Key points

◆We demonstrated that theta-band activity emerging shortly after odor onset (from around 0.1 s) encodes the chemical features of odorant molecules.
◆The precision of this early theta-band encoding correlated with individual differences in odor discrimination ability and predicted whether participants could correctly discriminate odors in an experimental task.
◆These findings suggest that low-level neural encoding occurring at the initial stage of olfactory processing provides a critical foundation for odor discrimination.

Summary

Odors are chemical molecules, and at the earliest stages of olfactory processing, their molecular structures are believed to be represented as neural activity patterns. However, the extent to which this early neural coding contributes to human olfactory ability has remained unclear. Associate Professor Masako Okamoto and colleagues at the Graduate School of Agricultural and Life Sciences, The University of Tokyo, analyzed high-density electroencephalography (EEG) recordings while participants smelled nine different odors. The results revealed that theta-band activity (around 4 Hz) occurring 80–640 ms after odor onset encoded the physicochemical features of odor molecules. Importantly, participants with higher coding precision in this theta range showed better odor discrimination ability. Trial-by-trial analyses further showed that decoding accuracy was higher for correctly discriminated odors, indicating behavioral relevance of this early neural code. These findings demonstrate that early brain activity following odor presentation encodes the chemical features of odor molecules, and that the precision of this encoding supports odor discrimination performance. The results open possibilities for applying odor-evoked neural patterns as objective indicators for diagnosing olfactory dysfunction or for developing evidence-based olfactory training programs.

Research details

　In everyday life, humans can rapidly extract diverse information from odors—for example, sensing the season from the scent of flowers or detecting danger from the smell of smoke. Such perceptual judgments rely on the brain’s ability to swiftly interpret the chemical properties of odor molecules. Although odors are physical molecules whose structures are thought to be represented by neural activity in early processing stages, the precise timing and behavioral relevance of this encoding in humans have remained largely unclear.
　In this study, we investigated what types of information the brain encodes within a few hundred milliseconds after smelling an odor and how this relates to individual olfactory performance. Thirty-two healthy adults (aged 19–28 years) were presented with nine odorants while their brain activity was recorded using a high-density EEG system (Note 1). Time–frequency analyses were combined with decoding (Note 3) and representational similarity analysis (RSA) (Note 4) to identify when and in which frequency bands odor information was represented. Participants also completed standardized olfactory tests (threshold, discrimination, and identification) and related questionnaires.
　Our analyses revealed that theta-band activity (approximately 4–7 Hz) from 80–640 ms after odor onset encoded the physicochemical features of odor molecules. Furthermore, individuals who exhibited more precise encoding around 300 ms performed better on the odor discrimination test (Note 6). In a separate experiment using a two-alternative forced-choice task conducted during EEG recording, decoding accuracy was higher in correct trials, confirming that theta-band activity directly contributes to odor discrimination behavior. In contrast, delta-band activity (1–3 Hz), which emerged after approximately 720 ms, encoded the pleasantness or unpleasantness of odors. Participants who showed greater precision in delta-band encoding also reported a stronger tendency to enjoy odors in their daily lives.
　Together, these findings suggest a temporal transformation in olfactory processing: early theta-band activity represents low-level molecular features, whereas later delta-band activity reflects subjective affective evaluation. This dynamic neural coding may underpin distinct aspects of odor perception and related behaviors. In the future, such early neural signatures could be applied to objective assessments of olfactory disorders or to the development of training methods aimed at enhancing olfactory function.
　This study was approved by the Ethics Committee of the University of Tokyo.

Researchers

・Mugihiko Kato — Doctoral Program, Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo (at the time of study)
・Toshiki Okumura — Doctoral Program, Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo (at the time of study)
・Kazushige Touhara — Professor, Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences / Principal Investigator, International Research Center for Neurointelligence (WPI-IRCN), The University of Tokyo
・Masako Okamoto* — Associate Professor, Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo

Publication Information

Journal: The Journal of Neuroscience (2025)
Title: Behavioral relevance of early neural coding of low-level odor features in humans
Authors: Mugihiko Kato, Toshiki Okumura, Kazushige Touhara, and Masako Okamoto*
DOI: 10.1523/JNEUROSCI.0203-25.2025
URL: https://www.jneurosci.org/lookup/doi/10.1523/JNEUROSCI.0203-25.2025

Funding

This work was supported by the Grant-in-Aid for JSPS Fellows (23KJ0377 to M.K.), the Grant-in-Aid for Scientific Research on Innovative Areas (21H05808, 23H04335, 25H00998 to M.O.), and the JST-MIRAI Program (JPMJMI17DC, JPMJMI19D1 to K.T.).

Glossary

(Note 1) High-density EEG – EEG recorded using 64–256 electrodes placed on the scalp. Compared to conventional low-density EEG, it provides finer spatial resolution of voltage distribution and allows more accurate estimation of neural sources.
(Note 2) Olfactory event-related potentials (OERPs) – Brain responses elicited by odor stimulation, typically showing characteristic components (N1, P2, P3) at 300–1300 ms post-stimulus. Although traditional studies require averaging many trials due to low signal-to-noise ratios, this study successfully analyzed single-trial responses using machine learning.
(Note 3) Decoding analysis – A machine-learning approach to predict stimuli, perceptions, or cognitive states from multivariate neural activity. Successful decoding indicates that specific information is encoded in the brain activity.
(Note 4) Representational similarity analysis (RSA) – A method for inferring what information is represented in the brain by comparing the similarity structure among neural patterns with that among stimulus or perceptual features.
(Note 5) Theta and delta bands – EEG frequency ranges associated with distinct neural processes. Theta (≈ 4–7 Hz) is linked to memory and olfaction, while delta (≈ 1–3 Hz) reflects large-scale integrative processing. This study showed that each carries different odor information (molecular vs affective).
(Note 6) Odor discrimination ability – The ability to distinguish between multiple odors, typically assessed by determining whether two odor samples are the same or different, or by identifying the odd odor among three samples.

Contact Information

For scientific inquiries:
Masako Okamoto, Ph.D.
Associate Professor, Laboratory of Biochemistry,
Department of Applied Biological Chemistry,
Graduate School of Agricultural and Life Sciences, The University of Tokyo
Tel: +81-3-5841-8043
Email: a-okmoto (at) g.ecc.u-tokyo.ac.jp

For institutional press inquiries:
Public Relations Office, General Affairs Team,
Graduate School of Agricultural and Life Sciences / Faculty of Agriculture, The University of Tokyo
Tel: +81-3-5841-5484 Fax: +81-3-5841-5028
Email: koho.a (at) gs.mail.u-tokyo.ac.jp
(Please replace “(at)” with “@” when sending email.)