October 21, 2013

The ISO settings of olfaction

Neurobiologists at the Friedrich Miescher Institute identify a novel feature of the olfactory system that resembles functions engineered for digital cameras. Interneurons in the olfactory bulb stabilize odor-encoding neural activity patterns when the intensity of an odor stimulus varies over a large range. This astonishing stability of “odor images” is achieved by an intriguing mechanism that is likely to facilitate the classification, recognition and memory of odors.

Sometimes the brain solves issues the same way as everyday gadgets do, and we only realize it, as we start to understand the brain better. In a recent publication in Nature Neuroscience, Rainer Friedrich, Group Leader at the Friedrich Miescher Institute for Biomedical Research, and his team show that the olfactory bulb, the site in the forebrain involved in the perception of odors, uses features analogous to the latest features of digital cameras.

In digital photography the most challenging situations are the ones in low light, such as after the sunset when you still see everything clearly but it is getting darker; or when there is a big difference between light intensities, like the Chagall windows illuminated by the sun in the rather dark Fraumünster in Zürich. Modern digital cameras offer two features in these situations: In the first instance you can increase film speed, the ISO setting, to increase the light sensitivity of the camera; in the second case you can use high dynamic range (HDR) imaging, where the camera takes differently exposed pictures of the same subject and combines them into one.

Friedrich and his group could now show, that very similar strategies are used in the brain as we distinguish scents. The brain represents scents by patterns of activity across many neurons that may be considered “pixels” in an “odor image”. It is important that low intensity pixels in these images are not lost in noise and that high-intensity pixels do not saturate. In the first stage of odor representation – the glomeruli of the olfactory bulb – this problem is addressed by bringing together the axons of millions of odorant receptor neurons, which collectively have a large dynamic range. At the next stage, however, odor information converges on a limited number of mitral cells that eventually transmit odor images to higher brain areas. Mitral cells must thus be able to adjust their sensitivity to the appropriate dynamic range very quickly to continuously optimize the odor image.

Using optogenetics, Friedrich and his collaborators could show that specific interneurons in the input layer of the olfactory bulb are coupled to mitral cells by both electrical connections and inhibitory chemical synapses. Interneurons thus boost responses of mitral cells through their electrical connections when the odor is weak, but attenuate responses of mitral cells through their inhibitory synapses when the odor is strong. This operation is similar to the automatic adjustment of ISO settings in a digital camera. Moreover, these interneurons increase the sensitivity of some output neurons but decrease the one of others. Output neurons therefore collect multiple “odor images” with different sensitivities, analogous to HDR imaging. The combination of these ISO and HDR strategies results in odor images that are nearly invariant to changes in odor concentration, which is highly important for the classification, recognition and memory of odors in higher brain areas.

“The intensity-invariance of odor images at higher stages of processing has been a mystery. We were surprised to find that this phenomenon can be explained by the combination of two basic strategies which in turn can be mapped onto a defined neuronal microcircuit,” comments Friedrich. “It is fascinating that the brain uses computational strategies similar to those invented by engineers. This suggests that it may be possible to decompose extremely complex higher brain functions into tractable elementary computations.”

Original publication
Zhu P, Frank T, Friedrich RW (2013) Equalization of odor representations by a network of electrically coupled inhibitory interneurons. Nature Neuroscience (advanced online publication)

On Rainer Friedrich
Rainer Friedrich is a group leader at the FMI and Professor at the University of Basel. He is interested in the strategies that the brain has developed to extract, store and utilize information about its environment. He and his team are using the olfactory system of zebrafish as models to study the function of neuronal circuits and their dysfunction in disease.
» More about Rainer Friedrich

About this site2018 © FMI Basel Switzerland