Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 May 19;11(1):10595.
doi: 10.1038/s41598-021-89969-9.

Data-science based analysis of perceptual spaces of odors in olfactory loss

Affiliations

Data-science based analysis of perceptual spaces of odors in olfactory loss

Jörn Lötsch et al. Sci Rep. .

Abstract

Diminished sense of smell impairs the quality of life but olfactorily disabled people are hardly considered in measures of disability inclusion. We aimed to stratify perceptual characteristics and odors according to the extent to which they are perceived differently with reduced sense of smell, as a possible basis for creating olfactory experiences that are enjoyed in a similar way by subjects with normal or impaired olfactory function. In 146 subjects with normal or reduced olfactory function, perceptual characteristics (edibility, intensity, irritation, temperature, familiarity, hedonics, painfulness) were tested for four sets of 10 different odors each. Data were analyzed with (i) a projection based on principal component analysis and (ii) the training of a machine-learning algorithm in a 1000-fold cross-validated setting to distinguish between olfactory diagnosis based on odor property ratings. Both analytical approaches identified perceived intensity and familiarity with the odor as discriminating characteristics between olfactory diagnoses, while evoked pain sensation and perceived temperature were not discriminating, followed by edibility. Two disjoint sets of odors were identified, i.e., d = 4 "discriminating odors" with respect to olfactory diagnosis, including cis-3-hexenol, methyl salicylate, 1-butanol and cineole, and d = 7 "non-discriminating odors", including benzyl acetate, heptanal, 4-ethyl-octanoic acid, methional, isobutyric acid, 4-decanolide and p-cresol. Different weightings of the perceptual properties of odors with normal or reduced sense of smell indicate possibilities to create sensory experiences such as food, meals or scents that by emphasizing trigeminal perceptions can be enjoyed by both normosmic and hyposmic individuals.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
Heat plot of the means of the ratings of odors (rows) for seven different properties (columns). The 40 odors used in the study were split into four sets of 10 odors each (indicated at the left). The matrix is sorted column wise, per odor set, to locate the lowest ratings at the bottom left corner and the highest ratings at the upper right corner. Marginal statistics are shown as boxplots, displaying the minimum, quartiles, median (solid line within the box) and maximum. The whiskers add 1.5 times the inter-quartile range (IQR) to the 75th percentile or subtract 1.5 times the IQR from the 25th percentile. Outliers and extreme values are omitted from the boxplots; therefore, please note that the scale of the axis includes scores [0,…,4], while the data range is [0,…,5] as indicated in the methods description section. The figure has been created using the R software package (version 4.0.3 for Linux; https://CRAN.R-project.org/), and the R library “ComplexHeatmap” (https://bioconductor.org/packages/release/bioc/html/ComplexHeatmap.html).
Figure 2
Figure 2
Results of a principal component analysis (PCA). Projection of the 80 × 7 data matrix obtained by averaging the ratings of perceptual odor properties, separately for each of the seven properties, the 40 odors and the 2 olfactory diagnoses. (A) Plot of the data projected on the space given by the first two principal components (Dim.1 versus Dim.2). The PCA plot shows the separation of olfactory diagnoses mainly to the right in Dim.1 and to the top in Dim2, see the thick arrow indicating the averages of the PCA coordinates between olfactory diagnoses. The same odors rated by either normosmic or hyposmic subjects are connected with arrows (paired data). (B,C) The marginal distribution plots show the segregation of the pain phenotype groups along the principal components. (D) Plots the Eigenvectors of a variable in PCA Dim1 versus Dim2. (C) Scree-plot of the amount of variance of the data captured by each principal component. (E) Bar graph of the explained variance by each principal component. (F) Bar graph of the contribution of each perceptual property to Dim.1. The dashed horizontal reference dashed corresponds to the expected value if the contribution where uniform. (G) Bar graph of the contribution of each perceptual property to Dim.2. (H) Sorted Euclidean distances between the same odors evaluated by either normosmic or hyposmic subjects, i.e., the lengths of the arrows in panel A. The vertical dotted lines show the decision boundaries obtained by ABC analysis of the distances. The figure has been created using the R software package (version 4.0.3 for Linux; https://CRAN.R-project.org/ (R Development Core Team, 2008)) and the libraries “ggplot2” (https://cran.r-project.org/package=ggplot2 (Wickham, 2009)) and “FactoMineR" (https://cran.r-project.org/package=FactoMineR).
Figure 3
Figure 3
Perceptual ratings of odors that possess properties that are informative for the distinction between normosmic and hyposmic subjects, or that lack such properties. (A) Tree map of odor property assessments. The figure is a structured representation of the results of the item categorization analysis in a hierarchical order. The first level represents the grouping of the odor property ratings in terms of the information they provide for distinguishing normosmic from hyposmic subjects. The subsequent levels show perceptual properties and odors separated from each other. The second level represents the perceptual properties. The size of the rectangles corresponds to the number of odors that possess the respective property in connection with the separation of normosmic and hyposmic test persons. The third level represents the individual odors that possess the respective property that has been found to be informative or non-informative for the separation of olfactory diagnoses. (B) Histograms of the pooled ratings of seven properties for all odors. The dotted and dashed perpendicular lines mark the medians of the respective ratings, separately for normosmic (dashed) and hyposmic (dotted) subjects. (C) Venn diagram showing the sets observed between two groups of odors identified to represent characteristics that best distinguish between olfactory diagnoses or, on the other hand, are least distinctive between olfactory diagnoses (Table 2). The figure has been created using the R software package (version 4.0.3 for Linux; https://CRAN.R-project.org/) and the R libraries RAM” (https://cran.r-project.org/package=RAM), “ggplot2” (https://cran.r-project.org/package=ggplot2) and “treemapify” (https://cran.r-project.org/package=treemapify).
Figure 4
Figure 4
Results of the exploration of differences between distinctive and non-distinctive odors with respect to chemically Advanced Template Search (CATS) 2D molecular descriptors. The boxplots of the top ten CATS2D variables used in random forests and bagged classification and regression tree models. The boxes have been constructed using the minimum, quartiles, median (solid line within the box), and maximum. The whiskers add 1.5 times the inter-quartile range (IQR) to the 75th percentile or subtract 1.5 times the IQR from the 25th percentile. The figure has been created using the R software package (version 3.6.1 for Windows; https://CRAN.R-project.org/) and the R library “ggplot2” (https://cran.r-project.org/package=ggplot2). (L lipophilic, A acceptor, D donor, N negatively charged, P positively charged).

Similar articles

Cited by

References

    1. Hummel T, Nordin S. Olfactory disorders and their consequences for quality of life. Acta Otolaryngol. 2005;125:116–121. doi: 10.1080/00016480410022787. - DOI - PubMed
    1. Rinaldi A. The scent of life The exquisite complexity of the sense of smell in animals and humans. EMBO Rep. 2007;8:629–633. doi: 10.1038/sj.embor.7401029. - DOI - PMC - PubMed
    1. Croy I, Nordin S, Hummel T. Olfactory disorders and quality of life–an updated review. Chem. Senses. 2014;39:185–194. doi: 10.1093/chemse/bjt072. - DOI - PubMed
    1. Santos DV, Reiter ER, DiNardo LJ, Costanzo RM. Hazardous events associated with impaired olfactory function. Arch. Otolaryngol. Head Neck Surg. 2004;130:317–319. doi: 10.1001/archotol.130.3.317. - DOI - PubMed
    1. Smeets MAM, et al. Sense of smell disorder and health-related quality of life. Rehabil. Psychol. 2009;54:404–412. doi: 10.1037/a0017502. - DOI - PubMed

Publication types

LinkOut - more resources