Stochastic Synchrony

A simple biophysical mechanism for generating beta/gamma network oscillations


Oscillatory activity in the beta (20-50 Hz) and gamma (30-90 Hz) frequency bands is a prominent feature of activity in many brain areas including the olfactory bulb. Concerning the olfactory system, one long standing hypothesis has been that gamma oscillations are caused by the recurrent dendrodendritic inhibition between mitral cells (projection neurons) and granule cells (local interneurons) that is a main characteristic of olfactory bulb (antennal lobe) circuitry. An apparent problem of this hypothesis is that mitral cells belonging to different glomeruli can synchronize even when they are not directly connected. Mitral cells of different glomeruli can however receive common inhibitory feedback from a given pool of granule cells, thus resolving this paradox. Nevertheless, the exact mechanism that accounts for synchronized activity in the beta/gamma frequency band remains unknown.

The inhibitory feedback into the mitral cells is a stochastic process, since GABA-release from granule cells is known to be random. We have studied this form of stochastic inhibitory feedback in computer simulations and in vitro electrophysiological experiments. Our studies reveal that stochastic inhibitory feedback can generate robust coherent activity in the beta/gamma range, as shown in the two figures below. Note the good agreement between simulations (figure 1) and experiments (figure 2).



Figure 1: Network Simulations. (First row panels from the top) Postsynaptic inhibitory feedback current into two mitral cells. (Second row panels) Reponse of the mitral cells to sensory input plus inhibitory feedback: Note the improvement of relative timing when the inhibitory feedback is correlated. (Third row panels) Local field potential calculated as the averaged response of 20 mitral cells. (Fourth row) Power spectral density of the local field potential as an estimator of coherent neural activity during 10 seconds.

stochastic synchrony in simulations



Figure 2: Experimental Data. Same as in Figure 1 but with electrophysiological recordings instead of simulated neurons. Note the agreement between both figures.

stochastic synchrony in real data

Back to research website



Last update: September 2009