Presentation Details
| The sense of smell is not idiosyncratic Michal Andelman-Gur, Tali Weiss, Noam Sobel. Department of Brain Sciences, Weizmann Institute of Science, Rehovot, Israel |
Abstract
Human olfaction is considered idiosyncratic. Here we tested the alternative hypothesis that people are not uniquely different in their sense of smell, but instead show individual variability comparable to that observed in vision. Sixty participants rated similarity between pairs of odorants and pairs of images. Stimuli spanned high, medium, and low similarity: odorants based on angular distance in physicochemical space, and images from the THINGS dataset. In a subset (n=30), the same stimuli were used in a 1-back fMRI paradigm. We quantified within-subject and between-subject variability in similarity judgments. Within-subject variability (mean within-pair range) was nearly identical for odors and images in both the full cohort and fMRI subset (all p≥0.32, |d|≤ 0.16, BF₀₁≈3.7–4.8). Two one-sided equivalence tests (TOST) with a ±5-unit margin confirmed equivalence in both cohorts (all p<0.001). For between-subject variability, a globally matched odor–image set was analyzed. Across-participant standard deviation did not differ significantly between modalities (paired tests p≥0.18, d≤0.57, BF₀₁≈0.87-2.1), and TOST supported equivalence within ±5 SD units in both cohorts (all p ≤ 0.033). The fMRI data dovetailed with the behavioral data: inter-subject variance of neural representational dissimilarity matrices (RDMs) within modality-specific cortical regions was similar across systems (mean variance: olfaction=0.04, vision=0.065, P=NS), and a universal similarity representation regardless of modality was uncovered in the angular gyrus. Taken together, these results imply that the concept of perceptual similarity stably spans sensory systems, and humans are not more different from each other in their olfactory representations than in their visual representations.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
Content Locked. Log into a registered attendee account to access this presentation.