In this section, we will first discuss two multiple-component models of arousal, each of which has received a good deal of empirical support. Second, we will discuss the action of three 'arousing' neurotransmitter systems, each arising in sub-cortical centers and potentially capable of modulating cortical activity. Our discussion will then focus upon one additional neurotransmitter system, the noradrenergic system, and we will review the evidence that suggests that this system in particular exerts an influence on attention. The section will conclude with a discussion of cortical influence on subcortical structures.
Several authors have followed Lacey's (1967) lead in proposing multiple component arousal systems in which two components are proposed to be mutually antagonistic. Two models in particular, those of Gray (1975) and Pribram and McGuinness (1975), have received a great deal of scrutiny and empirical support.
Gray's model can be thought of as having three major components:
1. A physiological arousal system (NAS) centered on the reticular formation which is responsible for the intensity dimension of behavior.
2. A behavioral activating system (BAS) centered on the medial forebrain bundle, lateral hypothalamus, rostral septal area and amygdala, responsible for the selection and organization of all active behavioral responses, and which responds to reward- conditioned stimuli.
3. A behavioral inhibition system (BIS) which is centered on the hippocampus and medial septal area, and which attempts to suppress all ongoing operant behavior whenever the organism is faced by punishment-conditioned or novel stimuli.
The BAS and BIS are mutually antagonistic, but both have direct positive input to the physiological arousal system. Imagine yourself walking down the sidewalk, and finding some cash. In Gray's terminology, cash is a reward-conditioned stimulus, and would serve to energize the BAS. You would be likely to exhibit approach behavior, that is, pick up the cash. As a result of the BAS activity, you would also receive a positive input to the NAS, that is, there would be some physiological excitement (no doubt dependent upon the amount of cash). Now, imagine yourself walking along, turning a corner, and finding a snarling dog in front of you. In this case, you would be faced with a punishment-conditioned stimulus, and your behavior (walking) would be likely to be inhibited. You would again receive positive input to the NAS, resulting in physiological excitement (in this case probably dependent upon the size of the dog).
Pribram and McGuinness (1975) have also proposed a model having three energetic components, these being:
1. An arousal system located in the reticular formation and anterior hypothalamus, which is controlled by the amygdala. This system controls all phasic physiological responses, and in particular is responsible for the orienting reflex.
2. An activation system which controls the organism's tonic readiness to respond, and which is located in the medial forebrain bundle, lateral hypothalamus and basal ganglia.
3. A coordinating system which demands effort on the part of the organism, and which is centered on the hippocampal and septal areas. This system is responsible for the coordination of arousal and activation in establishing the more difficult relationships between perception and action.
Hence, according to Pribram and McGuinness (1975), the effort mechanism both monitors and controls performance. It "asks questions" such as "Am I too excited or too fatigued to perform this task?", "Am I being distracted by irrelevant information?", "Am I locking on to one solution, and failing to consider all possible alternatives?", and "Am I trying as hard as I can?". Depending upon the answers to those questions, resources are reallocated accordingly.
From the preceding descriptions of these two models, we can see that the chief difference between them lies in the roles assigned to the amygdala and hippocampal circuit. For Pribram and McGuinness, the amygdala is primarily responsible for the controlling of all phasic responses to input, while Gray contends that the amygdala is involved in the selection and organization of behavioral responses, that is, that it is more closely tied to activation rather than arousal in Pribram and McGuinness's terminology. On the role of the hippocampus, Gray proposes it to have an inhibitory responsibility. Gray's argument is based upon a demonstrated association between inhibitory behavior and hippocampal theta (Gray, 1977), and the finding of response specificity in hippocampal theta (Gray, 1978). Pribram and McGuinness agree with Gray in ascribing an inhibitory role to the hippocampus, in the sense that it performs a coordinating role, but claim that it does so by suppressing inappropriate stimulus-response relationships.
Despite the idiosyncrasies of the models described above, a common characteristic is that both associate the reticular system with influence on physiological arousal responses. Since the identification of the reticular formation however, evidence has accumulated to specify a degree of neuro-chemical differentiation among ascending neurotransmitter systems. This has led to the notion that it is now perhaps more profitable to focus attention upon the location and type of influence exerted by these pathways, rather than to consider the potential influence of the undifferentiated reticular structure (Robbins, 1986).
Robbins has suggested that investigation of the multiplicity in ascending neurotransmitter pathways may illuminate the multidimensional nature of arousal. Our discussion here will focus upon the serotonergic (5-HT), cholinergic (ACh), and dopaminergic (DA) systems. The noradrenergic (NA) system will be discussed in greater detail separately.
The central serotonin (also called 5-hydroxytryptamine or 5-HT) projections arise in the dorsal and medial raphe nuclei (Azmitia, 1978). Terminal innervation, which is inhibitory, targets frontal cortex and neocortex, septal and hippocampal regions and the basal ganglia. With the exception of the latter, this topography is strikingly similar to that of the noradrenergic system. Indeed, it has been suggested (Zhang et al., 1995) that there is reciprocity in 5-HT and NA pathways in the support of alertness and responsivity to perceptual input. Similarly, electrophysiological activity in 5-HT neurons in the mesencephalon is directly related to behavioral arousal, a finding which is also common to NA (Robbins, 1986). Furthermore, as is the case with NA, the 5-HT system is responsive during stress (Feldman, Conforti & Weidenfeld, 1995), and has been implicated in behavioral inhibition resulting from anxiety-producing situations (Iversen, 1983). Interestingly, Gray (1982) emphasizes a role for NA in behavioral inhibition. There is one further commonality between the systems — in respect to the control of sleep. First, it has been demonstrated that lesions of the raphe (the central source of 5-HT) produce insomnia. Using cats, Jouvet and Renault (1966) destroyed 80 to 90 percent of the raphe, and observed complete insomnia for 3 to 4 days. Minimal restoration of slow wave sleep, but not REM sleep subsequently occurred. Jouvet (1968) further demonstrated that raphe lesions deplete cortical 5-HT, and that the amount of 5-HT present in the brain was positively correlated with the amount of time the animals spent sleeping. In addition, it has been found that administration of PCPA, which acts to limit 5-HT synthesis, also suppresses sleep (Mouret, Bobillier & Jouvet, 1968). NA mechanisms have also been implicated in the control of sleep, in particular, in regulating the cyclicity of REM sleep (Hobson, McCarley & Wyzinski, 1975) and in the control of waking mechanisms (Jouvet, 1972).
There are three primary sources of central ACh. The neocortex receives an extrinsic projection from the basal forebrain, while a projection to the hippocampus arises in the medial septum (Mesulam, Mufson, Levey & Wainer, 1983). In addition, intrinsic cortical innervation arises from cell bodies in cortical layers. Mesulam and Mufson (1984) have also described the afferents to the basal forebrain. In addition to subcortical inputs, they also found reciprocal pathways from the entorhinal cortex, the medial temporal cortex and the orbitofrontal cortex. Importantly, Robbins (1986) argues that these regions may be able to alter the activity of the ACh they receive and ACh innervation of the entire cortex.
ACh has been demonstrated to exert both excitatory and inhibitory effects upon the evoked potentials of many cortical regions, and appears to produce a relatively long-lasting effect on target cells, wherein the response of the target cell to its other inputs is exaggerated (Stone, 1972). ACh may also be involved in regulating discrimination performance in humans and other mammals (Bartus, 1980). ACh receptors are classified as being either nico-tinic or muscarinic, and hence, it is of further interest to note that nicotine has been shown to improve attention (Wesnes & Warburton, 1983), while scopo-lamine (a muscarinic antagonist) has been shown to detrimentally affect vigilance, attention and learning (Kopelman, 1985). Robbins (1986) argues that in combination, the findings reported above may be interpreted as suggesting a role for ACh in maintaining cortical arousal at an optimal level for cue discrimination. Robbins however also notes the work of Stanes, Brown and Singer (1976), which found extreme dose (anticholinesterase) specificity in the effects of ACh, suggesting that performance improvements occur within a narrow range of ACh activity.
In addition to being widespread in the cortex, ACh is also prevalent in the amygdala, where it appears to contribute to aggression. Hernandez-Peon, O'Flaherty and Mazzuchelli-O'Flaherty (1967) elicited both rage and flight by applying ACh to the amygdala. The affective reaction depended upon the precise locus of ACh application. On a final note, recent reports (LaBerge, 1995; Steriade, McCormick & Sejnowski, 1993) have implicated ACh path ways in the governance of slow-wave sleep, and possibly also in the consolidation of information acquired during the waking state.
Considerable overlap can be seen in the projection areas of NA, 5-HT and ACh. While dopaminergic projections bear some similarity to those of ACh, they are grossly dissimilar to those of NA and 5-HT. The primary central dopamine projection arises in the ventral tegmental region (Robbins & Everitt, 1982) and specifically targets the frontal cortex (the only cortical region to receive a substantial ascending DA input) and those elements of the limbic system having connections to the basal ganglia. A functional similarity to the ACh system may exist in that cortical feedback loops in the DA system could regulate mesencephalic DA activity (Nauta & Domesick, 1984).
There is some controversy in regard to the role of DA as a mediator of cortical arousal. Whereas Jacobs (1984) found that DA activity in the substantia nigra does not covary with changes in the waking state, and DA does not appear to play a major role in mediating activity in sensory systems (Ljungberg & Ungerstedt, 1976), other researchers (e.g., Trampus, Ferri, Adami & Ongini, 1993) have argued that DA is involved in the control of tonic cortical arousal, and as a mediator of stress (Imperato, Puglisi-Allegra, Casolini & Angelucci, 1991) .
No such controversy exists in regard to the importance of the DA system as a regulator of motor activation. It has long been known that the motor symptoms of Parkinson's disease result from the degeneration of dopaminer-gic neurons in a pathway from the substantia nigra to the caudate nucleus. Lesions of this pathway in rats have resulted in a response and postural bias and paw preference ipsilateral to the side of the lesion (Carli, Evenden & Robbins, 1985; Evenden & Robbins, 1985). Since the right hemisphere of the brain controls the left side of the body, and vice versa, the interpretation of this finding is that the ipsilateral preference results from a contralateral deficit.
We have chosen to consider the noradrenergic system in far greater detail than the 5-HT, ACh and DA systems. The reason for this being that a considerable amount of diverse evidence has converged on the idea of the NA system having specific effects upon arousal and attention.
Within the central nervous system, NA cells are distributed in the medulla of the hindbrain, and more densely in the locus ceruleus of the pons. Two primary ascending pathways originate in the locus ceruleus, the dorsal pathway innervating the neocortex, hippocampus, thalamus, cerebellum, certain portions of the hypothalamus and limbic system, and the ventral pathway which projects mainly to the limbic system and hypothalamus. Inputs to the locus ceruleus originate primarily in the visceral centers of the medulla, the reticular formation and the limbic system, including the septum. Locus ceruleus cells appear to respond to polymodal stimuli as a function of their intensive rather than spatial or temporal properties (Watabe, Nakai & Kasamatsu, 1982). For example, they respond to novel light flashes rather than to lines at particular orientations (as is the case with simple and complex cortical visual cells).
The wide-ranging influence of the dorsal pathway provides NA with diffusion sufficient to exert a modulatory effect on many diverse regions simultaneously. This is a topographic organization that would be expected from a system with some general function such as arousal. However, there is more precise evidence that associates NA with such a function.
The influence of NA in its terminal regions is to inhibit spontaneous activity in those areas, leading to an increase in the signal to noise ratio for inputs to a target cell, thereby resulting in a greater evoked potential to a sensory event (Segal, 1985). However, Woodward, Moises, Waterhouse, Hoffer and Freedman (1979) have provided strong evidence to suggest that the modulatory function of NA is to increase the effects of other inputs such that the current activity of a cell, whether facilitatory or inhibitory, is further accentuated. This view was originally proposed by Kety (1970) and has since been substantiated in a number of studies (e.g., Foote, Aston-Jones & Bloom, 1980; Kasamatsu & Heggelund, 1982).
In tests of NA influence on the performance of certain tasks, the most frequent conclusion has been an association of NA with attentional and arousing functions. Carli, Robbins, Evenden and Everitt (1983) found that an 84% reduction in cortical NA (due to 6-OHDA lesioning of the dorsal pathway) had no effect on rats' performance in spatial and brightness discrimination tasks when stimuli were presented slowly and regularly, but that performance was impaired when stimuli were presented at faster more unpredictable rates in the presence of white noise. The authors interpreted their results as implicating the ceruleo-cortical NA system in the preservation of discriminative accuracy under conditions of elevated arousal.
Feldman, Conforti and Weidenfeld (1995) have documented mechanisms by which NA may mediate the stress response. Robbins and Everitt (1982) have also reviewed evidence indicating that NA turnover increases following an organism's exposure to a variety of stressors. However, they emphasized that NA depletion occurs in rats faced with inescapable electric shock, that is in conditions of learned helplessness, but no such depletion is found when the shock is escapable (Anisman, Kokkindis & Sklar, 1981). Hence, the activity of the central NA system does not simply seem to be driven by environmental input, but is affected by efforts to cope on the part of the organism.
Aston-Jones (1985) has argued that the locus ceruleus system is responsible for the control of vigilance. This argument was based on a study of electrophysiological activity of NA neurons in the locus ceruleus of unanes-thetized rats. In addition, McCormick (1989) has reviewed evidence demonstrating that NA is at least partially responsible for the disruption of rhythmic oscillations in thalamocortical activity (which are associated with drowsiness) and for their replacement with a state of excitability that is consistent with cognition. In the following two sections we will consider studies of hemispheric specialization that lend further support to the argument that attention may be influenced by noradrenaline.
In experimental tasks employing warning signals, subjects display a number of characteristic physiological responses to the warning signal. Among these are heart rate deceleration, a galvanic skin response, and a slow negative EEG shift, termed contingent negative variation (CNV). It has been shown that right hemisphere (RH) lesions may result in the disruption of these responses. For example, Yokoyama, Jennings, Ackles, Hood and Boller (1987) have shown that heart rate deceleration in RH patients is less pronounced than in left hemisphere (LH) patients and controls. Similarly, Heilman, Watson and Valenstein (1985) showed that RH lesions in both humans and monkeys disrupt normal galvanic skin responses.
Other physiological data suggest a right hemisphere specialization for the sustaining of attention. Deutsch, Papanicolaou, Bourbon and Eisenberg (1987) reported greater blood flow to the right rather than to the left frontal regions during conditions of sustained attention, as did Cohen et al. (1987). A similar finding was reported by Pardo, Fox and Raichle (1990) who also found increased blood flow to the right but not left superior parietal cortex in both visual and somatosensory vigilance conditions. Hemispheric asymmetry in alpha desynchronization has also been observed. Heilman and Van Den Abell (1980) found right parietal alpha desynchronization following lateralized warning signals presented to either hemisphere, but left parietal desynchronization only after a lateralized warning signal had been presented to that hemisphere.
Behavioral data also suggest the right hemisphere's importance in conditions of sustained attention. Wilkins, Shallice and McCarthy (1987) demonstrated that right, but not left frontal patients were impaired in their ability to voluntarily sustain attention in a monotonous signal detection task. A similar result was obtained by Coslett, Bowers and Heilman (1987). Also, in studies of split brain patients, Dimond (1979) found that tactual, visual and auditory vigilance performance was substantially better when stimuli were presented to the right rather than to the left hemisphere.
Behavioral data from normals have also revealed hemispheric asymmetries. In a signal detection task, Dimond and Beaumont (1973) found that when targets were presented to the left hemisphere, vigilance deteriorated over the course of an eighty minute experimental session. Right hemisphere performance, although at a lower level (fewer signals detected) showed no such decrement. It should be noted however, that during the course of a session, targets were presented only to one hemisphere or the other. The results were interpreted as suggesting that the left hemisphere is the more sensitive "watchkeeper" but that its relatively high level of performance is susceptible to exhaustibility. The right hemisphere however, is more able to sustain vigilance, albeit at a lower level.
Whitehead (1991a) provided further behavioral evidence for the notion of right hemisphere superiority in maintaining the alert state. In a series of studies, normal subjects received a warning signal followed by a stimulus onset asynchrony of between 3 and 30 seconds. When stimuli were presented following delays of 12 seconds or more, subjects responded significantly faster to stimuli presented to their left rather than to their right visual field. During further investigation (Whitehead, 1991b), it was found that an external alerting stimulus (presented at the same time as the visual target) eliminated this effect, suggesting a common pathway between voluntary sustained attention and the more automatic alerting effects of external stimuli. It was also found that the act of sustaining alertness interacted with the covert orienting of attention. Specifically, when alertness was sustained for a long, as opposed to a short period, attentional engagement in the left visual field was increased. This result was interpreted as indicating some commonality between the structures affected by the act of sustaining alertness, and those responsible for the covert orienting of attention. Direct physiological evidence regarding this issue will be presented later.
Heilman and Van Den Abell (1979) have shown that warning stimuli projected to the right hemisphere reduce reaction times of the right hand to a greater extent than left hemisphere warning stimuli reduce left hand reaction times, and more importantly, to a greater extent than left hemisphere warning stimuli reduce right hand reaction times. Their interpretation was that although each hemisphere can mediate its own activation (Heilman & Valenstein, 1979), the right hemisphere is better able to activate the left hemisphere than the reverse. Subsequent to this study however, Heilman argued that the findings suggest that attention is entirely mediated by the right hemisphere. In support of his argument, he cited evidence indicating that neglect of the contralateral visual hemifield is more prevalent following right than left parietal lesions (Heilman & Van Den Abell, 1980). Weintraub and Mesulam (1987) adopted a similar position in claiming that right hemisphere lesions are more likely to lead to both contralateral and ipsilateral neglect, whereas left hemisphere lesions are more often associated with only contralateral neglect. Also, contralateral neglect is often less severe when the left rather than the right hemisphere is lesioned.
Although Heilman's data permit either an activational or attentional explanation, the majority of the extant data, including Heilman's, indicates right hemisphere dominance only when the task requires the sustaining of attention for long periods, either in situations where continuous processing is involved, or where subjects must remain alert in anticipation of a signal. When highly phasic attentional effects are studied, as for example in Derryberry's (1989) study, left hemisphere superiority may be evident. Also, even when vigilance is required, as was the case with Dimond and Beaumont's (1973) study, the left hemisphere displayed superior performance (albeit with impairment over time) in conditions where only one hemisphere received stimulation, that is, when there was little likelihood of the right hemisphere activating the left. One further piece of evidence is particularly damaging for Heilman's idea of attention itself being a right hemisphere function. Luck, Hillyard, Mangun and Gazzaniga (1989) have demonstrated quite convincingly that in split brain patients, visual search of each hemifield is performed independently by the contralateral hemisphere.
In summary, it would appear most plausible that demonstrations of hemispheric asymmetries of attention are not due to the fact that the visuo-spatial attention system resides solely or even primarily within the right hemisphere. Rather, some more general influence serves to sustain the activity of right hemisphere attentional mechanisms in conditions requiring continuous activity or vigilance, and that this influence is either less pronounced or is more phasic in its effect upon the left hemisphere.
2.3.2 Hemispheric specialization in noradrenergic pathways In addition to findings of hemispheric specialization in the control of attention, several lines of research have indicated hemispheric specialization in neu-rotransmitter systems. For example, Tucker and Williamson (1984) have proposed that lateralization in neurotransmitter pathways may be responsible for the production of subjectively meaningful affective states. Specifically, left lateralized dopaminergic innervation is associated with fluctuations in anxiety, and right lateralized NA pathways are responsible for variations in mood level.
The demonstration of a right hemisphere affinity for noradrenaline is of particular interest in view of the association between NA and arousal and attention, and demonstrations of right hemisphere specialization for sustained attention. Several lines of evidence converge on the supposition of right lateralization in NA pathways.
The dorsal NA pathway described earlier, in its innervation of the cortex is known to enter the cortex at the frontal pole, to ascend to the superficial layers, and to pass through a horizontal layer parallel to the outer layer of the cortex as it traverses toward posterior regions (Emson & Lindvall, 1979). Consistent with the notion of right hemisphere bias in this pathway, it has been shown that lesions of the right but not left hemisphere lead to depletions of NA bilaterally (Robinson, 1979). Furthermore, this effect is more pronounced when lesions are close to the frontal pole (Robinson, 1985). A behavioral consequence of these lesions is the production of spontaneous hyperactivity in rats. Oke, Lewis and Adams (1980) have shown that higher NA levels are found in the right rather than in the left thalamus of rats, and in humans, post mortem studies have shown a similar lateralization (Oke, Keller, Mefford & Adams, 1978).
Further indirect evidence contributes to the belief of right lateralized NA. It is thought that the effects of electro-convulsive therapy (ECT) on mood are mediated through the action of NA. Hence, the finding that right hemisphere activation is particularly facilitated by ECT (Kronfol, Hamsher, Digre & Waziri, 1978) is of relevance. Similarly, tricyclic antidepressant medication is NA mediated; following administration of this medication to children, it was found that improvements resulted in right but not left hemisphere cognitive performance (Brumback, Staton & Wilson 1980).
Since NA innervation of the cortex and thalamus is provided by the dorsal NA pathway, and since both display right lateralized NA concentrations, a parsimonious explanation of this distribution would be to suggest right lateralization in this pathway. Within the cortex, evidence suggests that NA innervation is provided by a branch of the dorsal pathway entering at the frontal pole, and which arborizes as it moves toward posterior regions.
If it is indeed the case that findings of right lateralization of attention and arousal are related, and that both are mediated by right lateralization of NA pathways, then it would be reasonable to expect an effect of NA manipulation on attention shifting. In a study of this question, Clark, Geffen and Geffen (1989) investigated the effects of an NA blocker, clonidine, on responses to visual targets which had been preceded by a directional cue (valid, neutral or invalid). They found that clonidine increased reaction times in general, but reduced the cost of invalid cueing — in essence, it made the subjects more distractible. This result does indeed suggest that NA exerts a specific action on the visuo-spatial attention system. In this study laterality effects were not investigated, however, using the same paradigm, Posner, Inhoff, Friedrich and Cohen (1987) did study hemispheric specialization in this system. They found that if a warning signal is omitted before a target, then right parietal patients are greatly slowed in their ability to respond to targets, whereas left parietal patients are not. This finding again supports the notion that the right hemisphere contains the mechanism responsible for the sustaining of alertness. When a patient who has damage to that mechanism must sustain alertness without an external aid (the warning signal), then he or she is poor at detecting the target. In contrast, left parietal patients suffer no comparable loss.
If one is to suppose that the shifting of attention is influenced by NA, then one would expect evidence of NA innervation of the structures involved. This has been shown to be the case. As mentioned earlier, Oke et al. (1978) have shown dense NA innervation of the thalamus. Of greater interest however, are the findings of Morrison and Foote (1986), wherein each of the structures of the visuo-spatial attention system (the posterior parietal lobe, superior collicu-
lus and lateral pulvinar of the thalamus) was shown to be densely innervated by NA, while much weaker NA innervation was found in the geniculo-striate and ventral pattern recognition pathways.
Finally, if NA is hypothesized to be responsible for the maintenance of the alert state, one would expect to find an inter-relationship between the physiological indices of that state and NA manipulation. Tackett, Webb and Privitera (1981) have indeed shown that following the release of NA, heart rate deceleration occurs, and Walker and Sandman (1979) have shown that evoked potentials are higher in the right than in the left cortical region during periods of heart rate deceleration.
2.4 Interactions between cortical and subcortical systems
Our discussion thus far has reviewed models of arousal which incorporate multiple component systems, and it has emphasized the diversity in potential neuromodulation among different ascending neurotransmitter systems. While these systems have been categorized as ascending, it is not the case that their influence on cortical structures and functions is simply unidirectional. Rather, there are many lines of research indicating reciprocity between cortical and subcortical systems in determining the ultimate influence exerted upon cognitive processing. Our review of this research will be less than comprehensive; instead, we will attempt to provide examples of the more concrete findings in this area. For a more complete treatment of this issue, the interested reader is referred to two excellent reviews by Tucker and his colleagues (Tucker & Williamson, 1984; Tucker & Derryberry, 1990).
The idea of executive cortical control over sub-cortical arousal mechanisms is not a new one, dating back at least to the work of Pribram and McGuinness (1975) who argued for "the involvement of the amygdala and related frontal cortical structures in the attentional control of the core brain arousal systems" (p. 119). Pribram and McGuinness based their argument upon extensive lesion evidence (e.g., Bagshaw & Benzies, 1968; Bagshaw, Kimble & Pribram, 1965) indicating that a frontolimbic circuit is involved in the control of the orienting response. Moreover, their evidence indicated some differentiation between a facilitatory system involving the dorsolateral frontal cortex, and an inhibitory system related to the orbitofrontal cortex.
Evidence also indicates a high degree of response specificity in amygdala neurons, in particular in their selection for reward- and punishment- conditioned stimuli (Ono, Tamura, Nishijo, Nakamura & Tabuchi, 1989) and in attributing motivational significance to incoming signals (Sarter & Markowitsh, 1985). The amygdala has also been ascribed a role in the mediation of anxiety, especially as a contributor to the fight or flight response (Sarter & Markowitsch, 1985). In addition, Applegate, Kapp, Underwood & McNall (1983) found that electrical stimulation of the amygdala results in feelings of anxiety in humans, and Davis, Hitchcock & Rosen (1987) eliminated conditioned fear responses through amygdaloid lesions.
The amygdala and orbitofrontal cortex share several characteristics in regard to responsivity. In addition to being implicated by Pribram and McGuinness in a system controlling the orienting response, cells in the orbito-frontal cortex, like those in the amygdala have been shown to respond to cues for reward and punishment (Thorpe, Rolls & Madison, 1983), to influence autonomic control (Mesulam & Mufson, 1982) and to be excessively active in patients with anxiety disorders (Baxter, Phelps, Mazziotta & Guze, 1987). Several authors (e.g., Nauta, 1971; Tucker & Derryberry, 1990) have argued that the orbital cortex has an executive role in anticipating the expected significance of events and in developing intentionality. Since the orbital cortex is reciprocally connected with the limbic structures, and also projects efferents to the amygdala via temporal regions, its potential sphere of influence is very extensive. For example, from the amygdala alone, influence may be exerted upon forebrain and brainstem circuits, temporal and occipital cortices, and (via the basal nucleus of Mynert) upon all cortical acetylcholine projections.
A second paralimbic circuit consists of a complex interaction between the posterior cingulate, entorhinal and parahippocampal cortices, the anterior thalamus, septum and hippocampus. As discussed earlier, Gray (1982) has implicated several of the components of this circuit in facilitating passive avoidance and behavioral inhibition, and also in modulating cortical and autonomic arousal. Gray views the Behavioral Inhibition System as a monitoring mechanism, one which responds to conditioned stimuli signifying forthcoming punishment or non-reward. In the event that such signals are detected, ongoing behavior is inhibited, cortical and autonomic arousal are adjusted and attentional resources are directed toward the avoidance of punishment. Hence, this system can also be characterized as exhibiting reciprocity between cortical and subcortical activation. In order for the BIS to perform its function, some evaluation of what constitutes a cue for punishment must have preceded, and in response to the evaluation, the BIS is apt to influence the adjustment of both the intensity and direction of behavior. In terms of brain activity, the progression may involve evaluation at the cortical level, BIS activity at the level of the limbic forebrain, arousal through the action of the reticular formation (with subsequent cortical repercussions), and a new direction of behavior, mediated again by cortical structures.
The control of motor actions constitutes another example of behavior being regulated by an interaction of cortical and subcortical systems. The association between dopaminergic pathways involving the basal ganglia and motor control is well established. There is massive interconnectivity between the frontal cortex and the basal ganglia, an area well established as contributing to the initiation of movements (Kornhuber, 1974). These connections take the form of multiple parallel loops (Groenewegen, 1988) in which, most importantly, activity may be modulated at both the cortical and subcortical level by ascending dopaminergic projections from the ventral tegmental region (Taber, Das & Fibiger, 1995). Since dopamine has also been associated with higher order motivational processes (Bunney & Aghajanian, 1977), and response readiness (Tucker & Williamson, 1984), we are again presented with a scenario in which cortical arousal, ostensibly arising in subcortical systems, is mediated not simply in a bottom-up fashion, but as the result of activity in cortico-subcortical feedback loops.
The control of attention is another area in which cortical and subcortical mechanisms interact. Mangun et al. (1994) have produced a compelling argument in favor of the idea that the right hemisphere's specialization for the control of attention is mediated in part by subcortical pathways. These researchers obtained findings suggesting an executive role for the right hemisphere in the control of attention, and importantly, did so in split-brain patients. Since the cerebral hemispheres of these patients are separated at the cortical level, one must suppose that the right hemisphere's influence over the left hemisphere is mediated, at least in part, by subcortical pathways. Furthermore, the cortical mechanisms of attention appear to involve both frontal and posterior mechanisms, with the former being characterized as having a regulatory control over the latter. For example, in a study of scalp potentials, Deeke, Kornhuber, Lang and Schreiber (1985) required subjects to make responses to visual targets. They documented a temporal progression of activity wherein subjects exhibited first frontal negativity, followed by occipital negativity and the resolution of the frontal activity, and finally, once a response was made, resolution of the occipital negativity. The implication here being that the frontal areas possess executive control, which in this task, was passed to the occipital cortex. In terms of personal experience, this would equate to the subjects deliberately shutting down on-going cognitive processing, since it might distract them from detection of the target, and assuming an 'empty-headed' state, in which attentional resources are directed only toward target detection. Once the target has been detected, conscious processing resumes, and activity in target detection mechanisms subsides. Similarly, Posner, Petersen, Fox and Raichle (1988) have found that in anticipation of a target, blood flow to the anterior cingulate gyrus is depleted, but once the target is presented and requires further processing, blood flow to the anterior cingulate is elevated above baseline. In addition, studies of event-related potentials (ERPs) in frontal patients further the supposition that an attentional mechanism resides within the frontal cortex. Frontally lesioned patients do not show the normal enhancement of temporal ERPs to attended auditory stimuli, but do exhibit abnormally large ERPs to unattended stimuli (Knight, Hillyard, Woods & Neville, 1981). The implication here is that the unattended stimuli are not being adequately screened out.
Finally, there is evidence suggesting that the frontal cortex is able to selectively inhibit posterior perceptual mechanisms. Ascending pathways from the thalamus to the posterior cortex are affected by activity in the reticular nucleus of the thalamus. The reticular nucleus has been argued to serve a central gating function (Skinner & Yingling, 1977; Yingling & Skinner, 1977). More importantly, the frontal cortex exerts an excitatory control over the reticular nucleus. Hence, frontal activation of the reticular nucleus encourages its gating function, resulting in inhibitory effects on posterior cortex. LaBerge (1995) has extended the Yingling and Skinner model to suggest that thalamo-cortical circuits operate as an attentional enhancement mechanism, wherein information (internally or externally generated) passing through the thalamus is enhanced for further higher-order processing. LaBerge elaborates on the earlier work of Yingling and Skinner by ascribing specific roles to each of the thalamic nuclei, and by detailing the anatomical connections between these nuclei and other brain areas.
We have made several references to the experiences of the bridge jumper in order to introduce various characteristics of arousal. By way of providing a final example of our understanding of the brain mechanisms of arousal, let us now speculate upon the bridge jumper's experiences from a neurophysiologi-cal perspective
As he stood on the bridge contemplating what he was about to do, he perceived cues for both punishment (threat of injury or death) and reward (anticipation of thrill). According to Gray's theory, these cues encouraged both behavioral inhibition and activation, hence energizing the hippocampus and medial septum (inhibition), and the medial forebrain, lateral hypothalamus, and amygdala (activation). As these two systems fought for control over the direction of behavior (to jump or not) each innervated the non-specific arousal system in the reticular formation. As a result of the amygdaloid, hypothalamic and reticular activity, the sympathetic branch of the ANS achieved dominance over the parasympathetic branch. Hence, the jumper exhibited a fight or flight response — heart rate, blood pressure and respiration increased, digestion was suspended. Also as a function of reticular activity, the locus ceruleus increased noradrenergic supply to other areas of the brain; in particular to attentional mechanisms. The jumper's mode of processing information became highly selective. As a result of thalamic activity, task-relevant information was engaged and enhanced. In the example we employed, the jumper did actually jump — therefore activation overcame inhibition. Once the jumper realized that he was out of danger, his anxiety diminished, but his still high state of autonomic arousal found a new outlet, in thrill or euphoria.
In conclusion, early views of arousal as unidimensional have been superseded by multidimensional conceptualizations. Multi-component structural models of arousal have become necessary in accounting for findings of situational specificity in human and animal behavior. Earlier views of the effect of arousal on information processing emphasized a one-way causal mechanism, wherein arousal was thought to impact upon higher-order processing but not the reverse. More recent thinking has emphasized reciprocity between higher- and lower-order processing, a view strengthened by anatomical findings of interactivity between cortical and subcortical structures. Neu-rophysiological studies of neurotransmitter systems, the NA system in particular, have further aided our understanding of energetical effects on behavior, and in particular, have helped us to accommodate arousal within a rapidly developing field — the neuroscientific study of human behavior.
Was this article helpful?