King M, Jentsch JD (2007) Prefrontal cortical norepinephrine depletion does not impair spatial working memory in rats. Neuroscience 2007 Abstracts 645.16/CCC18. Society for Neuroscience, San Diego, CA.
Summary: The midbrain dopamine neurons are thought to encode a reward prediction error signal (Schultz et al., 1997; Bayer & Glimcher, 2005). Parkinson’s disease (PD) is characterized by a loss of nigral dopamine neurons. Dopaminergic drugs including the dopamine precursor L-Dopa and D2 receptor agonists are taken to relieve disease symptoms. We hypothesized that patients with moderate PD (1) show atypical reinforcement learning off dopaminergic medication due to dopamine neuron loss, and (2) show more normal reinforcement learning on dopaminergic drug therapy. We developed a method to rapidly assess reinforcement learning in human subjects (Rutledge et al., SfN 2005) adapted from matching law tasks used in monkeys (Sugrue et al., 2004; Lau & Glimcher, 2005). On each trial, subjects choose one of two animated crab traps. Rewards (crabs worth $0.10) were scheduled for the two targets with different independent rates. Scheduled rewards remained available until the associated target was chosen, as in the original matching law experiments (Herrnstein, 1961). After a 5-minute training period, subjects completed 800 trials as we varied reward probabilities across blocks. PD patients (n=19) completed one session on and one off dopaminergic medication. Age-matched controls (n=21) and healthy young subjects (n=20) completed one session. We found that young and elderly control subjects had similar reinforcement learning rates, but learning rates were reduced in PD patients (when tested off medication). Learning rates in the same PD patients were restored to control levels when dopaminergic drugs were administered. We also found that the reinforcement-independent strategies of our subjects were influenced by dopamine. Young subjects tended to alternate targets independent of reward history. In contrast, elderly subjects (who suffer some dopamine neuron loss) had a tendency to perseverate in their choices. This tendency was increased in PD patients (off medication), but restored to control levels when dopaminergic drugs were administered. This effect on choice is not explained by existing models of dopamine function. These data support a role for dopamine in human reinforcement learning. Future models of decision making in reinforcement learning tasks must also account for a reward-independent effect of dopamine on choice behavior.
Related Products: Anti-DBH-SAP (Cat. #IT-03)