Decoding odor responses: universal patterns and individual signatures in psychophysiology using nonlinear models

Olfactory perception is a complex process driven by the chemical properties of odorants and shaped by a multitude of individual factors. As a result, predicting how an individual perceives a given odor remains challenging. We aimed to address this complexity by integrating individual response patterns, odorant properties, and psychophysiological responses into a unified model. Therefore, we tested perceptual dimensions (valence, temperature, and intensity) of 6 perceptually diverse monomolecular odorants with continuous time-series data from psychophysiological measures (respiration, heart rate, electromyography [EMG] corrugator, and EMG zygomaticus) in a sample of 41 participants. By simultaneously accounting for the odorant itself, individual rating tendencies, and both group-level and individual-specific physiological effect patterns in the nonlinear modeling process, we found that while the specific odorant and individual rating tendencies were the primary drivers of perception, the relative contributions varied significantly across perceptual dimensions. The inclusion of physiological signals significantly improved the predictive models, revealing that both generalizable (group-level) and highly individualized psychophysiological response patterns contributed to how an odor was perceived. Examination of the specific effect patterns revealed respiration and EMG corrugator as key group-level predictors for valence and intensity, while significant individual-specific effect patterns varied considerably across the perceptual dimensions. Our findings demonstrate that a comprehensive understanding of olfactory perception requires the consideration of the interplay between stimulus characteristics, idiosyncratic biases, and distinct universal versus person-specific physiological signatures, offering a more nuanced understanding of this sensory experience.