Visually induced nausea causes characteristic changes in cerebral, autonomic and endocrine function in humans

Abstract

An integrated understanding of the physiological mechanisms involved in the genesis of nausea remains lacking. We aimed to describe the psychophysiological changes accompanying visually induced motion sickness, using a motion video, hypothesizing that differences would be evident between subjects who developed nausea in comparison to those who did not. A motion, or a control, stimulus was presented to 98 healthy subjects in a randomized crossover design. Validated questionnaires and a visual analogue scale (VAS) were used for the assessment of anxiety and nausea. Autonomic and electrogastrographic activity were measured at baseline and continuously thereafter. Plasma vasopressin and ghrelin were measured in response to the motion video. Subjects were stratified into quartiles based on VAS nausea scores, with the upper and lower quartiles considered to be nausea sensitive and resistant, respectively. Twenty-eight subjects were exposed to the motion video during functional neuroimaging. During the motion video, nausea-sensitive subjects had lower normogastria/tachygastria ratio and cardiac vagal tone but higher cardiac sympathetic index in comparison to the control video. Furthermore, nausea-sensitive subjects had decreased plasma ghrelin and demonstrated increased activity of the left anterior cingulate cortex. Nausea VAS scores correlated positively with plasma vasopressin and left inferior frontal and middle occipital gyri activity and correlated negatively with plasma ghrelin and brain activity in the right cerebellar tonsil, declive, culmen, lingual gyrus and cuneus. This study demonstrates that the subjective sensation of nausea is associated with objective changes in autonomic, endocrine and brain networks, and thus identifies potential objective biomarkers and targets for therapeutic interventions.

Key points

• Nausea is a highly individual and variable experience. The reasons for this variability are incompletely understood although psychophysiological factors have been proposed.

• Herein we describe objective psychophysiological changes induced by the subjective sensation of motion sickness.

• In comparison to subjects who did not develop nausea, nausea-sensitive subjects demonstrated electrogastrographic and autonomic changes, which included an increase in sympathetic nervous system activity with a concomitant reduction in parasympathetic activity. Furthermore, differences were also evident in plasma ghrelin, and subcortical and cortical activity.

• These data have a number of important implications for future research examining the physiological mechanisms that underlie nausea:

  • The physiological, hormonal and cortical patterns identified herein represent potential biomarkers of the physiological mechanisms of nausea.

  • Reverse translation of the physiological factors identified may facilitate refinement of animal models used to investigate novel anti-emetic agents and emetic liability of candidate drugs, increasing their validity and translation of finding to humans.

Introduction

Nausea is a common and distressing experience that often precedes emesis (Andrews & Sanger, 2014). It is a highly individual experience, which is determined through sex, racial, psychological, physiological and neuroanatomical factors (Klosterhalfen et al. 2006; Stern et al. 2011). The relative contribution of each of these constituent factors remains incompletely understood. Previous studies have examined most of these factors in isolation, which has constrained our wider understanding of their co-relationships. Nevertheless, several lines of converging evidence support the view that the sensory experience of nausea and the associated physiological changes therein (e.g. pallor, sweating, gastric dysrhythmia), involve bidirectional interactions between the CNS, autonomic nervous system (ANS) and endocrine system (Himi et al. 2004; Muth, 2006; Napadow et al. 2013b; Andrews & Sanger, 2014; Farmer et al. 2014a).

While nausea-related changes in gastric myoelectrical activity and the ANS, the latter derived from cardiac cycle-derived metrics, have been relatively well documented, gaps exist in our understanding of endocrine and CNS responses to nausea. For instance, plasma arginine vasopressin is elevated by several stimuli inducing nausea (e.g. motion, chemotherapy, apomorphine), yet its physiological role is uncertain (Stern et al. 2011). Ghrelin, an orexogenic gastric hormone, stimulates appetite and food intake in humans (Wren et al. 2001). Exogenous ghrelin has an anti-emetic effect and enhances gastric emptying in animals when administered concomitantly with the cytotoxic chemotherapeutic agent cisplatin (Liu et al. 2006; Rudd et al. 2006). As anorexia is a prominent feature of nausea (Stern et al. 2011) it can be reasoned that nausea may be accompanied by a reduction in plasma ghrelin secretion. Moreover, the contemporaneous understanding of the CNS changes that accompany nausea is rudimentary. This is because in contrast to the plethora of functional neuroimaging studies that have focused on the identification of the cortical regions responsible for other gastrointestinal sensations, such as pain, there is a relative paucity of literature devoted to describing the areas involved in the genesis, maintenance and subsequent resolution of nausea (Napadow et al. 2013a,b). This paucity is probably due to significant methodological challenges in studying nausea in the confines of the brain imaging environment.

Whilst a number of animal models have provided important mechanistic insights into the physiology of vomiting, their absolute utility in studying nausea is uncertain (Stern et al. 2011; Andrews & Sanger, 2014). Although the existence of nausea in humans is not disputed, it represents a significant challenge to study effectively (Klosterhalfen et al. 2006; Holmes et al. 2009). Experimentally, nausea is frequently induced in humans using either pharmacological or physiological stimuli. Physiologically, the sensation of nausea can be induced as a component of motion sickness induced either by actual motion (Coriolis cross-coupled stimulus; Miller & Graybiel, 1970) or through illusory self-motion, known as vection (Hettinger et al. 2014). The latter involves placing a subject inside a spacious optokinetic drum, whose interior aspect is covered with an alternating pattern, e.g. alternating vertical black and white stripes (Stern et al. 2011). When the drum is rotated, it induces an ocularly mediated sensation of movement, giving rise to a sensory conflict with the vestibular system and additionally a conflict between the ocular and proprioceptive systems, thereby leading to nausea in susceptible subjects (Stern et al. 2011). Motion sickness, induced by vection, is a form of visually induced motion sickness (VIMS). In a manner similar to motion sickness induced by actual motion, it provokes nausea, dizziness, regional sweating and pallor; with intense stimulation emesis may occur (Muth, 2006; Shupak & Gordon, 2006; Kennedy et al. 2010; Keshavarz & Hecht, 2011). Neither the actual motion nor optokinetic drum (virtual motion) methods for induction of motion sickness described above are feasible for use in the functional neuroimaging environment due to their cumbersome nature and lack of technical compatibility including the movement of the subject for the Coriolis cross-coupled stimulus. However, the induction of motion sickness syndrome by visually perceived motion in the absence of actual movement by the subject (Kennedy et al. 2010) provides a stimulus that can be utilized in combination with functional neuroimaging.

The gaps in our understanding of the physiological mechanisms of nausea in humans cannot be overcome until a robust and reproducible method can be employed to reliably induce nausea in a convenient and easily administrable manner. Thus, the aims of the present study were threefold: (1) to utilize a readily administered and reliable stimulus to evoke VIMS, thereby enabling the study of nausea using functional brain imaging; (2) to characterize the changes in the ANS, plasma vasopressin and ghrelin concentration and gastric myoelectrical activity induced by the visually perceived motion stimulus in subjects stratified according to their susceptibility to develop nausea; and (3) using these stratifications, we aimed to characterize the changes in brain activity, using functional magnetic resonance imaging (fMRI), during the personal experience of nausea. Therefore, we hypothesized that objective physiological, subcortical and cortical differences would be evident in subjects who developed nausea while viewing a video intended to evoke VIMS in comparison to those that did not.

Methods

Ethical approval

All protocols were reviewed and approved by Queen Mary University of London (ref: QMREC2008/37) and King's College London Research Ethics Committee (ref: PNM/09/09-04). Informed written consent was obtained from all subjects and the studies performed herein conformed to the standards set by the Declaration of Helsinki 2008.

Subjects

In total, 98 healthy adult subjects (53 males, mean age 26 years, range 19–58 years) took part in the study and were recruited by local advertisement and from the staff, students and local residents of Queen Mary, University of London and King's College London. Subjects were non-smokers with no significant medical history and were not taking any medications. All subjects were right handed as assessed by a score >40 on the Edinburgh handedness inventory (Oldfield, 1971). Females were all studied in the follicular phase of their menstrual cycle. Subjects were asked to refrain from alcohol consumption for 24 h prior to the study. All subjects were screened for sub-clinical anxiety and depression using the validated Hospital Anxiety and Depression Scale (Zigmond & Snaith, 1983). All subjects were naïve to the experimental protocol.

Behavioural assessment of nausea

During the motion and control videos (vide infra), nausea and anxiety were each assessed using four-point visual analogue scale (VAS) where 1 represented no symptoms while a score of 4 represented severe nausea, and 1 represented no anxiety while a score of 4 represented severe anxiety, respectively. The validated motion sickness sensitivity score (MSSQ) (Golding, 1998) and motion sickness assessment questionnaire (MSAQ) (Gianaros et al. 2001a) were used to assess susceptibility to and symptoms associated with motion sickness, respectively. The MSSQ is a questionnaire asking subjects to rate their previous experiences of nausea as a child, and over the last 10 years, nine different situations, such as in a car, on buses, on a train, etc., on a five-point Likert scale ranging from not applicable to frequently felt sick. The MSAQ evaluates the experience of motion sickness across gastrointestinal, central, peripheral and sopite related dimensions by asking subjects to rate their sickness using a nine-point scale across 16 items. Based on mean nausea VAS rating during the whole period when the motion video was shown, subjects were placed into quartiles where the 1st quartile, i.e. not reporting any nausea, were considered nausea-resistant and where subjects in the 4th quartile, reporting at least moderate to severe nausea, were considered to be nausea-susceptible.

Personality and anxiety measures

The validated Big Five Inventory was used to measure the personality traits of neuroticism and extroversion (McCrae & Costa, 1987). State and trait anxiety was assessed using the validated Spielberger State-Trait Anxiety Inventory (Spielberger, 1983). These instruments were chosen based on our group's experience of using them in previous studies examining the role of personality factors and anxiety in visceral sensation (Paine et al. 2009; Farmer et al. 2013a) and also because neuroticism has previously been shown to be a modifying factor in nausea (Netter-Munkelt et al. 1972).

Motion video and control video

The visual stimulus was a 10 min video of a landscape (London Eye, Houses of Parliament) as seen from a point 2 m above the centre of Westminster Bridge, London, UK. The point of view rotated, panning the scene through 360 deg at a rate of 0.2 Hz about an axis tilted 18 deg from earth vertical. The tilted and rotating view visual display makes the viewer perceive that they are spinning round and round on the spot about a tilted axis due to vection. Viewing a moving tilted scene has been shown to enhance the onset of vection-induced nausea (Bubka & Bonato, 2003; Bos & Bles, 2004). The video was composed of a sequence of digital camera images taken from the viewpoint of a tall subject standing on the bridge. A similar stimulus has been previously used to induce motion sickness (Golding et al. 2012). The images were processed on a personal computer (Dell, Bracknell, UK) using programs by 3DSTATE (New York, USA) to provide a video sequence with a frame resolution of 1024 by 768 pixels at 16-bit colour. The motion video was then projected with an Acer H5360Eco (Acer, West Drayton, UK) projector at a refresh rate of 60 Hz onto a screen of 2.00 × 2.00 m placed 1.12 m from the subject. A 10 min video of a static scene of the above video was shown as the control condition. All subjects wore goggles during the stimuli to restrict peripheral vision.

ANS measures

Mixed measures – blood pressure and heart rate

Digital arterial blood pressure was measured non-invasively using the validated photoplethysmographic technique (Portapres, Amsterdam, Netherlands) (Benarroch et al. 1991; Eckert & Horstkotte, 2002). Electrocardiogram electrodes (Ambu Blue Sensor P, Ballerup, Denmark) were placed in right and left sub-clavicular areas and cardiac apex. The electrocardiogram was acquired at 5 kHz using a commercially available biosignals acquisition system (Neuroscope, Medifit Instruments, Enfield, UK) with heart rate (HR) derived from intervals between successive R waves. ANS parameters were recorded according to internationally agreed recommendations (European Society of Cardiology, 1996).

Parasympathetic nervous system measures – cardiac vagal tone

The Neuroscope derives a real time index of brainstem parasympathetic nervous system efferent activity, known as cardiac vagal tone (CVT), measured on a validated linear vagal scale (LVS), where 0 represents full atropinization (Julu, 1992). CVT is described in detail elsewhere (Farmer et al. 2014b), but in contrast to power spectral analysis of HR variability, it is validated for time epochs of less than 1 min (Julu, 1992).

Sympathetic nervous system measures – cardiac sympathetic index

R–R interval data were extracted from the Neuroscope recordings and were hand edited to remove any missed, or extra, beats, as per accepted recommendations as these can result in large artifacts (1996). In 96 out of 98 subjects, no editing of R–R intervals was required. In 2 out of 98 subjects, the data needed to be edited due to movement artifacts, which resulted in a short period (<5 s) of R–R intervals of <250 ms. Following this, the R–R data were reformatted and entered into the Cardiac Metric program (CMet, University of Arizona, AZ, USA) for calculation of the validated Toichi's cardiac sympathetic index (CSI) (Toichi et al. 1997). CSI is a ratio of R–R intervals and therefore has no units. Autonomic parameters were recorded according to internationally agreed guidelines (1996).

Electrogastrogram

The Medtronic Polygram NET electrogastrogram (EGG) system (Medtronic A/S, Copenhagen, Denmark) was used for multi-channel recordings, with four electrogastrographic signals being recorded simultaneously. The electrogastrographic system was configured to accept an electrode impedance of less than 11 kΩ after skin preparation. Signals were sampled at ∼105 Hz and then down-sampled to 1 Hz as part of the acquisition process with a low-pass filter of 15 cycles per minute (cpm) and high-pass filter of 1.8 cpm. Six electrodes (Ambu Blue Sensor P, Ballerup, Denmark) were placed on the subject's abdomen after skin preparation with an abrasive electrode paste (Nuprep, Weaver & Co, Aurora, CO, USA). The six electrodes consisted of four active recording electrodes, one reference electrode, and one ground electrode as per the method described by Parkman et al. (1996). EGG traces were manually screened for artifacts. Two epochs of 5 min in length (baseline, motion/control video) were analysed using a fast Fourier transformation procedure with a spectral resolution of 0.25 cpm. A frequency range of 2.5–3.75 cpm was regarded as normogastria and a range of 4.0–9.75 cpm as tachygastria. The percentage spectral power was then calculated from the total range of 0.75–15.0 cpm before being computed to ascertain the ratio between the percentage of the normogastria and the tachygastria band, as a surrogate indicator for nausea. A decrease in the ratio during exposure to the motion video would indicate a relatively greater proportion of tachygastria, which has previously been shown to be associated with circular vection-induced nausea (Stern et al. 1987).

Vasopressin and ghrelin assays

For assay of acylated ghrelin and vasopressin arginine-vasopressin, plasma was prepared in EDTA tubes (Vacutainer, Becton Dickinson, Helsingborg, Sweden) on ice. A protease inhibitor cocktail was added (SigmaFast S8830, Sigma Aldrich, St Louis, MO, USA). Samples were centrifuged at 2500 relative centrifugal force for 10 min at 4°C and stored at –80°C until assayed. All samples were run in duplicate. Ghrelin was assayed using a custom kit (MSD; Mesoscale Discovery, Gaithersburg, MD, USA). The lower limit of quantification was ∼9.7 ng l⁻¹. The intra- and inter-assay coefficients of variation were 8.5 and 10.4%, respectively. Vasopressin was quantified by radioimmunoassay (RIA) in a single run according to product insert (Vasopressin Direct, Cat no. RK-VPD, Bühlmann Laboratories, Schönenbuch/Basel, Switzerland). This RIA displays no cross-reactivity with lysine vasopressin, desmopressin, oxytocin or vasotocin. The lower limit of quantification was ∼1.3 ng l⁻¹. Sample intra-assay coefficient of variation was 12.3%.

fMRI

fMRI data (T2*-weighted images) were collected on a General Electric Signa Excite II 1.5 T HD scanner based at the Centre for Neuroimaging Sciences, Institute of Psychiatry, King's College London. Head movement was restricted using foam padding within the head coil and an eye movement tracker was mounted onto the head coil. Prior to the start of the fMRI experiment, a high-resolution gradient echo structural scan (43 × 3 mm slices, 0.3 interslice gap, TE 40 ms, TR 3000 ms, flip angle 90 deg, matrix 128², in plane voxel dimensions 1.875 × 1.875) was acquired in each volunteer to be used for Talairach normalization. During fMRI, 300 T2*-weighted images per slice (40 × 3 mm slices, 0.3 interslice gap, TE 25 ms, TR 3500 ms, flip angle 90 deg, matrix 64²), depicting blood oxygen level dependent (BOLD) contrast, were collected as subjects viewed the control and motion video.

Study design

The experimental design consisted of two separate studies, which are summarized in Fig.1.

Figure 1.

Flowchart of the overall experimental design in which psychophysiological responses to the motion and control video were measured

Nausea ratings derived from the motion video in Study 1 were used to stratify subjects into quartiles with the upper quartile being considered as nausea-susceptible and the lower quartile being nausea-resistant. Central responses to the motion video from subjects from these two groups were then measured using fMRI. Of the 20 ‘dropouts’, 12 subjects did not want to participate in the fMRI study, 6 subjects could not attend at a convenient time and 2 subjects did not attend their scanning appointment.

Study 1 – validation of motion video as a nauseogenic stimulus and psychophysiological differences between nausea-sensitive and nausea-resistant subjects

Study 1 consisted of two visits. All subjects were studied in the afternoon (from 12.00 to 14.00 h) in a temperature-controlled (20–22°C) quiet laboratory. Subjects completed the MSSQ and personality questionnaires and were reclined at 45 deg on a couch with their legs supported. Subjects were then allowed to relax for 30 min during which baseline ANS and EGG (resting/no stimulation) was acquired. After the baseline period subjects completed the state anxiety and MSAQ questionnaires. Subjects were then randomized (randomization schedule generated at www.randomization.com) to watch either the motion video or control video and asked to report VAS ratings of nausea and anxiety each minute during the video. Ten millilitres of venous blood was sampled from the left antecubital fossa at baseline and at the end of the motion video for serum vasopressin and ghrelin. Subjects who were exposed to the motion video at the first visit were then crossed over to watch the control video at the second visit and vice versa. The protocol was repeated at least 4 weeks after the first visit to ameliorate any potential carryover effect. Subjects were not specifically told whether they were/had watched the motion or control video and the investigator analysing the psychophysiological data generated from Study 1 was blinded to the intervention.

Study 2 – differences in brain processing between nausea-sensitive and nausea-resistant subjects

Subjects were recruited from the participants in Study 1 (vide supra) and subsequently their availability and willingness to participate. The mean time between completion of Study 1 and Study 2 was 5.6 weeks (range 1 day to 12.2 weeks). After an identical preparation as for Study 1, the subjects were asked to wear a pair of MRI-compatible goggles with a refresh rate of 120 Hz and a field of view equivalent to viewing a 42-inch TV screen from 1.83 m away, with a resultant horizontal angle of 28.5 deg and a vertical angle of 16.4 deg (CinemaVision, Lexington, NC, USA) and placed in the supine position in the MRI scanner. The goggles were used to project the motion video to the subject during scanning, allowing an unimpeded view of the stimulus. During the active scanning phase, the 10-min motion video was then projected onto the goggles, during which subjects were cued, on a minute-to-minute basis, to rate their nausea and anxiety on the aforementioned VAS using identical anchor points via a button box. During the study, subjects were instructed to remain still and to focus on the stimulus as much as possible whilst maintaining a constant respiratory rate to reduce inter-subject variability, as voluntarily controlled breathing can reduce the sensation of nausea (Yen Pik Sang et al. 2003).

Statistics

Psychophysiological data

Data distribution was analysed using the D'Agostino–Pearson omnibus K2 normality test. Quantitative data are presented either as median with interquartile ranges (IQRs) for non-normally distributed data, or mean ± standard deviation (SD) for parametric data. For quantitative data, differences between the groups were assessed using the Student's t-test for parametric data and using the Wilcoxon matched-pairs test for non-parametric data. The trapezoid area under the curve (AUC) was calculated for nausea-sensitive subjects using a repeated-measures ANOVA as per the method used by Willert et al. (2004). Agreement was assessed using a two-way, random effects, single measure intra-class correlational coefficient model and was interpreted according to Yen & Lo (2002). All tests were two-tailed, and P<0.05 was adopted as the statistical criterion. Analyses were performed using proprietary software (SPSS version 21, IBM, New York, NY, USA and GraphPad Prism 5, San Diego, CA, USA).

fMRI

fMRI data were analysed using the XBAM software (http://brainmap.co.uk), developed at the King's College London Institute of Psychiatry, which implements permutation-based methods to minimize the number of assumptions used in making statistical inference. fMRI data pre-processing, smoothing and individual brain activation mapping were performed according to the method described by Coen et al. (2009). Briefly, prior to time-series analysis, data were corrected for the effects of head motion in three dimensions and were smoothed and de-trended. The data were then analysed using a general linear model in which the task design was convolved with a mixture of two one-parameter gamma variate functions (peak responses at 4 and 8 s) to account for haemodynamic delay and dispersion. The time series at each voxel was regressed on the convolved design. The parameters obtained from the regression were used to calculate a ‘goodness of fit’ statistic (the ratio of the sum of squares of the model fit and the residual sum of squares (SS ratio)) at every voxel. The significance of the statistic was then assessed by a data-driven permutation approach. Group brain activation maps were generated by mapping the observed and randomized test statistics data for each individual into the standard space of Talairach and Tournoux. Differences in activation between groups were then tested for significance using an ANOVA at each voxel in standard space. ANOVA was undertaken to compare responses between groups (in this case nausea-sensitive versus nausea-resistant), by fitting an analysis of variance model to the standardized BOLD effect size at each intra-cerebral voxel as follows: Y = a + bX + e where Y is the vector of BOLD effect sizes for each individual, X is the contrast matrix for the particular inter-condition/group contrasts required, a is the mean effect across all individuals in the various conditions/groups, b is the computed group/condition difference and e is a vector of residual errors. The null hypothesis was tested by comparing coefficient b to critical values of its non-parametrically obtained null distribution (Bullmore et al. 1999). Critical values for a two-tailed test of size α (alpha can be set to any desired type I error rate for the test) are the 100*(α/2)th and 100*(1–α/2)th percentile values of this distribution. The detection of activated voxels was extended from voxel to cluster level using the method described in detail by Bullmore et al. (1999). The cluster probability under the null hypothesis can be chosen to set the level of expected Type I error clusters at an acceptable level (e.g. < 1 false positive cluster per whole brain); in this case all statistically significant differences between groups were corrected to P ≤ 0.001 so that there were fewer than one false positive/negative clusters per whole brain. As we used whole brain cluster statistics, all tests were intrinsically corrected for multiple comparisons.

Results

Study 1

Validation of motion video as a nauseogenic stimulus and psychophysiological differences between nausea-sensitive and nausea-resistant subjects

All subjects completed the study without vomiting. Based on nausea VAS scores, the motion video induced VIMS in 56 out of 98 (57%) subjects (27 male, mean age 24.6 years, range 21–54 years). The median nausea VAS scores were 3.0 (IQR 3–4) during the motion video and 1.0 (IQR 1–1) during the control video (mean difference 2.25, 95% confidence interval (CI) 2.0–2.57, P < 0.0001), demonstrating that the motion video was successful in inducing VIMS in susceptible subjects.

Baseline differences between nausea-sensitive and nausea-resistant subjects

The overall cohort was stratified into quartiles based on mean nausea scores during the motion video with the upper and lower quartiles being considered as ‘nausea-sensitive and -resistant’, respectively. Twenty-four subjects (12 male, mean age 24.5 years, range 19–37 years, mean nausea score 3.2 ± 0.7) were classified as nausea-sensitive and 24 subjects (13 male, mean age 27 years, range 19–58 years, mean nausea score 1 ± 0) as nausea-resistant. Differences in baseline characteristics, measured prior to the motion video, between the two groups are shown in Table1.

Table 1.

The baseline psychophysiological differences between nausea-sensitive and nausea-resistant subjects; see text for details of the methodology for selecting nausea-sensitive and -resistant subjects

		Nausea-sensitive (n = 24)	Nausea-resistant (n = 24)	P
Demographics	Age (years)	24.5 (19–37)	27 (19–58)	NS
	Sex (male/female)	12:12	13:11	NS
Personality questionnaires	State anxiety	29.0 ± 2.1	27.0 ± 1.3	NS
	Trait anxiety	36.0 ± 1.8	32.0 ± 1	0.04
	Neuroticism	2.9 (2-2.8)	2.3 (2-2.8)	NS
	Extroversion	3.8 (2.9-4.5)	3.7 (3.2–4.3)	NS
	MSAQ*	18.1 ± 14.3	5.8 ± 2.2	<0.0001
	MSSQ*	22.4 ± 2.5	8.7 ± 1.7	<0.0001
Autonomic parameters	Heart rate (bpm)	66.0 ± 1.9	66.0 ± 2	NS
	Mean blood pressure (mmHg)	84.4 ± 4.5	86.4 ± 3.2	NS
	Cardiac vagal tone (LVS)	9.6 (8.3–15)	9.8 (7.5–14.6)	NS
	Cardiac sympathetic index	2.45 ± 0.5	2.3 ± 0.3	NS
Electrogastrogram	Normogastria (%age spectral power)	32.3 ± 3	35.8 ± 3	NS
	Tachygastria (%age spectral power)	33.0 ± 2.3	28.3 ± 1.4	NS
Vasopressin and ghrelin	Vasopressin (ng l−1)	3.3 ± 1.3	4.5 ± 1.8	NS
	Ghrelin (ng l−1)	28.5 ± 21.7	27.3 ± 26.7	NS

^*MSAQ, motion sickness assessment questionnaire; MSSQ, motion sickness sensitivity questionnaire. NS, not significant.

Psychophysiological changes induced by the motion video in nausea-sensitive subjects

We sought to compare the psychophysiological changes induced by motion video, in comparison to the control video, in nausea-sensitive subjects across four time points of the experiment: at baseline and during early, middle and late periods of each of the videos, hereinafter termed P1, P2 and P3, respectively (Fig.2). Nausea ratings – as expected, nausea ratings were higher during the motion video than during the control video (AUC 5.5 vs. 3.2, ANOVA for interaction, F = 24.7, degree of freedom numerator (DFn) 3, degrees of freedom denominator (DFd) 184, P<0.0001). Anxiety ratings – similarly, anxiety ratings were higher during the motion video (AUC 4.9 vs. 3.1, ANOVA, F = 2.3, DFn 3, DFd 184, P = 0.03). EGG – the motion video resulted in a lower normogastria/tachygastria ratio value in comparison to the control video (AUC 1.1 vs. 1.3, ANOVA, F = 4.0, DFn 1, DFd 92, P = 0.048). Autonomic variables: mixed measures – the motion video had no significant effect on HR but caused an increase in mean blood pressure (AUC 227 vs. 208.4, ANOVA, F = 17.4, DFn 3, DFd 184, P < 0.0001). Sympathetic and parasympathetic measures – the motion video resulted in an increase in CSI (AUC 7.4 vs. 6.1, ANOVA, F = 9.5, DFn 3, DFd 184, P = 0.0002) with a concomitant significant reduction in CVT (AUC 27.6 vs. 33.6, ANOVA, F = 5.5, DFn 3, DFd 184, P = 0.008). Plasma vasopressin and ghrelin – there was a trend for an increase in plasma vasopressin concentration (mean difference 0.97 ng l⁻¹, 95% CI −0.56 to 5.7, P = 0.1) in nausea-sensitive subjects during the motion video. Post-motion video vasopressin concentration positively correlated with mean nausea score (r = 0.66, P = 0.02). Plasma ghrelin concentration decreased significantly over this time period (mean difference −5.5 ng l⁻¹, 95% CI −9.9 to −0.1, P = 0.009). Post-motion video ghrelin negatively correlated with mean nausea score (r = −0.6, P = 0.03).

Figure 2.

Effects of the motion and control videos

The effect of the motion video (•) and the control video (▪) nausea ratings (A), anxiety ratings (B), EGG normogastria/tachygastria ratio (C), mean blood pressure (D), cardiac sympathetic index (E) and cardiac vagal tone (F) in nausea-sensitive subjects. Data shown are group mean ± SEM. ***P < 0.001, **P < 0.01, *P < 0.05.

Study 2

Differences in brain processing between nausea-sensitive and nausea-resistant subjects

Twenty-eight subjects (17 nausea-sensitive (7 male, mean age 25 years, range 19–32), 11 nausea-resistant (7 male, mean age 23 years, range 19–32)) completed the study. No subjects vomited during the fMRI scanning procedure.

Reproducibility of nausea ratings during motion video

The reproducibility of nausea VAS ratings was good between Study 1 and Study 2 in the 28 subjects who participated in both studies, indexed by an intra-class correlational coefficient of 0.97 (95% CI 0.97–0.98).

Brain activity in nausea-sensitive versus nausea-resistant subjects

A two-way interaction analysis of group (sensitive, resistant) × time bin (P1, P2, P3) revealed a single region of the brain, the left anterior cingulate cortex, to be differentially activating between the groups (peak Talairach coordinates (x, y, z) 0, 11, −3; P = 0.04; see Figs3 and 4).

Figure 3.

Differential activation of the left anterior cingulate cortex from a two-way interaction analysis group × time (cluster P = 0.041)

Figure 4.

Plot of the mean BOLD response in the right anterior cingulate cortex showing differential activations between the groups and the time bins

NS, nausea-sensitive; NR, nausea-resistant.

Correlations between nausea scores and brain activity

In nausea-susceptible subjects, nausea VAS scores correlated positively with left inferior frontal and middle occipital gyri activity (see Fig.5 and Table2). In contrast, nausea scores correlated negatively with brain activity in the right cerebellar tonsil, declive, culmen, lingual gyrus, cuneus, as well as left cuneus and posterior cingulate gyrus (Table2).

Figure 5.

Positive correlation of nausea VAS scores with brain activity in the left inferior frontal and middle occipital gyri in nausea-susceptible subjects

Table 2.

Correlations of brain activity and nausea reporting in nausea-sensitive subjects

	Peak Talairach
3D cluster size	coordinates (x, y, z)	P	Brodmann area	Region
Positive correlation between nausea reporting and brain activity in nausea-sensitive subjects
213	−40, −70, 10	0.0046	BA19	L middle occipital gyrus
450	−54, −26, 10	0.0038	BA45	L inferior frontal gyrus
Negative correlation between nausea reporting and brain activity in nausea-sensitive subjects
40	36, −59, −33	0.0094	n/a	R cerebellar tonsil
185	29, −56, −17	0.0050	n/a	R declive
186	18, −37, −10	0.0032	n/a	R culmen
269	7, −74, −7	0.0051	BA18	R lingual gyrus
265	7, −78, 17	0.0040	BA18	R cuneus
119	−22, −81, 20	0.0032	BA18	L cuneus
170	0, −52, 23	0.0063	BA23	L posterior cingulate

n/a, not applicable.

Discussion

Nausea is a common and aversive experience, the psychophysiological mechanisms of which are incompletely understood. To gain insight into the mechanisms involved we compared the psychophy-siological responses in subjects sensitive or resistant to VIMS. Brain activity was investigated in a subset of sensitive and resistant subjects using the same motion video but delivered via MRI-compatible goggles and, to the best of our knowledge, this is the first time that this particular approach has been reported. The results provide evidence to support three principal findings: (1) the motion video induces VIMS with nausea as the principal reported sensation and it is reproducible, (2) in response to the motion video nausea-sensitive subjects displayed psychophysiological changes that differed from resistant subjects and (3) differential brain activity was demonstrated in the nausea-sensitive group compared with the resistant group.

In comparison to the control video, the motion video induced the sensation of nausea in 57% of subjects, an incidence comparable to other methods for induction of VIMS (Koch, 1999; Benson et al. 2012). Moreover, the motion video reproducibly induced nausea, as demonstrated by the intra-class correlational coefficient in subjects who participated in both studies.

Overall, the behavioural evidence presented herein supports the hypothesis that viewing the rotating, tilted landscape video evokes VIMS, of which nausea is the predominant sensation. A VIMS stimulus provides a unique setting for the study of nausea, particularly in healthy subjects (Koch, 1999), and can be utilized in neuroimaging studies using MRI-compatible goggles. A previous brain imaging study used a custom-built head coil, which allowed subjects to experience vection, leading to nausea (Napadow et al. 2013b). However, such customized head coils are not yet commercially available and do not allow a fully unimpeded view of the stimulus. In contrast, our results show that the motion video, projected either onto a screen or via MRI-compatible glasses, giving subjects a completely unimpeded view of the stimulus, reliably induces VIMS in sensitive subjects.

In agreement with previous studies of ANS tone during VIMS, we have demonstrated during the subjective sensation of nausea an increase and decrease in sympathetic and parasympathetic tone, respectively, using cardiometric indices (Gianaros et al. 2003; Ohyama et al. 2007). Whilst the present study does not identify the origin of fluctuations in ANS ‘tone’, we propose two distinct, albeit not mutually exclusive possibilities. First, the autonomic changes may reflect the generally stressful/arousing nature of the experience of nausea, and similar changes would be observed with any similar unpleasant stressful/arousing stimulus (Farmer et al. 2013b) as the default response to any threat, in this case VIMS, is activation of sympatho-excitatory circuits. Secondly, the ANS changes may reflect pattern alterations, which may be ‘specific’ to nausea. Napadow et al. (2013b) demonstrated that nausea was preceded by phasic activation in the amygdala, a critical component of the limbic system, which regulates autonomic and endocrine function, particularly in response to stimuli that have an emotional valence. The present study showed an increase in activity in the left anterior cingulate cortex in nausea-sensitive as compared to nausea-resistant individuals while watching the video. The anterior cingulate cortex has been described as the ‘visceromotor cortex’ because of its involvement in autonomic responses (particularly cardiovascular) to behavioural challenges (Seth et al. 2011; Critchley & Harrison, 2013). A meta-analysis of human brain imaging studies of autonomic function identified the anterior cingulate cortex as a region associated with sympathetic, but not parasympathetic, regulation (Beissner et al. 2013). Confirmation of the precise mechanistic link between ANS changes and nausea requires further study but our finding of anterior cingulate activation is consistent with the physiological indices (e.g. CSI and EGG) of an increase in sympathetic nervous system activation before/during nausea.

Similar to changes in ANS tone, gastric myoelectrical activity of a dysrhythmic nature has been demonstrated in both healthy and clinical populations with nausea. For instance, in response to VIMS, Stern et al. (1987) observed an association between tachygastria and symptoms in a proportion of healthy subjects who developed motion sickness. In those with chemotherapy-induced nausea, gastric myoelectrical perturbations have been reported, albeit poorly correlated with symptom burden (Gianaros et al. 2001b). It is proposed that gastric dysrhythmia and tachygastria in VIMS are sequelae of a combination of increased sympathetic and decreased parasympathetic (vagal) cardiac activity (Stern et al. 2011), which is hypothesized to reflect changes in ANS outflow to other systems (e.g. stomach). Such a proposal is supported by evidence from murine studies demonstrating that animal vagotomy abolishes gastric myoelectrical disturbances in response to motion sickness (Percie du Sert et al. 2010).

Vasopressin, secreted from the posterior pituitary gland, is released in response to VIMS in susceptible subjects via a cholinergic prostaglandin-independent pathway (Kim et al. 1997), although large inter-individual variations in values are reported (Koch et al. 1990) as in our study, which probably accounted for only a trend for an increase in the plasma concentration of vasopressin. However, there was a moderate correlation between post-stimulus vasopressin concentrations and self-reported nausea intensity. Similarly weak associations between nausea and vasopressin have been reported in other studies of motion sickness (Xu et al. 1993) but it is not a universal finding (Koch et al. 1990), again reflecting inter- and intra-individual variability. The role of the increase in vasopressin in nausea is not known but proposals include: water retention ahead of anticipated vomiting; reduction of mesenteric blood flow; genesis of gastric dysrhythmias; and activation of central emetic pathways via the area postrema (Stern et al. 2011).

In our study, plasma levels of ghrelin decreased in nausea-sensitive individuals, although a previous small study, published in abstract form, found no change in response to VIMS (Klose et al. 2006). A decrease in ghrelin is consistent with the hypothesis that the full psychophysiological expression of nausea requires factors that increase (e.g. vasopressin, sympathetic activity) and a withdrawal of others (e.g. ghrelin, vagal influences on enteric cholinergic neurons) that would otherwise be antagonistic to the genesis of behavioural and physiological changes such as reduced appetite and delayed gastric emptying. A similar role has been argued for the acute reduction in plasma cortisol following administration of the cytotoxic drug cisplatin (Morrow et al. 2002) and a negative association between cortisol and chemotherapy-induced emesis (Hursti et al. 1993). A decrease in plasma ghrelin associated with nausea is consistent with the demonstration that a ghrelin receptor agonist reduced nausea associated with diabetic gastroparesis (Ejskjaer et al. 2013) and the anti-emetic effects of exogenous ghrelin in animal models (Liu et al. 2006; Rudd et al. 2006).

Changes in cortical activity during motion sickness have been reported using electroencephalography (Chelen et al. 1993), functional magnetic source imaging (Miller et al. 1996) and fMRI (Napadow et al. 2013,b). In agreement with our findings, although using a different but comparable technique, Miller et al. (1996) demonstrated an association between activity in the inferior frontal gyrus and nausea ratings induced by natural vestibular stimulation or by ingestion of ipecac. More recently Napadow et al. (2013b), using fMRI in nausea-sensitive female subjects who were exposed to a nausea-inducing video, demonstrated heightened activity in higher cortical areas such as the medial pre-frontal cortex and anterior cingulate cortex. These results contrast with the findings in our larger study, which demonstrated a positive nausea intensity-dependent effect in the inferior frontal gyrus yet a negative effect in the cerebellar tonsil, declive, lingual gyrus, culmen, cuneus and posterior cingulate cortex. The Napadow et al. (2013b) study did not identify any brain areas with a decrease in activity associated with increasing nausea perception. These differences are probably due to differences in methodology and sex inclusion. Furthermore, the Napadow et al. study evaluated a smaller number of participants, did not compare nausea-sensitive to nausea-resistant subjects and recruited only females. The reduction in activity in a number of brain regions in our study is an unexpected novel finding. The dorsal posterior cingulate cortex is implicated in parasympathetic regulation (Beissner et al. 2013) so arguably a decrease in activity here could account for the overall decrease in parasympathetic activity associated with nausea so as not to oppose the effects of increased sympathetic activity probably caused by anterior cingulate activation (vide supra). An fMRI (BOLD) study investigating brain areas where vestibular and visual information interact interestingly identified a number of the same regions as the present study (Della-Justina et al. 2015). Although the stimuli used did not induce nausea, both visual (a black and white checkerboard alternating every 62 ms) and vestibular (galvanic stimulation) stimulation individually induced a positive signal in several regions, including the middle occipital and inferior frontal gyri and the anterior cingulate cortex and simultaneous stimulation additionally activated the lingual gyrus. Negative BOLD responses were evoked in the lingual gyrus, cuneus and middle occipital gyrus but only in response to the visual or combined stimuli and not the vestibular stimulus. Although the present study and that (Della-Justina et al. 2015) differ in many respects, both point to the occipital and frontal cortex as sites at which visual and vestibular signals interact and in our study as the stimulus is intended to generate a conflicting signal, i.e. a sensory conflict between vestibular and visual aspects, leading to nausea.

Taken together with previous studies (Miller et al. 1996; Napadow et al. 2013b) our results indicate that nausea is associated with changes in brain activity (region-specific increases and decreases) in a complex manner in a diverse array of regions concerned with the sensory discriminative, cognitive evaluative and affective motivational aspects of this obdurate symptom (Stern et al. 2011).

Our study is not without its limitations. For example, we did not stratify subjects according to taster status as it has been demonstrated that non-taster status to bitterness influences sensitivity to VIMS (Benson et al. 2012). In addition, whilst the number of participants in the brain imaging study was small, the numbers exceeded other previous studies and we recruited both male and female participants (Miller et al. 1996; Napadow et al. 2013b).

In summary, our study provides novel evidence to validate the motion video as a VIMS stimulus and demonstrates a cluster of objective physiological changes that are associated with the subjective sensation of nausea. Moreover, we have demonstrated both increases and decreases in neuronal activity in a broad central network. This work has a number of potential implications for future physiological research in nausea. (1) Reverse translation of the combination of physiological factors identified will allow refinement of animal models used to investigate novel anti-emetic agents and emetic liability of candidate drugs, increasing their validity and translation of findings to humans (Percie du Sert & Andrews, 2014). (2) Studies are now required in patients experiencing nausea to investigate the extent to which the findings from visually induced nausea can be generalized to other causes. (3) We have identified three cortical regions and associated physiological changes that could be targeted as biomarkers to treat nausea.

Glossary

Abbreviations

ANS

autonomic nervous system

AUC

area under the curve

BOLD

blood oxygen level dependent

CI

confidence interval

CSI

cardiac sympathetic index

CVT

cardiac vagal tone

DFd

degrees of freedom denominator

DFn

degrees of freedom nominator

EGG

electrogastrogram

fMRI

functional magnetic resonance imaging

HR

heart rate

IQR

interquartile range

LVS

linear vagal scale

MSAQ

motion sickness assessment questionnaire

MSSQ

motion sickness sensitivity questionnaire

VAS

visual analogue scale

VIMS

visually induced motion sickness

Additional information

Competing interests

None of the authors have any competing interests to declare.

A.D.F.: Drafting of the manuscript; statistical analysis; critical revision of the manuscript for important intellectual content. D.L.W., P.M.H.: Acquisition of data; analysis of peptide hormones; critical revision of the manuscript for important intellectual content. S.J.C., V.F.B., V.P.G.: Acquisition of functional MRI data; analysis of fMRI data; critical revision of the manuscript for important intellectual content. G.B., M.A.G., S.C.W., G.J.S.: Study concept and design; critical revision of the manuscript for important intellectual content. P.L.R.A.: technical or material support; study supervision; critical revision of the manuscript for important intellectual content. Q.A.: Obtained funding; study supervision; critical revision of manuscript for important intellectual content.

Funding

These studies were funded by an NC3Rs project grant (ref - G0900797).