Review of: Chowdhury R, Guitart-Masip M, Lambert C, Dayan P, Huys Q, Düzel E, Dolan RJ. Dopamine restores reward prediction errors in old age. Nat Neurosci 2013;16:648-653.
Clinicians working in the movement disorders field are quite familiar with the effects of dopaminergic agents on the symptoms of Parkinson's disease. Basal ganglia circuitry is involved in many cognitive functions as well, and L-DOPA conceivably could be useful in conditions degrading these functions. Chowdhury et al have now explored this in the context of decision making in an older population. These authors studied fMRI signatures of decision making in the nucleus accumbens in the context of L-DOPA administration, and found that older people had abnormal fMRI signatures of the expected value during a choice task, which could be corrected in a subset of subjects with L-DOPA. Specifically, part of the function of the nucleus accumbens (NA) is to provide a reward prediction error (RPE) signal to other brain structures, and this RPE signal is abnormal in older adults in the choice task used in the study. This signal was subsequently corrected with the administration of L-DOPA.
Chowdury et al studied 32 older adults and 22 younger adults performing a type of choice task having a two-armed bandit design. In this test, two fractal images were displayed on a computer screen, and the participant chose one of them. The subsequent display of a green up arrow indicated a “win,” while a yellow horizontal bar indicated no win. The same two fractal images were used in each trial iteration, and the probability of a win associated with one of the two images varied on a trial-by-trial basis according to a Gaussian random walk distribution. The older subject group performed the task twice on separate days under two conditions, either after taking 150 mg of L-DOPA or after a placebo. The younger age group did not undergo pharmacologic manipulation. Diffusion-tensor MRI imaging was also performed on the participants to investigate connectivity between the substantia nigra/ventral tegmental area (SN/VTA) and the ventral striatum as a function of inter-individual differences in the expression of the expected reward value signal.
In the two-armed bandit choice task, the young adults were able to earn significantly more money than older adults who received placebo. This difference was corrected when the older adults were administered L-DOPA. Not all older adults improved however, and in a median split analysis, the older adults with higher baseline performance actually performed worse after receiving L-DOPA (the “win-less” group) as compared to the subgroup performing better after L-DOPA (the “win-more” group). Not only was task performance improved in the adults with lower baseline performance, but their learning rate also improved after L-DOPA administration as demonstrated with the application of a parametrized reinforcement learning model. fMRI analysis of the nucleus accumbens demonstrated reward prediction errors consistent with RPE signals biased by a more negative expected reward in the two-armed bandit choice task when imaging was performed in the older subjects after L-DOPA administration. This effect was not seen during placebo administration, and was accentuated when studying specifically the median subgroup that benefited more in general with L-DOPA administration (the “win-more” group), implying even more negative expected values in this group represented in the nucleus accumbens while on L-DOPA.
The authors also used DTI to derive a measure of predominantly dopaminergic connectivity between the SN/VTA and striatum (whose ventral portion incudes the NA). Greater connectivity strength correlated with the “win-less” older adult group (the group not benefiting as much from L-DOPA administration), implying higher base-line integrity of this circuitry compared to the older group that benefited significantly from L-DOPA. In summary, Chowdhury et al found that dopamine administration restored error signals in the brain associated with reward prediction, implying the possibility of clinically improving age related decision making problems in older adults. The study further points out the importance of dopaminergic modulation in subcortical circuitry in aspects of brain functioning not related to motor control, and could further research efforts at new neuromodulatory targets.