Premium
Detection Thresholds of 10 Odor‐active Compounds Naturally Occurring in Food Using a Replicated Forced‐Choice Ascending Method of Limits
Author(s) -
Jaeger Sara R.,
Silva H. Nihal,
Lawless Harry T.
Publication year - 2014
Publication title -
journal of sensory studies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.61
H-Index - 53
eISSN - 1745-459X
pISSN - 0887-8250
DOI - 10.1111/joss.12085
Subject(s) - odor , detection threshold , two alternative forced choice , statistics , population , psychology , toxicology , food science , mathematics , chemistry , audiology , medicine , environmental health , computer science , biology , organic chemistry , real time computing
A series of threshold measurements for aroma detection of 10 compounds naturally occurring in foods were collected on 113 individuals in quadruplicate, using a 3‐alternative forced‐choice ascending method of limits. Mixed‐model analyses examined differences among compounds, subgroups of individuals tested as cohorts, order effects and several covariates such as age, gender, propylthiouracil sensitivity and body mass index. Differences among compounds were observed, as expected. Women had lower thresholds (higher sensitivity) than men for three of the 10 compounds, cis‐3‐hexenol, isovaleric acid and 4‐methyl octanoic acid. A consistent warm‐up effect was noted for the second compound tested on any given day. The overall panel was tested in cohorts of approximately 10 panelists each, and significant residual variation was noted for β‐ionone and isobutyraldehyde, for which bimodal threshold distributions were observed. The latter result suggests that testing small groups of panelists (e.g. n = 10), as suggested by the ASTM E ‐679 procedure, may not accurately capture the population threshold variability, nor provide an accurate estimate of the mean detection threshold. The research contributes to guidelines regarding good practice for odor detection testing. Practical Applications This research reports on odor detection thresholds for 10 compounds. The data were obtained using a 3‐alternative forced‐choice ascending method of limits which was executed in accordance with the specifications in ASTM E ‐679. Mean and standard deviation ratio values obtained from 113 individuals are reported for each odorant. For researchers studying foods, beverages and personal/household care products where these odorants are odor‐active, the data have practical significance. Further, the research has implications for others measuring odor thresholds. Specifically, we find that odor threshold testing is prone to warm‐up effects and we caution that threshold testing using small groups may be prone to misestimation of population thresholds.