Bayesian Optimization of Interval Timing and Biases in Temporal Memory as a Function of Temporal Context, Feedback, and Dopamine Levels in Young, Aged and Parkinson's Disease Patients
© 2016 by Koninklijke Brill NV, Leiden, The Netherlands. Interval timing behavior and its sensitivity to both temporal context and changes in dopamine (DA) levels has recently received considerable attention. Nevertheless, the exact manner in which those interactions occur is far from clear. We examined temporal reproduction with feedback in the supra-seconds range as a function of DA levels using two well-studied timing procedures. Healthy young and aged participants were studied as well as Parkinson's disease (PD) patients tested ON and OFF their dopaminergic medication. The findings confirm the hypothesis that the 'migration effect' (e.g., 'short' durations are over-produced and 'long' durations are under-produced) in PD patients and the closely related Vierordt's effect are largely influenced by the effective level of DA and in the case of the 'migration effect' by the probability of feedback as well. Using a Bayesian model seeking optimal timing under conditions of uncertainty, we were able to accurately simulate the distorted patterns of temporal reproduction in all groups of participants. As DA levels decrease across groups, optimal timing behavior shifts towards a greater reliance on a statistical representation of all of the durations reproduced within a specific temporal context rather than on the representation of a single duration being timed on any one trial. This analysis demonstrates the utility of Bayesian models of interval timing and highlights the importance of DA levels on clock speed and the associated uncertainty that contributes to temporal distortions.
Gu, BM; Jurkowski, AJ; Shi, Z; Meck, WH
Volume / Issue
Start / End Page
Electronic International Standard Serial Number (EISSN)
International Standard Serial Number (ISSN)
Digital Object Identifier (DOI)