The Influence of Circadian Timing on Olfactory Sensitivity

Abstract

Olfactory sensitivity has traditionally been viewed as a trait that varies according to individual differences but is not expected to change with one's momentary state. Recent research has begun to challenge this position and time of day has been shown to alter detection levels. Links between obesity and the timing of food intake further raise the issue of whether odor detection may vary as a function of circadian processes. To investigate this question, 37 (21 male) adolescents (M age = 13.7 years) took part in a 28-h forced desynchrony (FD) protocol with 17.5 h awake and 10.5 h of sleep, for 7 FD cycles. Odor threshold was measured using Sniffin' Sticks 6 times for each FD cycle (total threshold tests = 42). Circadian phase was determined by intrinsic period derived from dim light melatonin onsets. Odor threshold showed a significant effect of circadian phase, with lowest threshold occurring on average slightly after the onset of melatonin production, or about 1.5○ (approximately 21:08 h). Considerable individual variability was observed, however, peak olfactory acuity never occurred between 80.5○ and 197.5○ (~02:22-10:10 h). These data are the first to show that odor threshold is differentially and consistently influenced by circadian timing, and is not a stable trait. Potential biological relevance for connections between circadian phase and olfactory sensitivity are discussed.

Keywords: adolescents; food intake; forced desynchrony; individual differences; odor threshold; trait-state.

Figure 1.

Schematic of study protocol. Study day is on the y axis and clock time on the x axis. Black cells represent scheduled sleep during the at-home portion of the study (21:30–07:30). Dark gray cells represent scheduled sleep during in-lab protocol and light gray scheduled wake. Of note, sleep is scheduled for 4 h later each day, as part of the FD protocol.

Figure 2.

Predicted circadian function for olfactory sensitivity. Sample average circadian function with 95% confidence interval estimated using multilevel cosinor model. Twenty-four-hour clock time was pegged to the average initial dim light melatonin onset phase (20:59). Shaded bands represent biological night. Mesor, mid-value of a circadian cycle; Acrophase, phase of circadian peak; Amplitude, height of the circadian cycle (distance from Mesor to Acrophase). Higher amplitude indicates greater olfactory sensitivity.

Figure 3.

Fit of Cosinor model. The cosinor model predicted function was compared to data estimated using 60° circadian bins. Marginal Mean Estimates and 95% confidence interval for each circadian bin were estimated using a linear mixed effect model. Higher amplitude indicates greater olfactory sensitivity.

Figure 4.

Radial plot of predicted circadian functions for each participant. Radial plot depicts Amplitude (distance from center) and Acrophase (clockwise angle from 0°). Sample average and 95% confidence interval are equivalent to those depicted in Figure 2. Individual Predicted Values were estimated using the best linear unbiased predictor for each participant estimated from the multilevel cosinor model. Shaded area represents the biological night.