Perspectives on Behavioural Function of Dopamine in the Nucleus Accumbens

John Salamone
salamone@psych.psy.uconn.edu

1. Reinforcing stimuli have activating effects, and accumbens DA is clearly involved in these effects. This is an important aspect of reinforcement, but I think the term "reward" does not accurately describe this particular function of DA.

2. Bindra talked about "instigation of actions" as an important aspect of incentive motivation.  Accumbens and striatal DA are involved in the "instigation of actions", which is hardly different from saying that DA is involved in the "initiation of movement".  This process of initiating instrumental actions is a fascinating and complex process, which is an integral part of reinforcement or incentive motivation mechanisms, but it is misleading to use the term "reward" to describe this.

3. For any reinforcer, directional aspects of motivation are a critical aspect of the reinforcement process. I did not invent this.  One has only to read Thorndike, Premack, Bindra, Timerlake, Allison and Dickenson, and it is evident that this is a major line of thought in the behavioral literature on reinforcement (see Salamone, 1992; Salamone et al. 1997, for a discussion of the "motivational" or "regulatory" view of reinforcement). In addition, it is evident that this view has profoundly influenced the DA/reward literature.  Well, considerable evidence indicates that fundamental aspects of food motivation are intact after interfering with accumbens DA (see my presentation here, or Salamone et al., 1997).  The CRF schedule is profoundly dependent upon food motivation, and represents the most fundamental and simple schedule of food reinforcement. CRF performance is dramatically affected by extinction and by prefeeding. Yet, this schedule is relatively insensitive to the effects of accumbens DA depletions. Considerable evidence indicates that the primary, or unconditioned reinforcing effects of food are intact after accumbens DA depletions (see presentation and Salamone et al., 1997). Animals with accumbens DA depletions are still directed towards the aquisition and consumption of food provided that work requirements are very low.  It seems to me that if this fundamental aspect of reinforcement remains intact, the term "reward deficit" is not appropriate.

4. Accumbens DA seems to be involved in aspects of learning, particularly procedural learning.  Of course, the term "procedural learning" is not the same as "reward", and procedural learning involves higher-order aspects of sensorimotor integration.  

5. Too much of the support for the DA hypothesis of reward comes from the stimulant self-administration literature.  One cannot use evidence for dopaminergic invovlement in drug self administration in general or stimulant abuse in particular as evidence for dopaminergic involvement in primary food reinforcement.  I think that NIDA needs to move beyond the general anhedonia model in order to understand drug abuse.  The evidence simply does not support the notion that cocaine is turning on the "reward system" for natural reinforcers such as food.
    I have observed two very intelligent and successful researchers follow the logic of the DA/reward hypothesis squarely into the realm of fantasy (I will not name them). One, at a neuroscience poster a couple of years ago, attributed the "aphagia" seen after DA depletions to the destruction of the DA reward system in accumbens. Of course, several papers have shown that chow feeding is intact after accumbens DA depletions, and that aphagia is due to lateral striatal DA depletions.  A second researcher was at a presentation, and was asked if accumbens DA depletions in humans could be used to suppress drug abuse. He replied that based upon the animal literature, that would be dangerous because it would result in a complete block of food motivation, and the people would starve to death. Of course, anyone who as ever worked with animals that have accumbens DA depletions, or has ever read this literature, would recognize this statement as completely erroneous. These encounters illustrate the danger of a scientific hypotheses that seems so plausible that it leads you to ignore findings to the contrary.  Although I am not personally a drug abuse researcher, I do feel that the DA reward hypothesis has had an unhealthy effect on that area. We certainly have learned alot about drug abuse based upon this first generation of research, but it is probably time to move on.