Voluntary, purposive behavior requires that we extract information about the world, formulate plans for action, and then execute the movements required to bring about desired outcomes. Our lab studies a critical nexus in the mammalian brain where sensory information and motor planning come together to subserve volition - the basal ganglia.
Our recent paper described a behavioral task in which mice learn to maximize reward by adopting optimal waiting times in the presence of highly variable reward delays.
Our new paper describing how the inhibitory microcircuit in the midbrain determines the gain of the basal ganglia output is now in press.
The work in our lab aims to elucidate the neurobiology of purposive behavior, as well as its pathological disruption in Parkinson's disease and addiction. See all the updates from the lab.
Our lab spans a wide range of technical approaches from electrophysiology during behavior to two-photon imaging. For our work we seek talented people with skills in physiology, behavior, imaging, or, computation, and a strong desire to combine multiple technical approaches. Inquiries can be sent directly to the lab by contacting us.
For interested students, we have a small number of slots for exceptional graduate students through the Janelia Graduate Program or the Graduate Research Fellowship program. For undergraduates, our lab is a regular participant in the Janelia Undergraduate Scholars program.
Over the past few years we have focused on studying the circuits of the basal ganglia in the context of reward seeking behaviors in mice. Previous work has spanned a range of techniques and questions from behavioral measures of disease progression in human patients to the atomic structure of glutamate receptors.
Data derived from: Google Scholar.
"...The problem of actions, in other words the effect which the organism is striving to achieve, is something which is not yet, but which is due to be brought about. The problem of action, thus, is the reflection or model of future requirements (somehow coded in the brain); and a vitally useful or significant action cannot be either programmed or accomplished if the brain has not created a prerequisite directive in the form of future requirements that we have just mentioned."
-N. Bernstein (1962)
The major input nucleus of the basal ganglia - the striatum - is composed largely (>90%) of a class of projection neurons called 'medium spiny neurons'. This image is one such neuron filled with fluorescent dye through a small glass pipette and imaged on a two-photon microscope.
Multiple classes of neocortical neurons provide input to the dorsal striatum. Using reagents developed at Janelia we can label multiple classes (green vs. red) selectively.
A rendering of the anatomy of the basal ganglia in the mouse brain. The major input nucleus, the striatum, is shown in pink. Cortical and thalamic regions that provide input are shown in blue and major output targets are shown in green.
Retrogradely labeled dopamine neurons (green) and anterogradely labeled axonal projections (red) from the dorsal striatum.
All proteins are subject to brownian motion, including ion channels in the membranes of neurons. Thermal fluctuations in protein confirmation lead to stochastic opening and closing of the conducting pore of the ion channel. Although these fluctuations are assumed to be small we used computational modeling to show that in a very specific aspect of neuron function - the time when the membrane is very close to the threshold for firing an action potential - these fluctuations can dominate the spiking behavior of the neuron. As a consequence complicated patterns of spiking can emerge exclusively in stochastic models (but not in deterministic models which are generally used). This image graphically illustrated the unstable, but balanced state of the membrane around spiking threshold.
We used a series of retrograde tracing experiments to attempt to estimate the covariance of inputs to the dorsal striatum. The structure revealed a surprisingly important contribution of convergent targeting (~40% of variance) superimposed on a topographic organization (~60% of variance). This is similar to the estimate produced by a large anterograde tracing dataset by the Allen Brain Institute.