Nucleus basalis neurons exhibit axonal branching with decreased impulse conduction velocity in rat cerebrocortex

Brain Research, 325 (1985) 271-285 Elsevier

271

BRE 10478

Nucleus Basalis Neurons Exhibit Axonal Branching with Decreased Impulse Conduction Velocity in Rat Cerebrocortex GARY ASTON-JONES, ROBERT SHAVER and TIMOTHY G. DINAN*

Centerfor Neurobehavioral Sciences, State Universtty of New York at B,nghamton, Bmgharaton, NY 13901 (U.S.A.) (Accepted April 26th, 1984)

Key words: nucleus basahs-- basal forebrain-- acetylcholine-- cholinergic-- unit recordings - antidromlc activation - - cortical afferents

Smgle neurons m the basal forebrain (nucleus basalis area) were antidromicaUy activated from the frontal or parietal cortex in anesthetized rats. Wide ranges of antidromic latencies were observed overall, with frontal and panetal stimulation yielding values ranging from 1.0 to 26.0 ms and 1.6-24.0 ms, respectively. Individual neurons often exhibited multiple antldromic latencies, such that deeper sites of stimulation or greater stimulation amplitudes generally yielded discretely different, shorter latencies than more superficial sites or lower amplitudes of stimulation. Single neurons were also often driven from neighboring sites (1-2 mm apart) within the frontal cortex, but no cell was coactivated from both frontal and parietal cortices. Finally, patterns and rates of spontaneous actwity varied markedly among these cortically projecting neurons, with some cells being non-spontaneous and others exhibiting tonic rates of 30-40 Hz. Impulse waveforms also differed among driven cells, from relatively low-amplitude, negative spikes to large-amplitude, entirely positive spikes in unfiltered signals. These results indicate that cortically projecting, putatively cholinergic neurons m the basal forebrain form a physiolog~caUyheterogeneous population in terms of impulse conduction velocity, spontaneous discharge, and spike waveforms. Our finding of multiple antldromic latencies and driving from neighboring sites indicate that these fibers may be highly branched m local terminal fields, but that individual cells may project exclusively to a single cortical area. Faster conduction velocities for deep compared to superficial cortical stimulation sites imply that these fibers may become non-myelinated upon entering cortical terminal fields, or that they may become markedly thinner as they travel within the cortex. This system of cholinergic cortical afferents differs in many physiologic aspects from the other non-thalamic cortical input systems of catecholamine or indoleamme neurons.

INTRODUCTION While most inputs to the c e r e b r a l cortex originate in the thalamus, a few very diffusely projecting systems of cells providing n o n - t h a l a m i c cortical after-. ents have been described over the past few decades. Noradrenergic, serotoninergic and d o p a m i n e r g i c innervation of c e r e b r a l cortex, from s o m a t a in midbrain and pontine nuclei, have b e e n the subject of intense research and debate17,22. 36. Their extensive terminal fields imply that these systems m a y play a global role in brain function, p e r h a p s serving to m o d u l a t e activity in other systems with m o r e limited a n a t o m i c targets3.8,16.17,48,49,57.

M o r e recently, a n a t o m i c and biochemical s t u d i e s have defined a fourth system of non-thalamic cortical afterents (in addition to c a t e c h o l a m i n e and indoleamine tracts), originating from basal forebrain cholinergic neurons. Schute and Lewis 47 first described acetylcholinesterase ( A C h E ) - c o n t a i n i n g cells in the basal forebrain with direct p r o j e c t i o n s to the cerebral cortex and o t h e r telencephalic areas. T h e y p r o p o s e d that these cells were part of a diffusely projecting cholinergic system in the brain. Their results have been substantiated with n e w e r anatomical techniques able to m o r e definitively establish source cells for fiber terminals and the cholinergic nature of specific neurons. Thus, Kievit and K u y p e r s 26, using ret-

* Present address: Departments of Psychiatry and Pharmacology, St. George's Medical School, University of London, Tooting, London, UK Correspondence" G Aston-Jones Present address' Department of Biology, New York Umversity, Room 1009 Main Building, Washmgton Square, New York, NY 10003, U.S.A. 0006-8993/85/$03 30 (~) 1985 Elsevier Science Publishers B.V.

272 rograde transport techniques, demonstrated direct cortical afferents from basal forebram neurons in the monkey. Experiments utilizing excltotoxin or electrolytic lesion techmques have demonstrated that the majority of markers for acetylcholine (ACh) in the cortex originate from cells in the basal forebrain 23,24,28,54. Studies utilizing retrograde transport combined with AChE hlstochemistry have demonstrated direct cortical projections from AChE-positive somata located in the basal forebrain of the rat 7,28 and the monkey 32-34. More recently, similar results have been obtained for cortically projecting neurons that stain immunocytochemically for choline acetyltransferase (ChAT)34,4a,58. In the primate, cortical cholinergic projections originate from the nucleus basalis of Meynert (nBM) 25,26,33,34.In the rat, the cortical cholinergic innervation originates from cells located primarily in the area of the ventral globus pallidus, nucleus preopticus magnocellularis and the lateral hypothalamus7,23,24, 28.3L54, collectively denoted nucleus basalis (NB) 7,3s. Additional interest in these neurons stems from recent evidence linking changes in this cortically projecting cholinergic system to behavioral and CNS anatomic anomalies associated with aging 4z. In particular, patients with presenile dementia of the Alzheimer type exhibit a pronounced loss of cholinergic markers in cerebral cortex and a significant reduction in numbers of nBM cells 12,5556. Some investigators have proposed that degeneration of the cholinergic cells in the nBM is specifically associated with this presenile dementia 12.55. In view of the anatomical and clinical significance of these neurons, relatively few studies of their physiology have appeared. DeLong and colleagues have recorded single unit activity in the nBM of waking monkeys. These cells exhibited rapid discharge rates that varied in association w~th food reward 14,35. Similar results were obtained in studies of the substantia inominatalO, 43 A recent study in the cat 52 found ventral basal forebrain cells that exhibited highest discharge rates during non-rapid-eye-movement sleep. However, not all neurons in the substantia innominata-nBM area appear to project to the cerebrocortex or to be chohnergic25.32-34,38. We have recently reported certain physiologic attributes of NB neurons antidromically activated from the frontal cortex6 The present study sought to determine spontaneous

impulse activity, fiber conduction velooty, refractoriness and other physiological charactenstLcs of NB neurons that project directly to the frontal or parietal cortex as identified by antidrom~c activation Recent anatomic results indicating that all or nearly all cortically projecting NB cells in the rat contain markers for ACh 7,28,44.58 allow us to tentatively conclude that driven cells reported here are chollnerg~c. Some of the present results have appeared in short communicationsS,6. MATERIALS AND METHODS Forty-six adult (2-3 months old) male albino rats were used in the present study. Animals were anesthetized with 400 mg/kg chloral hydrate intraperitoneally, and additional booster injections were given as required during experimental sessions. The skull was exposed and holes drilled to allow placement of stimulation electrodes in frontal or parietal cortices, as well as a recording electrode aimed at the NB area. Stimulation electrodes consisted of arrays of two or four 250-/~m-diameter, stainless steel, insulated wires in the frontal cortex, about 1 mm apart, uninsulated for 0.25-0.50 mm from the bluntly cut tips (9 animals), or twisted pairs of similar wire (fully insulated except for the bluntly cut tips) in frontal (30 animals) or parietal cortex (23 animals) Stereotaxic coordinates for frontal cortex stimulation sites were 2-5 mm rostral to the bregma and 2 mm lateral to midline (1-4 mm lateral for electrode arrays). Coordinates for parietal cortex placements were 0.5 mm rostral to the bregma and 5 mm lateral to midlme. In initial experiments, arrays of stimulation electrodes were glued in place with dental acrylic (0.5 -1.0 mm below the cortical surface). In other experiments stimulation electrodes were placed 0.5 mm below the cortical surface (under microscopic control) and then raised or lowered in 0.2-0 3-mm steps with a microdrive during the course of testing individual driven cells. The exposed cortical surface was periodically moistened and at least I min intervened after such movements before further testing in order to allow recovery from tissue compression. Bipolar stimuli were presented with a square pulse stimulator (Grass $44) through a stimulus isolation unit (Grass SIU5). Pulse duration was 0.2-0.5 ms and stimuli ranged m amplitude from 0.05 to 4.0 mA. Testing for

273

Fig. 1 Photomicrograph of a coronal section through the globus pallidus from an experimental rat brain (Neutral Red stain). Pontamine Sky Blue (spot m ventral globus palhdus, at arrow) was iontophoresed at the location of a cell driven antldromically from the panetal cortex. For cahbration, arrow length = 420/~m

high-frequency activation and collision of driven spikes with spontaneous impulses was conducted with stimulation amplitudes of 1.5-2.0 times threshold for driving. Recordings were obtained from glass micropipettes with a tip diameter of 2 - 3 p m , filled with a 2% solution of Pontamine Sky Blue (PSB) in 0.5 M sodium acetate (approximately 10 MQ impedance). Stereotaxic coordinates for recording penetrations were 0.5-1.5 mm caudal to the bregma, 2.5-3.5 mm lateral to midline, and 4.0-8.0 mm ventral to skull surface. Micropipette signals were amplified and filtered (Grass P16) to display both unfiltered and filtered (500 Hz to 10 kHz bandpass) traces. Unit impulses were digitized with a waveform discriminator and fed into a computer for on-line generation of histograms, as well as to a tape recorder for later analysis. Microelectrode placement was marked at the end of successful penetrations by iontophoresing PSB

from pipette tips (6/~A for 5 min, negative polarity). Microelectrode penetrations were separated by at least 0.5 mm to permit unambiguous histological reconstruction, and usually no more than two successful penetrations were made in each animal. At the end of recording sessions, animals were deeply anesthetized and then perfused transcardially with 10% formalin in phosphate buffer. Brains were postfixed with a 25% sucrose in 10% formahn solution for at least 1-2 days. Frozen 50 pm-thick sections were taken through stimulation and recording sites and subsequently stained with Neutral Red Electrode penetrations were reconstructed by correlating depths of recorded cells noted during the experimental session with distances from PSB spots under microscopic examination using a measuring eyepiece reticle. All cells reported here were localized m this way (see Fig. 1).

274

Fig. 2. Oscilloscope photographs dlustrating coihslon test and high-frequency act~vatlon for an NB cell driven from the frontal cortex (left side) and another driven from the parietal cortex (right side). Left side" Upper panel - - Stimulation of frontal cortex (arrow) 8.9 ms after spontaneous spikes (at far left of traces) elicits driven spikes (star) at 8 5 ms latency Middle panel - - Driven spikes are occluded for simdar stimuh dehvered 8.4 ms after spontaneous impulses, indicating collision between spontaneous and driven spikes. Lower panel - - Driven spikes (stars) elicited by each of paired stimuli (arrows, 3 ms interpulse intervals), indicating frequency-following for thiscell at 333 Hz Rightstde. Upper panel - - Stimulation of parietal cortex (arrow) 7.0 ms following spontaneous lmpulses (at far left of traces) drives spikes (star) at 6.5 ms latency. Middle panel - - Driven spikes are occluded for similar sumuh (arrow) presented 6.2 ms after spontaneous spikes, indicating collismn between spontaneous and driven impulses Lower panel - - Driven sp~kes (stars) elicited by each of paired stimuli (arrows, 2 5 ms mterpulse interval), indicating frequency-following for this cell at 400 Hz Ten superimposed sweeps m each panel Upper trace m each panel is unfiltered mlcroplpette signal; lower traces are the same correspondmg signals filtered (500 Hz to 10 kHz bandpass) Hortzontal cahbratton 2 ms Vertzcal cahbrat~on Right side, all traces = 0 1 mV. left side, unfiltered traces = 0 5 mV, filtered traces = 0 1 mV For these and all subsequent oscdloscope photographs, posmvlty i~ upwards

275 sistently drive two spikes at constant latencies with double pulse stimulation at a frequency of 200 Hz or greater, and occlusion of driven spikes by spontaneous impulses used to trigger stimuli within a critical period approximately equal to the latency for driving plus the absolute refractory period (collision testiS). For those cells that were either non-spontaneous or insufficiently isolated from activity in neighboring cells, antidromicity was established using only the

RESULTS

A total of 1159 cells was studied in the area of NB in this and a previous report6. Of 858 cells examined with frontal cortex stimulation, 66 were antidromically activated (7.7%), while 13 of 357 cells (or 3.6%) were antidromically driven from the parietal cortex. Criteria for antidromicity included constant latency driving at threshold for activation, the ability to con-

(

~, :'-.

(

S

"...~

',

,,

"~

"

"

: rtl

Fig 3. Locations of antidromically driven cells plotted on coronal hemisections from the rat brain atlas of Konlg and Klippel Consecutive sections in rostral to caudal order (left to right, top to bottom), from A6790 to A5660/zm from interaural line. Filled circles denote cells driven from the frontal cortex, and crosses indicate locations of cells driven from the parietal cortex. Note that most dnven cells were found in the globus pallidus (GP), mechal and ventral to the caudate nucleus (lateral border of GP delimited by dashed line near center of sections), lateral to the internal capsule (medial border of GP illustrated by vertical scalloped line), and dorsal to the lateral hypothalamus rostrally, and amygdala caudally. Note also that cells dnven from the parietal cortex are located more caudally overall than those driven from the frontal cortex. Besides GP, driven cells were also localized in the entopeduncular nucleus (within internal capsule, medial to GP) and in the magnocellular preoptic area (ventral to GP), as indicated on these plots. These sites correspond to locations of intensely AChE-staining or ChAT-positive somata that are reported to project directly to the cerebrocortex in previous anatomic studies; these cells are referred to as nucleus basalis (NB) in the rat (see text).

276 I

_z

'~

I

14a L

:

I II 11 Jill I Ij

3"

r

r

,,

I

+

II

~

I |

TI

II

I

tl

TTTT]Ir

!11/ /ill/!l

140Z

~ 100-

a

60" u

30 0 Hz

20" !

.50

I.O0 TIME (sec)

1.50

13

2.00

2's TIME (see)

35

70 z m ,, 5 0 +

z m ,,

lZ

8u

o o

.25

.50 TIME ( x ¢ )

I 0 . 5 Hz

IZ

4.6Hz

.75

I O0

.13

.25 TIME (se©)

.311

.50

Fig. 4 Illustrations of spontaneous acuvlty exhibited by NB cells driven from the frontal cortex (upper panels and lower left panel) or from the parietal cortex (lower right panel) Each panel consists of an mterspike-interval histogram (ISH) accumulated over about 1000 spikes, as well as a sample 5 s oscilloscope record of the corresponding cell's impulse activity (inset above each ISH; 500 Hz to 10 kHz bandpass). Discharge rates (in Hz) are given for each neuron in corresponding panels. A wide variety of discharge rates and patterns were obtained from rat NB as illustrated here, from slow, bursty cells (upper left panel; stars indicate spike doublets generating counts m short bins of ISH) to fast, regularly active neurons (upper right panel). In addition, many non-spontaneously active NB neurons were also found driven from the cortex.

first two criteria. Examples of collision testing and high-frequency activation of neurons in the present study are presented in Fig. 2. As illustrated in Fig. 3, antidromically driven cells were localizod to the globus pallidus, internal capsule (entopeduncular nucleus) or preoptic magnocellular area, corresponding to previous anatomic maps of cells retrogradely labeled from the cortex that also stain for A C h E or C h A T 7,28.44,58. Neurons driven antidromically from the parietal cortex were located more caudally overall than those driven from the frontal cortex, although there was considerable overlap. Seven of 13 cells driven from parietal sites were localized to the caudal one-third of the giobus paflidus area, while only 17 of 66 cells driven from the frontal cortex were similarly caudal.

Impulse waveformsand spontaneous activay All data reported here are derived from impulses having characteristics of recordings from somata rather than fibers (e.g. spike durations ~> 1 ms). A variety of different spike waveforms were exhibited by these cortically projecting neurons (see Figs. 2, 4, 5, 8 and 9). Many recordings yielded relatively small (75-100/zV) impulses, entirely negative in the unfiltered signal. While this could be a consequence of electrode position relative to a cell, most of these neurons were recorded for substantial periods ot time (over 30 min), and such spikes did not usually increase in amplitude with m o v e m e n t of the electrode. Other cells were found to exhibit large (0.5-2.0 m V ) impulses, either entirely positive or positive followed by a smaller negative component in unfiltered sig-

277 nals. A few of these larger spikes resembled impulses elicited by putative dopaminergic neurons in the substantia nigra zona compacta9,19. 20, with large positive-negative waveforms m unfiltered records (triphasic, positive-negative-positive, in filtered records) about 2 ms in duration. Three of the 5 cells exhibiting such waveshapes were localized to the nucleus preopticus magnocellularis. This nucleus also contained the largest percentage of large, initially positive spike waveforms (5 of 9 cells). Spontaneous discharge rates varied from nonspontaneously active to tonic rates of 30-40 Hz, with most cells yielding rates between 5 and 20 Hz. Patterns of spontaneous discharge also varied markedly among driven cells, as illustrated for some representative neurons in Fig. 4. Of the 12 cells that were non-spontaneous or fired at rates less than 1.0 Hz overall, 8 were locahzed to the ventromedml aspect of globus pallidus or the entopeduncular nucleus just medial to the ventral globus pallidus, 3 were in the ventral lateral globus pallidus and 1 was in the dorsolateral globus pallidus. Thus, it appears that a high percentage of cortically projecting cells in the area of the ventromedial GP (8 of 27) are not (or virtually not) spontaneously active. For 18 NB neurons driven antidromically from the frontal cortex, we also qualitatively examined re-; sponsiveness to noxious stimuli (pressure applied to the tail, a rear paw or wound edges). Five of these cells mcreased discharge in response to such stimuli; responsiveness in some cells apparently increased with decreasing levels of anesthesia. We also noted during the course of these experiments that spontaneous activity often increased as the depth of anesthesia decreased, and decreased again shortly following a booster injection of chloral hydrate. Overall, we found no characteristics of either impulse waveforms, sensory responsiveness or spontaneous activity that allowed us to predict the likehhood of any cell being antidromically driven from the cortex. These properties varied markedly among both driven and neighboring non-driven neurons.

Axonal conductton velocines and refractoriness In rats implanted with stimulation electrodes in the frontal cortex at fixed depths of 0.5-1.0 mm, a wade range of conduction latencies was obtained. In this group of 13 animals, antidromic latencies ranged

from 1.0 to 13.0 ms 6. At certain stimulation amplitudes, some cells exhibited two or more discrete latencies, 'jumping' back and forth between different values with successive stimuli (see Fig. 5). This phenomenon, considered in light of the wide range of latencies obtained overall, led us in the present study to investigate antidromic latencies of individual neurons as a function of stimulation electrode depth in the cortex. We found that most cells exhibited more than one antidromic latency, varying with the dorsoventral position of the cortical electrode as well as with stimulation amplitude. Driven spikes at each discrete latency for a single cell were shown to be antidromic by collision, either by yielding appropriate collision intervals (with corresponding stimulation amplitudes to produce one or the other latency consistently, illustrated in Fig. 5), or by collision of all driven activity when stimuli were triggered by spontaneous spikes at delays slightly less than the shortest driven latency. Threshold for activation at a given latency generally decreased as the stimulation electrode was moved ventrally through the cortex, reaching a minimum at a particular cortical depth. Threshold would then either begin increasing with more ventral cortical electrode positions, or gave rise to a discretely shorter latency whose threshold for activation was lower than the previous longer latency, thereby occluding it. The shorter latency also typically decreased in threshold with more ventral stimulation loci, obtaining a minimum (usually at about 1.7-2.0 mm below cortical surface), at which point more ventral stimulation sites yielded gradually increased thresholds for this latency as well. In nearly every case examined, longer latencies for a given cell were the lowest threshold events obtained with shallow cortical stimulation sites, while deeper cortical stimulation placements preferentially yielded shorter latencies. Increasing stimulation intensity in shallow cortical sites also generally resulted in discretely shorter latencies Thus, for typical antidromically driven cells in NB, threshold stimulation within the superficial 0.5 mm of the cortex would yield driven activity at one constant latency, while movement of the stimulation electrode deeper in cortex, or raising the stimulation amplitude, would result in a discretely shorter antidromic latency of up to 11 ms (generally 2-7 ms) Usually, a stimulation intensity could be found at one stimula-

278

Fig. 5. Oscilloscope photographs of an NB neuron driven from the frontal cortex with two discrete antich'omic |atencies. Upper pa~l: Oscilloscope sweeps triggered by sequential stimuh (91 V, at arrow) to frontal cortex. Note two discrete latencies, at 2.8 (dosed star) and 7 9 ms (open star). At higher stimulation amplitudes this cell was driven consistently at 2.8 ms, while lower stimulus intensities yielded the longer latency of 7.9 ms. Lower leftpanel: Upper trace - - Stimulation (130 V, at arrow) 3.4 ms following spontaneous impulses (at far left of trace) yielded driven spikes at a constant 2.8 ms latency (closed star) Lower trace - - Driven spikes are occluded for similar stimuli delivered 3 0 ms following spontaneous impulses, indicating collision between spontaneous and driven activity at this latency. Lower right panel: Upper trace - - Stimulation (81 V, at arrow) 8 8 ms after spontaneous i ~ driven spikes at a constant 7.9 ms latency'(ol~n star). Lower trace - - Driven spikes are occluded for similar stimuli presented 8.2 ms after spontaneous spikes, indicating collision between spontaneous and driven activity at this latency also. Traces in lower panels are 10 superimposed sweeps each. Vertwal calibratton: All traces = 1 inV. Horizontal calibration" Upper and lower left panels = 1 ms; lower right panel = 2 ms. Unfiltered recordings

279

°i

~.1 > 2 I...

cn .D

2

4

6

8

10

12

14

2

4

6

8

10

12

14

26

0

,f

"~ 8-1 Z

6 . 4 2

Latency (msec) Fig. 6. Frequency histograms of antidromlc latencies observed for NB neurons with lowest thresholds for activation in superficial (0 2-1.2 mm deep, upper histogram) or deep (1.3-2.3 mm deep, lower histogram) frontal cortex. These two latency distributions differ s]gnificantly (t-test, P < 0.01), indicating that I-roger latencles are most easily obtained from the superficial cortex, while shorter latencies have lowest thresholds in deep cortical s~tes.

tion site that resulted in two'discrete latencies, alternating (or 'jumping') in an all-or-none fashion with different trials of sequential stimuli (as shown for one example in Fig. 5). Many cells exhibited more than two discrete latencies with various cortical stimulation locations or intensities. For 22 cells in the present study, antidromic latencies with lowest thresholds for activation were determined at various depths in the frontal cortex with stimulation electrode placements 0.2-2.3 mm from the cortical surface. Latencies were considered discretely different (i.e. not due simply to more proximal stimulation along the same fiber branch) if either 'latency jumping' was observed, or a movement of 0.2-0.3 mm yielded a 2 ms or greater change in latency. Fig. 6 is a plot of latencies obtained whose thresholds were least within the superficial most 1.2 mm of frontal cortex, as well as for those with thresholds lowest in deeper (1.3-2.3 mm) stimulation sites.

Analysis by t-test revealed that these latency distributions are significantly different (P < 0.01), indicating that latencies for spikes to reach the superficial cortex (mean + S.E.M. = 7.9 + 0.9 ms) were longer than those for fiber branches in the deep frontal cortex (4.4 + 0.6 ms). Together with cells antidromically driven with fixed electrodes 6, latencies for antidromic activation overall from the frontal cortex ranged from 1.0 to 26.0 ms, with a mean value of 5.4 + 0.4 ms. In addition, the antidromic latency for one cell tested with 10-Hz trains in the frontal cortex increased from 8.5 ms to 9.6 ms, resembling fluctuations in conduction velocity previously reported for noradrenergic locus coeruleus neurons 4. Similar results were obtained for 10 cells antidromically driven from the parietal cortex, although latencies were longer overall than with frontal cortex stimulation. Fig. 7 is a latency histogram for parietal cortex stimulation sites. As seen in this figure, these latencies ranged from 1.6 to 24.0 ms and the overall mean was 8.0 + 1.2 ms. When latencies were analyzed for stimulation depth yielding lowest threshold, these latencies formed two significantly different populations (P < 0.01, t-test) such that fibers with lowest thresholds in the superficial cortex (0.2-1.0 mm) yielded longer latencies than those most easily driven from deeper cortical sites (1.1-2.0 mm). Six of the cells subjected to this depth analysis were found to exhibit discrete, all-or-none 'jumps' in antidromic latency similar to those seen with frontal cortex stimulation (see Fig. 8). For cells exhibiting more than one latency, shorter latencies always had lowest thresholds at deeper stimulation sites than longer latencies. Refractory periods were measured for 57 cells

2

~

4

6

8

10

12

14

16

18

20

22

24

26

Superfical L a t e n c y (rnsec) ueep No Depth A n a l y s i s

Fig. 7 Frequency histogram of antldromlc latencies observed for the parietal cortex sUmulation from various depths, as indicated. Note that longer latencies have lowest thresholds in superficial cortex (< 1.0 ram), while shorter iatencies are more readily obtained from deep cortical stimulation sites (> 1 0 mm)

280 some 1-2 mm apart in the medlolateral dimens~on" Sample records for one such cell are presented tn Fig. 9. Antidromic activation of lndwldual neurons from different mediolateral locations In the frontal cortex usually yielded latencles that differed between sites by up to 1.5 ms. In contrast, we were unable to demonstrate antidromic activation of any NB neuron from both the frontal and the parietal cortex m 15 rats tested. DISCUSSION

Ng 8 Oscilloscope photograph showing two discrete antidromlc latencles for an NB neuron driven from the parietal cortex (consecutive sweeps from top to bottom). Stimuli (at arrow) drive spikes at 15 ms (closed star) or 21 ms latencles (open star) Higher amplitude stimulation (or stimulation at deeper sites) elicited the shorter latency consistently, which yielded a positive result for collision (not shown). Conversely, lower stimulation intensities (or more shallow stimulation sites) ehcited the longer latency consistently, which was also shown to be antidromic by colhslon critena. Compare to similar results obtained with frontal cortex stimulation (Fig 5). Horizontal calibration = 5 ms; vertical calibration = 0 2 mV. Recordings filtered 500 Hz to 10 kHz bandpass driven from the frontal cortex (including 66 latencies). Refractory periods were defined with double pulse stimulation as the interpulse interval that resulted in approximately 50% activation of the second driven spike (see Fig. 2). For frontal cortex stimulation, refractoriness ranged from 0.9 to 5.0 ms, with a mean value of 2.2 + 0.1 ms. For cells driven antidromically at more than one latency, the refractory period was usually shorter for the shorter latency when tested. For 10 cells driven from the parietal cortex, refractoriness ranged from 1.8 to 6.4 ms, with a mean value of 2.8 + 0.5 msec. We found that cells yielding longer latencies generally exhibited longer refractory periods than shorter latency neurons for both stimulation sites

Antidromic activation from multiple sites In the experiments using arrays of stimulation electrodes implanted in frontal cortex, we found that 11 of 13 NB neurons tested could be driven from different points of cathodal stimulation in frontal cortex

The present study details physiologic properties of cortically projecting basal forebrain neurons in chloral hydrate-anesthetized rats. The main results are: (1) Single NB neurons could be antidromically activated from either the frontal or the parietal cortex, yielding wide ranges of conduction latencies. (2) Antldromically driven cells were recorded in the globus pallidus, entopeduncular nucleus, or preoptic magnocellular area. Previous anatomic reports that all or nearly all cortically projecting neurons in this region (denoted NB) contain A C h E or ChATT,28.44,58 indicate that the driven cells reported here are presumably cholinergic. (3) The patterns and rates of spontaneous activity, as well as impulse waveforms, varied markedly among cortically projectmg NB neurons. (4) Individual NB neurons often exhibited multiple antidromic latencies, with values depending on the relative depth of stimulation site in the cortex and stimulation amplitude. (5) Single NB neurons could often be driven from neighboring sites within the frontal cortex, but not from both the parietal and the frontal cortex. Mean (+ S.E.M.) latencies for antidromic activation from the frontal or parietal cortex were 5.4 + 0.4 ms and 8.0 + 1.2 ms, respectively. Assuming fibers to the frontal stimulation site project rostrally around the genu of the corpus callosum 45-47.54, and therefore have straight-line lengths of about 8 mm, these latencies roughly translate to 0.3-8.0 m/s (mean = 1.5 m/s). Calculation of fiber length to parietal stimulation sites is somewhat less certain, as lateral cortical areas are reported to be innervated by NB axons m a rather oblique fashion, passing through the external capsule 45-47. Measuring approximately 11 mm for this fiber pathway yields conduction velocities for

281 70-

t

Z1 ,.n ~, 5 0 (/1 IZ = 30o U I0-

k,,dlL .,,,llJ~k,~,d I,.II/,.L d ~ ' u

•., 5 0 u') I'" 0 ¢.)

25Hz

25 7.7

30-

" ,_,t.. . . . . . .

.5'0 TIME (sec)

L, k. . . . . . . . . . . . . . . .;,,

.75

I O0

Hz

.25

50 TIME

75

I O0

( sec )

!

) i

¢z

I

Fig. 9. Intersplke-mterval histograms (ISHs) (upper left panel) and oscilloscope photographs dlustratmg two neurons recorded simultaneously in NB, driven antldromically from the lateral (left, middle and lower panels; upper ISH) or medial frontal cortex (right panels; lower ISH). Left stde: Middle panel - - Stimulation of lateral frontal cortex (arrow) 3.6 ms following spontaneous positive-polarity impulses (at far left of trace) drives this cell (star) at 2.8 ms latency. Lower panel - - Driven spikes are occluded for similar stlmuh (arrow) presented 3.2 ms after spontaneous spikes, in&cating collision between spontaneous and driven actwity. Right side: Upper panel - - Stimulation of medial frontal cortex (arrow) 1 5 ms following spontaneously occurring negative-polarity impulses drives this cell (open star) at 1.0 ms latency. Posntive-polarity spike (closed star), same as cell m left panels (&fferent sweep speed), driven at 3 6 ms latency. Middle panel - - Driven negatwe-polarity spikes are occluded for sirmlar stimuli (arrow) delivered 1.0 ms following spontaneous impulses from this cell, indicating collisxon between spontaneous and driven spikes. Posmve-polanty spike (closed star) remains driven. Lower panel - - Driven positive-polarity spikes are occluded for similar stlmuh presented 4.2 ms after spontaneous impulses from this cell (at far left of trace), indicating collision between spontaneous and driven spikes. Negative-polarity impulses remain driven (open star) All traces contain 10 superimposed sweeps. Verticalcalibration. All traces = 0 5 mV. Horizontal cahbratton" Left traces = 2 ms; right traces = 1 ms. Unfiltered recordings

282 these axons of about 0.5-6 9 m/s (mean = 1 4 m/s). This indicates that the axons traveling to the frontal or parietal cortex have roughly similar impulse conduction velocities. Previous studies of conduction velocity characteristics of central axons51. 53 estimate that fibers conducting impulses at less than about 1 m/s are probably non-myehnated, while those conducting at speeds above about 3 m/s are likely to be myelinated. Many fibers in the present study yielded impulse conduction velocities between these values, consistent with either small myelinated or non-myelinated fibers. However, it also seems clear from this analysis that cortical afferents from NB are morphologically heterogeneous, probably consisting of small, non-myelinated (slowly conducting) as well as myelinated (faster conducting) axons. Similar pronounced variation among cells in impulse conduction latencies to cortex is also found for noradrenergic neurons in locus coeruleus4,17, although with slower speeds of impulse flow overall. Results of our experiments on conduction latencies of cells driven from superficial or deep cortical placements indicate that it may be more meaningful to analyze conduction velocities of these axons separately for intracortical and subcortical fiber projections. The mean value (+ S.E.M.) for antidromic latencies with lowest thresholds in the superficial 1 2 mm of the frontal cortex was 7.9 + 0.9 ms, while the corresponding value for those with lowest thresholds in the deep frontal cortex was 4.4 + 0.6 ms Assuming a roughly radial course of fibers within the cortex (reported for putatively cholinergic fibers in many cortical laminae)21,24. 27, these results indicate that approximately 3.5 ms is required for impulses to traverse about 1 mm in the cortex, yielding an approximate conduction velocity of 0.3 m/s, while subcortlcal fibers conduct impulses at about 1.8 m/s. Simdar results occur when latencies are compared within individual cells exhibiting different latency values as a function of stimulation depth placement in the frontal cortex. In this analysis, latencies with lowest thresholds in the superficial 0.1-1.0 mm of the frontal cortex yielded a mean value of 9.1 + 1.0 ms, while lowest threshold latencies at depths of 1.4-2.3 mm yielded a mean value of 4.6 +_ 0.8 ms(n = 10). Thus, our data reveal that NB fibers may possess markedly faster conduction velocities for their subcortical extensions than for their intracortical projections. The

values obtained herem imply that many NB fibers may be thin and myehnated below the cortex, losing their myelin sheaths upon entering cortical termination areas, or that NB fibers often become markedly thinner upon entering the cortex Although this analysis accounts for some of the overall variation among NB neurons in cortical conduction latencles observed with fixed stimulation electrodes at various depths m the frontal cortex 6, there still remains about a 6-fold difference between shortest and longest latencies observed from either superficial or deep stimulation sites (see Fig. 6), again suggesting marked morphologic heterogeneity among these fibers m at least their subcortical projections The present results of multiple discrete antidromic latencies for single NB neurons as a function of stimulation intensity or depth in the cortex imply that fibers of these cells may be highly branched within the cortex, consistent with previous anatomic reports 21,27. It has previously been found 29 that main (larger) axons are more easily excited at a distance than fine axonal branches. Therefore, as either stimulation intensity is increased or the stimulation electrode is lowered in the cortex, a larger, faster conducting fiber branch may become preferentially activated instead of interbranch segments closer to the stimulating electrode, yielding a discretely shorter conduction latency. Similar results of all-or-none discrete 'jumps' in antidromic latency have previously been reported for substantia nigra dopammerglc projections to stnatum 11, and for certain hypothalamic neurons30, 40. These results have similarly been interpreted as indicating that the activated fibers either follow a very tortuous course through the area of stimulation or that the fiber projections are highly branched in the area of stimulation, with branches of different distances from the stimulating electrode having different thresholds for activation 40. Our observation that cells with multiple latencles preferentially yielded discretely shorter latencies as the stimulation electrode was lowered through the cortex implies that fiber morphology may change significantly even within the cortex, such that superficial branches may be thinner (or less well-myelinated) than deeper fiber lengths. The finding that individual NB neurons could often be driven antidromically from neighboring sites in the frontal cortex also supports the notion that these fibers are highly branched

283 in terminal fields, with horizontal extensions of at least some 1-2 mm. The absence of cells driven from both the frontal and the parietal cortex in the present study implies that individual NB neurons do not project to different cortical regions, but rather terminate within one specific region of the cortex only. This finding is at variance with one previous anatomic report that a significant percentage of NB neurons are doubly labeled from retrograde tracer injections in the frontal and parietal cortex 31, but supports other similar studies finding more restricted terminal fields for individual NB neurons 7,39,45,46. Our result that cells driven from the parietal cortex tend to be more caudally located overall than those antldromically activated from frontal sites supports previous anatomic reports of a similar topographic organization in the NB7,28,44.54. The present study also reveals that, unlike other non-thalamic cortical afferent systems (i.e. catecholamine or indolamine neurons), these putative cholinergic NB neurons form a physiologically heterogeneous population in terms of discharge rates and patterns of spontaneous activity, as well as impulse waveforms. While the physiologically more uniform catecholamine and indolamine neurons9.17,22 are also relatively homogeneous in.terms of somata size and shape within individual cell groupsl3,36,37,50, the cytology of cholinergic NB neurons is less agreed upon at present. Most authors emphasize large isodendritic neurons as the predominant feature of intensely AChE-positive profiles in the NB area that project to the cerebrocortex in the rat7.15.24,2s. However, Mesulam et al. in a recent study in the monkey34 noted the presence of substantial cytologic heterogeneity in putatively cholinergic, cortically projecting nBM neurons, including large fusiform, pyramidal, and multipolar neurons, as well as many smaller fusiform neurons. If physiologic attributes relate to neuronal morphology (especially as expected for impulse waveform and somata-dendritic cytology), then the present results indicate that cortically projecting, cholinergic NB neurons in rat may be more anatomically

heterogeneous (perhaps similar to the above observations in monkey) than currently appreciated. It is also possible, however, that the present study reveals populations of non-cholinergic, cortically projecting NB neurons not observed with present anatomic techniques, with differing physiologic characteristics reflecting different functional roles played by various transmitter substances. The functional significance of the observed pronounced physiologic heterogeneity for cortically projecting cholinergic neurons is not clear at present. It is interesting to note that the physiologically homogeneous noradrenergic locus coeruleus neurons appear to have more divergent efferent projections, such that individual neurons are found to project to widely different brain areas 17. It may be that the more restricted terminal fields of NB cholinergic neurons correspond to their physiologic heterogeneity, such that more restricted target areas are differentially controlled by individual neurons to a greater extent than in the locus coeruleus system. Thus, the noradrenergic locus coeruleus system may act more or less in unison to have a global concerted influence on brain activity1-3. In contrast, NB neurons, by virtue of their physiologic heterogeneity and more restricted terminal fields, may exert greater differential control of select target areas. These results demonstrate the feasibility of physiologic identification of putatively cholinergic, cortically projecting NB neurons. This approach should facilitate further investigations of the role of this nonthalamic cortical afferent system in brain and behavioral activities. Initial studies from this laboratory 41 indicate that NB fiber conduction velocity is decreased subcortically in aged subjects. ACKNOWLEDGEMENTS We thank R. L. Isaacson, J. Rogers and J. Hedreen for helpful discussions. This work was supported by NINCDS Grant 19360, BRSG Grant SO7RR7149-09, and SUNY Research Foundation Grant 0140-03.

284 REFERENCES 1 Aston-Jones, G and Bloom, F E , Activity of noreplnephrlne-contalnlng locus coeruleus neurons in behaving rats anticipates fluctuations in the sleep-waking cycle, J Neuroscl., 1 (1981) 876-886 2 Aston-Jones, G and Bloom, F E , Norepinephrlne-containmg locus coeruleus neurons in behaving rats exhibit pronounced responses to non-noxious environmental stlmuh, J Neurosct. 1 (1981) 887-900 3 Aston-Jones, G., Foote, S. L. and Bloom, F. E , Anatomy and physiology of locus coeruleus neurons, functional implications In M. G. Zlegler and C. R Lake (Eds.), Noreptnephrme Clinical Aspects, Williams and Wilkins, Baltimore, In press 4 Aston-Jones, G , Segal, M and Bloom, F E , Brain amlnergic axons exhibit marked vanability in conduction velocity, Bram Research, 195 (1980) 215-222 5 Aston-Jones, G., Shaver, R and Dinan, T., Physiology of cortically projecting neurons m rat basal forebraln, Soc Neurosct. Abstr, 9 (1983) 355 6 Aston-Jones, G , Shaver, R and Dlnan, T , Cortically projectmg nucleus basahs neurons in rat are physiologically heterogeneous, Neurosct Lett., 46 (1984) 19-24 7 BIgl, V , Wootf, N. J and Butcher, L L , Chollnerglc projections from the basal forebram to frontal, parietal, temporal, occipital, and cingulate cortices' a combined fluorescent tracer and acetylchohnesterase analysis, Brain Res. Bull, 8 (1982) 727-749 8 Bloom, F E., Chemical integrative processes m the central nervous system In F O Schmltt and F G Worden (Eds.), The Neurosciences Fourth Study Program, MIT Press, Cambridge, MA, 1979, pp. 51-58. 9 Bunney, B S , Waiters, J R., Roth, R. H and Aghajanian, G. K., Dopaminerglc neurons effect of antipsychotlc drugs and amphetamine on single cell actwlty, J Pharmacol exp Ther, 185 (1973) 560-571. 10 Burton, M J , Mora, F and Rolls, E. T , Visual and taste neurons in the lateral hypothalamus and substantm inomlnata: modulation of responsiveness by hunger, J. Physiol. (Lond), 252 (1975) 50-51 11 Colllngndge, G L , James, T A and MacLeod, N K , Antidromic latency variations of nlgral compacta neurons, Expertenua, 36 (1980) 970-971 12 Coyle, J T , Price, D L and DeLong, M R , Alzhelmer's disease, a disorder of cortical cholinerg~c lnnervation, Science, 219 (1983) 1184-1190 13 Dahlstrom, A and Fuxe, K , Evidence for the existence of monoamme-contalnlng neurons in the central nervous system I Demonstration of monoamlnes in the cell bodies of brain stem neurons, Acta phystol scand, 62, Suppl 232 (1964) 1-55 14 DeLong, M R , Activity of palhdal neurons during movement, J Neurophysiol., 34 (1971) 414-427 15 Dlvac, I., Magnocellular nuclei of the basal forebram project to neocortex, brain stem, and olfactory bulb. Review of some functional correlates, Brain Research, 93 (1975) 385-398 16 Foote, S. L , Freedman, R. and Oliver, A P., Effects of putauve neurotransm~tters on neuronal actmty m monkey auditory cortex, Brain Research, 86 (1975) 229-242 17 Foote, S. L , Bloom, F E. and Aston-Jones, G , Nucleus locus coeruleus, new evidence of anatomical and physiological speoficlty, Phystol Rev, 63 (1983) 844-914

18 Fuller, J and Schlag, J , Determination of antldromlc exoration by the collision test problems of lnterl~retat~on, Brain Research, 112 (1976) 283-298 19 Grace, A A and Bunney, B S , Nlgral dopamme neuron~ lntracellular recording and Identification with i-Dopa injection and hlstofluorescence, Science, 210 (1980) 654-656 20 Guyenet, P G and Aghajanian, G K , Anttdromlc identification of dopamlnerglc and other output neurons of the rat substantla nlgra, Bram Research. 150 (1978) 69-84 21 Houser, C R , Crawford, G D , Barber, R P Salvaterra, P M and Vaughn, J. E , Organization and morphological characteristics of chohnerglc neurons an lmmunohistochemical study with an antibody to chohne acetyltransferase, Brain Research. 266 (1983) 97-119 22 Jacobs, B L and Gelperin, A (Eds), Serotomn, Neurotransmtsston and Behavior, MIT Press, Cambridge, MA, 1981 23 Johnston, M V., McKmney, M and Coyle, J T , Ewdence for a chohnergic projection to neocortex from neurons in basal forebraln, Proc. nat Acad Sct U S A . 76 (1079) 5392-5396 24 Johnston, M V , McKlnney, M and Coyle, J T , Neocort~cal chohnerglc innervat~on a description of extrinsic and Intrinsic components in the rat, Exp Brain Res. 43 (I981) 159-172 25 Jones, E G , Burton, H., Saper, C B and Swanson, L W , Midbraln, dtencephahc and cortical relationships of the basal nucleus of Meynert and assocmted structures m primates, J comp Neurol, 167 (1976) 385-420 26 Klewt, J and Kuypers, H G J. M , Basal forebrain and hypothalamlc connections to frontal and parietal cortex in the rhesus monkey, Science, 187 (1975) 660-662 27 Krnjevlc, K and Silver, A , A hlstochemlcal study of cholinerglc fibers in the cerebral cortex, J Anat, 99 (1965) 711-759 28 Lehmann, J , Nagy, J I , Atmadla, S and Flblger, H C., The nucleus basahs magnocellularls the origin of a chohnerg~c projection to the neocortex of the rat, Neurosctence, 5 (1980) 1161-1174 29 LIpskL J , Antidromlc actwatlon of neurons as an analytic tool in the study of the central nervous system, J Neurosct Meth., 4 (1981) 1-32 30 MacLeod, N K and Mayer, M L., Electrophyslologlcal analysis of pathways connecting the medial preoptlc area with the mesencephalic central grey matter m rats, J Phystol (Lond), 298 (1980) 53-70 31 McKlnney, M , Hedreen, J and Coyle, J T , Topographic analysis of the lnnervatlon of the rat neocortex and hippocampus by the basal forebraln chollnerglc system, Soc Neuroscl Abstr, 8 (1982) 518. 32 Mesulam, M - M , A horseradish peroxldase method for the ldentfflcatmn of the efferents of acetyl cholinesterase-contaming neurons. J Htstochem Cytochem, 24 (1976) 1281-1286. 33 Mesulam, M -M and Van Hoesen, G W., Acetylchohnesterase-rIch projections from the basal forehrain of the rhesus monkey to neocortex, Brain Research, 109 (1976) 152-157 34 Mesulam, M.-M , Mufson, E J , Levey, A. I and Warner, B H , Chohnerglc mnervat~on of cortex by the basal forebrain cytochemistry and cortical connections of the septal area, diagonal band nucleL nucleus basahs (substantm mominata), and hypothalamus in the rhesus monkey, J comp Neurol. 214 (1983) 170-197

285 35 Mitchell, S J., Richardson, R. T. Baker, F H. and DeLong, M. R., Electrophyslological and functional charactenstlcs of neurons in the nucleus basalis of Meynert in macaque monkeys, Soc. Neurosci. Abstr., 8 (1982) 212. 36 Moore, R Y and Bloom, F. E., Central catecholamlne systems anatomy and physiology of the dopamine systems, Ann Rev. Neurosct., 1 (1978) 129-169. 37 Moore, R Y. and Bloom, F. E , Central catecholamine neuron systems' anatomy and physiology of the noreplnephrine and epinephrine systems, Ann Rev. Neurosct., 2 (1979) 113-168 38 Parent, A., Gravel, S and Oliver, A , The extrapyramldal and llmbic systems relationship at the globus palhdus level a comparative histochemlcal study in the rat, cat and monkey. In L J Poirler, T. L. Sourkes and P. J Bedard (Eds.), Adv. Neurol, Vol 24, Raven Press, New York, 1979, pp. 1-11. 39 Price, J. L and Stern, R., Individual cells In the nucleus basahs-dlagonal band complex have restricted axonal projections to the cerebral cortex in the rat, Brain Research, 269 (1983) 352-356 40 Renaud, L P., Blume, H W and Plttman, Q. J , Neurophysiology and neuropharmacology of the hypothalamic tuberoinfundibular system. In W F Ganong and L. Martini (Eds.), Frontiers m Neuroendocrmology, Vol. 5, Raven Press, New York, 1978, pp. 135-162. 41 Rogers, J., Aston-Jones, G., Dinah, T., Shaver, R and Moss, D., Conduction velocity of cortically projecting nucleus basalis neurons decreases with age, Soc Neuroscl Abstr., 9 (1983) 925 42 Rogers, J. and Bloom, F. E., Neurotransmitter metabolism and function in the aging central nervous system. In C E. Finch and E Schneider (Eds), Handbook of the Biology of Aging, Plenum, New York, in press. 43 Rolls, E T., Sanghera, M. K. and Roper-Hall, A., The latency of activation of neurones m the lateral hypothalamus and substantia inominata during feeding in the monkey, Brain Research, 164 (1979) 121-135 44 Rye, D. B , Wainer, B. H., Mesulam, M -M. and Saper, C B., Identification of chohnergic projections to cortex using combined retrograde tracing and immunohistochemical locahzatlon of choline acetyltransferase: cholinerglc pathways IV, Soc Neurosci. Abstr, 9 (1983) 963. 45 Saper, C. B , Organization of the nucleus basalis projection to cerebral cortex in the rat' chohnergic pathways VIII, Soc. Neurosct. Abstr., 9 (1983) 963. 46 Saper, C B , Organization of cerebrocortical afferent sys-

47

48

49

50

51

52

53

54

55

56

57

58

tems in rat: I. Magnocellular basal nucleus, J. comp. Neurol., In press. Schute, C. C. D. and Lewis, P. R., The ascending cholinergic reticular system: neocortical, olfactory and subcortlcal projections, Brain, 90 (1967) 497-520. Segal, M., The effects of brainstem priming stimulation on interhemispheric hlppocampal responses in the awake rat, Exp. Bram Res., 28 (1977) 529-541. Segal, M. and Bloom, F E , The action of norepinephnne in the rat hlppocampus IV. The effects of locus coeruleus stimulation on evoked hlppocampal unit activity, Bram Research, 107 (1976) 513-525. Steinbusch, H. W M., Distribution of serotonin immunoreactivity in the central nervous system of the rat - - cell bodies and terminals, Neuroscience, 6 (1981) 557-618. Swadlow, H A. and Waxman, S. G., Variations in conduction velocity and excitability following single and multiple Impulses of visual callosal axons in the rabbit, Exp. Neurol., 53 (1976) 128-150. Szymusiak, R and McGlnty, D., Basal forebrain unit activity across the sleep-waking cycle in freely moving cats, Soc Neuroscz. Abstr., 9 (1983) 1203. Waxman, S. G. and Bennett, M V L., Relative conduction velocities of small myehnated and non-myelinated fibers in the central nervous system, Nature New Biol., 238 (1972) 217-219. Wenk, H., Bigl, V. and Meyer, U , Chohnergic projections from magnocellular nuclei of the basal forebrain to cortical areas in rats, Bram Res Rev., 2 (1980) 295-316. Whitehouse, P. J., Price, D L., Clark, A. W., Coyle, J T. and DeLong, M. R , Alzhelmer's &sease: evidence for selective loss of chohnergic neurons in the nucleus basalis, Ann Neurol., 10 (1981) 122-126. Whitehouse, P J , Price, D. L., Struble, R G , Clark, A. W., Coyle, J T and DeLong, M R., Alzhelmer's disease and senile dementia: loss of neurons in the basal forebrain, Science, 215 (1982) 1237-1239. Woodward, D J , Moises, H. C , Waterhouse, B. D., Holler, B. J and Freedman, R., Modulatory actions of noreplnephrine in the central nervous system, Fed Proc., 38 (1979) 2109-2116 Woolf, N J., EckensteIn, F. and Butcher, L. L., Chohnergic projections revealed by choline acetyltransferase (CHAT) immunohlstochemistry and fluorescent tracer histology performed on the same tissue section, Soc. Neurosct Abstr., 9 (1983) 967.