Olfactory Bulb Network Model - O'Connor, Angelo and Jacob 2012

OB Mitral Cell Network Wiki

Development of O’Connor et al. (2012) olfactory bulb model to provide realistic input for the Vanier (2001) olfactory cortex model


An understanding of how significance is attributed to percepts of odour within the olfactory system is intimately tied to how the system is wired. The evidence from anatomical tracing studies has revealed random connection patterns between the olfactory bulb and olfactory cortex (Haberly and Price 1977, Scott et al. 1980). It has therefore been assumed that the system is auto-associational and all relationships are learned (Haberly and Bower, 1989). However, Vanier (2001) created a model of the olfactory system in which different connection patterns were explored. This exploration was both between the olfactory bulb and olfactory cortex and internally in the olfactory cortex. The goal was to replicate current source density contour plots for weak and strong shock stimulation to the lateral olfactory tract (Ketchum and Haberly, 1993, Figure 1). This work suggested that the only way of replicating the data was with structured wiring pattern of sub networks. This would allow intrinsic odour correlations to be embedded in the structure of the olfactory systems. It has been suggested that such a system might be evolved to track metabolic pathways associated with the ripening of fruit, nuts, and other potential important food sources (Chee-Ruiter 2000).

Figure 1 – Experimental current source density contour plots (Ketchum and Haberly, 1993).

There has recently been renewed interest in odour percepts fed by the expanding knowledge of the olfactory receptor genome and sensitivity to ligands types of individual receptor types. This can be seen in the construction of the Sense Lab database and the first Odor Spaces workshop held in Hannover, Germany, September, 2013. So we feel the time is right to resume development of the Vanier and Bower project.

The Vanier (2001) model could not completely replicate the current source density contour plots due to the simple nature of the olfactory bulb model. There is a recent model of the olfactory bulb (O’Connor et al. 2012, Figure 2) that can be adapted to provide the kind of oscillatory input (Figure 3) that would be required to provide closer replication of the current source density contour plots with the Vanier (2001) model.

Figure 2 – Olfactory bulb gap-junction connected network model created using neuroConstruct .

Figure 3 - Somatic voltage recordings from the O’Connor et al. olfactory bulb model.

The aim of this project is to allow collaboration on the development of the O’Connor et al. (2012) model so that it is capable of providing the input for Vanier (2001) olfactory cortex model. It has been placed on Open Source Brain to allow anybody who wishes to collaborate in the development of this model and the join in the debate about how the model should be developed. If you wish to collaborate in this project and or have data that may be useful for constraining this development, please make contact at simon.oconnor@btinternet.com.

A forum for general discussion of this project can be found on Github issue here: https://github.com/Simon-at-Ely/OlfactoryBulbMitralCell/issues/7


Chee-Ruiter 2000 California Institute of Technology, PhD Thesis.
Haberly and Bower, 1989 http://www.ncbi.nlm.nih.gov/pubmed/2475938
Haberly and Price 1977 http://www.ncbi.nlm.nih.gov/pubmed/68803
Ketchum and Haberly, 1993 http://www.ncbi.nlm.nih.gov/pubmed/8381858
O’Connor et al. 2012 http://www.frontiersin.org/computational\_neuroscience/10.3389/fncom.2012.00075/abstract
Scott et al. 1980 http://www.ncbi.nlm.nih.gov/pubmed/7451680
Vanier (PhD Thesis 2001) http://www.cs.caltech.edu/\~mvanier/extra/Vanier-thesis.pdf

Current road map

Prelude: Debugging the neuroConstruct implementation of the olfactory bulb model

Changes to neuroConstruct since the implementation of the olfactory bulb model O’Connor et al. (2012) require some limited debugging to synchronise model results for GENESIS and NEURON.

Phase 1: Improve understanding of ion channel shaping of mitral cell response

It is clear that the modelling of calcium ion kinetics in the calcium ion channel and the calcium dependent potassium channel requires verification as discussed in O’Connor et al. (2012). Burst firing in mitral cells is only observed in 50%\ of\ the\ population\ Ma\ and\ Lowe\ .\ We\ need\ to\ be\ able\ to\ understand\ the\ factors\ controlling\ this.
It\ would\ be\ advantageous\ to\ develop\ protocols\ for\ obtaining\ this\ information,\ including\ the\ fitting\ methods\ to\ be\ used,\ to\ guide\ the\ interaction\ with\ experimentalists.
A\ forum\ for\ discussion\ of\ Phase\ 1\ can\ be\ found\ here:\ https://github.com/Simon-at-Ely/OlfactoryBulbMitralCell/issues/3

Ma\ and\ Lowe\ 2010\ http://www.ncbi.nlm.nih.gov/pubmed/20600657
O’Connor\ et\ al.\ 2012\ http://www.frontiersin.org/computational\_neuroscience/10.3389/fncom.2012.00075/abstract\\

Phase\ 2:\ Incorporate\ timing\ modulation\ into\ olfactory\ bulb\ model
Inhibitory\ circuits\ and\ glutamate\ auto-excitation\ influence\ firing\ rhythms\ in\ olfactory\ bulb\ mitral\ cells.\ This\ can\ be\ simply\ modelled\ as\ mechanisms\ attached\ to\ the\ mitral\ cells\ .\ Having\ the\ ability\ to\ have\ a\ simple\ tuning\ control\ on\ the\ model\ will\ allow\ investigation\ of\ these\ effects\ within\ a\ connected\ olfactory\ bulb\ –\ olfactory\ cortex\ model\ and\ develop\ an\ insight\ into\ the\ influence.\
A\ forum\ for\ discussion\ of\ Phase\ 2\ can\ be\ found\ here:\ https://github.com/Simon-at-Ely/OlfactoryBulbMitralCell/issues/4

Christie\ and\ Westbrook\ 2006\ http://www.ncbi.nlm.nih.gov/pubmed/16495454\\
Mori\ et\ al.\ 1977\ http://www.ncbi.nlm.nih.gov/pubmed/198060\\
Saftenku\ 2005\ http://www.ncbi.nlm.nih.gov/pubmed/15784271\\
Schoppa\ and\ Westbrook\ 1999\ http://www.ncbi.nlm.nih.gov/pubmed/10570488\\
Schoppa\ and\ Westbrook\ 2002\ http://www.ncbi.nlm.nih.gov/pubmed/12379859\\

Phase\ 3:\ Incorporate\ stochastic\ input\ from\ modelling\ of\ olfactory\ epithelium
Recordings\ of\ olfactory\ receptor\ neurons\ overlaid\ on\ stimulus\ response,\ exhibit\ sniffing\ cycles\ and\ basal\ firing.\ The\ convergence\ of\ the\ olfactory\ receptor\ neurons\ on\ to\ apical\ dendrite\ tufts\ of\ a\ single\ glomerulus\ is\ large30,000:~80 mitral cell population.

For the purposes of the olfactory bulb model, thought will have to be given how this stochastic input might be modelled. This will involve making a choice between modelling olfactory receptor neuron morphology and modelling a mass of point source neurons.
The Vanier 2001 model found that stochastic background firing input from the olfactory bulb was needed to elevate the membrane potential in the olfactory cortex model sufficiently so there response to shock stimuli had the short latency found in experimental recordings (Ketchum and Haberly, 1993).

This requirement for background stochastic firing agrees with experimental recordings in the olfactory bulb (Vanier 2001 derived from Bhalla and Bower 1993, Figure 4).

Figure 4 – Raster plot of background spiking in the olfactory bulb .

A forum for the discussion of phase 3 can be found here: https://github.com/Simon-at-Ely/OlfactoryBulbMitralCell/issues/5


Bhalla and Bower, 1993 http://www.ncbi.nlm.nih.gov/pubmed/9257234
Ketchum and Haberly, 1993 http://www.ncbi.nlm.nih.gov/pubmed/8381858
Reisert, 2010 http://www.ncbi.nlm.nih.gov/pubmed/20974772
Reisert and Zhao 2012 http://www.ncbi.nlm.nih.gov/pubmed/21875979
Vanier (PhD Thesis 2001) http://www.cs.caltech.edu/\~mvanier/extra/Vanier-thesis.pdf

Phase 4: Expand model from a single glomerulus to multiple glomeruli and connect to piriform cortex

Within the rat olfactory bulb there are about 1000 glomeruli, while the current model represents a single glomerulus. Therefore to connect to the olfactory cortex model parallel copies of the model will need to be used as an input source.

N.B. The reconstructions of mitral cells with fitted passive parameters used in the olfactory bulb model are likely to be from different glomeruli (see orientation of mitral cells in Figure 2). The passive parameters differ over a range and produce asynchronous sub threshold currents and firing thresholds that are not synchronised by gap junctions (see Figure 3). It would be interesting to see if mitral cells reconstructed with the passive parameters fitted for the same glomerulus have tuned passive parameters to allow then to work more as a single population.

A forum for the discussion of phase 4 can be found here: https://github.com/Simon-at-Ely/OlfactoryBulbMitralCell/issues/6