Dynamic dopaminergic activity controls the timing of self-timed movement.

Abstract

What makes us move? Human movement disorders like Parkinson’s disease have long suggested a vital role for dopamine in movement initiation, but there exists surprisingly little evidence connecting endogenous dopaminergic activity to a causal role in this process. Although “standard” stimulus-response paradigms have not resolved this question, most natural movements are not short-latency reactions to abrupt stimuli. Intriguingly, Parkinson’s patients struggle to self-initiate movement more than to react to sensory cues, suggesting a role for dopamine might be found by examining movements that rely on a spontaneous, internal process rather than an abrupt cue to decide when to move. One class of self-initiated movement is self-timed movement, in which subjects must decide when to move in the absence of overt cues. Pharmacological studies have suggested that exogenous dopamine can affect the timing of these movements. We hypothesized that dopaminergic activity influences movement initiation by guiding its moment-to-moment timing. We recorded dopaminergic signals with fiber photometry in gender-balanced cohorts of mice as they prepared to initiate self-timed movements (GCaMP6f in genetically-defined dopamine neurons; the novel dopamine indicator dLight1.1 or DA2m in striatal neurons). We controlled for optical/motion artefacts with co-expressed tdTomato photometry, neck EMG, back-mounted accelerometer, and high-speed video. We found that even when accounting for ongoing movements, task variables, and trial history, dopaminergic signals were highly predictive of single-trial movement timing via a seconds-long “ramp-up” that unfolded between the start-timing cue and the self-timed movement. The rate of this ramping was correlated with the pre-cue level of dopaminergic signaling, with higher signals preceding relatively early movement, and lower signals portending later movement. Consistently, optogenetic activation of dopamine neurons systematically early-shifted movement timing, whereas inhibition caused late-shifting, whereas no-opsin stimulation caused no consistent behavioral effect. Optogenetic stimuli were subthreshold for generating/preventing movement outside the task, suggesting dopaminergic activity influenced movement onset by adjusting the moment-to-moment probability of a planned movement. Applying a reward-prediction error model for dopaminergic activity recapitulated the dynamic signals in our task and explained puzzling findings from a recent perceptual timing task, suggesting a unifying interpretation for the role of dopaminergic activity across timing paradigms.