Psychological states affecting initial pupil size changes after olfactory stimulation in healthy participants

Abstract

Odor perception affects physiological and psychological states. Pupillary light reflex (PLR) parameters can be affected by olfactory stimulation and psychological states, although it remains unclear whether the olfactory stimulation-induced psychological changes can associate with PLR parameter changes. This study aimed to investigate effects of olfactory stimulation-induced psychological changes on PLR parameter changes with repeated measurements. We collected data on six mood subscales of the profile of mood states, and on five PLR parameter measurements from 28 healthy participants. Participants underwent a 10-min olfactory stimulation on different days with six odorants available with the T&T olfactometer. As obtained data were clustered, we used linear mixed-effects models for statistical analyses. The olfactory stimulation using the no-odor liquid did not affect mood states and the initial pupil size (INIT). The sweat odorant worsened all mood subscales including fatigue-inertia (Fatigue)/Vigor-Activity (Vigor), and decreased INIT compared to the no-odor liquid. When comparing INIT responses related to changes in mood subscales between the no-odor liquid and the sweat odorant, worsened states of Fatigue/Vigor were associated with decreased INIT in the sweat odorant. Fatigue/Vigor can be used as mental fatigue indicators. Thus, mental fatigue can be associated with decreased INIT in the olfactory stimulation.

Figure 1

PLR parameters adopted, olfactory stimulation using a facemask, and study design. (A) Five PLR parameters were used for statistical analyses: ACV, DELTA, INIT, LAT, and MCV. (B) A 2 × 2 cm sterile gauze was soaked in an odorant in a petri dish. The soaked sterile gauze (dotted square) was sandwiched between folds of a commercially available facemask. The participant was asked to put on the mask so that the side sandwiching the gauze was on the face side. In a dim room, PLR measurement was performed using a commercially available pupillometer. To prevent stimulation by ambient light to the eyes, the sides of both eyes were covered with the participant’s hand or the rubber cup of the pupillometer. (C) The time schedule and the collected data are shown. For each time point of PLR measurement, eight data points (2 eye sides × 4 light stimulus intensities) were collected for each participant. The 1st and 2nd PLR data were defined as pre-1 PLR and pre-2 PLR, respectively. The 3rd PLR data were defined as post-PLR. ACV averaged constriction velocity, DELTA constriction ratio, INIT initial pupil size, LAT constriction latency, MCV maximum constriction velocity, PLR pupillary light reflex, POMS profile of mood states.

Figure 2

Time-course changes in six mood subscale scores with olfactory stimulation using the no-odor liquid. (A) Anger, (B) Confusion, (C) Depression, (D) Fatigue, (E) Tension, and (F) Vigor. All values represent means and 95% confidence intervals (n = 56 data points: 28 participants × 2 time-points). A Wilcoxon signed-rank test is used for statistical analyses. Anger Anger–Hostility, Confusion confusion–bewilderment, Depression depression–dejection, Fatigue fatigue–inertia, POMS profile of mood state, Tension tension–anxiety, Vigor Vigor–Activity.

Figure 3

Forest plot of regression coefficients showing time course changes of PLR parameters in the olfactory stimulation using the no-odor liquid. The estimated regression coefficients are obtained by linear mixed-effects models. Black circles indicate significant regression coefficients. Pre-1 PLR data are set as references. *regression coefficient × 10⁻²; **regression coefficient × 10⁻¹. CI confidence interval, ACV averaged constriction velocity, LAT constriction latency, DELTA constriction ratio, INIT initial pupil size, MCV maximum constriction velocity, PLR pupillary light reflex, Pre 1, Pre 2, and Post pre-olfactory stimulation at 1st, 2nd, and post-olfactory stimulation, respectively.

Figure 4

Forest plot of regression coefficients showing changes of mood subscales at the post-olfactory stimulation using six odorants. For the six odorants, changes in mood subscales calculated by subtracting values at pre-olfactory stimulation from those at post-olfactory stimulation are shown by Δ. Estimated regression coefficients and p values calculated by linear mixed-effects models are shown. The no-odor liquid is set as a reference. Black circles indicate significant regression coefficients. Anger Anger–Hostility, CI confidence intervals, Confusion confusion–bewilderment, Depression depression–dejection, Fatigue fatigue–inertia, Tension tension–anxiety, Vigor Vigor–Activity.

Figure 5

Forest plot of regression coefficients showing changes of INIT at the post-olfactory stimulation among six odorants. For each of six odorants, INIT changes are calculated by subtracting values at pre-olfactory stimulation from those at post-olfactory stimulation, shown by Δ. Regression coefficients and p values calculated by linear mixed-effects models are shown. The no-odor liquid is set as a reference. CI confidence interval, INIT initial pupil size.

Figure 6

Forest plot of regression coefficients showing initial pupil size responses due to mood subscale changes at the post-olfactory stimulation using the no-odor liquid and the sweat odorant. The change values calculated by subtracting values at pre-olfactory stimulation from those at post-olfactory stimulation are shown by Δ. Regression coefficients and p values calculated by linear mixed-effects models are shown. The no-odor liquid is set as a reference. Anger Anger–Hostility, CI confidence interval, Confusion confusion–bewilderment, Depression depression–dejection, Fatigue fatigue–inertia, Tension tension–anxiety, Vigor Vigor–Activity.

Figure 7

Comparison of initial pupil size responses due to three mood subscales’ changes between the no-odor liquid and the sweat odorant. The change values calculated by subtracting values at pre-olfactory stimulation from those at post-olfactory stimulation are shown by Δ. The graphs show the slopes reflecting the mean ΔINIT to Δmood subscales with 95% confidence intervals: (A) ΔAnger, (B) ΔFatigue, and (C) ΔVigor. Regression coefficient values of interaction effects calculated using linear mixed-effects models are shown in each figure. The interaction effect represents the differences of slopes between the no-odor liquid and the sweat odorant. The no-odor liquid is set as a reference. Anger Anger–Hostility, Fatigue fatigue–inertia, INIT initial pupil size, Vigor Vigor–Activity.