Electrophysiological evidence for the involvement of the locus coeruleus in alerting, orienting, and attending

C.D. Barnes and 0. Pompeiano (Eds.) Prugres.5 in Brain Research, Vol. 88 0 1991 Elsevier Science Publishers B.V.

521

CHAPTER 36

Electrophysiological evidence for the involvement of the locus coeruleus in alerting, orienting, and attending S.L. Foote, C.W. Berridge, L.M. Adams and J.A. Pineda Department of Psychiatry (0603), School of Medicine, University of California, Sun Diego, La Jolla, CA, U.S.A.

In this chapter, we describe recent observations from our laboratory which support the thesis that the locus coeruleus ( L O , via its massively divergent efferent projections, participates in generating a generalized brain state that can be characterized as “alertness.” The first of these observations suggests that LC activation can convert the electroencephalographic (EEG) activity of the forebrain from patterns characteristic of a non-alert state to those characteristic of an alert state. The second observation indicates that LC activation alters sensory responses of individual neocortical neurons in a way that is compatible with the general thesis presented here, suggesting that LC-induced alterations in cortical neuronal activity may be an integral component of a hypothesized

participation of the LC in cortically mediated attentional processes. The third observation indicates that LC may modulate forebrain components of orienting responses that are indexed by event-related potentials (ERPs). Thus, the experiments described below involve electrophysiological assessment of forebrain information processing at three different levels of organization: activity of individual neurons in the millisecond range, neuronal ensemble activity persisting for 10-200 msec as indexed by ERPs, and ensemble/regional activity sustained for seconds to minutes as indicated by E E G measures. These observations suggest that alterations induced in forebrain function by manipulations of LC activity are evident at all three of these levels.

Key words: attention, cortex, hippocampus, EEG, arousal

Introduction

A number of observations suggest that the noradrenergic nucleus locus coeruleus (LC) participates in processes such as arousal and alerting. Many anatomical and physiological characteristics of the LC-noradrenergic system, including its massively divergent efferent projections and the ability of norepinephrine (NE) to enhance stimulus-elicited responses in target neurons, are compatible with this hypothesis (see Foote and Morrison, 1987). In addition, LC neurons exhibit en-

hanced discharge rates just before and during periods of arousal as indicated by behavioral and forebrain electroencephalographic (EEG) measures (Hobson et al., 1975; Foote et al., 1980; Aston-Jones and Bloom, 1981). However, these physiological and anatomical observations are correlational and do not provide strong demonstrations of a causal relationship between LC and forebrain activity. This chapter describes the effects of experimental manipulations of LC activity on various types of cortical electrophysiological measures.

522

These experiments begin to address the issue of whether changes in LC activity are necessary and/or sufficient to induce changes in these forebrain measures.

LC activation alters forebrain EEG characteristics In halothane-anesthetized rats, cortical EEG (ECoG) and hippocampal EEG (HEEG) typically exhibit activity similar to that of slow-wave sleep. However, periods of EEG activity closely resembling that seen during waking are sometimes observed spontaneously and are always observed following a stimulus such as a tail-pinch, even though the animal is still at a surgical level of anesthesia and does not overtly respond to any such stimulation (all of these phenomena are also observed in humans). We have utilized this preparation to ask whether activation of the LC would reliably alter forebrain E E G status (Berridge and Foote, in press). One advantage of the halothane-anesthetized preparation for such a study is that the “baseline” EEG status of the

LC TRIGGER

-

LC 2OGCOUNTS lOSEC 0-

,,L

animal is stable for prolonged periods, so that the effect of a particular manipulation can be reliablq assessed. Also, after a period of “arousal,” EEG signs usually return to their baseline levels so thal repeated tests can be performed. The LC is difficult to activate selectively bj means of electrical stimulation because it is small and surrounded by major fiber tracts originating from other nuclei. A method we had previouslq developed (Adams and Foote, 1988) was utilized in this study. A recording/infusion probe consisting of a microelectrode and infusion cannula wa$ lowered, using physiological landmarks, so thal the microelectrode was in the LC and the infusion cannula was 200-400 microns lateral to its lateral edge with the beveled tip oriented toward the LC. As in our previously published study, infusion of the cholinergic agonist bethanechol was used to repeatedly activate the LC for a few minutes at a time, as verified by recordings of the electrophysiological activity of individual or multiple LC neurons through the microelectrode. LC activation typically consisted of a peak increase in discharge rates of 3-5 times resting discharge

“,, /

*

Fig. 1. Relationship of LC activity to ECoG (top panel) and, in a separate experiment, HEEG (bottom panel) before, during, and after peri-LC bethanechol infusions. Bethanechol was infused at a constant rate throughout the interval indicated. EEG activity is shown in the top trace of each panel, the raw trigger output from LC activity in the middle trace, and the integrated trigger output (counts per 10-sec interval) in the bottom trace. In the top panel, LC activity is seen to increase during the latter part of the infusion, and several seconds later reduced amplitude and increased frequency become evident in the ECoG trace. As LC activity begins to decrease following the infusion, ECoG amplitude begins to increase and its frequency decreases. In the bottom panel, enhanced LC activity becomes evident in the latter part of the infusion period, and several seconds later theta rhythm begins to dominate the HEEG trace. For the remainder of the trace, LC activity remains elevated and theta rhythm predominates (From Berridge and Foote, in press.)

523

LC activation as a crucial mediating event in producing the EEG effects that follow the bethanechol infusions. These EEG changes have been quantified and statistically verified using power-spectrum analyses (see Fig. 2). It has long been known that LC electrophysiological activity is correlated with measures of forebrain EEG activation (see Foote et al., 1983 and Aston-Jones et al., this volume, for reviews). The present observations are especially interesting because they provide evidence that LC activation may initiate changes in cortical and hippocampal EEG patterns. Specifically, LC activation may be sufficient, although possibly not necessary, to induce ECoG desynchronization and a predominance of theta rhythm in HEEG activity. Obviously, however, certain caveats must apply to the interpretation of these results since they have been obtained from anesthetized preparations.

rates for a 3-5 min period. Our experience indicates that the infusion volumes, drug concentrations, and the infusion rates used activate a substantial majority of the LC neurons ipsilateral to the infusion. ECoG activity was recorded from sites in frontal neocortex (HEEG) and dorsal hippocampus. This experiment has now been performed in 28 animals with the following findings: (1) LC activation is consistently followed, within 5-20 sec, by ECoG desynchronization ( i e . , decreased amplitude and increased frequency) and a preponderance of theta activity in the HEEG (see Figs. 1 and 2); (2) if the recording/infusion probe is located so that the infusion is not effective in activating LC neurons (e.g., is placed more than 500-600 p M outside the LC), no such forebrain EEG effects are produced by the infusion; (3) following infusion-induced activation, forebrain EEG returns to pre-infusion patterns with about the same time course as the recovery of LC activity (10-20 min for complete recovery); (4) whether infusions are made from sites medial or lateral to LC, forebrain EEG changes invariably follow LC activation. These observations point to

t-

LC activation alters sensory responses of neocortical neurons This recent experiment (Adams and Foote, in preparation) was motivated by numerous studies,

PSA+

t---PSA

08

9.8

18.8

27.9

HZ

369

459

+

0 20

0 20

08

98

188

279 HZ

36.9

459

08

9.8

18.8

27.9

36.9

459

HZ

Fig. 2. Raw EEG data and power-spectrum analyses from the pre-infusion, post-infusion, and recovery periods from carbachol of a typical experiment. The exact ll-sec interval utilized for each power-spectrum analysis is indicated by the bar marked “PSA’ below each trace. There is an 80-second gap between the pre- and post-infusion traces, and a 5-min 20-sec gap between the post-infusion and recovery traces. The most striking post-infusion changes are the loss of activity at approximately 1 Hz in both traces and the emergence of dominant theta activity in the HEEG trace and power spectrum. (From Berry and Foote, in press.)

524

by ourselves and others, of the effects of norepinephrine (NE), the putative neurotransmitter released by LC axons, on presumed LC target neurons in neocortex, hippocampus, cerebellum, and numerous other brain sites (reviewed in Foote et af., 1983; Foote and Morrison, 1987). For example, in our own previous work, the effects of NE were tested on auditory cortex neurons which were activated acoustically by species-specific vocalizations in awake squirrel monkeys (Saimiri sciureus) (Foote et al., 1975). Five-barrel glass electrodes were used to record the activity of individual neurons in the superior temporal gyrus and to microiontophoretically apply NE or other putative neurotransmitters. Peristimulus-time histograms and raster displays of spontaneous activity and responses to the vocalizations were computed before, during, and after iontophoresis. With NE application, dose-dependent reduction of spontaneous and stimulus-evoked discharge was observed. For a given dose, the percent reduction of spontaneous activity was greater than the percent reduction of stimulus-elicited activity. Quantitative analysis also revealed that within periods of evoked activity, the percent reduction was inversely related to discharge rate, i.e., the percent reduction was least for episodes of most intense activity. We proposed that the presumed tonic inhibition resulting from synaptically released NE “would enhance the difference between background and stimulus-bound activity,” i.e., NE would act to enhance the “signal-tonoise” ratio of the cells. Extensive studies of the effects of NE on evoked and spontaneous activity have now been made in numerous other brain sites, both in uivo and in vitro, and other chapters in this volume detail much of this work. We wished to extend our previous observations by evaluating the effects of LC activation, rather than NE application, on sensory responses of neocortical neurons. These studies, conducted in halothane-anesthetized rats, involved the pharmacological activation of the LC, using the recording/infusion probe described above, during the simultaneous

recording of spontaneous and stimulus-elicited activity from neurons in primary somatosensory cortex. LC activation effects might be expected to differ from those of iontophoretically applied NE because the amount and spatial distribution of synaptically released NE would differ from that resulting from iontophoretic application. Also, LC activation would release NE in all the brain areas to which the LC projects, for example, somatosensory thalamus, and the effects measured at cortex would be the net result of NE release at all these sites. Presumably, LC activation more realistically mimics the effects of physiological activation of LC neurons than does NE iontophoresis. Our neocortical recordings were obtained from the hindlimb region of primary somatosensory cortex. Air-puff or electrical stimulation of the appropriate receptive field was used to activate somatosensory neurons. Peristimulus-time histograms were computed before, during, and after the repeated LC activations. Under baseline conditions, somatosensory neurons exhibited the type of stimulus-elicited activity previously reported by others. A shortlatency (15-22 msec), consistent, brief activation was followed by a longer duration (50-70 msec) suppression of firing (to below spontaneous levels) and a gradual return to baseline discharge rates (see Fig. 3). The effect of LC activation on each response component was consistent over infusions for a given animal and was also very similar in different animals. The initial, brief activation was reduced, the subsequent, longer-duration pause was converted into an activation, and spontaneous discharge rates were reduced. All of these cortical effects exhibited recovery when LC activity returned to baseline levels. In experiments designed to control for nonspecific effects of the LC infusion, it was found that lowering the infusion probe to a site 400 p m below the LC and doing the same infusions did not result in any effect on somatosensory responses. Also, systemic infusion of the a-adrenergic agonist clonidine, which inhibits LC activity,

525 RAT 60

RAT 90

RAT 00

RAT 230

LLL I

BASELINE

INFUSION

RECOVERY

1

I 1000

I

2000

1

~ ~ ~ ~ ( m s e c )

I 1000

I

2000

1

TIME (msec)

I 1000

I

2000

~ ~ ~ ~ ( m s e c )

TIME (msec)

Fig. 3. Peristimulus-time histograms of cortical activity during baseline, infusion, and recovery conditions from NE. Note that although the overall levels of activity differ between animals, the nature of the baseline response is similar in each case, and LC activation affects each response similarly: the short-latency activation is reduced, the post-activation pause is changed to a sustained activation, and spontaneous activity is reduced. The S below the trace indicates the time of occurrence of the somatosensory stimulus. (From Adams and Foote, in press.)

in doses sufficient to neutralize the activation of LC also negated the infusion effects on cortical activity. Quantitative analyses revealed that the suppression of the initial peak, the reversal of the normal reduction in activity into a sustained enhancement of activity (above spontaneous levels), and the reduction in spontaneous activity were statistically significant (see Fig. 4). The absolute magnitude of the total response was increased, and the ratio of evoked activity to spontaneous activity was enhanced to an even greater extent. These observations indicate that LC activation and NE iontophoresis effects in cortex exhibit certain similarities and certain differences. The overall “signal-to-noise” effect is present with LC activation, but there is a loss of the temporal specificity of the phasic sensory response. The sharp onset and post-activation inhibition which clearly circumscribe the response under baseline conditions are reduced and reversed, respectively. This results in a response which is of greater

magnitude in terms of the absolute number of spikes elicited, and in an even greater enhancement of the ratio of evoked to spontaneous activity. These electrophysiological effects may have a behavioral counterpart in that there is sometimes a “trade-off” in arousal states whereby stimulus

-100‘

Short-Latency Excitation

N

Long-Latency

I

Total Evoked Total Evoked Response Spontaneous

PRE-DRUG 0 1-MIN AFTER 0 6-MIN AFTER N 12-MIN AFTER

Fig. 4. Changes induced by NE in somatosensory responses and in spontaneous cortical activity by LC activation. All rates are expressed as percentages of pre-infusion, spontaneous discharge rates except the far-right bar which is expressed as a percentage of post-infusion spontaneous rate. (From Adams and Foote, in preparation.)

526

detectability is enhanced, but aspects of discriminability are simultaneously reduced (e.g., see Posner, 1978). LC lesions alter monkey P3 event-related potentials

Event-related potentials (ERPs) are another electrophysiological index of forebrain information processing. They have been extensively studied in humans because they can be recorded non-invasively and offer the hope of sampling brain activity with temporal resolution in the millisecond range. In order to study the neural substrates of these potentials, we and others have developed animal models of particular ERP components. For example, we have previously demonstrated that monkeys exhibit certain ERP components with latency, polarity, and contingency similarities to those observed in humans (Neville and Foote, 1984; Pineda et al., 1987; 1988). These compounds will be indicated as positive (P) or negative (N) values following numbers corresponding to their latency in msec. One of the most intensively studied components of the human ERP is P300 or P3 which is elicited in response to novel and/or task-related stimuli. Two types of hypothesis regarding the neural origins of P3-like potentials, particularly P3b, are evident in the literature: those implicating a single neural source whose field potentials are volume conducted throughout large regions of the brain and those suggesting the existence of multiple sources. Single-source hypotheses have been challenged by recent evidence that steep potential gradients and polarity inversions in human P3-like activity can be recorded intracranially in

several sites, such as the frontal cortex (Wood and McCarthy, 19861, hippocampus and amygdala (Halgren et al., 1980; Wood et al,, 1980; McCarthy et al., 19821, and thalamus (Yingling and Hosobuchi, 1984). The alternative view of multiple active sites is also supported by the wide distribution of P3-like potentials recorded intracranially in animals (cingulate cortex: Gabriel et al., 1983; suprasylvian and marginal gyri: O’Connor and Starr, 1985). While the multiplesource hypothesis could account for a widespread distribution of P3 activity, it does not account for the uniformity of latencies at such spatially distributed sites. Therefore, some modification of the multiple-source proposition would appear to be necessary to formulate a more adequate hypothesis of P3 neurogenesis. One possibility is that a widely distributed neural system synchronously impacts on a number of forebrain areas. Several such candidate systems have been characterized, among them is the LC which exhibits the anatomic, physiologic, and functional properties (reviewed in Foote et al., 1983; Foote and Morrison, 1987) necessary to subserve such a role. The evidence concerning widely divergent LC projections, homogeneous activity of source neurons, transmitter effects on target neurons, and LC effects on EEG activity suggests that LC is activated during alerting or arousal, which leads to NE release onto target neurons in many brain regions. Thus, our hypothesis suggests that novel, surprising, and attentioneliciting events increase LC discharge activity and enhance “attentiveness.” In the context of the present experiment, we further hypothesize that such LC activation may be a necessary preccndition for the elicitation or enhancement of slow

Fig. 5. Sagittal sections through the brain stem of a control and a lesioned monkey (SM24). A. Section reacted with anti-dopaminep-hydroxylase (anti-DpH) from a normal monkey shows the darkly labeled noradrenergic cells of the LC nucleus and its anterior pole (AP). B. Section from lesioned monkey SM24, reacted with anti-DpH, is in approximately the same plane as the control section (A). The dashed lines demarcate the boundaries of the electrolytic lesion. C. Adjacent Nissl-stained section also showing the extent of the electrolytic lesion. The dorsal and rostra1 portions of the sections are to the top and left, respectively. BC, brachium conjunctivum; IV, trochlear nerve; dIV, decussation of trochlear nerve; mV, tract of mesencephalic nucleus of trigeminus. (From Lewis et al., 1987.)

521

528

interruption of dorsal bundle (DB) fibers (Pineda et al., 1989). Stimuli consisted of 2 and 6 kHz tone pips (40 msec duration, 60 dB above nHL) presented at 1 Hz in random order. In most sessions, one tone constituted 90% of the stimuli and the other tone lo%, while in some sessions

endogenous potentials or may directly produce synchronized slow electrical activity that may be evident as surface-recorded P3-like potentials. ERPs were recorded from untrained squirrel monkeys (Saimiri sciureus) twice a week for 4 weeks before and after bilateral LC lesions and

SM12

SM16

SM24

N106 N250-900

P372

-Pre LC Lesion 0

200

400

600

800

............ Post LC Lesion

msec Fig. 6. Averaged ERPs from individual subjects recorded at FZ, P3, and P4 in response to infrequent (10% probability) tones before (solid lines) and after (dotted lines) lesions. Subjects (SM12, SM16, SM24) with extensive cell-body and fiber lesions are shown. (Modified from Pineda et a/., 1989.)

529

tones were made equiprobable to test the effects of manipulating stimulus probability. LC and DB lesions were made under general anesthesia by first localizing the nucleus with a microelectrode, using electrophysiological landmarks, and then creating an electrolytic lesion. The electrode was then placed at the anterior pole of the nucleus and a knife cut effected. The extent of damage to LC perikarya and ascending axons was assessed by reconstructing lesions from Nissl-stained sagittal sections through the brain stems (see Fig. 5). The efficacy and selectivity of lesions in eliminating cortical noradrenergic axons were immunohistochemically verified utilizing antisera directed against dopamine-P-hydroxylase and tyrosine hydroxylase to label noradrenergic and dopaminergic axons, respectively. Prelesion ERPs resembled those previously reported for squirrel monkey in this paradigm (Neville and Foote, 1984; Pineda et al., 1987). As illustrated in the grand average ERPs for individual subjects shown in Figure 6, a large triphasic response centered over midfrontal areas was evident in the first 200 msec following either frequent or infrequent tones. This response consisted of a large positivity (mean latency, 52 msec) followed by a large negativity (mean latency, 106 msec). At longer latencies, and most prominently at lateral parietal electrodes, a broad positivity with a bipeaked morphology (mean latencies of 239 and 372 msec, respectively) was elicited in response to infrequent tones. A long-duration negativity was also recorded over frontal cortex (N250-900), temporally overlapping with the late positive components and peaking in the 500-600 msec range. Analyses of P52, P176, and N250-900 areas did not reveal any effects of, or interactions with, lesion type. In contrast, monkeys with damage to LC cell bodies and dorsal bundle fibers showed decreased P239 and P372 areas in the 90-10 blocks relative to their prelesion measures, while monkeys in which damage was restricted to DB fibers did not show such a marked decrease, (P239; lesion type X lesion F(1,3) = 10.44, P < 0.05; P372; lesion type X lesion, F(1,3) = 49.57,

P < 0.01). Monkeys with 40-70% of the LC damaged exhibited P239 and P372 area decreases greater than 60%. In contrast, little (14%) or no damage to the LC resulted in small decreases ( < 42%) or even increases in area. A Spearman rank-order correlation coefficient based on all subjects’ data indicated a statistically significant relationship between the extent of LC damage and the percentage decreases in P239 and P372 area ( r S= 0.90, P < 0.05). Several factors suggest caution in interpreting these data. First, despite the homogeneity of LC neurons and the small size of the electrolytic lesions, it is clear that areas in the trajectory of the microelectrode (e.g., cortex, colliculi), areas adjacent to the LC nucleus, and fibers of passage were also affected. Thus, systems other than the one targeted have been damaged, including systems possibly involved in the perception of sound. Second, the inherent variability in the extent of lesions plus the small number of subjects used reduces the power of any statistical evaluation. Several means of controlling for these factors are available and await future investigations. Investigations of these ERP observations suggest that the LC plays a modulatory role in the regulation of cortical responsiveness to sensory stimulation. LC activity may contribute to the production of a behavioral state in which novel sensory stimuli are more effectively processed, and this behavioral state may be a necessary precondition for the elicitation of long-latency, P3-like potentials. Alternatively, volleys of LC activity may directly produce slow waves that are recorded as P3-like activity. The decreased magnitude of monkey P3-like responses following lesions of the LC and ascending NE fibers is consistent with either of these hypotheses. Also, these results do not indicate whether NE-LC is both necessary and sufficient for the generation of such activity. Results from the present investigation are also consistent with behavioral studies, involving lesions of the LC or of DB fibers in rodents, which have reported decreases in startle responses

530

(Adams and Geyer, 1981), enhanced behavioral responsiveness to novelty (Britton et al., 19841, and increased distractibility and exploratory behavior toward novel objects (Roberts er d., 1976; Oke and Adams, 1978; Koob et al., 1984). The demonstration that P3-like generation or modulation may be dependent upon the integrity of the NE-LC system could also link the observations that patients with clinical disorders such as dementia, alcoholism, schizophrenia, and affective disorders (Schildkraut and Kety, 1967; Dongen, 1981; Mair and McEntee, 1983) exhibit pathology of the NE-LC, with demonstrations that these conditions are also associated with altered P3 components (Diner et al., 1985; Mirsky and Duncan, 1986; Polich et al., 1986). Interpretations and hypotheses

The possibility that LC activation can induce ECoG desynchronization and HEEG activity dominated by theta rhythm is compatible with a proposed involvement of the LC in “alerting.” Certainly these experiments must be extended to unanesthetized preparations in order for this proposal to receive more solid support. However, even were similar results obtained in unanesthetized preparations, two immediate complications would present themselves. First, dominant forebrain EEG “arousal” is evident during rapideye-movement (REM) sleep, a state during which LC neurons are silent (see Foote et al., 1983, for review). Second, there may be certain behavioral states characterized by the ECoG and/or HEEG, signs that appear to be induced by LC activation even though LC neurons are silent during these states (see Aston-Jones and Bloom, 1981). These observations possibly indicate that LC activation is sufficient to initiate these EEG signs, at least under some circumstances, but LC activation is not necessary for these signs to occur, possibly because other systems are also sufficient, but not necessary, for these processes. The effects of LC lesions on monkey P3 activity can be interpreted in light of our previous

neuroanatomical studies (Morrison and Foote, 1986) demonstrating a very dense NE innervation of monkey parietal cortex. Electrodes over this cortical region exhibited the largest P3 activity prior to LC lesions and showed the largest decrease following LC lesions. It is of interest to note the congruence between these monkey observations and the localization of an important alerting component of a “posterior attention system” in humans to a presumably homologous neocortical area using PET scanning methods (Posner and Petersen, 1990). Our studies of the effects of LC activation on neocortical sensory responsiveness also offer a possible mechanism for certain observations in studies of human attention. In studies of the effects of alertness on responding to discriminated stimuli, Posner (1978) has observed that enhanced alertness leads to more rapid responding, but that incorrect responses are also facilitated. This may be a correlate of our observation that neuronal responses are enhanced in magnitude but at the cost of some of the temporal and perhaps spatial resolution of these responses. The most obvious hypothetical scheme that would incorporate all three of the observations described above is that the LC participates in inducing alertness throughout many brain regions, as manifested by its ability to initiate ECoG desynchronization and intense theta rhythm in HEEG activity. This alert state is composed of changes in responsiveness of individual neurons throughout the brain, including sensory cortices, an example being the changes observed in the spontaneous discharge activity and sensory responses of somatosensory cortical neurons. The alert state is also a precondition, perhaps in some complex way, for subsequent appropriate orienting, attending or other information processing tasks. Acknowledgements

This research was supported by PHS Grant MH40008, AFOSR Grant 90-0325 and by a grant

531

from the John D. and Catherine T. MacArthur Foundation. References Adams, L.M. and Foote, S.L. (1988) Effects of locally infused pharmacological agents on spontaneous and sensoryevoked activity of locus coeruleus neurons. Bruin Res. Bull.. 21: 395-400. Adams. L.M. and Foote, S.L. Effects of locus coeruleus activation on spontaneous and sensory-evoked activity of somatosensory neocortical neurons. In preparation. Adams. L.M. and Geyer, M.A. (1981) Effects of 6-hydroxydopamine lesions of locus coeruleus on startle in rats. Psychopharmacvlv~,73: 394-398. Aston-Jones, G. and Bloom, F.E. (1981) Activity of norepinephrine-containing locus coeruleus neurons in behaving rats anticipates fluctuations in the sleep-waking cycle. J. Neurosci., 1: 876-886. Berridge, C.W. and Foote, S.L. (1991) Effects of locus coeruleus activation on electroencephalographic activity in neocortex and hippocampus. Neuroscience, in press. Britton, D.R., Ksir, C., Britton, K.T., Young, D. and Koob, G.F. (1984) Brain, norepinephrine depleting lesions selectively enhance behavioral responsiveness to novelty. Physiol. Behui>.,33: 473-478. Diner, B.C., Holcomb, P.J. and Dykman, R.A. (1985) P300 in major depressive disorder. Psychiat. Res., 15: 175-184. Dongen, P.A.M. (1981) The human locus coeruleus in neurology and psychiatry. Progr. Neurvbiol. 17: 97-139. Foote, S.L. and Morrison, J.H. (1987) Extrathalamic modulation of neocortical function. Ann. Rec. Neurosci. 10: 67-95. Foote, S.L., Freedman, R. and Oliver, A.P. (1975) Effects of putative neurotransmitters on neuronal activity in monkey auditory cortex. Brain Res., 86: 229-242. Foote, S.L., Aston-Jones, G. and Bloom, F.E. (1980) Impulse activity of locus coeruleus neurons in awake rats and monkeys is a function of sensory stimulation and arousal. Proc. Natl. Acad. Sci. USA, 77: 3033-3037. Foote, S.L., Bloom, F.E. and Aston-Jones, G. (1983) The nucleus locus coerleus: new evidence of anatomical and physiological specificity. Physiol. Rec., 63: 844-914. Gabriel. M., Sparenborg, S.P. and Donchin, E. (1983) Macropotentials recorded from the cingulate cortex and anterior thalamus in rabbits during the “oddball” paradigm used to elicit P300 in normal human subjects. Svc. Neurvsci. Ahstr., 9: 1200. Halgren, E., Squires, N.K., Wilson, C.L., Rohrbaugh, J.W., Babh, T.L. and Crandall, P.H. (1980) Endogenous potentials generated in human hippocampal formation and amygdala by infrequent events. Science, 210: 803-805. Hobson. J.A., McCarley, R.W. and Wyzinski, P.W. (1975) Sleep cycle oscillation: Reciprocal discharge by two brainstem neuronal groups. Science 189: 55-58. Koob, G.F., Thatcher-Britton, K., Britton, D.R., Roberts, D.C.S. and Bloom, F.E. (1984) Destruction of the locus coeruleus or the dorsal N E bundle does not alter the

release of punished responding by ethanol and chlordiazepoxide. Physiol. Behau., 33: 479-485. Lewis, D.A., Campbell, M.J., Foote, S.L., Goldstein, M. and Morrison, J.H. (1Y87) The distribution of tyrosine hydroxylase-immunoreactive fibers in primate cortex is widespread but regionally specific. J. Neurosci., 7: 279-290. Mair, R.G. and McEntee, W.J. (1983) Korsakoffs psychosis: Noradrenergic systems and cognitive impairment. Behau. Brain Res., 9: 1-32. McCarthy, G., Wood, C.C., Allison, T., Goff, W.R., Williamson, P.D. and Spencer, D.D. (1982) Intracranial recordings of event-related potentials in humans engaged in cognitive tasks. Svc. Neurosci. Abstr., 8: 976. Mirsky, A.F. and Duncan, C.C. (1986) Etiology and expression of schizophrenia: Neurobiological and psychosocial factors. Ann. Rec Psychol., 37: 291-319. Morrison, J.H. and Foote, S.L. (1986) Noradrenergic and serotonergic innervation of cortical, thalamic, and tectal visual structures in old and new world monkeys. J. Cvmp. Neurol. 243: 117- 138. Neville, H.J. and Foote, S.L. (1984) Auditory event-related potentials in the squirrel monkey: Parallels to human late wave responses. Bruin Res. 298: 107-1 16. O’Connor, T.A. and Starr, A. (1985) Intracranial potentials correlated with an event-related potential, P300, in the cat. Bruin Res., 339: 27-38. Oke, A.F. and Adams, R.N. (1Y78) Selective attention dysfunctions in adult rats neonatally treated with 6-hydroxydopamine. Pharmacol. Bivchem. Brhai;., 9: 429-432. Pineda, J.A., Foote, S.L. and Neville, H.J. (1987) Long-latency event-related potentials in squirrel monkeys: Further characterization of waveform morphology, topography, and functional properties. Electroencephalogr. Clin. Neurophysid., 67: 77-90. Pineda, J.A., Foote, S.L., Neville, H.J. and Holmes, T. (1988) Endogenous event-related potentials in monkeys: The role of task relevance, stimulus probability, and behavioral response. Electroencephalogr. Clin. Neurophysiol., 70: 155171. Pineda, J.A., Foote, S.L. and Neville, H.J. (1989) Effects of locus coeruleus lesions on auditory, long-latency, event-related potentials in monkey. J. Neurosci., 9: 81-93. Polich, J., Ehlers, C.L., Otis, S., Mandell, A.J. and Bloom, F.E. (1986) P300 latency reflects the degree of cognitive decline in dementing illness. Electroencephalogr. Clin. Neurvphysiol., 63: 138-144. Posner, M.I. (1978) Chronometric Explorations of Mind, Erlbaum, Englewood Heights, New Jersey, 271 pp. Posner, M.I. and Petersen, S.E. (1990) The attention system of the human brain. Ann. Rec. Neurosci., 13: 25-42. Roberts, D.C.S., Price, M.T.C. and Fibiger, H.C. (1976) The dorsal tegmental noradrenergic projection: Analysis of its role in maze learning. J. Cvmp. Physivl. Psycho/., 90: 363-372. Schildkraut, J.J. and Kety, S.S. (1967) Biogenic amines and emotion. Science, 156: 21-30. Wood, C.C. and McCarthy, G. (1986) A possible frontal lobe contribution to scalp P300. In J.W. Rohrbaugh, R. John~

532 son, Jr. and R. Parasuraman (Eds.), Eighth International Conference on Event-Related Potentials of the Brain (EPIC CIII): Research Reports, Stanford, CA. Wood, C.C., Allison, T., Goff, W.R., Williamson, P.D. and Spencer, D.D. (1980) On the neural origin of P300 in man. Prog. Brain Res., 54: 51-56.

Yingling, C.D. and Hosobuchi, Y.A. (1984) Subcortical correlate of P300 in man. Electroencephalogr. Clin. Neurophysiol., 59: 72-76.