I have been working as an intern during Summer 2010 on a research project with the Pittsburgh Supercomputing Center in conjunction with Dr. Nathan Urban (http://www.cmu.edu/bio/faculty/urban.html) on developing a biologically realistic computational model of mitral cells in the mouse olfactory bulb. My work has resulted in a model of a mitral cell using the computational package NEURON (http://www.neuron.yale.edu/neuron/). In the process, I have familiarized myself with the NEURON package, which is increasingly being used by electrical engineers as they interface with neuroscientists and biologists. Additionally, neuron is being increasingly integrated with a host of programming platforms (including python), which is likely to make it a powerful tool in simulating multi-compartment electrical models, like neurons. Neuron treats models of cells as assemblages of capacitors with nonlinear resistors, drawing from early experimental and theoretical work that demonstrated that the biophysical excitability of a cell could be treated using this framework. Many such neurons can be connected to form networks. Consequently, microcircuits in the brain can be modeled as a series of nonlinear electro-chemical circuits. The model itself is implemented using partial differential equations. The research thus involves nonlinear circuits, dynamical systems theory, and numerous concepts from the area of modeling, analysis, mathematical neuroscience and control of systems. I would like to continue the work as a formal undergraduate research course during Fall 2010 and beyond. Towards that end, I have outlined the proposal in more detail below and seek approval to work on this project under Dr. Urban's guidance, for my undergraduate research, and get course credits for the same
The objective of the research is to understand the firing patterns of mitral cells in the olfactory bulb of the brain. The olfactory bulb is composed of several modules called glomeruli. Within each glomerulus, olfactory receptor neurons (neurons that sense chemical compounds in the nose and transform this information about chemical identity into an electrical signal) synapse onto mitral cells, the principal neurons of the olfactory bulb. These cells in turn interact with other cells in the olfactory network and transmit a signal to the olfactory cortex. Mitral cells, even those connected to the same glomerulus (these cells receive input signals with a correlation of near 1) are not homogeneous, despite structural similarities. Notably, their responses to various electrical inputs are markedly varied. For instance, in response to a DC current injection, some cells show a tonic-spiking pattern while in others a bursting pattern is visible. The Urban lab is interested in finding out why this happens, and the problem lends itself well to computational studies. The variability in firing patterns could be due to two factors: (Note that the condition 2 is unlikely as these networks are pharmacologically isolated)
1) Differences in intrinsic properties of mitral cells Differences in certain properties in mitral cells, especially differences in channel concentrations could be the cause of the differences in firing patterns seen in different cells. Bursting in cells can be caused by a slow acting hyperpolarizing current, which takes a while to activate (during which time tonic spiking is observed in the cell), and which, after activating, prevents the membrane potential from reaching threshold. Such a current could be present with higher conductances in some mitral cells while having lower channel concentrations in other mitral cells, displaying the differences in firing patterns that are observed. The presence of other currents, such as depolarizing Calcium currents coupled with a Calcium activated hyperpolarizing Potassium current could also provide the same effect.
2) Differences in neurons may confer to them unique responses to incoming stimuli that distinguish them from the other cells in the network. However, their connectivity and the properties in the network are likely to play an important role in shaping those responses. Some interactions, like the emergence of stochastic synchrony (Galan et al) in the bulb may help to facilitate increased similarity in mitral cell firing patterns, while other slow decorrelative mechanisms mediated through lateral inhibition (Arevian et al, Friedrich et al) may make mitral cell responses more different. As a result, understanding how networks can shape the responses of intrinsically diverse cells in like to be a key question in neural coding.
Interneurons called granule cells have reciprocal dendrodendritic synapses with mitral cells of several different types (coming from different glomeruli), causing them to be able to take in information from one type of mitral cell, and use that information to affect mitral cells of other glomeruli. For example, granule cells can be involved in lateral inhibition of mitral cells, taking information from a firing mitral cell and then inhibiting the other mitral cell type(s) that it is connected to. This allows only one type of mitral cell to fire at one time.
There is one more type of interneuron, called a periglomerular cell. Periglomerular cells form dendrodendritic synapses with multiple mitral cells within a single glomerulus, therefore ensuring that they only allow communication between multiple mitral cells of the same type. Periglomerular cells also can form axodendritic connections outside glomeruli and dendroaxonic connections to olfactory receptor neurons, which they inhibit.
Mitral cells are also coupled electrically through gap junctions Different mitral cells may be wired very differently into these intricate networks, causing differences in firing patterns and other behavior between cells.
Single cell Models: The mitral cell model will be started as a single compartment of a mitral cell. Parameters will be modified based on biological data from the literature. Such data include the types of current channels present in such cells, and electrophysiological data such as resting membrane potential, input resistance, and responses to various current injections. This model will then be expanded to include several compartments, simulating an entire mitral cell in isolation.
Network model: After the single cell model is developed, we will proceed to a model of the glomerulus that will include mitral cells, interneurons, and olfactory receptor neurons.
With models at both the levels, we can begin addressing the question of variability in mitral cell firings. The models themselves will continue to be refined based on data from the Urban lab and other labs working in the area.
Expected Results: These compartmental models can provide valuable insights into how individual glomeruli function as units and how specific parameters in mitral cells affect their individual firing patterns. We will find out exactly how different currents operate in mitral cells and what excitatory/inhibitory networks in glomeruli have the greatest and most important effects on the signal transmission through the olfactory bulb. They will thus form computational 'test beds' to complement the wet-lab experiments, and help to obtain additional insights into the underling physiological mechanisms responsible for the variability in mitral cell firing patterns, and other problems of interest to the Urban Lab.
Arevian, Armen C., Vikrant Kapoor, and Nathaniel N. Urban. "Activity-dependent Gating of Lateral Inhibition in the Mouse Olfactory Bulb." Nature Neuroscience 11.1 (2008): 80-87. Print.
Friedrich, Rainer W., and Gilles Laurent. "Dynamic Optimization of Odor Representations by Slow Temporal Patterning of Mitral Cell Activity." Science 291 (2001): 889-94. PubMed. Web. 24 June 2010.
Galan, Roberto F., G. Bard Armentrout, and Nathaniel N. Urban. "Stochastic Dynamics of Uncoupled Neural Oscillators: Fokker-Planck Studies with the Finite Element Method."Physical Review E 76.5 (2007). Print.
Hayar, Abdallah, Sergei Karnup, Michael T. Shipley, and Matthew Ennis. "Olfactory Bulb Glomeruli: External Tufted Cells Intrinsically Burst at Theta Frequency and Are Entrained by Patterned Olfactory Input." The Journal of Neuroscience 24.5 (2004): 1190-199. PubMed. Web. 18 June 2010.
Heyward, Philip, Matthew Ennis, Asaf Keller, and Michael T. Shipley. "Membrane Bistability in Olfactory Bulb Mitral Cells." The Journal of Neuroscience 21.14 (2001): 5311-320.PubMed. Web. 24 June 2010.
Hines, Michael, Ted Carnevale, and Gordon Shepherd. Using the NEURON Simulation Environment. 2003. Print.
Onoda, N., and K. Mori. "Depth Distribution of Temporal Firing Patterns in Olfactory Bulb Related to Air-Intake Cycles." Journal of Neurophysiology 44.1 (1980): 29-39. PubMed. Web. 28 June 2010.
Schoppa, Nathan E., and Nathan N. Urban. "Dendritic processing within olfactory bulb circuits." TRENDS in Neuroscience 26.9 (2003): 501-06. Print. Wilson, Donald A., and Regina M. Sullivan. "The D2 Antagonist Spiperone Mimics the Effects of Olfactory Deprivation on Mitral/Tufted Cell Odor Response Patterns." The Journal of Neuroscience 15.8 (1995): 5574-581. PubMed. Web. 21 June 2010.