Harvard University - Department of Molecular & Cellular Biology

FEEDBACK CIRCUITS IN OLFACTION [MURTHY LAB]

by David H Gire and Dan Rokni

December 19th, 2012

(l to r) Dan Rokni, David Gire, and Venki Murthy   

What is the function of a sensory system? One straightforward idea is that it strives to create a complete and accurate internal representation of the external world, which enables organisms to react to the changing environment. Yet our everyday experience clearly shows that the representation is not complete, and perhaps not even accurate in the literal sense of the word. Our senses filter out information that is of little importance by mechanisms of selective attention. What these mechanisms are is not clear. A common hypothesis is that feedback projections from higher brain areas such as the cortex to more peripheral sensory regions actively suppress or enhance responses to stimuli based on their importance at any given moment. Importantly, such feedback will depend on the cortex having a good model of what to expect in the world.  In a recent paper published in Neuron, the Murthy lab reported the first direct functional examination of feedback connections within the olfactory system, shedding light on a century old mystery in neural processing.  
Feedback projections in the olfactory system were first demonstrated anatomically in the early 1900s by Santiago Ramon y Cajal, a towering figure in neuroscience. He noted that axons from different olfactory cortical structures feed back onto the first stage of olfactory processing in the brain – the olfactory bulb (OB). These and other anatomical data indicated that the feedback projections target local inhibitory neurons in the OB suggesting that they may indeed serve to inhibit OB output. However, due to an inability to selectively manipulate these connections, their role in neural processing has been unsolved for the last century.  
In order to specifically activate the projections from the cortex, we expressed the light-sensitive cation channel, channelrhodopsin-2 (ChR2) in neurons within a region of the rat olfactory cortex called the anterior olfactory nucleus (AON).  Shining blue light on the OB resulted in specific activation of the feedback projections. We found that feedback fibers from the AON indeed activate diverse types of local inhibitory interneurons, resulting in strong inhibition in the output neurons of the OB, the mitral cells. Surprisingly, we also found that cortical feedback provides direct excitation to mitral cells, which precedes temporally the more indirect inhibition through interneurons. The combination of excitation and inhibition is capable of controlling the timing and the number of action potentials in mitral cells, and thus information transfer from the OB to the rest of the brain.
Interestingly, the excitatory effect of feedback fibers on mitral cell output is only evident when cells are mildly active, but not when they are highly active or completely silent. This activity-dependent effect of feedback inputs led us to postulate that feedback from AON to the OB could be involved in detection of low concentration odorants within a noisy background of masking odors. One could imagine that when a mouse is searching for food in an environment that is rich with high concentration masking odorants unrelated to food, many mitral cells are firing at high rates in response to the masking odorants. When the mouse will first encounter odorants from a food source they will likely be at low concentrations and therefore elicit only mild activity in mitral cells. Activation of feedback axons in this context will temporarily increase the firing of the mildly active cells (responding to food odors) and decrease the highly active cells (responding to the masking environment). This temporary reversal of the relationship between firing rate and odor concentration may be used for the cortex to “query” the OB about any low concentration odors of interest that may be masked by the high concentration background.
Ongoing experiments using optogenetic strategies to transiently and reversibly silence feedback connections during odor-guided behavior will allow us to test hypotheses generated by our current study.

Read more in Neuron or download the PDF

Read more in the Harvard Gazette:  Sniff Mechanics: Research explores how odors are processed in the brain

View Venkatesh Murthy's Faculty Profile