Are Pleasant and Unpleasant Odors Handled Separately in the Brain?

From a deeply unpleasant odor, through a neutral one, to a fragrance you love: pleasantness seems expressible on a single scale. But does the brain use a single mechanism to handle it? In 2026, Vivek Sagar and colleagues investigated this question using an existing dataset of detailed brain recordings collected while people smelled odors.

Their results indicated that pleasant and unpleasant odor information is represented separately in olfactory regions, while parts of the frontal brain show information shared between them. That common feature was not the direction of liking or disliking, but how far an odor lay from neutrality.

Approximately 18 hours of recordings for each of three people

The analysis used fMRI data from three healthy people aged 23–24 who smelled 160 monomolecular odors. Each person underwent approximately 18 hours of scanning, with 4,320–4,560 odor-presentation trials per person. Rather than measuring many people briefly, this study measured a small number repeatedly and in detail.

Participants also rated odor characteristics, scoring pleasantness from −10, dislike very much, through 0, neutral, to +10, like very much. Each odor’s pleasantness was rated at least three times. The researchers divided pleasant and unpleasant odors according to each individual’s own ratings.

They focused on four regions: the piriform cortex and amygdala, involved in olfactory processing, and the orbitofrontal cortex and ventromedial prefrontal cortex. fMRI does not record individual neurons; it captures signals related to blood flow and oxygenation in small spatial units. Here, the researchers analyzed the activity patterns formed by groups of these units.

The direction of preference and distance from neutrality

The researchers derived two perspectives from pleasantness ratings. One was a directional scale running from dislike to like. On this scale, −8 and +8 lie on opposite sides.

The other measured distance from zero. Both −8 and +8 are eight units from neutrality. The authors treated this absolute value as salience: both very pleasant and very unpleasant odors have substantial emotional significance.

Salience here is not odor concentration or intensity itself. Nor is it a direct measurement of conspicuousness, arousal, or the direction of attention. Although salience has different meanings in different fields, this paper defines it as the absolute value of pleasantness.

Does a relationship learned from one odor category also apply to the other?

First, when pleasant and unpleasant odors were analyzed together, activity patterns in all four regions yielded information related to pleasantness and salience ratings. Pleasantness could also be predicted within each category when pleasant and unpleasant odors were analyzed separately.

This alone, however, does not establish whether liking and disliking share a representational mechanism. The researchers therefore trained a model on the relationship between activity and ratings for pleasant odors and tested whether that relationship also applied to unpleasant odors. They also performed the reverse analysis, training on unpleasant odors and testing on pleasant ones.

In the piriform cortex and amygdala, predictions transferred from one category to the other did not perform significantly above chance. Although each region contained information about both pleasant and unpleasant odors, rules learned from one did not readily carry over to the other. The authors interpret this as suggesting that pleasantness for pleasant and unpleasant odors is represented by separate activity patterns.

This does not mean that the piriform cortex is devoted to liking and the amygdala to disliking. Both regions contained pleasantness-related information for both kinds of odor, but the form of that representation was not necessarily shared.

In the orbitofrontal cortex, strong pleasantness and strong unpleasantness connect

The orbitofrontal cortex showed a different kind of commonality. When predictions learned from one odor category were transferred to the other, they showed an inverse relationship with actual pleasantness ratings. In the authors’ framework, this is the expected result if strong pleasantness and strong unpleasantness are represented through their shared distance from neutrality.

When the region was divided more finely, this relationship was significant in orbitofrontal area 11. By contrast, this cross-prediction did not produce evidence of a single pleasantness scale running in the same direction from dislike to like.

The authors interpret this as integration of separate olfactory information into salience spanning pleasantness and unpleasantness in frontal regions. Rather than making pleasant and unpleasant odors feel the same, the brain may represent their shared emotional importance despite their opposite directions.

Can we now read preferences accurately from the brain?

These predictions used part of an individual’s recordings for training and other recordings for evaluation. The researchers did not create a perfume-preference diagnostic that works for everyone. When predicting overall pleasantness or salience, correlations between predicted and actual ratings were approximately 0.07–0.16 across regions. A significant association is not the same as accurately reading a single experience.

Even with thousands of trials, there were only three participants. The ability to characterize relationships within one person must be distinguished from generalizability to many people. Each fMRI spatial unit also includes contributions from many neurons and blood vessels. Failure to generalize predictions to the other odor category does not prove completely independent circuits at the cellular level.

The researchers examined influences including odor intensity, breathing, and the number of measured spatial units, but did not stimulate or suppress brain regions to test causality. This analysis alone cannot establish the actual pathways information travels, when it is integrated, or what behavior results.

Summary

This analysis suggests that odor liking and disliking are not placed on one common scale everywhere in the brain. Separate representations of pleasant and unpleasant odors were suggested in the piriform cortex and amygdala, while the orbitofrontal cortex, especially area 11, represented their shared distance from neutrality.

The authors suggest that integrating separate information offers the advantage of adapting meaning to context. While distinguishing odors we want to approach from those we want to avoid, the brain can also represent which matters to us now. This study offers concrete clues for the hypothesis that such flexible integration supports behavior.

References

Sagar, V., Zelano, C. M., & Kahnt, T. (2026). Separable and integrated pleasantness coding for appetitive and aversive odors across olfactory and ventral prefrontal cortices. Nature Communications, 17, 6732. DOI: 10.1038/s41467-026-73001-7 / Full original article.

Original article license: CC BY 4.0. Contributions by US NIH authors are treated as works of the United States Government.