blocks of crossovers during the dark phase. Magnitude of response to the first l0 air puff stimuli was positively correlated with the extent of B8 damage and negatively correlated with hippocampal tryptophan hydroxylase, but not significantly correlated with tyrosine hydroxylase. Hippocampal tyrosine hydroxylase was negatively correlated with the first startle response and better related, positively, to the index of sensitization over the first 10 trials. Furthermore, a stepwise regression analysis revealed that hippocampal tyrosine hydroxylase was correlated with the sensitization slope even when the variance due to the first response was extracted (partial r : +0.396, F [1,37] = 6.72, P ~ 0.025). Our measure of the habituation slope correlated positively with B8 damage and negatively with hippocampal tryptophan hydroxylase, as might be expected from the significantly increased initial responses in B8-1esioned animals. The correlational analysis was repeated after removal of the B8-1esioned rats so we could estimate whether our measures of sensitization and habituation were reliably correlated w~th the regional enzyme levels in the animals exhibiting a normal range of startle responding (n : 31), independent of the manipulation (i.e. B8 lesions) that altered behavior and hippocampal and septal tryptophan hydroxylase. Without the B8 animals the index of habituation was no longer correlated with hippocampal tryptophan hydroxylase (r : --0.076, n.s.). Similarly, when the variance due to the B8 damage was extracted in a stepwise regression analysis, the partial correlation o hippocampal tryptophan hydroxylase with the index of habituation was negligible (r : --0.074). Thus, the rate of habituation is not clearly related to our measure of serotonergic activity in the hippocampus except in B8-1esioned animals. In contrast, the correlations between enzyme levels and the sensitization index were still sigmficant after exclusion of the B8 animals. In the 31 rats exhibiting a normal range of behavior the degree of sensitization correlated positively with hippocampal tyrosine hydroxylase (r----+0.461, P < 0.01) and negatively with septal tryptophan hydroxylase (r = --0.389, P < 0.05). DISCUSSION

The results of our experiments suggest that the mesolimbic serotonergic pathway originating in the median raphe nucleus is primarily responsible for several behavioral alterations that have been reported as consequences of manipulation of whole brain 5-HT. Hyperactivity, hyper-reactivity, and increased locomotor response to amphetamine, which have been observed after combined lesions of B7 and B829,a4

266 i njectaon of the neurotoxins 5,6- or 5,7-dihydroxytryptamme2,4, or systemic treatment with parachlorophenylalanine 4,6,14, were all produced an our ammals by B8 lesions, but not by B7 or B9 lesions. The behavioral changes are apparently related to the relatively specific reduction of serotonerglc innervation in limbic structures (septal nuclei and hippocampus) which results from discrete lesions m the median raphe nucleus. In contrast to the effects of B8 lesions, lesions of the dorsal raphe nucleus selectwely reduced the tryptophan hydroxylase levels in the striatum and thalamus, without affecting limbic tryptophan hydroxylase. Both lesions lowered the enzyme levels in cortex and hypothalamus. Nevertheless, B7 lesions had no demonstrable effect on any behavioral variable assessed in our experiments. Srebro and Lorens 41 recently reported somewhat similar findings with respect to open field activity m rats with B7, B8, or combined lesions, as did Jacobs et al. 27. Our lesions destroyed a substantial proportion of the bilateral serotonergic cell bodies identafied by fluorescence hlstochemlstry as B9. Nevertheless, no reliable alterations in tryptophan hydroxylase actwlties were demonstrable following our B9 lesions. Similarly, B9-1esioned rats were comparable to controls in all the behavioral measures. This group did serve as a useful control for non-specific tissue damage, mole so than the sham-lesioned animals, because the total damage with the bilateral B9 lesaons was roughly twice that caused by either B7 or B8 lesions. We observed the behavioral alterations produced by B8 lesions in diverse experimental situations that have been used for years to assess various expressions of behavioral inhibition. Our findings are consistent with a general relationship between inhibitory functions and brain 5-HT, at least with respect to the median raphe nucleus. Such limbic forebrain structures as the septal nuclei and hippocampus, which are innervated by B8, have also been implicated m response inhibition without reference to the monoammergic neuronal systems 23. An extenswe hterature describes ammals with either septal or hippocampal lesions as hyperactive 19,28, hyper-responslve to stimuli 28,36, and perseveratave 9,1°. Converging evidence suggests that some expressions of behavioral inhibition are mediated by the serotonergic innervation of the hippocampus and perhaps the entire limbic system. For example, hippocampal lesions preclude the locomotor hyperactivity normally induced by parachlorophenylalanine or median raphe lesions in rats 26. Functional relationships have yet to be established for the striatal and thalam~c projections originating m the dorsal raphe nucleus. One of our most striking results was that animals with B8 les~ons perseverated when given two unremforced trials m a Y-maze (or T-maze). This phenomenon has been attributed to an animal's inablhty to inhibit his initial alley preference '~s. Dalland 9,10 has shown that septal-lesioned rats perseverate on the basis of stimulus parameters, while hippocampal-lesioned rats perseverate with respect to responses. Since our B8 lesions reduced serotonergic innervation to both those structures, further experimentation is required to see whether raphe lesions parallel hippocampal or septal deficits. It is generally agreed that 5-HT reduction by raphe lesions 13 or parachloro-

267 phenylalanine administrationT,s increases the magnitude of startle responses in rats. Our data confirm those observations and suggest, again, that impairment of the mesolimbic system originating in B8 was primarily responsible for the hyper-responsivity, since neither B7 nor B9 lesions significantly affected any measure of startle behavior. The mechanism of the 5-HT effect on startle responding is the subject of some controversy. Treatment with parachlorophenylalanine has been reported to attenuate habituation over trialsT,s, while large raphe lesions appear to some to increase sensitization without affecting habituation13. Our results with selective raphe lesions are somewhat inconsistent with either interpretation. In our experiments, B8 lesions increased the magnitude of the startle responses over the first 10 trials without significantly altering either the first response or the degree of sensitization. However, the mean differences between groups on the initial response were proportional to those for the first 10 responses, which suggests that B8lesioned rats may be generally hyper-reactive. As mentioned, in a subsequent replication experiment the increase in the magnitude of the first response was significant after B8 lesion (Geyer et al., manuscript in preparation). Sensitization p e r se appears from our correlational analyses to be better related to the noradrenergic hippocampal innervation than to the serotonergic systems. Independent of the magnitude of the first response, hippocampal tyrosine hydroxylase levels were directly correlated with the sensitization slope. Further studies have shown that lesions of the locus coeruleus, where the noradrenergic projections to the hippocampus originate, reduce sensitization to startle stimuli without affecting the first response (Geyer et al., manuscript in preparation). Lesions of the median raphe also increased our index of the habituatory decrement in responding over trials (Fig. 3), an effect opposite to that reported with parachlorophenylalanine7,s. However, because of the significant augmentation of the first 10 responses produced by B8 lesions, it is difficult to interpret this effect as an increased rate of habituation. Furthermore, if we exclude the B8 animals from the analysis, there is no correlation between hippocampal tryptophan hydroxylase and the habituation slope. Further studies may clarify whether such procedural differences as stimulus modality contributed to the disparate conclusions of these studiesT,S,13. At present we conclude at least that reductions in the serotonergic projections from B8 to limbic forebrain increase startle responsivity. We find little evidence for an increase in sensitization after raphe nuclei lesions and no convincing relationship between forebrain serotonergic activity and rate of habituation. In a study of the effects of intraventricular infusions of 5-HT on responses to air puff stimuli, we observed an overall decrement in response magnitude that was not demonstrably attributable to either sensitization or habituationis. We also reported that infusions of the putative hallucinogen bufotenin (N,N-dimethyl-5-HT) increased the startle response to air puffsis. The evidence relating the actions of hallucinogens to inhibition of central serotonergic neurons is extensive and has been reviewed recently1,3,88. The hyper-responsivity observed after B8, but not B7, lesions may indicate that many of the perceptual effects of hallucinogens reflect inhibition of the serotonergic cells in the median raphe nucleus. If so, selective lesions of the raphe nuclei

268 should preclude some of the behavioral alterations produced by mdoleamine hallucinogens. The serotonergic inhibitory system originating m the me&an raphe nucleus necessarily inter-relates functionally with other central neuronal circuits. Brodie and Shore 5 originally proposed that 5-HT and the catecholamines exert opposing effects on the levels of behavioral arousal, with 5-HT being inhibitory and norepinephrme bemg excitatory. Others have reported evidence supporting this simple dichotomyTM 31,42. By measuring both tyrosine hydroxylase and tryptophan hydroxylase, the ratelimiting enzymes m catecholammergic and serotonerglc biosynthesis, from several brain regions of the same ammals, we tested the notion of functional antagonism between the transmitter systems, assuming that enzyme levels d~rectly or at least consistently reflect functional activity. Behavioral arousal, as reflected by spontaneous locomotor activity, appears to be inhibited normally by the serotonerglc neurons originating in B8. When they were introduced into the activity chambers and during their normally active period in the dark phase of the light cycle, B8-1esioned rats were hyperactive, and their actwlty levels correlated negatively with hippocampal but not striatal tryptophan hydroxylase levels. These behavioral variables were also correlated, positively, with striatal but not hippocampal tyroslne hydroxylase. Because dopamme is the predominant catecholamlne in the stnatum, these data suggest a reciprocal relationship in the influence of stnatal dopamine and h~ppocampal 5-HT on spontaneous locomotor activity. As with the locomotor actiwty measures, startle responses were, during the first 10 trials, markedly augmented by B8 lesions and negatively correlated with hippocampal tryptophan (but not tyrosine) hydroxylase. Conversely, hippocampal tyrosme hydroxylase was posltwely correlated with a different measure of startle behavior - sensitization to the initial stimuli, which is presumably related to arousal processes 11,24. Inhibition of the hippocampus has long been thought to be involved in the so-called 'orienting' response to novel stimuh z°, and recently the occurrence of theta rhythms m the hippocampus has been temporally related to both orienting responses and the firing patterns of the raphe nuclei and the locus coeruleus, which provJdes the noradrenergic innervatton to the hippocampus 32,4°,44. Furthermore, the tyrosme hydroxylase levels, especially in the smatum, were positively related to gross locomotor activity (see Table I). We have reported that increases in the activity of catecholaminergic systems increase the behavioral arousal reflected in locomotor actiwty15,16,as. If we assume that high levels of tyrosine hydroxylase reflect relatively high functional activity of synaptic catecholamines, the positive correlations we have observed are consistent with the generally accepted role of brain catecholammes m arousal a7. The results of our multivariate correlational analysis are thus compatible with Brodie and Shore's 5 early hypothesis that behavioral arousal is a joint function of serotonergic inhibition and catecholaminergic excitation. Moreover, some provocative distinctions as to which expressions of arousal are most influenced by either the dopamlnerglc striatal system or the noradrenergic hippocampal circuit are apparent in our correlational data. Because of the possibility for erroneous correlations inherent

269 in such analyses, these relationships require confirmation, especially in u n t r e a t e d rats, before any definitive conclusions m a y be drawn. The intrinsic p o w e r o f m u l t i v a r i a t e studies o f individual a n i m a l s s h o u l d p r o v i d e some u n d e r s t a n d i n g o f the c o m p l e x interr e l a t i o n s h i p s a m o n g the various n e u r o t r a n s m i t t e r systems a n d their c o m b i n e d influences on b e h a v i o r . ACKNOWLEDGEMENTS

This w o r k is s u p p o r t e d b y U n i t e d States Public H e a l t h Service G r a n t D A 00265-03; D a v i d Segal is the recipient o f U S P H S R e s e a r c h Scientist A w a r d N o . MH-70183-02. W e t h a n k Skip M c C u n n e y a n d J e r r y C a h i l l for their care in r u n n i n g the beh a v i o r a l e x p e r i m e n t s a n d B a r b a r a B l o m g r e n for helping p r e p a r e the m a n u s c r i p t .

REFERENCES 1 AGHAJANIAN,G. K., ANU HAIGLER,H. J., Studies on the physiological activity of 5-HT neurons. In G. ACSESON(Ed.), Pharmacology and the Future of Man, Vol. 4, Karger, Basel, 1973, pp. 269-285. 2 BAUMGARTEN,H. G., ANDLACHENMAYER,L., 5,7-Dihydroxytryptamine: improvement in chemical lesionmg of indolamine neurons in the mammalian brain, Z. Zellforsch., 135 (1972) 399-414. 3 BRAWLEY, P., AND DUFFIELD, J. C., The pharmacology of hallucinogens, PharmacoL Rev., 24 (1972) 31-66. 4 BREESE,G. R., COOPER,B. R., AND MUELLER,R. A., Evidence for involvement of 5-hydroxytryptamine in the actions of amphetamine, Brit. J. Pharmacol., 52 (1974) 307-314. 5 BROUIE,B. B., ANDSNORE,P. A., A Concept for a role of serotonin and norepinephrine as chemical mediators in the brain, Ann. N.Y. Acad. Sci., 66 (1957) 631-642. 6 BRODY, J. F., JR., Behavioral effects of serotonin depletion and of p-chlorophenylalanlne (a serotonin depictor) in rats, Psychopharmacologia (Serl.), 17 (1970) 14-33. 7 CARLTON, P. L., AND ADVOKAT, C., Attenuated habituation due to parachlorophenylalanine, Pharmacol. Biochem. Behav., 1 (1973) 657-663. 8 CONNOR, R. L., STOLK, J. M., BARcrL~s, J. D., AND LEVINE, S., Parachlorophenylalanine and habituation to repetitive auditory startle stimuh in rats, Physiol. Behav.. 5 (1970) 1215-1219. 9 DALLAND,T., Response and stimulus perseveration in rats with septal and dorsal hippocampal lesions, J. comp.physiol. Psychol., 71 (1970) 114-118. 10 DALLAND,T., Stimulus perseveration of rats with septal lesions, Physiol. Behav., 12 (1974) 10571061. 11 DAVIS,M., Signal-to-noise ratio as a predictor of startle amplitude and habituation in the rat, J. comp. physiol. Psychol., 86 (1974) 812-825. 12 DAVIS,M., AND BEAR,H. D., Effects of N,N-dimethyltryptamine on retention of startle response habituation in the rat, Psychopharmacologia (Bed.), 27 (1972) 29--44. 13 DAvis, M., ANDSHE~,D, M. H., Habituation and sensitization of the rat startle response: effects of raphe lesions, Physiol. Behav., 12 (1974) 425-431. 14 FIBIGER, H. C., AND CAMPBELL,B. A., The effect of parachlorophenylalanine on spontaneous locomotor activity in the rat, Neuropharmacology, 10 (1971) 25-32. 15 GEYER, M. A., AND SEGAL,D. S., Differential effects of reserpine and alpha-methyl-p-tyrosineon norepinephrine and dopamine induced behavioral activity, Psychopharmacologia (Bed.), 29 (1973) 131-140. 16 GEYER, M. A., SEGAL,D. S., AND MANDELL,A. J., Effect of intraventricular infusion of dopamine and norepmephrine on motor activity, Physiol. Behav., 8 (1972) 653-658. 17 GEYER, M. A., PUERTO,A., DAWSEY, W. J., KNAPP, S., BULLARD,W. P., AND MANDELL, A. J., Histologic and enzymatic studies of the mesolimbic and mesostriatal serotonetgic pathways, Brain Research, 106 (1976) 241-256.

270 18 GEYER, M. A., WARBRITTON, J. D , MENKES, D. B., ZOOK, J. A , AND MANDELL,A. J., Opposite effects of mtraventrlcular serotonin and bufotenm on rat startle responses, Pharmacol Biochem Behav., 3 (1975) 687-691 19 GOMER, F. E., AND GOLDSTEIN, R., Attentional rigidity during exploratory and simultaneous discrimination behavior in septal lesioned rats, Physiol. Behav., 12 (1974) 19-28. 20 GRASTY.g,N,E , LISS~,K, K., MADAR.~SZ,I., AND DONHOFFER,H., Hlppocampal electrical actwity during development of condJtloned reflexes, Electroenceph chn Neurophystol, l 1 (1959) 409-430. 21 GROVES, P. M , AND THOMPSON, R F , Habituation: a dual-process theory, Psycho/. Rev, 77 (1970) 419-450. 22 HORLINGTON, M., Startle response circadian rhythm in rats: lack of correlation with motor activity, Phystol. Behav., 5 (1970) 49-53. 23 ISAACSON,R. L. The Limbtc System, Plenum Press, New York, 1974, pp. 1-292 24 ISON, J. R , AND HAMMOND,G. R., Modification of the startle reflex m the rat by changes m the auditory and visual environments, J. comp. physiol. Psychol., 75 0971) 435-452 25 JACOBS,~B L , Evidence for the functional interaction of two central neurotransmltters, Psychopharmacologla (Bet/.), 39 (1974) 81-86. 26 JAcoas, B L., TRIMBACH, C., EUBANKS, E. E., AND TRULSON, M., Hlppocampal mediation of raphe lesion and PCPA-induced hyperactivity in the rat, Brain Research, 94 (1975) 253-261. 27 JACOBS, B. L., WISE, W D , AND TAYLOR, K. M , Differential behavioral and neurocbemical effects following lesions of the dorsal or median raphe nuclei m rats, Brain Research, 79 (1974) 353-361. 28 KIMBLE, D. P., The effects of bdateral hippocampal lesions in rats, J. comp physiol. Psychol., 56 (1963) 273-283 29 KOSTOWSKI,W., GIACALONE,E., GARATT1NI,S., AND VALZELLI,L., Studies on behavioural and biocbemlcal changes in rats after lesions of the midbrain raphe, Europ. J. Pharmacol., 4 (1968) 371-376. 30 LORENS, S. A., SORENSEN,J. P., AND YUNGER, L. M , Behavioral and neurochemical effects of lesions in the rapbe system of the rat, J. comp. physiol. Psychol., 77 (1971) 48-52. 31 MABRY,P. D., AND CAMPBELL,B. A., Ontogeny of serotonergic inhibition of behavioral arousal in the rat, J. comp physiol. Psychol., 86 (1974) 193-201. 32 MACADAR,A. W., CHALUPA,L. M., AND LINDSLEY,D. B., Differentiation of brain stem loci which affect hippocampal and neocorUcal electrical activity, Exp. NeuroL, 43 (1974) 499-514. 33 MANDELL,A. J., AND GEYER, M. A., Hallucinations: chemical and physiological. In R. GRENELL AND S. GABAY(Eds.), Biological Foundations of Psychiatry, Raven Press, New York, 1975, in press 34 NEILL, D B , GRANT, L. D., AND GROSSMAN,S. P., Selective potentiation of locomotor effects of amphetamine by midbram rapbe lesions, PhysioL Behav., 9 (1972) 655-657. 35 PERSIP, G. L., AND HAMILTON,L. W , Behavioral effects of serotonin or a blocking agent applied to the septum of the rat, PharmacoL Biochem. Behav., 1 (1973) 139-147. 36 SCHWARTZBAUM,J. S , GREEN, R. H., BEATTY,W. W , AND THOMPSON, J. B., Acqmsition of avoidance behavior following septal lesions m the rat, J. comp. physiol. Psychol., 63 (1967) 95-104. 37 SCHILDKRAUT, J J., AND KETY, S. S, Biogenic amines and emotion, Science, 156 (1967) 21-30. 38 SEGAL,D. S , AND MANDELL,A. J., Behavioral activation of rats during intraventrlcular infusion of noreplnephrlne, Proc. nat Acad. Sci. (Wash.), 66 (1970) 289-293 39 SEGAL, O. S., AND MANDELL, A. J , Long-term administration of D-amphetamine: progressive augmentation of motor activity and stereotypy, Pharmacol. Biochem. Behav., 2 (1974) 249-255. 40 SHEU, Y -S., NELSON, J. P., AND BLOOM, F E., Discharge patterns of cat raphe neurons during sleep and waking, Bram Research, 73 (1974) 263-276. 41 SREBRO,B., AND LORENS, S. A., Behavioral effects of selective midbram raphe lesions in the rat, Brain Research, 89 0975) 303-325. 42 STEIN, L., AND WISE, O. C , Serotonin and behavioral inhibition. In E. COSTA, G. L. GESSAAND M. SANDLER(Eds.), Serotonin - - New Vistas, VoL 1l,/tdvanc. Biochem. Pharmacol., Raven Press, New York, 1974, pp. 281-292. 43 SWONGER,A. K., AND RECH, R. H., Serotonergic and chotinergic involvement in habituation of activity and spontaneous alternation of rats in a Y maze, J. comp. physiol. Psychol, 81 (1972) 509522. 44 UNGERSTEDT, U., Stereotaxlc mapping of the monoamme pathways in the Iat brain, Acta physiol scand., Suppl. 367 (1971) 1-48