Skip to content


Method Development for Studying Human Olfaction

Magnetic resonance imaging (MRI) is one of the most powerful tools to study the human olfactory system in vivo. Due to the small size of olfactory brain regions and their proximity to the sinuses, standard MRI techniques often contain significant artifacts. Our group is developing a series of imaging protocols to optimize both functional and structural signals from olfactory regions. For better structural imaging of the olfactory bulb and olfactory tracts, we are developing protocols to achieve sub-millimeter resolution of human in-vivo images with minimal artifacts. For improved functional imaging, we are taking advantage of multi-echo multi-slice EPI acquisition protocols to significantly improve sensitivity to BOLD signals. Finally, we are using ultra-high field strength at 7T to capture signals in small nuclei of the brainstem and basal forebrain to assess their role in olfactory processes. Collectively, these novel approaches have the potential for allowing unprecedented access to the human olfactory system.

While MRI approaches have provided crucial insights into the central mechanisms of olfactory processing, relatively little is known about the peripheral mechanisms of human olfaction. Historically, investigations of human olfactory receptor class have been woefully understudied compared to mouse models, but in collaboration with Dr. Tom Bozza at Northwestern University, we have begun to establish a novel repertoire of antibodies that target human olfactory receptors.

Olfactory Cognitive Maps

It has been proposed that items, locations or events can be organized in map-like mental representations, also referred to as cognitive maps. We recently started exploring the concept of cognitive maps in the olfactory domain. In a previously published paper (Bao et al., 2019 in Neuron) [PDF link], our lab showed that human subjects map two-component odor mixtures of varied intensity ratios onto a two-dimensional space, the axes of which represent the intensity gradients of individual odor components. The formation of these “abstract” olfactory cognitive maps was accompanied by the emergence of grid-like representations in ventromedial, piriform and entorhinal cortices. Building on these initial findings, we are now testing whether similar maps also support odor-guided navigation in a virtual environment, offering a more ecological approach. To this end, we are combining virtual reality software alongside olfactometry methods and fMRI techniques to test whether grid-like representations support navigation based on olfactory landmarks, and how these representations are modulated by reward context.

How does the olfactory system contend with variations in odor stimulus features?

The natural odors we encounter in daily life are complex, multi-component mixtures that change dynamically across space and time. Furthermore, the same scent can appear across a wide variety of contexts and may face interference from other odorants. These variations introduce significant perceptual noise, making it challenging for the olfactory system to reliably identify odors to support robust decision-making and behavior. However, we know very little about whether and how the olfactory system might account for such noise during odor identification and learning. What types of neural mechanisms might enable accurate olfactory perception, even in the face of uncertainty? We combine computational modeling with ultra-high field 7T fMRI and intracranial EEG to address these questions.

Temporal Processing in the Olfactory System

Earlier work in the human olfactory system has focused on understanding how different brain regions are engaged in processing odor information. How odor responses unfold over time, and how variations in stimulus features such as odor concentration influence temporal dynamics of neural activity, remain poorly understood. To address these questions, we use intracranial EEG recordings in patients with medically refractory epilepsy, enabling us to obtain electrophysiological signals at millisecond resolution from olfactory regions such as piriform and orbitofrontal cortices. 

Scents by their very nature are also dynamic stimuli that change rapidly over time. However, not every time point of an odor signal necessarily carries the same information about the stimulus. For example, molecules making up a complex scent might have different volatilities, influencing their relative time of arrival at the nose, and consequently the precise stimuli composition experienced at any given time point. Our lab is employing both behavioral and fMRI imaging approaches to explore how the human olfactory system exploits such olfactory stimulus dynamics to inform odor percepts.