implicit motivation

Vertebrates are remarkable for their ability to select and execute purposive actions: motor skills that are critical for thriving in complex, competitive environments. Although we have often focused on the ability of a skilled movement to be executed rapidly and with precisely replicated kinematics; motor skill is also characterized by the incredible flexibility with which an action can be executed while reliably obtaining its goal. In this latter sense, the ability to act with a range of vigor is an essential aspect of skill. Several lines of evidence suggest that the dorsal basal ganglia control movement vigor in rodents. For example, the loss of dopaminergic input to dorsal striatum is thought to lead to the profound enervation of voluntary movement in Parkinson's disease. To study the circuit mechanisms of such changes we developed behavioral paradigms to study precise movement kinematics in mice performing voluntary, reaching movements. We recently used this paradigm to demonstrate that dopamine neuron loss in a mouse model of Parkinson's disease leads to a dramatic, but rapidly reversible, change in representation of movement vigor in striatum. Ongoing work in the lab is focused on identifying strategies for remediation of dopamine neuron loss and understanding the role of dopamine in adaptive changes in movement vigor.

reward learning

To make optimal decisions in the presence of uncertainty requires the inference of probabilistic models. For example, financial decision theory posits that selection of optimal portfolios requires information about both the mean and variance of expected returns. In the timing literature it is often suggested that rodents learn the delay until reward delivery with a fixed relative uncertainty. However, in a natural environment the timing of events may have an arbitrary uncertainty. Thus, we asked whether mice could infer the extrinsic variance of reward timing even in highly variable environments. Our recent paper argues that mice can infer extrinsic variance over at least two orders of magnitude by accumulating information from ~50 trials. Ongoing work in the lab seeks to address how these computations are performed. For example, is a heuristic that approximates optimal computations used? And if so, how is this heuristic implemented in cortico-basal ganglia circuits?

Mice infer probabilistic models for timing

PNAS 110(42): 17154-17159

population dynamics

Vertebrates, and mice in particular, adapt their behavior to be in the right place at the right time and expend the correct amount of effort to efficiently collect rewards. Moreover, for skilled mice this precision can be obtained despite a variance in initial conditions and interference during action execution. To reliably obtain a goal in the face of this variability we suggest that basal ganglia circuits must track progress towards a goal, be that the elapsed time or evolving plans for action or the kinematics of complex movements. To monitor basal ganglia activity we use simultaneous recording of distinct, intermingled neuronal cell types during behavior. For example, we have recently observed that a conditioned stimulus (CS) recruits a precise sequence of activations in distinct cell-types in the ventral midbrain. The first two populations to respond are GABAergic neurons (confirmed in vivo using optogenetic tagging) followed by the phasic response of dopamine (DA) neurons. The recruitment of this sequence of cell types is contingent upon learned associations between CS and a subsequent reward delivery. In ongoing work we have demonstrated that the precise timing of sequential recruitment of midbrain populations can be used to decode a mouse's approach to a source of reward providing direct evidence that the basal ganglia contains dynamic representations of evolving actions.

circuit mechanisms

The canonical circuit of the basal ganglia is present in all vertebrates. From lampreys to primates the basic cell types, neurochemical signaling pathways, and mesoscopic connectivity appear to be highly conserved. However, the basal ganglia is also, to some degree an oddity relative to many other central neural circuits. For example, several nuclei including the globus pallidus and substantia nigra appear to be devoid of local interneurons. In many circuits, especially neocortical circuits, primary sensory, and hippocampal circuits that have been studied extensively local interneurons play critical functional roles in the computations implemented by those circuits. Over the past several years we have been interested in the extent to which local circuitry, albeit different in many respects, can perform similar computational roles to those identified in circuits with a greater diversity of cell types. Recently, we demonstrated that feedback gain control can be implemented by the microcircuit of the substantia nigra even though it lacks interneurons and is composed exclusively of tonically active, inhibitory projection neurons. Gain control effect is implemented by a combination of unique biophysics in the projection neurons and properties of microcircuit connectivity. Ongoing work in the lab is trying to understand the role of this microcircuit computation in the control of behavior by basal ganglia.

hardware & software

The work in our lab combines optical recording, large-scale extracellular electrophysiology and intracellular electrophysiology with techniques to identify and perturb the activity of specific cell types in the mouse brain. This constellation of technical approaches requires significant development and refinement of hardware and software for data acquisition and analysis. In addition, we are actively involved in developing and refining molecular tools that allow us to target and manipulate specific cell-types in concert with physiology. We collaborate with several groups at Janelia Research Campus to develop these tools. In addition, we draw heavily upon the expertise of many staff scientists and engineers in the Instrument Design & Fabrication, Molecular Biology Core, Transgenic Mouse Facility, and Scientific Computing. Some of the fruits of our labor can be found on the resources section of the website where there are details about how to recreate the hardware developed for our experiments and software available for download.