Research Interests :
I am fascinated by the complexity of individual neurons. In the central nervous system, a typical neuron receives thousands of inputs from other neurons, resulting in thousands of input events per second. Some of those input events may represent a particular stimulus (a sensation or an intended action, for example). Many others, however, carry information describing the surrounding context—general things like emotional state, alertness and attention, and also specific competitive interactions from alternative stimuli whose comparative importance must be considered. The neuron’s task is to condense these many and disparate inputs, continuously and in real time, into a single output stream that reflects both the stimulus and the broader context. This is not to suggest that the output from a single neuron could tell you everything you needed to know about a situation—quite the contrary! But however small the information content of a single neuron’s output, arriving at that output is a complex process that takes into account many contextual variables.
Given the complexity of their task, it seems fitting that neurons are themselves highly specialized and complex. Neurons often have elaborate, highly branched dendritic trees that allow them to process inputs in a partially compartmentalized fashion. Different types of inputs can activate distinct intracellular signaling pathways, thereby gaining unique signatures, and different signaling pathways have widely divergent temporal and spatial characteristics, meaning some inputs affect the entire cell while others affect only a local region and interact with only a few other inputs. Interactions between different pathways can be cooperative or antagonistic, linear or non-linear, and can have long-lasting effects on cellular properties. As scientists discover more new pathways, signaling molecules and regulatory mechanisms, the number of variable combinations that a single neuron can take into account steadily increases.
My hope as a neurobiologist is to contribute to our understanding of how single neurons translate their many inputs into a single output, and of how information about the stimulus and context are encoded in that output. I believe that a good way to address these goals is to combine controlled neurobiological experiments with computer simulations: experimental information is used to constantly refine the computer model, which is in turn used to help formulate new hypotheses and guide future experiments.
I am presently working with Dr. Dieter Jaeger to determine how neurons in the globus pallidus (GP) process synaptic inputs. The GP is a collection of neurons deep within the brain that forms a part of a larger circuit called the basal ganglia. The basal ganglia circuit is involved in movement control, and is the subject of a great deal of scientific research because deficits in the circuit are associated with Huntington’s and Parkinson’s Diseases. The GP occupies a central position within the circuit, as it is directly interconnected with both the primary input areas (striatum and subthalamic nucleus) and output areas (substantia nigra pars reticulata, entopeduncular nucleus / internal segment of globus pallidus). Previously, Jesse Hanson, Yoland Smith and Dieter Jaeger demonstrated that neurons in the GP have voltage-dependent sodium channels expressed at excitatory synapses and that moderate levels of synaptic excitation can trigger dendritic action potentials that propagate to the soma (Hanson, Smith & Jaeger, 2004). Dendritic action potential initiation has been found in only a few other types of mammalian neurons and suggests that GP neurons process synaptic inputs in a unique way.
To figure out what this unusual specialization of GP neurons means for their input-output properties, I am pursuing two complimentary lines of investigation. First, I use computer models of GP neurons to systematically explore the influence of dendritic action potentials and active dendritic conductances on input processing in GP neurons; second, I use whole-cell and cell-attached patch clamp recordings from GP neurons in the brain slice preparation to study local synaptic interactions in GP dendrites.
The computer models capture the detailed three-dimensional structure of real GP neurons. They are based on biocytin-filled rat GP neurons characterized and reconstructed by Jesse Hanson. The models include 11 types of voltage-dependent ion channels based on published electrophysiology data, and accurately reproduce the basic electrical properties of GP neurons in brain slices. Simulations are carried out using the GENESIS software package (http://www.genesis-sim.org/GENESIS/).
Electrophysiology experiments are done using standard brain slice procedures. Recorded neurons are filled with a fluorescent dye to allow visualization of the dendrites. Once the dendrites are visible, stimulation electrodes can be positioned to activate synaptic inputs locally. The same electrodes that are used to activate synaptic inputs can also be used to apply antagonists of receptors or ion channels specifically to the sites of input, and synaptic interactions can be probed using multiple stimulation electrodes.