Long-lasting effects of lavender exposure on brain resting-state networks in healthy women

Abstract

Introduction: Most brain imaging studies on olfaction focus on short-term odorant stimuli, with few examining long-lasting odor exposure or its after-effects. In this study, we utilized resting-state fMRI (rsfMRI) to investigate the effects of prolonged odor exposure to lavender on brain activity and whether these persist post-exposure.

Methods: Fourteen healthy women underwent two fMRI sessions, conducted one week apart, in a randomized order. Both sessions included rsfMRI scans before, during, and up to 2 h after a 14 min exposure to either lavender essential oil or a non-odorant control.

Results: An Independent Component Analysis identified the salience network (SAL) and default mode network (DMN) as the most consistent resting-state networks. A two-factorial ANOVA revealed significant time-varying interaction effects between the SAL and DMN. During odor exposure, functional connectivity (FC) increased within the SAL, and a negative correlation between the SAL and DMN appeared, which intensified immediately after exposure. Two hours post-exposure, the FC between SAL and DMN turned positive.

Discussion: These findings suggest that prolonged odorant exposure to lavender can induce long-lasting brain effects detectable up to 2 h afterwards in women. This proof-of-concept study should be extended to other odorants and to men, and offers new possibilities for exploring the effects of aromatherapy or other odor exposure interventions on brain activity.

Keywords: Independent Component Analysis (ICA); default mode network (DMN); olfactory connectome; resting state functional MRI (rsfMRI); salience network (SAL).

FIGURE 1

Experimental study design. Each participant went twice through the procedure, once while being exposed to lavender, and once during exposure to a non-odorant substance. The order of the two sessions was counter-balanced across participants. The image to the upper left corner shows the odor administration procedure with the cotton patch attached to the head coil of the magnetic resonance imaging (MRI).

FIGURE 2

Results of the seed-based correlation (SBC) approach using the posterior cingulate cortex (PCC) and right anterior insula as seed regions. Results are shown for the three time points: before, during and after odor exposure. (A) SBC maps based on PCC (x = 0, y = -56, z = 26) as seed region. (B) SBC maps based on the right anterior insula (x = 40, y = 14, z = -2) as seed region. All maps are thresholded using the FDR-correction method. The two seed regions are indicated with white arrows in the upper part of the figure.

FIGURE 3

Probabilistic map of the salience network (SAL), thresholded at 60% overlap between subjects, for the two conditions (odor and control) and the five time points. The map confirmed that the proper regions of interest were consistently selected for the conditions and time points. The color coding shows areas of overlap between the different maps.

FIGURE 4

Two-factorial analysis of variance (ANOVA) of the salience network (SAL) maps with the factors odor and time point. (A) Main effect of the factor odor, collapsed over all time points. The significant regions coincide with large parts of the default mode network (DMN), suggesting a potential interaction between the SAL and DMN networks. Significant areas included clusters in the posterior cingulate cortex, perigenual anterior cingulate, inferior prefrontal cortex (BA 47) and superior prefrontal cortex (BA 6, BA 8). (B) Main effect of the factor time, pooled over the two odor conditions. Two clusters were observed, a large one in the posterior cingulate cortex, just anterior to the cluster observed for the factor odor, and a second one in the perigenual anterior cingulate cortex. Results were thresholded using the cluster threshold correction method. All maps shown at P < 0.005.

FIGURE 5

Analysis of variance (ANOVA) results of the functional connectivity maps of the salience network (SAL), contrasting lavender with no odorant control (Lavender > Control) at four time points. Data are presented on axial slices. The numbers at the bottom refer to the z-coordinates in MNI space. Images are presented in radiological convention (left part of the brain is shown to the right). Cluster correction was applied to correct the maps. All clusters shown are at q(FDR) < 0.001.