Human responses to propionic acid. I. Quantification of within- and between-participant variation in perception by normosmics and anosmics.

The objective of this study was to fully characterize normosmic perception of stimuli expected to cause widely varying degrees of olfactory and nasal trigeminal stimulation and to directly evaluate the possible role of olfactory nerve stimulation in nasal irritation sensitivity. During each of four identical test sessions, four anosmic and 31 normosmic participants were presented with a range of concentrations extending from peri-threshold for normosmics to supra-threshold for anosmics. For each session, odor (O) and nasal irritation (NI) sensitivities were summarized in terms of the concentrations required to produce four sensation levels ('iso-response' concentrations). Within-participant variation in these iso-response concentrations was < 10-fold for 95% of normosmics, for both O and NI. For O but not NI, these apparent fluctuations in sensitivity were largely accounted for by the uncertainty surrounding the iso-response concentrations calculated for each session. Anosmics exhibited minimal within- and between-participant variation in NI and required, for all but the highest perceptual level, a higher concentration than almost all normosmics. Between-participant variation, expressed in terms of 90% confidence interval widths, was approximately 0.5 log units for both O and NI for the highest perceptual level, but increased to approximately 0.8 and 1.8 log units, respectively, for the lowest (peri-threshold) level. Our findings suggest that: (i) most apparent variation over time in O sensitivity is actually a reflection of the uncertainty surrounding estimates of sensitivity obtained for each session; (ii) within- and between-participant variation in O sensitivity is far less than is commonly reported; and (iii) low to moderate levels of NI in normosmics are the result of relatively weak trigeminal stimulation combined with much greater olfactory activation.