Full-length reviewThe role of nucleus accumbens dopamine in motivated behavior: a unifying interpretation with special reference to reward-seeking
Introduction
Brain dopamine (DA) has been linked to rewarding processes in the brain for three decades. Two decades ago, Wise [299]and Wise et al. [304]suggested a pleasure or hedonic role for brain DA. Since then, extensive research on the functions of DA has refined such ideas, and our thinking about DA functions has evolved substantially 16, 21, 68, 136, 162, 181, 208, 219, 290, 303. Although there is no robust evidence to support direct involvement of DA in hedonic effects, which resemble the consummatory aspects of reward phenomena, many lines of evidence suggest that the DA neurons from the ventral tegmental area (VTA) that innervate the nucleus accumbens septi (NAS), referred to as the meso-accumbens DA system, play an important role in mediating reward-seeking effects by presently unknown mechanisms.
However, some workers have questioned such a role of NAS DA because of its involvement in other functions, especially aversive functions. For example, Gray et al. [88]stated that āWe believeā¦ that there is no special relationship between dopamine release in the nucleus accumbens and positive reinforcementā (p. 1548). Hence, an essential question is whether or not NAS DA is involved in both reward-seeking effects as well as other effects in negative contexts that seem apparently incompatible with the former effects. The aim of the present paper is to review findings regarding the behavioral functions of NAS DA and to examine whether a unified conception of NAS DA in both reward-seeking and yet unidentified functions in aversive contexts can be constructed. The present paper shall focus on the functions of DA in the NAS as clarified primarily through research on laboratory rats.
Studies employing systemic manipulations of DA systems are not comprehensively reviewed in the present paper. Obviously, systemic manipulations can modify activity not only in the meso-accumbens DA system but in other systems as well, particularly the physically more robust DA system of the nigro-striatal continuum. Although DA synapses in the NAS (a portion of the ventral striatum) and caudateāputamen (the dorsal striatum) share similarities in cytoarchitectural organizations, these regions receive distinct afferent inputs and send distinct efferent projections [91]. Such differential anatomical connections reflect functional differences between the NAS and caudateāputamen (e.g., Refs. 7, 8, 39, 40, 47, 59, 127, 210, 220, 273), including reward-related effects being clearly elaborated by NAS circuitry but not so in various other DA terminal regions 31, 32, 34, 37, 126, 247, 260.
The major aim of this paper is to focus on the specific functions of NAS DA, cultivating the emerging recognition that different components of ascending DA systems may govern quite different aspects of psychobehavioral integration within the brain, while not denying that such diverse functions may all be subsumed under the broad conceptual umbrella that all brain DA systems promote widespread sensoryāmotor arousal and competence within the brain. The major claim of this paper is that available evidence indicates that NAS DA is an essential neurochemical system for animals to flexibly approach various rewards, both positive as well as avoidance of negative states, and to construct learned incentive structures in the brain. Our goal is also to provide an extensive literature review embedded in relevant historical perspectives to allow new investigators to become conversant with the many lines of work and those that are relevant for this rich field of inquiry. However, before discussing specific functional issues, we shall provide a brief summary of relevant methodological and anatomical considerations.
Several behavioral test procedures provide major methodological frameworks for modern neurobiological investigations of DA function. These methodologies allow investigators to quantify behavioral responses and to define responses in relation to specific environmental changes. Of course, existing conceptual frameworks constrain the ways in which empirical findings are interpreted.
Pavlov [180]developed an ingenious experimental protocol that enabled investigators to study the learning of stimulusāresponse association using a simplified methodology. In the Pavlovian (or classical) conditioning procedure, biologically important stimuli are defined as unconditioned stimuli because they can trigger unconditioned responses, that are `inborn' or `species-typical' reflexes [180]. Examples of unconditioned stimuli are food, water, and various noxious stimuli. When other comparatively neutral sensory stimuli precede the presentation of such unconditional stimuli, and this pairing is repeated, conditioning occurs. Conditioned responses that were not present prior to such pairings can be now observed when the previously neutral sensory stimuli are presented alone. The sensory stimuli are now referred to as conditioned stimuli. An important feature of this paradigm is that the resulting learned responses of organisms (i.e., conditioned responses) have no consequence on the presentation of unconditioned stimuli (even though they may have consequences on how organisms cope with such stimuli). This paradigm can be used to study factors and neural mechanisms involved in the formation of central states that anticipate future events [102](i.e., even subjectively experienced causal relationships among environmental stimuli).
Another important methodology is the operant procedure. Reinforcers are those stimuli or events that increase the future probability of the occurrence of responses, when those responses are associated with the presentation or removal of some biologically important stimuli (i.e., unconditioned stimuli in the Pavlovian procedure). For example, animals increase responses that result in the presentation of incentive stimuli such as food and sexual stimuli (positive reinforcers), as well as responses that result in the removal of noxious stimuli (negative reinforcers). Skinner [237]made a distinction between responses that are elicited by stimuli (i.e., the Pavlovian procedure) and responses that are emitted by organisms (as measured by the operant or instrumental procedure). Skinner refers to the former as respondents and the latter as operants, with both yielding conditioned responses. On this foundation, Skinner created a rigorous methodology to study operant or instrumental responses, which, to this day, are contrasted with conditioned responses of the Pavlovian variety (see Ref. [204]).
Still another procedure that has been utilized frequently for studying neural mechanisms of rewarding effects is the place-conditioning paradigm (see Ref. 36, 99, 223, 268). During the conditioning phase, animals receive a stimulus in one compartment of the place-conditioning chamber, and typically receive a control stimulus in another compartment. For the testing phase, no stimulus is presented and animals are free to go to either compartment. Place-preference can be produced by positive reinforcers such as food, sexual stimuli and drugs of abuse; typically, animals spend more time in the compartment paired with positive reinforcers. Place-avoidance can be obtained by negative reinforcers; animals spend less time in the compartment paired with the aversive events. Thus, the conditioning phase of this procedure appears to be based on long time-frame Pavlovian contingencies, but the responses measured during testing are more akin to instrumental response; thus, this procedure can be viewed as a blend of instrumental and Pavlovian conditioning.
The same stimuli such as food, sexual, and noxious stimuli may be labeled somewhat differently depending on the research paradigm used. To unify related concepts, this paper refers to these biologically relevant stimuli as unconditioned stimuli. Moreover, those unconditioned stimuli that elicit approach responses are defined as rewards. The reward concept is useful because NAS DA appears to be involved in the acquisition of Pavlovian conditioning 211, 274, operant conditioning (see Section 6.3) and perhaps place conditioning as well. There are, however, some serious concerns in the use of the reward concept. First, the term `reward' has been defined quite differently depending on theoretical positions or phenomena being studied (e.g., Refs. 286, 293). Second, it can carry additional excess meaning for certain subjective effects such as pleasure. In the present paper, reward is simply used to refer to unconditioned stimuli that can evoke approach behavioral effects as defined above, without necessarily implying any subjective positive hedonic effects.
DA in the NAS is released by the neurons whose cell bodies are located in the ventromedial mesencephalon (A10) (Refs. 50, 65, 77, 139, 270; for review, see Ref. [164]), primarily in a zone commonly known as the VTA [193]. Fig. 1A depicts the location of the meso-accumbens DA system in relation to other structures. The NAS can be divided into two major sub-regions: the shell (the ventromedial part) and the core (the dorsolateral part) which have different connectivities 92, 312. The shell sends efferent projections to the ventromedial ventral pallidum, extended amygdala (including the bed nucleus of stria terminalis, central amygdaloid nucleus, and interconnecting sublenticular area), lateral preoptic area, lateral hypothalamus, entopeduncular nucleus, VTA, mediodorsal substantia nigra pars compacta, mesopontine reticular formation, and periaqueductal gray. The core sends major efferent projections to the dorsolateral ventral pallidum, entopeduncular nucleus, lateral part of VTA, and substantial nigra. Fig. 1B summarizes efferent projections of the NAS.
The VTA and NAS receive afferent inputs from a variety of regions throughout the brain (see Fig. 1C and D). The inputs from these structures influence the transmission properties of NAS DA circuitry. The VTA 171, 192receives afferent inputs from many forebrain regions: the prefrontal cortex, NAS, bed nucleus of stria terminalis, diagonal band of Broca, substantia innominata, lateral preoptic area, and lateral hypothalamus. The lower brainstem projections to the VTA include the superior colliculus, substantia nigra, dorsal raphe, parabrachial nucleus, and dentate nucleus of cerebellum. The NAS [91]also receives afferent inputs from forebrain structures (including the medial prefrontal cortex, amygdala, hippocampus, thalamus) as well as mesopontine areas (including VTA, dorsal raphe, and mesopontine reticular formation) (see Ref. [157]for differential afferents between the shell and core).
Section snippets
A brief overview of empirical findings on NAS DA and behavior
Many lines of evidence support the role of NAS DA in reward-seeking processes. However, we presently need to identify more precisely those psychobehavioral processes in which NAS DA are specifically involved, since emerging evidence indicates that NAS DA is also involved in some types of aversive processes.
A brief conceptual history of brain DA and behavioral function
The aim of this section is to provide a brief overview of classical literature within which the specific functions of NAS DA will be discussed. In other words, to further clarify the specific role of NAS DA in the control of behavior, it is necessary to specify the types of natural brain functions in which NAS DA normally participates in animals' every-day affairs. We will first review several traditional concepts for the basic functional organization of the central nervous system (CNS) in the
An updated hypothesis of specific NAS DA functions
Fig. 3 summaries a conceptual model that highlights the role of NAS DA in behavioral control. A key feature of the model is that it has distinct sensorimotor paths for approach and consummatory responses. It is assumed that the same approach response system is engaged when animals escape from negative incentives and develop the ability to avoid such stimuli. Key sub-systems underlying the approach responses include declarative perceptions, habits, and incentive-cue formation systems. The
Behavior/environment correlates of NAS DA
Three major techniques have been used to explore correlational relationship between the meso-accumbens DA system and behavior. Unit recordings have identified DA neuron activity changes in behaving animals. Measurement of action potentials in millisecond time frames provides an opportunity to explore detailed temporal relationships between DA neuron activities and behavior. Some caution needs to be exercised in interpreting such results because it is not certain how well activities of the DA
Behavioral effects of direct NAS DA manipulations
Two major interventions have been used to elucidate functions of NAS DA: the brain microinjection technique and the 6-OHDA lesion approach. The microinjection technique allows experimenters to manipulate discrete regions of the brain pharmacologically without significantly affecting other parts of the brain directly. An advantage of this technique is that behavior can be assessed immediately, giving a minimal time for the nervous system to compensate for manipulations. A disadvantage of the
Issues and implications
Considering the well-accepted role of DA imbalances in the genesis of schizophrenia and obsessiveācompulsive disorders, the present views should have implications for those types of psychiatric disorders (e.g., Ref. [83]). Indeed, the development and solidification of human belief system (as well as core delusions) may be promoted by the functional dynamics of these NAS DA meaning-construction mechanisms [174]. Because the clinical issues are not directly relevant to the present coverage, those
Coda: the function of NAS DA and beyond
As should be evident from this review, a massive amount of perplexing evidence is now encouraging investigators to utilize more subtle concepts than those that have been used before ā in the present case to envision the meso-accumbens DA system, along with closely related brain systems, as a generalized approach-seeking system that was designed by evolution to allow organisms to generate efficient goal-directed activities in response to a large number of positive and negative incentives.
Acknowledgements
We would like to thank Drs. David Highfield, Roy Wise and Jeff Witkin for their helpful comments on earlier versions of the manuscript. The first author would like to express appreciation to Roy Wise for his discussions on conceptual issues and support on this review and related projects.
References (313)
- et al.
Nucleus accumbens dopamine depletions make rats more sensitive to high ratio requirements but do not impair primary food reinforcement
Neuroscience
(1999) - et al.
Suppression of exploratory locomotor activity and increase in dopamine turnover following the local application of cis-flupenthixol into limbic projection areas of the rat striatum
Brain Res.
(1987) - et al.
Stress-induced analgesia plays an adaptive role in the organization of behavioral responding
Brain Res. Bull.
(1988) - et al.
The effects of ibotenic acid lesions of the nucleus accumbens on spatial learning and extinction in the rat
Behav. Brain Res.
(1989) - et al.
Psychobiology of novelty-seeking and drug-seeking behavior
Behav. Brain Res.
(1996) - et al.
Differential effects of intraventricular administration of 6-hydroxydopamine on behavior of rats in approach and avoidance procedures: reversal of avoidance decrements by diazepam
Pharmacol. Biochem. Behav.
(1975) The role of dopamine in locomotor activity and learning
Brain Res. Rev.
(1983)- et al.
Organization of brainstem behavioral systems
Brain Res. Bull.
(1976) - et al.
What is the role of dopamine in reward: hedonic impact, reward learning or incentive salience?
Brian Res. Rev.
(1998) - et al.
Mesocorticolimbic dopaminergic systems and emotional states
J. Neurosci. Methods
(1990)