We know that the ultimate site of dopamine activity caused by drugs is the ventral tegmental area, or VTA, and an associated structure, the nucleus accumbens. But dopamine neurons in the VTA actually perform two distinct functions. They discriminate acutely between the expectation of reward, and the actual reward itself. Pavlov showed how these dual functions are linked, but the manner in which dopamine neurons computed and then dealt with the differences between expectation and reward—a controversial concept known as reward prediction error—was not well understood.
We all know about reward and punishment, however. Years ago, behaviorism’s emphasis on positive and negative reinforcement demonstrated the strong connection between reward, punishment, and learning. As Michael Bozarth wrote in “Pleasure Systems in the Brain, ” addictive drugs “pharmacologically activate brain reward mechanisms involved in the control of normal behavior. Thus, addictive drugs may be used as tools to study brain mechanisms involved in normal motivational and reward processes.”
But how does the evolutionary pursuit of pleasure or avoidance of punishment that guarantees the survival of an organism—fighting, fleeing, feeding, and… fornicating, in the well-known “4-F” configuration—become a pathological reversal of this function? To begin with, as Bozarth writes, “the direct chemical activation of these reward pathways does not in itself represent any severe departure from the normal control reward systems exert over behavior…. Simple activation of brain reward systems does not constitute addiction!”
What does, then? Bozarth believes addiction results from “motivational toxicity,” defined as deterioration in the “ability of normal rewards to govern behavior.” In an impaired reward system, “natural” rewards don’t alter dopamine function as strongly as drug rewards. “Direct pharmacological activation of a reward system dominates the organism’s motivational hierarchy at the expense of other rewards that promote survival,” Bozarth writes. The result? Drug addicts who prefer, say, methamphetamine to food.
How does an addict’s mind become so addled that the next hit takes precedence over the next meal? A group of Harvard-based researchers, writing in Nature, thinks it may have a handle on how the brain calculates reward expectations, and how those calculations go awry in the case of heavy drug and alcohol use.
The dopamine system somehow calculates the results of both failed and fulfilled expectations of reward, and uses that data in future situations. Cellular biologists, with some exceptions, believe that dopamine neurons effectively signal some rather complicated discrepancies between expected and actual rewards. Dopaminergic neurons were, in effect, computing reward prediction error, according to the theory. They were encoding expectation, which spiked when the reward was better than expected, and fell when the reward was less than expected. As Scicurious wrote at her blog, Neurotic Physiology “If you can’t predict where and when you’re going to get food, shelter, or sex in response to specific stimuli, you’re going to be a very hungry, chilly and undersexed organism.” (See her excellent and very readable post on dopamine and reward prediction HERE . )
But nobody knew how this calculation was performed at the cellular level.
Enter research mice.
As it turns out, dopamine is not the whole story. (A single neurotransmitter rarely is.) Dopaminergic neurons account for only about 55-65% of total neurons on the VTA. The rest? Mostly neurons for GABA, the inhibitory transmitter. “Many addictive drugs inhibit VTA GABAergic neurons,” the researchers note, “which increases dopamine release (called disinhibition), a potential mechanism for reinforcing the effects of these drugs.” By inhibiting the inhibitor, so to speak, addictive drugs increase the dopamine buzz factor.
The researchers used two strains of genetically altered mice, one optimized for measuring dopamine, the other for measuring GABA. The scientists conditioned mice using odor cues, and offered four possible outcomes: big reward, small reward, nothing, or punishment (puff of air to the animal’s face). Throughout the conditioning and testing, the researchers recorded the activity of neurons in the ventral tegmental area. They found plenty of neurons with atypical firing patterns. These neurons, in response to reward-predicting odors, showed “persistent excitation” during the delay before the reward. Others showed “persistent inhibition” to reward-predicting odors.
It took a good deal of sorting out, and conclusions are still tentative, but eventually the investigators believed that VTA dopamine neurons managed to detect the discrepancy between expected and actual outcomes by recruiting GABA neurons to aid in the dendritic computation. This mechanism may play a critical role in optimal learning, the researchers argue.
Furthermore, the authors believe that “inhibition of GABAergic neurons by addictive drugs could lead to sustained reward prediction error even after the learned effects of drug intake are well established.” Because alcohol and other addictive drugs disrupt GABA levels in the brain’s reward circuitry, the mechanism for evaluating expectation and reward is compromised. GABA, dopamine’s partner in the enterprise, isn’t contributing properly. The ability to learn from experience and to accurately gauge the likelihood of reward, so famously compromised in active addiction, may be the result of this GABA disruption.
Naoshige Uchida, associate professor of molecular and cellular biology at Harvard, and one of the authors of the Nature paper, said in a press release that until now, “no one knew how these GABA neurons were involved in the reward and punishment cycle. What we believe is happening is that they are inhibiting the dopamine neurons, so the two are working together to make the reward error computation.” Apparently, the firing of dopamine neurons in the VTA signals an unexpected reward—but the firing of GABA neurons signals an expected reward. Working together, GABA neurons aid dopamine neurons in calculating reward prediction error.
In other words, if you inhibit GABA neurons through heavy drug use, you screw up a very intricate dopamine feedback loop. When faced with a reward prediction error, such as drug tolerance—a good example of reward not meeting expectations—addicts will continue taking the drug. This seems nonsensical. If the drug no longer works to produce pleasure like it used to do, then why continue to take it? It may be because dopamine-active brain circuits are no longer accurately computing reward prediction errors. Not even close. The research suggests that an addict’s brain no longer registers negative responses to drugs as reward errors. Instead, all that remains is the reinforcing signals from the dopamine neurons: Get more drugs.
[Tip of the hat to Eric Barker (@bakadesuyo) for bringing this study to my attention.]
Cohen, J., Haesler, S., Vong, L., Lowell, B., & Uchida, N. (2012). Neuron-type-specific signals for reward and punishment in the ventral tegmental area Nature DOI: 10.1038/nature10754
Photo Credit: http://www.zazzle.com