National Institutes of Health

Eunice Kennedy Shriver National Institute of Child Health and Human Development

2015 Annual Report of the Division of Intramural Research

Olfactory Coding and Decoding by Ensembles of Neurons

Mark Stopfer
  • Mark Stopfer, PhD, Head, Section on Sensory Coding and Neural Ensembles
  • Zane Aldworth, PhD, Postdoctoral Fellow
  • Yu-Shan Hung, PhD, Postdoctoral Fellow
  • Subhasis Ray, PhD, Postdoctoral Fellow
  • Kazumichi Shimizu, PhD, Postdoctoral Fellow
  • Kui Sun, MD, Technician

All animals need to know what is going on in the world around them. Brain mechanisms have thus evolved to gather and organize sensory information in order to build transient and sometimes enduring internal representations of the environment. Using relatively simple animals and focusing primarily on olfaction and gustation, we combine electrophysiological, anatomical, behavioral, computational, genetic, and other techniques to examine the ways in which intact neural circuits, driven by sensory stimuli, process information. Our work reveals basic mechanisms by which sensory information is transformed, stabilized, and compared, as it makes its way through the nervous system.

A temporal channel for information in sparse sensory coding

Brain circuits encode sensory information into a variety of neural representations ranging from dense, time-varying patterns of spikes in overlapping sets of neurons to sparse spikes in a few selective neurons. Sparse codes are used by nearly all sensory systems, including vision, audition, somato-sensation, and olfaction, and are thought to be advantageous for distinguishing between similar stimuli and for learning associations. In dense codes, the timing of spikes has been shown to contain sensory information, but the role of timing in the relatively few spikes in sparse sensory codes is unclear.

We used the olfactory system of awake locusts to test whether the timing of spikes in Kenyon cells, a population of neurons that respond sparsely to odors, carries sensory information to, and influences, follower neurons. First, we characterized two major classes of direct followers of Kenyon cells in a brain area called the β-lobe. With paired intracellular and field-potential recordings made during odor presentations, we found that the followers portray odor identity in the temporal patterns of their spikes but not in the spike rate, the spike phase, or the identities of active neurons. We also found that subtly manipulating the relative timing of Kenyon cell spikes, using temporally and spatially structured electrical microstimulation, reliably altered the responses of the followers. Our results show that even the remarkably sparse spiking responses of Kenyon cells provide information through odor-specific variations in timing on the order of tens to hundreds of milliseconds and that the variations determine responses downstream.

Sparse coding in sensory areas has been viewed as an outcome of, rather than as a substrate for, temporal processing. Stimulus-specific variations in spike timing provide a useful channel to increase the coding capacity of neurons while retaining the benefits of sparseness. Our results establish the importance of spike timing in sparse sensory codes.

Functional analysis of the lateral horn, a higher olfactory center

Understanding how information is processed through a neural circuit requires characterizing its structures and functions. The lateral horn of the insect brain is a prominent area widely thought to play several important roles in olfaction. These include maintaining the sparseness of responses to odors in Kenyon cells by means of feed-forward inhibition and encoding preferences for innately meaningful odors. However, relatively little was known about the structure and function of lateral horn neurons (LHNs), making it difficult to test these ideas. We surveyed more than 250 LHNs in locusts, using sharp-electrode recordings to test their responses to sensory stimuli, dye-fills to define their morphologies, and immunostaining to characterize their neurotransmitters.

We found a great diversity of LHNs, which we organized into ten highly distinct morphological classes, thus establishing a useful dataset. Surprisingly, we found no evidence to support a long-suspected role for these neurons in the feed-forward inhibition proposed to mediate olfactory response sparsening. Instead, through paired recording experiments and phase analyses of the odor-elicited responses of several types of neurons, we found that a different mechanism, feed-back inhibition from a newly identified giant GABAergic neuron (GGN), plays this role. Further, we found that all tested LHNs responded to all odors we tested, making it unlikely these neurons serve as "labelled" lines mediating specific, innate behavioral responses to specific odors. Instead, our results point to three other possible roles of LHNs: extracting general stimulus features such as odor intensity; mediating bilateral integration of sensory information; and integrating multimodal sensory stimuli. This work was recently published (Gupta N, Stopfer M, Curr Biol 2014;24:2247-2256).

Tradeoff between information format and capacity in the olfactory system

How does the nervous system 'decide' what format to use to encode information? The brain’s internal representation of the environment is built from patterns of neural activity bearing information about sensory stimuli. As this activity travels through the brain, a succession of neural circuits manipulates it into a series of coding formats, each thought to provide specific advantages for processing information. But each coding format has advantages and disadvantages. How the benefits of a given format are balanced against its costs is largely unknown.

One common coding format uses periodic inhibition to coordinate neural spiking into synchronous oscillations. Oscillatory synchrony has been proposed to offer several benefits, including enhancing the discriminability of sensory representations, and 'binding' diverse stimulus features into coherent percepts. We examined how neural oscillatory synchronization affects another measure of coding quality: the rate information is transmitted. We evaluated this potential tradeoff between coding format and information rate in the olfactory system of the locust. To test the possible tradeoffs imposed by synchrony, we needed a richly structured olfactory stimulus appropriate for the measurement of information properties. We decided to focus on stimulus timing because we could provide, with a synthetic odor plume, a broad sample of an environmentally meaningful stimulus space. Using artificial plumes based on the statistical structures of temporal variability measured outdoors in real odor plumes allowed us to calculate lower bound estimates of information in the temporal structure of an ethologically relevant stimulus. Thus, we delivered odorants as controlled, repeatable plumes while recording responses from populations of projection neurons as they transmitted information about the plume’s temporal structure. We evaluated the information content of neurons in terms of the mutual information rate between the temporal dynamics of the odorant stimulus and the neuronal response by finding the difference between the unconditional and stimulus-conditioned response entropies.

Surprisingly, our results showed that pharmacologically blocking synchronization by locally injecting picrotoxin led to a significant increase in information rate. Thus, the use of a synchronous coding scheme introduces a tradeoff: synchrony allows correlation coding and fine olfactory discrimination; however, by reducing the number of spikes and spike positions available for encoding information, it also reduces the ability of the system to rapidly transmit information about the stimulus. The inhibition-induced reduction in transmission capacity that we observed in the olfactory system is likely to occur in any neural circuit using periodic inhibition. Our results suggest reformatting to an oscillatory structure comes at a cost and represents a fundamental tradeoff between coding capacity and other aspects of format utility.

Spatiotemporal coding of individual chemicals by the gustatory system

Four of the five major sensory systems (vision, olfaction, somatosensation, and audition) are thought to be encoded by spatiotemporal patterns of neural activity. The only exception is gustation. Gustatory coding by the nervous system is thought to be relatively simple: every chemical ('tastant') is associated with one of a small number of basic tastes, and the presence of a basic taste, rather than the specific tastant, is represented by the brain. In mammals as well as insects, five basic tastes are usually recognized: sweet, salty, sour, bitter, and umami. The neural mechanism for representing basic tastes is unclear. The most widely accepted proposal is that, in both mammals and insects, gustatory information is carried through labelled lines: separate channels, from the periphery to sites deep in the brain, of cells sensitive to a single basic taste. An alternate proposal is that the basic tastes are represented by populations of cells, with each cell sensitive to multiple basic tastes.

Testing these ideas requires determining, point-to-point, how tastes are initially represented within the population of receptor cells and how this representation is transformed as it moves to higher order neurons. However, it has been very challenging to deliver precisely timed tastants while recording cellular activity from directly connected cells at successive layers of the gustatory system. Using a new moth preparation, we designed a stimulus and recording system that allowed us to fully characterize the timing of tastant delivery and the dynamics of the tastant-elicited responses of gustatory receptor neurons and their monosynaptically connected second order gustatory neurons, before, during, and after tastant delivery.

Surprisingly, we found no evidence consistent with a basic taste model of gustation. Instead, we found that the moth’s gustatory system represents individual tastant chemicals as spatiotemporal patterns of activity distributed across the population of gustatory receptor neurons. Further, we found that the representations are transformed substantially as multiple types of gustatory receptor neurons converge broadly upon follower neurons. The results of our physiological and behavioral experiments suggest that the gustatory system encodes information not about basic taste categories, but rather about the identities of individual tastants. Further, this information is carried not by labelled lines, but rather by distributed, spatiotemporal activity, which is a fast and accurate code. The results provide a dramatically new view of taste processing.

An identified gustatory second-order neuron in the Drosophila brain

Little is known, in any species, about neural circuitry immediately following gustatory sensory neurons, which makes it difficult to know how gustatory information is processed by the brain. By genetically labeling and manipulating specific parts of the nervous system, we identified and characterized a bilateral pair of gustatory second-order neurons in Drosophila. Previous studies had already identified gustatory sensory neurons that relay information to distinct parts of the gnathal (subesophageal) ganglia. To identify candidate gustatory second-order neurons, we took an anatomical approach. We screened about 5,000 GAL4 driver strains for lines that label neural fibers innervating the gnathal ganglia. We then combined GRASP (GFP reconstitution across synaptic partners) with presynaptic labeling to visualize potential synaptic contacts between the dendrites of the candidate gustatory second-order neurons and the axonal terminals of Gr5a–expressing sensory neurons, which have been shown to respond to sucrose. Results of the GRASP analysis, followed by a single cell analysis by FLP-out recombination, identified a specific pair of neurons that contact Gr5a axon terminals in both brain hemispheres and send axonal arborizations to a distinct region within the gnathal ganglia. To characterize the input and output branches, respectively, we expressed the fluorescence-tagged acetylcholine receptor subunit (Dα7) and active-zone marker (Brp) in the gustatory second-order neurons.

We found that input sites of the gustatory second-order neurons overlaid GRASP–labeled synaptic contacts to Gr5a neurons, while presynaptic sites were broadly distributed throughout the neurons’ arborizations. GRASP analysis and further tests with a new version of GRASP that labels active synapses suggested that the identified second-order neurons receive synaptic inputs from Gr5a–expressing sensory neurons, but not Gr66a–expressing sensory neurons, which respond to caffeine. The identified second-order neurons relay information from Gr5a–expressing sensory neurons to stereotypical regions in the gnathal ganglia. Our findings suggest an unexpected complexity for taste-information processing in the first relay of the gustatory system. We are presently following up on this work to identify additional second-order neurons and with optical imaging and intracellular electrophysiology experiments to characterize their functions and information-coding strategies.

Additional Funding

  • Additional funding from the Japan Society for the Promotion of Science to Kazumichi Shimizu


  1. Reiter S, Campillo Rodriguez C, Sun K, Stopfer M. Spatiotemporal coding of individual chemicals by the gustatory system. J Neurosci 2015; 35:12309-12321.
  2. Aldworth Z, Stopfer M. Tradeoff between information format and capacity in the olfactory system. J Neurosci 2015; 35:1521-1529.
  3. Huston SJ, Stopfer M, Cassenaer S, Aldworth ZN, Laurent G. Neural encoding of odors during active sampling and in turbulent plumes. Neuron 2015; 88:1-16.
  4. Miyazaki T, Lin TY, Ito K, Lee CH, Stopfer M. A gustatory second-order neuron that connects sucrose-sensitive primary neurons and a distinct region of the gnathal ganglion in the Drosophila brain. J Neurogenet 2015; 29(2-3):144-55.
  5. Kee T, Sanda P, Gupta N, Stopfer M, Bazhenov M. Feed-forward versus feedback inhibition in a basic olfactory circuit. PLoS Comput Biol 2015; 11:e1004531.


  • Maxim Bazhenov, PhD, Howard Hughes Medical Institute, The Salk Institute for Biological Studies, La Jolla, CA
  • Kei Ito, PhD, University of Tokyo, Tokyo, Japan
  • Chi-Hon Lee, MD, PhD, Program in Cellular Regulation and Metabolism, NICHD, Bethesda, MD


For more information, email or visit

Top of Page