Serotonin and Behavior

A General Hypothesis

Barry L. Jacobs and Casimir A. Fornal

Serotonin is an enigma. It is at once implicated in virtually everything, but responsible for nothing. Its name also adds to this mystery. Unlike other neurotransmitters, such as glutamate, acetylcholine, and gamma-aminobutyric acid (GABA), whose names evoke images of test tubes and tiled laboratories, serotonin (5-hydroxytryptamine; 5-HT) conjures up visions of the esoteric and the exotic. This characterization was fostered by early theories and research which linked it to hallucinogenic drug action, schizophrenia, and depression. Modern neuroscience has added to the mystique surrounding 5-HT: Axon terminals containing 5-HT are found in even the remotest reaches of the central nervous system (CNS); the release of 5-HT may not be subject to classical synaptic physiology; and 5-HT acts at a bewildering diversity of pre- and postsynaptic receptor subtypes.

Research over the past several decades has led to the development of specific theories regarding the function of the related brainstem-originating neurotransmitter systems, dopamine and norepinephrine. Thus, as detailed in other chapters in this volume, dopamine appears to be involved primarily in response initiation and reward, whereas norepinephrine is believed to be concerned primarily with vigilance. By contrast, hypotheses regarding 5-HT's role in the CNS remain wide-ranging and include regulation of processes as diverse as cardiovascular and respiratory activity, sleep, aggression and sexual behavior, nutrient intake, anxiety, mood, motor output, neuroendocrine secretion, and nociception and analgesia (67).

This chapter sets forth a general theory of 5-HT function within the CNS. It is based largely upon our own single-unit studies of 5-HT neuronal activity in behaving cats. We believe that the primary function of 5-HT neurons is to facilitate gross motor output in both the tonic and repetitive modes. Concurrently, the system acts to inhibit sensory information processing and to coordinate autonomic and neuroendocrine function with the demands of ongoing motor output. Under certain conditions, when the 5-HT system is inactivated, these relationships are reversed: Tonic motor output is disfacilitated and sensory information processing is disinhibited.

We propose that the diversity of behavioral and physiological processes in which 5-HT has been implicated can be subsumed within this integrative perspective. Additionally, we believe that the elucidation of 5-HT's role in human psychopathology is embedded within its basic biology. Accordingly, this chapter focuses upon the motor aspects of the 5-HT system, with some attention also given to its role in sensory and physiological processes (see also Anatomy, Cell Biology, and Maturation of the Serotonergic System: Neurotrophic Implications for the Actions of Psychotropic Drugs and Electrophysiology of Serotonin Receptor Subtypes and Signal Transduction Pathways).

From the time of its initial description with fluorescence histochemistry, the anatomy of the CNS 5-HT system was known to be extremely widespread (see Anatomy, Cell Biology, and Maturation of the Serotonergic System: Neurotrophic Implications for the Actions of Psychotropic Drugs). Of all the neurotransmitter systems within the vertebrate CNS, the 5-HT system is the most expansive. Nonetheless, this innervation pattern is neither ubiquitous nor nonspecific. Even within a particular target site some portions may receive a very dense 5-HT input, whereas neighboring regions may be only sparsely innervated. Additionally, some brain regions are almost devoid of 5-HT innervation.

Of special note regarding the distribution of 5-HT axon terminals is a preferential targeting of primary and secondary motor areas in the CNS (56). For example, in the rat there is a very dense innervation of the ventral horn, the motor nucleus of the trigeminal (MoV), the facial motor nucleus (MoVII), the substantia nigra, and the globus pallidus. As many as 13% of all the synaptic contacts in MoV of the rat are 5-HT-immunoreactive (54). Another interesting feature of the input to motor areas is its lack of homogeneity. In the ventral horn of the spinal cord, for example, the 5-HT input preferentially innervates motoneurons projecting to axial rather than distal musculature (56). In the brainstem of a variety of species there is dense innervation of motoneurons projecting to the large muscles of the jaw, face, and neck, but the extraocular muscles receive only sparse 5-HT input (56, 60). Also noteworthy is the fact that the input to the cerebellum is relatively sparse. This pattern of differential innervation of motor structures reveals something about the function of the 5-HT system. It implies that 5-HT should be more strongly associated with movements employing gross skeletal muscles rather than those utilizing fine or discrete muscles. Thus, as we have seen, there is a denser input to the medial portion of the ventral horn, where axial motoneurons serving the trunk and limbs are found, as compared to the lateral portions, where distal motoneurons serving paws and digits are found. Similarly, in the brainstem, there is a much denser 5-HT input to MoV and MoVII, controlling jaw and facial muscles, respectively, as compared to the nuclei controlling eye movements. Consistent with this is the relatively weak 5-HT projection to the cerebellum, a structure associated more with the smoothing or adjustment of movements rather than with their direct execution.

If 5-HT neurons in the CNS projected exclusively to motor nuclei, this fact alone would constitute a strong case for the primacy of motor function for this system. There are, however, a multitude of other projections, some quite dense, to non-motor targets such as the hippocampus, the dorsal horn, dorsal column nuclei (DCN), and so on. This anatomical pattern is part of a larger picture that interrelates motor outflow to sensory information processing and to autonomic and neuroendocrine regulation. A couple of examples may help to clarify this point. It is known that 5-HT plays an important role in the regulation of the theta rhythm in the hippocampus and that this rhythm is often related to different types of motor patterns (35, 62). Although the DCN are primary sensory nuclei, the dense 5-HT input to these sites in cats and monkeys selectively innervates those portions of the nuclei involved in motor function (i.e., projections to cerebellum, pretectum, inferior olive, etc.) rather than those portions involved in fine somatosensory discrimination (i.e., projections to VPL of the thalamus) (6).

Finally, it is worth noting that the basic plan of 5-HT cell bodies and axon terminals is a primitive one, found in the simplest vertebrate brains, and one that remains remarkably conserved across phylogeny (45).

One of the earliest findings regarding 5-HT and behavior came from studies employing its biosynthetic precursors L-5-hydroxytryptophan (5-HTP) and L-tryptophan. When these compounds are administered to any of a variety of mammals, a distinctive and complex syndrome, comprised of tonic and repetitive motor outputs, is produced. Its most conspicuous signs are tremor, rigidity, hindlimb abduction, Straub tail, head shakes or "wet dog" shakes, lateral head weaving, and reciprocal treading of the forepaws (20). (Similar effects are also seen in infra-mammalian vertebrates.) It is also clear that most of these motor signs can be elicited by drug treatment very early in ontogeny (e.g., in 3- to 4-day-old rat pups), thus implying that the system is at least partially functional at or near birth (50).

Equally relevant for the present discussion is the less-well-known fact that this "serotonin syndrome" is also seen in human patients administered 5-HT drugs (58). Most important is the fact that the effects observed are restricted almost exclusively to motor signs (myoclonus, tremor, shivering), with few, if any, indications of significant sensory alterations. Consistent with this, various drugs influencing 5-HT neurotransmission have also been reported to produce repetitive chewing or bruxism (tooth grinding) in rats and humans (46, 59).

The actions of 5-HT drugs on more discrete aspects of behavior, such as treadmill-induced locomotion in spinal cats, have also been examined (3). The most prominent action of 5-HT here is to increase the flexor and extensor burst amplitude (a smaller increase is seen in burst duration) during locomotion. In this preparation 5-HT cannot by itself trigger locomotion and is assumed to act by increasing motoneuron excitability (4). In paralyzed rabbits, 5-HTP can enhance or even evoke reciprocal flexor and extensor hindlimb nerve activity, with the dominant effect seen in the flexors (64). We have examined the effects of injecting 5-HT directly into MoV in awake cats, and we have observed increases in both the amplitude of the tonic electromyogram (EMG) of the masseter muscle and the amplitude of an externally elicited jaw-closure (masseteric) reflex (49). Finally, 5-HT agonists or precursors evoke repetitive swallowing in anesthetized rats (5). These latter effects continued to be seen in decerebrate preparations, demonstrating that brainstem neural mechanisms were sufficient for their manifestation.

The varied species in which similar motor effects of 5-HT are seen supports the position that this 5-HT system subserves a common functional role across the vertebrates (20).

The effects of 5-HT on motoneurons has been examined in the rat spinal cord and brainstem (36, 68). By itself, 5-HT produces little or no change in neuronal activity (however, it is important to bear in mind that such studies are typically carried out on animals with reduced excitatory drive to these neurons—that is, either those under anesthesia or in reduced preparations). However, when 5-HT is combined with direct application of excitatory amino acids or with electrical stimulation of dorsal roots or motor cortex, it produces a strong facilitation of neuronal activity. This effect has been characterized as a bistability with a 5-HT-induced shift from a stable hyperpolarized state, with little or no neuronal activity, to a new stable depolarized "plateau" state, with tonic neuronal activity (19).

Similar analyses have also been carried out in a more complex situation, where cortical stimulation is used to elicit rhythmic masticatory-like activity in anesthetized guinea pigs (27). The activity of digastric (jaw opener) motoneurons is directly facilitated by the iontophoretic application of 5-HT, but, as above, only in the presence of glutamate or electrical excitation of these neurons. Additionally, iontophoretic application of 5-HT can facilitate and bring to threshold rhythmic digastric motoneuronal discharges during subthreshold repetitive cortical stimulation. A similar picture is seen when one examines the influence of 5-HT upon the neuronal mechanisms mediating fictive locomotion (swimming) in the isolated spinal cord of the lamprey (15). When applied to the spinal cord or to reticulospinal neurons, 5-HT elicits a depression of the afterhyperpolarization (AHP) that normally follows the action potential. Because the AHP is the primary factor in determining discharge frequency, this depression produces an increase in motoneuron discharge. If 5-HT is applied to the solution bathing the isolated spinal cord during fictive locomotion, motoneuronal bursts become more intense and longer, and the burst rate increases. 5-HT also modulates the intersegmental phase delay in the lamprey spinal cord, an integral component of coordinated swimming (18).

Although the literature is not as extensive and the data perhaps not as clearcut as its effects on motoneurons, 5-HT consistently is reported to inhibit primary sensory neurons in the forebrain. Once again, these studies are conducted primarily in anesthetized rats (see ref. 38 for a review). Iontophoretic application of 5-HT to relay neurons in the rat dorsal lateral geniculate nucleus (LGN) suppressed spontaneous activity as well as that evoked by light flashes or electrical stimulation of the optic chiasm (51). In related experiments, high-frequency stimulation of 5-HT cell bodies produced an inhibition of LGN activity lasting for periods as long as several tens of seconds (28).

Comparable effects of 5-HT are seen in neocortex of anesthetized rats. When neurons are activated by somatosensory stimuli, iontophoretically applied 5-HT consistently suppresses this evoked activity to a greater degree than background or basal activity, resulting in a decrease in signal/noise ratio (66). Similar effects are seen when 5-HT's actions are examined on primary visual cortex neurons. The responses evoked by stimuli moving across the visual field were suppressed by 5-HT to a greater extent than was the basal activity of these cells (65). Of particular interest is the fact that norepinephrine consistently produces a complementary effect to that of 5-HT in both the LGN and neocortex—that is, an increase in signal/noise ratio (see ref. 38 for a review).

In summary, 5-HT's effect on alpha motoneurons is one of excitation, whereas its effect on primary sensory neurons is that of inhibition.

Over the past 20 years, much of the basic neurophysiology of 5-HT neurons in the various brainstem raphe nuclei—dorsalis (DRN), medianus (NRM), magnus (MRN), and pallidus (NRPa)—has been worked out (22). The neurons are autoactive, discharging in a stereotyped, almost clock-like manner, with an intrinsic frequency of 1–5 spikes/sec. The membrane properties that accompany this slow regular activity have been described, as have the ionic currents and channels mediating it (see Electrophysiology of Serotonin Receptor Subtypes and Signal Transduction Pathways). Additionally, at least in the rat, these basic neuronal properties are manifested early in development (3–4 days prenatal). Finally, there is a negative feedback mechanism which limits 5-HT neuronal activity. As activity increases and 5-HT is released locally from dendrites or axon collaterals, it acts upon somatodendritic 5-HT "autoreceptors" to inhibit neuronal activity. The mechanism functions only under physiological conditions in the sense that it is inoperative with low levels of neuronal activity, but becomes increasingly engaged as neuronal activity increases (see below) (23). Dysfunction of this regulatory mechanism may be implicated in some forms of human pathology, and therefore drugs targeted at these autoreceptors provide a potentially important site of therapeutic intervention.

One of the first significant discoveries about brain 5-HT neurons was that their activity was dramatically altered across the sleep–wake–arousal cycle (23, 39). From a stable, slow, and regular discharge pattern of, for example, 3 spikes/sec during quiet waking, neuronal activity displays a gradual decline as the animal becomes drowsy and enters slow-wave sleep. A decrease in the regularity of firing accompanies this overall slowing of activity during sleep. During rapid eye movement (REM) sleep, 5-HT neuronal activity falls silent, but in anticipation of awakening, neuronal activity returns to its basal level, or above, several seconds prior to the end of the REM sleep epoch. During an aroused or active waking state, discharge rate may increase to 4 or 5 spikes/sec.

Because brain 5-HT has been implicated in such a variety of physiological processes (thermoregulation, cardiovascular control, respiration, etc.) and behaviors (aggression, nutrient intake, sleep, etc.), it was deemed imperative to examine 5-HT neuronal activity under a wide diversity of conditions. Accordingly, while recording the activity of 5-HT neurons in the DRN, MRN, or NRM, we exposed cats to the following conditions: loud noise, physical restraint, a natural enemy (dog), a variety of mildly painful stimuli, a heated environment or systemic administration of a pyrogen, drug-induced increases or decreases in blood pressure, or insulininduced glucoprivation.

Exposing a cat to 100 dB of white noise for 15 min elicits strong sympathetic activation, as indicated by significant increases in tonic heart rate and plasma norepinephrine levels. It also evokes a stereotyped behavioral response of crouching, with ears flattened. Despite this, during the presentation of this stimulus, the activity of DRN 5-HT neurons was not significantly different from that observed during an undisturbed active waking baseline (70).

Similarly, when cats were physically restrained for 15 min, this also evoked a strong sympathetic activation. Struggling and vocalizations during the restraint provided additional behavioral evidence for the stressful nature of the stimulus. Once again, despite the behavioral and physiological activation produced by restraint, the activity of DRN neurons was not significantly different from that observed during the pre-stress baseline condition (70).

In the final experiment in this series, a dog was brought into proximity of the cat for 5 min. This evoked the typical stereotyped feline defense reaction of arched back, facing broadside, piloerection, often growling and hissing, and physiologic indices of sympathetic activation. However, despite this behavioral and physiological activation evoked by the dog, the discharge rate of DRN neurons was once again unchanged from that observed during an undisturbed active waking baseline (70).

A related series of studies examined the response of 5-HT neurons in the NRM to a variety of phasic or tonic painful stimuli. There was no change in neuronal activity in response to any of these stimuli, when compared to an active waking baseline (2). There was also no change in activity in response to the systemic administration of morphine in a dose that produced analgesia (2). These results are consistent with a recent study reporting that identified NRM 5-HT neurons in the rat are not activated by painful stimuli eliciting the withdrawal reflex (47).

Thus, as a whole, these data indicate that the activity of 5-HT neurons cannot easily be driven above the level observed during an undisturbed active waking baseline. This is in spite of the fact that the stimuli employed in these studies evoked a variety of different forms of behavioral arousal and strong sympathetic activation. Finally, contrary to the present results with 5-HT neurons, the same stimuli were effective in strongly activating a neighboring brainstem neurochemical system, the noradrenergic neurons of the locus coeruleus (LC) (reviewed in ref. 21).

Paralleling these studies of behavioral/environmental challenges, we also examined the response of DRN 5-HT neurons during perturbation of several physiological regulatory systems.

The activity of DRN neurons was examined in response to both increased ambient temperature and pyrogeninduced fever, stimuli eliciting opposite thermoregulatory responses (11). For environmental heating, the temperature in the experimental chamber was raised to 43°C. The activity of DRN neurons remained unaffected during the interval when ambient temperature was increased from 25°C to 43°C. During this initial phase of heating, no appreciable behavioral or physiologic responses were seen. However, following prolonged heat exposure, cats displayed intense continuous panting, relaxation of posture, and a progressive rise in body/brain temperature (range: 0.5°C to 2.0°C). Once again, however, no change in DRN single-unit activity was observed. DRN neuronal activity was also examined during the febrile response elicited by systemic administration of a synthetic pyrogen (muramyl dipeptide). Following drug administration, body/brain temperature began to increase within 30 min, reached a peak at 1–2 hr, and returned to predrug level by 6 hr. The peak elevation of body temperature that was attained was typically 1.5°C to 2.5°C. Once again, no change in DRN single-unit activity was observed during any phase of the pyrogen-induced febrile response.

The response of DRN neurons was also studied in relation to changes in blood pressure induced by peripherally acting drugs: Phenylephrine induces increases in blood pressure, sodium nitroprusside phasically decreases blood pressure, and hydralazine produces prolonged hypotension (13). Over a range of 20- to 70-mmHg increases and 10- to 50-mmHg decreases in mean arterial pressure, there was no change in the discharge rate of DRN 5-HT neurons. This is despite the fact that these blood pressure changes were of sufficient magnitude to produce significant reflexive changes in heart rate and plasma catecholamines.

The final study in this series manipulated blood glucose levels in both directions. The activity of DRN 5-HT neurons in behaving cats was not significantly altered by bolus injection of glucose (500 mg/kg, i.v.) that elevated blood glucose levels threefold (12). Likewise, the activity of these neurons was not significantly affected by the administration of a dose of insulin (2–4 IU/kg, i.v.), which lowered blood glucose by 50% or more, or following the rapid reversal of this hypoglycemia by subsequent glucose administration.

As we have seen with environmental stressors, physiologic challenges to the animal also fail to significantly activate brain 5-HT neurons above the level seen during an undisturbed active waking state. Once again, this is in spite of the fact that these manipulations produce behavioral arousal as well as activate the organism's sympathetic nervous system. And as with environmental stressors, these same physiologic challenges do significantly activate noradrenergic neurons in the cat LC (reviewed in ref. 21).

A fundamental feature of REM sleep is a paralysis mediated by inhibition of motoneurons controlling antigravity muscle tone. Because the activity of 5-HT neurons is totally suppressed during REM sleep, we examined the possibility that there might be a relationship between these two phenomena. Lesions of the dorsomedial pons produces a condition which permits investigation of this issue. Cats with this lesion enter a stage of sleep which by all criteria appears to be REM sleep except antigravity muscle tone is present and the animals are thus capable of movement and even coordinated locomotion (26) [this condition has also been observed in humans (34)].

During both waking and slow-wave sleep, the activity of DRN 5-HT neurons in these pontine-lesioned cats was similar to that in normal animals (61). However, when these animals entered REM sleep, neuronal activity increased instead of displaying the decrease typical of this state. Those animals displaying the greatest amount of muscle tone and overt behavior during REM sleep showed the highest levels of neuronal activity, with some of their 5-HT neurons discharging at a level approximating that of the waking state.

If the cholinomimetic agent carbachol is microinjected into this same pontine area, a condition somewhat reciprocal to non-atonia REM sleep can be produced. These animals were awake, as demonstrated by their ability to track stimuli visually, but were otherwise paralyzed. However, the activity of their DRN 5-HT neurons was silent (57). In the same study, we found that peripheral paralysis, induced by blocking transmission at the neuromuscular junction, had no effect on 5-HT neuronal activity, but that a centrally acting muscle relaxant also completely suppressed 5-HT neuronal activity.

These data suggest that a strong relationship exists between tonic motor activity and 5-HT neuronal discharge. More recently, we observed a relationship between 5-HT neuronal activity and another general type of motor output. When cats engage in a variety of types of central pattern generator (CPG)-mediated oral–buccal activities, such as chewing/biting, licking, or grooming the body surface with the tongue, approximately one-fourth of DRN and MRN 5-HT neurons increase their activity by as much as two- to fivefold (22, 48). (The remaining 5-HT neurons simply maintain their state-related or tonic-motor-related clock-like activity.) The increased neuronal activity often precedes the onset of movement by several seconds, but typically terminates coincident with the offset of the behavior. It is also occasionally phase-locked to the repetitive responses. Some of these neurons are also activated by somatosensory and proprioceptive stimulation of the head and neck area. During a variety of other purposive episodic or phasic movements, even those involving the oral–buccal area, no increase in neuronal activity is observed; in fact a slight decrease is often seen. Under some conditions involving gross behavior, a dramatic decrease in neuronal activity is observed. For example, if an arousing stimulus elicits an orienting response (evidenced by suppression of overt behavior and foveation toward the source of the stimulus), 5-HT neuronal activity in the DRN or MRN may fall silent for several seconds and then resume its normal activity (25).

A somewhat complementary picture emerges when one examines the activity of NRP 5-HT neurons, the primary source of 5-HT innervation of ventral horn motoneurons. These 5-HT neurons are activated (two- to threefold) in association with repetitive behaviors mediated by spinal cord CPGs, such as treadmill locomotion or hyperpnea (induced by exposure to CO2) (63). In some cases there is a strong positive correlation between neuronal activity and speed of locomotion or rate/depth of respiration. Neuronal activity may also increase in association with tonic motor changes such as postural shifts. As with DRN neurons, the activity of NRP 5-HT neurons is occasionally phase-locked to the repetitive responses.

These cellular data from behaving animals have led us to the following conclusions. There is a general relationship between level of tonic motor activity and 5-HT neuronal activity. Superimposed upon this in some neurons is an additional relationship in which a further, often dramatic, neuronal activation is seen in association with repetitive CPG-mediated behaviors. Reciprocally, during the active inhibition of gross behavior (e.g., during orientation), 5-HT neuronal activity is suppressed. We hypothesize that the processing of sensory information is inhibited during the activation of tonic or repetitive motor activity but is disinhibited during the suppression of gross motor outflow.

The study of motor and sensory systems is typically carried out separately. However, there is abundant information that these two systems often interact to influence each other. For example, in both animals and humans, there is evidence that sensory transmission is suppressed during gross bodily movements (8, 9). This is illustrated by the fact that just prior to and during arm movements in the monkey, somatosensory evoked potentials elicited by an irrelevant peripheral stimulus are depressed by as much as 60–70% at the lemniscal, thalamic, and cortical levels (9). Additionally, the single-unit response of thalamic and cortical neurons to somatosensory stimuli is suppressed during locomotion in the rat (55). Relevant to these physiological results, the anatomy of 5-HT inputs to sensory neurons indicates that they exert a direct influence at an early stage of processing. For example, in the primate visual cortex, 5-HT preferentially innervates layer IV, suggesting an effect on those cortical neurons that are the direct recipients of the inputs deriving from the LGN (10).

We believe that a similar motor–sensory interaction mechanism may explain 5-HT's well-established involvement in analgesia. As noted above, 5-HT neurons in NRM are neither activated by a variety of painful stimuli nor activated by an analgesic dose of morphine (2, 47). Thus, the NRM 5-HT system does not constitute a pure primary antinociception or analgesic system per se. However, under physiological conditions, the suppression of nociception by 5-HT, at both forebrain and spinal levels, may occur as a concomitant of tonic or repetitive motor outflow. Reciprocally, as discussed above, sensory transmission may be enhanced during periods when the activity of 5-HT neurons is suppressed—for example, during orientation. Consistent with this, it is well known that ponto-geniculate-occipital cortex (PGO) waves are held under tonic 5-HT inhibition (53) and that these potentials can be evoked by exposing the behaving animal to strong phasic stimuli which elicit orienting responses (7). Additionally, under pharmacologic conditions, when 5-HT neurotransmission is compromised, this disinhibition of sensory processing is manifested as increased responsiveness or enhanced excitability or sensitivity in a variety of paradigms, including nociception (31, 69) (see further discussion of this issue in the final section of this chapter).

In a similar manner, we hypothesize that 5-HT neurons facilitate the well-known sympathetic activation that accompanies, and even anticipates, motor activity (16). For a recent review of the sympathoexcitatory effects of 5-HT in the intermediolateral column of the spinal cord, see ref. 37. 5-HT has also been shown to facilitate the activity of respiratory (phrenic nerve) motor neurons in the rat spinal cord (e.g., see ref. 42). In this context, recall, as described above, that 5-HT neuronal activation often precedes increases in motor activity or muscle tone. Thus, an important ancillary role of 5-HT neurons may be the activation of physiological regulatory systems, such as those controlling cardiovascular activity and respiration, in the service of increasing motor demands.

5-HT function in invertebrates provides striking parallels to many aspects of the data from vertebrates described in this chapter. This is impressive, in view of the enormous differences in their gross bodily morphology, ecological niche (terrestrial versus aquatic for the invertebrates in these studies), and general organizational pattern of their nervous systems (brain versus ganglionic). Several examples will help to make this point. First, direct injection of 5-HT into the systemic circulation of several arthropods results in a general motor change (29, 33). Second, when it has been examined, for example in lobsters, 5-HT neurons are found to be endogenously active (33). In both lobsters and aplysia, 5-HT neurons discharge with a slow and regular pattern (0.5–1.0 Hz) that can increase to 2–5 Hz during feeding or with postural changes (30, 33). Furthermore, in Aplysia the 5-HT metacerebral cell alters its somewhat regular firing pattern to become phase-locked to oral–buccal movements during feeding (30). Third, in several molluscs, arthropods, and annelids, 5-HT modulates, rather than mediates, motor outflow, often by acting on CPGs (17). The involvement in motor control appears to be with both tonic (e.g., posture) (29) and repetitive (e.g., swimming, biting, etc.) outputs (17, 43). Finally, in Aplysia and leeches, 5-HT exerts its effects on behavior at multiple levels (e.g., directly on muscles, on CPGs, on the cardiovascular system, etc.) (30, 32).

As detailed in other chapters in this volume, 5-HT has been implicated strongly in the etiology and/or treatment of several forms of psychopathology, most notably depression and obsessive–compulsive disorders (OCDs). Do these results from basic research on brain 5-HT provide any insights into these clinical disorders? Recall that the activity of the brain 5-HT neurons is at an elevated level during increased tonic motor output and that for a subgroup of neurons it achieves an even higher level of activity during repetitive motor acts. Thus, if there is a deficit in 5-HT neurotransmission in at least some forms of depression, then it might be beneficial for such patients to increase their tonic motor activity or to engage in some form of simple repetitive motor task, such as riding a bicycle or jogging. Consistent with this, there are scattered reports of the salutary effects of jogging or other forms of exercise for depressed patients (e.g., see refs. 14, 41, and 44) and, more generally, reports of exercise exerting mood-altering effects in nondepressed subjects (e.g., see refs. 40 and 52).

On the basis of our research we have also arrived at a novel way of viewing OCDs. Because repetitive or compulsive motor acts increase 5-HT neuronal activity, we believe that patients with OCD may be engaging in such behaviors as a means of self-treatment. In other words, they are activating their brain 5-HT system in a physiological manner in order to derive some (as yet unknown) benefit or rewarding effect. Treating them with drugs that block the reuptake of 5-HT into the presynaptic neuron accomplishes the same neurochemical endpoint and thus allows them to disengage from time-consuming, socially unacceptable, and often physically harmful behaviors. (The same case could also be made for repetitive obsessional thoughts, but this, obviously, is difficult to examine in the laboratory.)

Finally, because 5-HT neuronal activity, and therefore neurotransmitter release, is under negative feedback control, this provides a potentially productive avenue for drug intervention in the clinic. Administration of precursors of 5-HT, such as tryptophan or 5-HTP, are of limited value for elevating synaptic levels of brain 5-HT because they produce a compensatory decrease in neuronal activity through this negative feedback mechanism (1). However, if these treatments were combined with a low dose of an autoreceptor (5-HT1A) antagonist drug, they might be of therapeutic value and thus capitalize on the advantage of employing natural biological precursors rather than synthetic drugs. Additionally, because of the ratedependency of the feedback mechanism, our data suggest that autoreceptor antagonist drugs might be ineffective in quiescent, lethargic, or somnolent patients, but quite effective in spontaneously active patients or perhaps those activated by artificial means.

During an undisturbed waking state, brain 5-HT neurons discharge in a slow and rhythmic manner that is a manifestation of their endogenous pacemaker activity. This regular firing during waking creates a steady synaptic release of 5-HT which provides a tonic excitatory drive that modulates motor system neuronal activity. During gross repetitive motor behaviors that are mediated by brainstem and spinal cord CPGs, subpopulations of 5-HT neurons are activated (Fig. 1), attaining discharge levels several times greater than that observed during undisturbed waking. This activation, seen in association with chewing, grooming, running, and so on, is sometimes phase-locked to the cycling motor output. The distribution of 5-HT axon terminals in the spinal cord and brainstem is consistent with 5-HT's involvement in patterned movement employing gross skeletal muscles rather than those movements utilizing fine or more discrete muscles.

Several important functions may be served by these 5-HT inputs to motor structures. They may smooth motor outputs and may also obviate the need for continuous repetitive excitatory inputs to maintain a continuous output in motor systems. By augmenting weak or polysynaptic inputs, 5-HT may also bring motoneurons to threshold. The anticipation of motor activity by 5-HT neurons suggests that they may serve a priming function for motor output. 5-HT may also serve a timing or integrative function. The simultaneous inhibition of "irrelevant" sensory information processing acts to suppress inputs that might disrupt motor output (Fig. 1). Reciprocally, when 5-HT neuronal activity is phasically decreased, for example, during orientation, this serves to sharpen sensory function while disfacilitating tonic or repetitive motor output and thereby preventing it from disrupting sensory processing. Furthermore, motoneurons are now poised to respond phasically to discrete excitatory inputs. Finally, 5-HT's involvement in autonomic and neuroendocrine regulation serves a support function for the demands of changes in the level of motor output (Fig. 1), such as (a) increased oxygenation of the blood and increased blood flow to skeletal muscles or (b) increased carbohydrate consumption for maintaining a stable glucose supply to the brain.

It is parsimonious to hypothesize that 5-HT serves an integrative and overarching function in the CNS, rather than to assume that it is discretely and separately involved in a diversity of behavioral and physiological processes. The apparent involvement of 5-HT in this variety of functions is attributable to the fact that it is a widely projecting system that exerts a biasing influence over its target structures. In most experimental studies in this field the level of 5-HT synaptic transmission is perturbed far beyond the physiological range achieved under environmental or biological conditions, such as those described above. This comes about because, typically, the manipulations grossly influence 5-HT either by destroying 5-HT neurons, inhibiting its synthesis, blocking its receptors, preventing its reuptake into the presynaptic terminal, or by precursor loading. This, in turn, either by causing a general increase in motor activity or skeletal muscle tone, or by making the organism generally overreactive to environmental stimuli, biases whatever behavioral or physiological output is under examination (Also see Mesocorticolimbic Dopaminergic Neurons: Functional and Regulatory Roles and Central Norepinephrine Neurons and Behavior for behavioral overviews of other monoamine systems.)

The authors' research described in this chapter was supported by grants from the AFOSR (90-0294) and the NIMH (MH 23433). Special thanks to Ms. Arlene Kronewitter for preparing this manuscript.

published 2000