Hops, the immature inflorescences of the female hop plant ( Humulus lupulus L.) are one of the main components of beer and provides flavor and bitterness. Β-Eudesmol, an oxygenated sesquiterpene, is reported to accumulate in a particular hop cultivar.
![]()
Recently, we revealed that β-Eudesmol ingestion affected autonomic nerve activity in an animal model. The effect on humans has not been elucidated, therefore, we investigated the effects of β-Eudesmol on reducing objective and subjective markers related to sympathetic nerve activity after the application of mental stress in healthy participants. Fifty participants (male and female aged 20 to 50 years) were randomly assigned to two groups. Five minutes before taking the Trier Social Stress Test (TSST) as a mental stressor, participants in each group ingested a beverage containing β-Eudesmol, the active beverage, or a placebo beverage that did not contain β-Eudesmol. Saliva 3-methoxy-4-hydroxyphenylglycol (MHPG), a major product of noradrenaline breakdown and a representative marker of sympathetic nerve activity, was significantly lower just after the TSST in the active group compared with the placebo group. Saliva cortisol, a marker of the endocrine stress response system, was not significantly different between the two groups. No adverse events related to test beverage ingestion were observed.
This is the first experimental evidence of β-Eudesmol effect for mental stress in human. IntroductionMental stress has been defined as “a set of events in the social milieu which modify steady state conditions so as to activate adaptive mechanisms”.
Mental stress itself is not a disorder; however, it affects numerous biological systems, such as autonomic nervous system , hypothalamic-pituitary-adrenal (HPA) axis , the immune system , and increases the risk of mental illnesses such as anxiety, nervousness, symptoms of depression , post-traumatic stress disorder , and irritable bowel syndrome. The prevalence of mental stress has rapidly increased around the world and intensive research has been conducted into the development of drugs and functional foods which can prevent or reduce mental stress. Natural medications are thought to be mild and safe methods in comparison to conventional medicines. Some natural products have been reported to be effective in reducing mental stress responses in humans ,. For example, it was reported that L-theanine intake resulted in a reduction in the heart rate responses to an acute stress task, suggesting that it affects autonomic nervous system. However, evidence from clinical trials of natural products is still limited.β-Eudesmol, an oxygenated sesquiterpene, has been found in medicinal plants, such as Atractylodes , Teucrium , and Anaxagorea , and accumulates in edible plants, especially a particular hop cultivar and Eucalyptus.
Some of the physiological effects of β-Eudesmol, determined using a mammal model, have been reported; for example, anti-angiogenic activity, anti-tumor activity, blocking action of succinylcholine on acetylcholine-activated channel activity, and activation of transient receptor potential channels , however, effectiveness for humans has not been well studied. It was recently shown that oral administration of β-Eudesmol could affect autonomic nerve activity, resulting in suppression of the adrenal efferent sympathetic nerve system in rats. The effects of β-Eudesmol on autonomic nerve activity response to stressors, and its effectiveness in humans, have also not yet been elucidated in a clinical trial.To confirm the effect of β-Eudesmol on mental stress in humans, we performed a placebo-controlled double-blind cross-over study. Healthy adult male and female volunteers ingested water or β-Eudesmol-containing water, just before taking the Trier Social Stress Test (TSST), which is a laboratory-based mental stress test. TSST is utilized as an acute mental stress protocol to study the stress response in human, which was designed in the early nineties. Briefly, participants are introduced a role-playing scenario after a resting period in the laboratory.
In the test, a speech (job interview) and a mental arithmetic task (serial subtraction) are performed in front of assessors. Following these tasks, participants rest and post-stressor measurements are carried out.
Mental stress related biomarkers are affected due to mental stress caused by TSST ,. In this study, saliva samples before and after stress exposure were analyzed for concentrations of 3-methoxy-4-hydroxyphenylglycol (MHPG) a marker used for sympathetic nerve activity and cortisol, a marker of hypothalamic-pituitary-adrenal (HPA) axis activity. Subjective evaluations were also carried out by questionnaire. In addition, we evaluated the safety of β-eudesmol ingestion in this study. Study ProceduresThe protocol (Protocol No. HR-2016-KR03) was approved by the Institutional Review Boards of the Ethical Committees at the Oriental Ueno Detection Center, General Incorporated Association and Oriental Occupational Health Association, Tokyo Branch (Tokyo, Japan), in accordance with the ethical standards established in the Helsinki Declaration and the ethical guidelines for epidemiological research of the Ministry of Education, Culture, Sports, Science, and Technology, and the Ministry of Health, Labor, and Welfare in Japan.
This study was registered with the UMIN Clinical Trials Registry as UMIN000020896 (title: Study of the effects of beverage containing plant ingredient on mental stress), and was conducted in compliance with the protocol. Written informed consent was obtained from all participants. This study was performed by a contract research organization, TES holdings Co., Ltd.
(Tokyo, Japan) from January 2016 to June 2016 at the TKP Ichigaya Conference Center (Tokyo, Japan).Study ID number and website: UMIN000020896,. ParticipantsThe entry criteria were as follows: (1) male and female subjects aged between 20 to 50 years old at the time of registration, (2) no history of chronic disease, (3) has feelings of mental stress, such as tension, blushing, increased heartbeat when talking in front of an audience, (4) agreeing to participate voluntarily and in writing, after receiving sufficient explanation of the purpose and content of this study, (5) able to attend the designated examination date and receive examination, and (6) judged able to participate by the study physician. Study DesignThe TSST was conducted according to previous reports.
![]()
Appendix A – Dundee Stress State Questionnaire. The Dundee Stress State Questionnaire (DSSQ; Matthews et al, 1999, 2002) will be used in the current.
A randomized double-blind, placebo-controlled cross-over study was conducted over five weeks, consisting of a screening day, a first beverage ingestion day, a two-week washout period, and a second beverage ingestion day. The screening test for study eligibility was performed over 16 days before the beginning of the first beverage ingestion period and eligible participants visited the TKP Ichigaya Conference Center for all of the tests.
Measurements of the anthropometric and circulatory parameters were only conducted at the screening test. Interviews by a physician were conducted at every test.
To minimize the circadian variations of biological markers, all experimental sessions started after 11:00 and ended before 17:30. Subjects arrived at least 30 min before the start of the study and this was named the habituation period. Subjects ate a standardized meal as lunch during the habituation period.The study controller randomly assigned participants in a 1:1 ratio into two groups with random numbers, and stored the assignment list in a locked container. All participants, investigators, and study personnel, except for the controller, were blinded to the assignment list throughout the study.
Participants consumed one test beverage just before the TSST on ingestion day. Participants were instructed to continue their usual eating, exercise, sleeping, smoking and drinking habits throughout the study. Use of oral medication, dietary supplements and functional foods affecting mental stress were prohibited. Job interviews and public speaking were prohibited during the test period.
On the day before the test, participants were prohibited from drinking alcohol and were asked to finish their evening meal 12 h before the test commenced, after which eating (except for the prepared standardized meal) and drinking (except water) were prohibited until the test was complete. Smoking was also prohibited on the test day until the test was complete.Discontinuance criteria for the study participants were as follows: (1) injury to the participant, (2) difficulty in continuing the study due to a serious adverse event or accident, (3) continuous or serious noncompliance of the protocol, (4) pregnancy, or (5) anything which the site investigator judged necessary to cause discontinuance of the study.
EndpointsThe primary endpoints were improvement in salivary MHPG and cortisol levels after TSST, and secondary endpoints were improvements in salivary chromogranin A , State-Trait Anxiety Inventory-Form JYZ , and Dundee Stress State Questionnaire III. Chromogranin A is an acidic glycoprotein released with catecholamine from the adrenal medulla and the sympathetic nerve endings and has been used as a stress marker in saliva.
Safety endpoints were any occurrence of an adverse event. Evaluation of safety was conducted by doctor’s inquiries to participants in test days, and diary during the examination period written by participants. When an adverse event occurred, the site investigators conducted a follow-up survey until the event disappeared or recovery began (last date of the follow-up survey: 20 March 2016). Measurements of Salivary ParametersSaliva was collected using a swab and centrifuge tube (Salimetrics Oral Swab (SOS), SALIMETRICS, Carlsbad, CA, USA) at Ichigaya Conference Center. And the resulting saliva was divided and kept frozen at −80 °C until use at Kirin Company and Kurume University, respectively.
Saliva MHPG was measured by gas chromatography coupled with mass spectrometry (Hitachi–M80B, Hitachi, Japan) as described previously at Kurume University. Standard solutions containing 1, 5, and 10 ng of free–MHPG in 10 mL of Ringer’s solution were prepared, and used to derive the calibration curve. GC-MS analysis was carried out using a Hitachi M–80B (Hitachi, Tokyo, Japan) double focused mass spectrometer interfaced to a data acquisition system. Helium was used as a GC carrier gas (40 mL/min).
The GCMS interface oven and transfer line were set at 260 °C. A megabore column coated with DB-1 (0.53 mm i.d. × 15 m, J and W, Co., Santa Clara, CA) was used. The oven temperature was increased at a rate of 25 °C/min from 125 °C to 220 °C. The electron emission current was 100 mA and the electron energy was 70 eV.
For the electron ion monitoring study, m/z 472 and m/z 475 were monitored. The ratio of the peak areas was used for estimation of the amount of MHPG from the calibration curve. The data are expressed as ng/mL of MHPG. Saliva cortisol and chromogranin A were measured by commercially available kits according to the manufacturer’s instructions (Salivary cortisol EIA kit, (SALIMETRICS) and Human Chromogranin A EIA, (Yanaihara Institute Inc., Shizuoka, Japan), respectively at Kirin Company (Tokyo, Japan).
Study Design and Background of the ParticipantsThe study design is presented in, and a detailed explanation is given in the materials and methods section. The time schedule of the test day is shown in B; the TSST start time is defined as 0 min. The flow of participants enrolled in this study is shown in. Out of 106 participants which underwent the screening test, 56 participants were excluded so that 50 participants (33 male and 17 female) were randomized into two groups (25 participants in each group). Four participants dropped out of the study due to personal problems not related to the study.
Among the 46 remaining participants who finished the study, eight participants were excluded by the site investigators on the basis of their primary and secondary endpoint evaluations. Exclusion was performed before the release of the allocation list and was in accordance with the exclusion criteria determined before the study. Exclusion was due to: failure to obey compliance rules (five participants) and insufficient amount of saliva for MHPG analyses (three participants). The baseline characteristics of the 38 participants are shown in, indicating that age met entry criteria, and blood pressure was lower than exclusion criteria. Statistical analyses of primary endpoint, MHPG evaluation, and secondary endpoints (State-Trait Anxiety Inventory-Form JYZ, and Dundee Stress State Questionnaire III) were conducted on 38 participants (25 male and 13 female).
Among the 38 participants, nine participants were excluded due to in sufficient amount of saliva for analysis on the basis of their saliva cortisol evaluation. Of the remaining 29 participants, 16 were also excluded due to insufficient amount of saliva for chromogranin A evaluation. Consequently, statistical analyses of saliva cortisol and chromogranin A were conducted on 29 and 13 participants, respectively. Primary EndpointA shows the change in concentration of saliva MHPG levels during the tests. Saliva MHPG was significantly increased in both groups just after taking the TSST when compared with before the test started (−5 min). Saliva MHPG levels were reduced in both groups during the 10 min resting period (25 min in A), and no significant differences were observed between the groups.
Saliva MHPG levels were significantly decreased between the −5 min and 20 min resting time points in both groups (35 min in A). Saliva MHPG was significantly lower in the β-Eudesmol group compared with the placebo group just after the TSST; however, levels were significantly increased at the 15 min time point when compared to −5 min in both groups (A, p.
Effects of β-Eudesmol on saliva 3-methoxy-4-hydroxyphenylglycol (MHPG) and cortisol. ( A) Effects of β-Eudesmol on saliva MHPG. Saliva MHPG levels were measured at −5, 15, 25, and 35 min. ( B) Area under the curve (AUC) of saliva MHPG. ( C) Effects of β-eudesmol on saliva cortisol. Saliva cortisol levels were measured at −60, −5 15, 25, 35, 45, 75, and 135 min.
( D) AUC of saliva cortisol. Data are expressed as means ± SEM. Β-Eud, β-Eudesmol; n.s., not significant; TSST, Trier Social Stress Test., p. Secondary EndpointsThe change in saliva chromogranin A levels throughout the study was shown in. No significant difference was observed between the β-eudesmol group and placebo group during all time points tested. Saliva chromogranin A was significantly increased in the placebo group at 45 and 135 min after the TSST when compared with −5 min; however, there were no significant differences between −5 min and the measured time points in the β-eudesmol group.The State-Trait Anxiety Inventory-Form JYZ, and Dundee Stress State Questionnaire III were used to evaluate anxiety and stress state, respectively. There were no significant differences in the −5 min values of any parameter in the two questionnaires between the two groups.
As shown in, the TSST significantly increased the scores for state anxiety in both groups; however, no significant differences were observed between the β-Eudesmol group and placebo group. Trait anxiety scores were not significantly affected by the TSST in either group. The Dundee Stress State Questionnaire III showed that the TSST significantly increased “Unpleasant stress” and “Anxiety” scores in both groups; however, no significant differences were observed between the groups. A significant difference in the “Concentration on tasks” score was observed between the placebo group and β-Eudesmol group at 135 min. The β-Eudesmol group showed a significant difference in “Concentration on tasks” score between −5 min and 135 min. Safety EndpointsSafety endpoint evaluation was conducted on 46 participants who ingested test beverages at least once.
No clinically problematic findings were noted from urinalysis throughout the study. During the study period, 15 adverse events involving 11 participants were reported. The most common adverse event was pollen allergy-like symptoms or headache. Adverse events reported during the study period were as follows: three cases of pollen allergy, three cases of headache, two cases of stomach ache, two cases of muscle ache, two cases of menstrual pain, and one case each of shoulder discomfort, dental pain, and dry eye. All cases were judged to be mild and to have no relation to the test beverages by the site investigators.
DiscussionWe investigated the effect of β-Eudesmol ingestion prior to exposure of mental stress via the TSST model in healthy adult participants. Saliva MHPG and cortisol were significantly increased by the TSST in both groups. Saliva MHPG increased faster than cortisol which may reflect the difference between the two parameters: saliva MHPG indicates autonomic nerve activity , and saliva cortisol indicates endocrine system activities. The response pattern against a mental stressor was consistent with previous reports ,. In addition, scores from the State-Trait Anxiety Inventory-Form JYZ showed that the TSST significantly increased anxiety in both groups.
These results suggest that mental stress was successfully loaded onto participants by our experiments.MHPG reflects central nervous system activity, and is relevant to symptoms of anxiety or depression. Β-Eudesmol significantly lowered saliva MHPG concentrations after the TSST in adult participants when compared with the placebo group. The level of saliva MHPG is reported to highly correlate with plasma MHPG levels and is therefore an established marker of central noradrenergic metabolism. It is suggested that β-Eudesmol could suppress sympathetic nerve activity which responds to acute mental stress. However, no significant difference in saliva cortisol level was observed between the β-Eudesmol and placebo groups although saliva cortisol was increased in both groups after the TSST.
Saliva cortisol was higher at −60 min when compared with the TSST starting time (0 min) in both groups. Saliva cortisol is reported to be affected by dietary components , and the higher cortisol levels at −60 min may be due to the standardized meal during the habituation period. These results indicate that β-Eudesmol mainly affects autonomic nerve activities and not HPA-axis activities. Participants were instructed to maintain their normal lifestyle i.e., eating, exercise, sleeping, smoking, and drinking habits, during the study period. Saliva MHPG and cortisol concentrations did not differ between the two groups before the start of test beverage intake, indicating that ingestion of β-Eudesmol reduces saliva MHPG secretion after mental stress induced by the TSST in healthy participants.
A significant difference was observed in the MHPG response after the TSST between the groups; however, no significant difference in the subjective evaluations was found between the groups, except the “Concentration on task” from the Dundee Stress State Questionnaire III after 135 min of β-Eudesmol intake. The significant decrease in “Concentration on task” score in the β-Eudesmol group suggests a relaxation effect on subjective feeling, possibly as a result of a reduction in sympathetic nerve activity by β-Eudesmol. Further research is required to reveal any relationship between the “Concentration on tasks” score lowering effect and the physiological effects of β-Eudesmol.Our previous study revealed that oral-administration of β-Eudesmol affects autonomic nerve activity in an animal model , which supports the mechanism involving β-Eudesmol intake lowering the saliva MHPG concentration after mental stress in this study. In an animal model, adrenal efferent sympathetic nerve activity is suppressed after orally administered β-Eudesmol; however, it is not suppressed by subcutaneous administration. It is reported that the scent threshold of β-eudesmol is 10,000 ppb : the active beverage contained 5 ppb β-Eudesmol in this study and could not be distinguished from the placebo beverage by the controller. These results suggest that the organ in which β-Eudesmol first acts may be the digestive tract and reflects autonomic nerve activity. In previous report, β-Eudesmol could stimulate transient receptor potential ankyrin 1 (TRPA1).
The TRPA1 is a calcium-permeable non-selective cation channel which plays a crucial role in the susceptibility to various stimuli. TRPA1 is mainly expressed in sensory nerves and plays an important role in the susceptibility to various stimuli and is also expressed in the digestive tract ,. Reduction of sympathetic nerve activity by oral-administered β-Eudesmol was eliminated in TRPA1 knockout rat , suggesting that TRPA1 activation by β-Eudesmol may be involved for reduction of MHPG in this human study. In previous study, l-theanine intake reduced the heart rate responses to an acute stress task, suggesting that l-theanine affects autonomic nerve activity against mental stress. Suggesting mechanism of its effect is to increase γ-aminobutyric acid (GABA).
Whereas the relation of l-theanine and TRPA1 was not reported, it is a future work whether β-eudesmol shows the same effect as the l-theanine, such as reduction of heart rate or increase of GABA.The TSST has been widely used in the basic understanding of psychological endocrinology. In addition, the TSST has been used to assess the effects of both therapeutic and nonmedical drugs , and has also been used to verify the effect of foods and beverages which may relieve mental stress ,. Previously, food ingredients with stress response effects have been elucidated using the TSST and measuring saliva and serum cortisol, or adrenocorticotropic hormone (ACTH) elevation as stress markers ,. Our study showed that the TSST increased saliva MHPG as well as saliva cortisol, indicating that the TSST is also able to detect activation of the sympathetic nervous system due to mental stress. Since saliva can be collected noninvasively, saliva MHPG is thought to be a useful marker in developing functional foods or beverages to mitigate acute physiological responses against a stressor.
Beverages containing β-Eudesmol may suppress the activation of the sympathetic nervous system and it can be taken by the simple method of a single oral intake just before mental stress. Especially, it may be effective against mental stress caused by socially evaluated environment, since the TSST is a well-established method that can induce social-evaluative threat in the laboratory.All adverse events reported during the study period were judged to be mild and to be unrelated to the test beverages by the site investigators. Thus, it may be assumed that single ingestion of β-Eudesmol at the dose used in this study is safe. In addition, the effects of β-Eudesmol on reducing saliva MHPG was mitigated at the end of the test, indicating that the effect of a single dose of β-Eudesmol is transient and does not persistently affect autonomic nerve activity.
ConclusionsIn conclusion, our study revealed that single ingestion of β-Eudesmol reduces the acute mental stress response, in particular, saliva HMPG secretion indicating sympathetic nerve activity in healthy humans. Intake of β-Eudesmol (950 ng) was sufficient to achieve this effect. Therefore, β-Eudesmol could be a useful and safe tool to prevent mental stress responses and related disorders. The effectiveness of β-Eudesmol on the acute mental stress response was confirmed in this study; the effects on chronic mental stress will be investigated in a future study.
This study demonstrates that rumination is reflected in two behavioural signals that both play an important role in face-to-face interactions and provides evidence for the negative impact of rumination on social cognition. Sixty-one students were randomly assigned either to a condition in which rumination was induced or to a control condition. Their task was to play a speech-based word association game with an Embodied Conversational Agent during which their word associations, pitch imitation and eye movements were measured. Two questionnaires assessed their ruminative tendencies and mind wandering thoughts, respectively. Rumination predicted differences in task-related mind wandering, polarity of lexical associations, pitch imitation, and blinks while mind wandering predicted differences in saccades. This outcome may show that rumination has a negative impact on certain aspects of social interactions. IntroductionWe all experience our thoughts drifting away while attempting to concentrate on a task, whether it be reading an article, listening to a lecture, writing a paper, or even having a conversation.
Indeed, such mind wandering (MW) thoughts comprise up to half of our daily thoughts ,. MW has been defined as self-generated thought which is active and independent of perceptual input, often unrelated to the task at hand and directed to goals that extend beyond the here and now –. However, in some cases, MW may also be task-related (e.g.:“I wonder how long it will take me to finish reading this article.”) ,. The past decade has seen a substantial increase in the understanding of how MW thoughts emerge and the reasons for their occurrence. Their costs include decreased text comprehension , , higher variability in reaction times , increased number of errors in both memory, working memory and choice reaction time tasks , , as well as lower measures in general aptitude and increases in negative mood. Indeed, MW can have negative consequences; however, it also provides freedom from immediacy and has been associated with creativity in unusual uses tasks and future planning ,.
Given its pervasiveness in our lives, it may reflect an essential adaptation of the mind, serving to maintain a coherent sense of self by integrating the past and present self with future experiences –. MW has been studied in the laboratory during various interactive tasks as well as in daily-life; however, most of these studies have not focused on the social nature of our mind wandering thoughts and their role in our daily interactions. The present study aims to fill this gap in research on MW by investigating social and cognitive cues to ruminative and MW episodes and exploring the implications for social interactions. Personal goals, negative mood, and mind wanderingMW thoughts are often directed towards personal goals that are not directly related to the task at hand , but are associated with current life concerns , which suggests that the adaptive function of MW can facilitate problem-solving in daily life ,. Within the context of personal goals, a negative mood may indicate personal problems that individuals may try to solve during MW.
Both the induction of personal concerns , , as well as induction of negative moods have been shown to increase MW.When MW about problems and concerns leads to effective problem solving, this may have positive consequences on both cognition and affect. However, when ineffective, MW heightens the salience of the current problem or concern, which can in turn have negative affective and cognitive consequences. Studies indicate that it is unclear whether negative mood precedes MW , , yet it has been proposed that negative mood influences the affective content of MW, making mood congruent thoughts more accessible ,. The type of MW hence depends on the format and content of MW thoughts, which dictates whether the thoughts will have (mal)adaptive outcomes. Rumination and wanderingRumination is characterised in terms of persistent and recurring self-reflective thoughts about a particular theme that deviate attention away from relevant themes and current tasks in the immediate external environment ,. Whenever MW becomes rigid and inflexible in the form of ruminative, perseverative cognition, it may become a risk factor ,.
That is, when MW loses its expansive, adaptive form, it may under certain circumstances lock into a repetitive spiral of homogeneous negatively-valenced thoughts. The first objective of the current study is to corroborate previous research by examining if rumination increases frequencies of MW. On the one hand, MW and rumination may be seen as antithetical concepts, where MW is seen as a form of free, unguided, internal thought, while rumination is thought that is fixed around a single theme ,. On the other hand, rumination may be seen as a style of thinking which may take hold of a wandering thought, lock it into a spiral of repetitive self reflection, and impede individuals from focusing on the task at hand , –. As MW and rumination are inevitably linked , this study aims to integrate these concepts, motivated by the fact that one cannot discuss rumination without taking into consideration that it is a style of MW.
Nor can one discuss MW without taking into account the possibility that a particular portion of these self-generated thoughts may fall captive to a ruminative style of thinking. Although the majority of research presupposes MW to refer to task-unrelated thoughts alone, other methodological perspectives deviate from such a view and consider there to be two types of self-generated thought, namely, interfering thoughts concerning appraisal of one’s performance on a task (task-related interference, TRI), and thoughts directed towards information that is unrelated to the current environment or to the current task (task-unrelated thinking, TUT). The rigid and inflexible quality of ruminative thought appears to be closely related to the rigid characteristics of task-related interferences. At the same time, rumination has the potential to exacerbate the relationship between pre-existing dyspohoria and increases in task-unrelated thoughts.Rumination can be divided into two major subcategories, namely, reflective pondering and brooding. Self-reflection suggests a purposeful inward focus aimed at cognitive problem solving. Brooding, on the other hand, involves a comparison of one’s current situation with some unachieved standard. Often, brooding is associated with a decreased controllability of thoughts.
Rumination can be adaptive when self-reflection serves to solve problems, however, it may become maladaptive when either reflection or brooding lead an individual passively think about their problems and feel helpless in finding solutions. Such maladaptive rumination has been found to be associated with the excessive elaboration of negative information. Moreover, high trait rumination has been associated with enhanced recollection memory for negative words in young females, also after controlling for negative mood. Hence, the second objective of this study is to expand upon previous research and investigate whether rumination is related to the production of more negative lexical associations in a task. Ruminative self-focus in social cognitive mechanismsA significant proportion of MW thoughts concerns others , ; i.e., it is interpersonally-focused and social in nature. Various studies have investigated MW in interactive settings, such as in the classroom, both online and offline , as well as with intelligent tutoring systems ,. Although most of this research is social and interactive by nature, it has primarily investigated performance-related consequences of mind wandering in these interactive settings.
However, the social and relational consequences of the mind wandering have been largely neglected. The current study addresses this gap in the literature and highlights the need to investigate behavioural social cues to MW in interactive environments. Based on past research, we hypothesise that MW thoughts might be detectable through social cognitive mechanisms such as pitch imitation and eye movements, which provide valuable information about individuals’ emotions, mental states, and behaviours during social interactions ,. Pitch imitationSpeakers have frequently been shown to accommodate to one another’s pitch patterns, as pitch is perhaps the most important indicator of the emotions and attitudes of a speaker ,. The ability to correctly perceive pitch in another’s speech and to adapt one’s pitch according to one’s goals is an essential communicative and social skill (Communication Accommodation Theory; –). Pitch convergence may then be an indicator of awareness towards one’s environment and of desire for social approval and acceptance –.Neuroimaging evidence suggests that reduced imitation is associated with self-related processing and independent self-construals. Moreover, experimentally induced self-focus inhibits imitation ,.
Additionally, depressed individuals, who tend to have more negative thoughts, have also been found to show less behavioural imitation , while dysphoric and depressed individuals have failed to express normal facial imitation of both positive and negative facial expressions. Consolidating the association between self-focus and reduced imitation, the third objective of this study is to investigate whether rumination, characterised by recurring, self-focused thoughts, is associated with reduced (pitch) imitation. Gaze behaviourAn important cue signalling joint attention between interaction partners is gaze behaviour. It provides important information about people’s social and cognitive behaviours and is indicative of visual attention processes ,. Only few studies have investigated eye movements in relation to rumination , ; and various studies have investigated eye movements and MW in a range of tasks including reading, driving, tasks of sustained attention, and learning with an intelligent tutoring system , –.During social interactions, eye contact serves as a signal of joint attention and interest between interaction partners. In MW studies, fixations and saccades decrease and average fixation duration becomes longer , , indicating that eye movements are both slower and less frequent during MW episodes, which may be indicative of increased cognitive inflexibility.
In a study by Rauthmann and colleagues , individuals who scored high on neuroticism, a personality trait highly tied to rumination, had less and longer fixations and spent more time dwelling on an abstract animation. With respect to blinks, Smilek and colleagues found that participants blinked more during MW than during on-task episodes. More blinks have also been shown to indicate the exchange of attention from the external task at hand to internal thoughts, and were thus associated with reduced attention and increased error in the processing of external information ,. In sum, less active saccades, less and shorter fixations, and more blinks may represent reduced attention towards the external environment during MW. The third objective of this study is then to build upon and complement previous literature on MW and eye movements by investigating to what extent rumination affects eye movements in an interactive setting.
Current studyThe present study was designed as a speech-based word association game with an Embodied Conversational Agent in order to investigate how rumination affects participants’ lexical associations. A word association game with an Embodied Conversational Agent was used in this study because of its relative simplicity, enabling us to control for any noise that might arise from more ecologically valid contexts, such as a free conversation with an interaction partner. At the same time, it is less constrained than a simple speech shadowing task in a non-interactive setting , , in which participants are limited to only repeating isolated words. In contrast, participants have the freedom to come up with their own word associations in this game. This study is innovative in that it explores two social and cognitive mechanisms, namely, pitch imitation and eye behaviour as possible behavioural cues to rumination and mind wandering, and addresses their implications during social interactions.
The guiding question in this research then is: To what extent does self-focused rumination affect MW, negative lexical associations, pitch imitation and eye behaviour? We propose the following hypotheses: H1) Rumination will result in increases in MW; H2) Rumination will increase the occurrence of negative lexical associations; H3) Rumination will negatively affect imitation, and finally; H4) Rumination will affect eye movement behaviours; more specifically, it will be associated with more blinks, less saccades, less fixations, and longer average fixation duration. Participants and designSixty-two English speaking students were recruited from the student population at Tilburg University. Participants were aged 18 to 33 years; mean age was 23.82 years ( SD = 2.50). This sample size was sufficient for detecting the smallest effect size of interest (SESOI) according to Simonsohn (2015)’s recommendation to set the smallest effect size of interest so that the original experiment had 33% power to reject the null hypothesis if this effect size was true. The SESOI for this experiment was based on Smallwood and O’Connor (2005) for MW, Mattheij and collegues (2015) for pitch, and lastly, Smilek and colleagues (2010) and Rauthmann and colleauges (2012) for eye movements ,.
Thirty-one males and 31 females participated, but data from one male participant had to be excluded due to a procedural error. The study was approved by the Tilburg University Institutional Review Board, and written informed consent was obtained from each participant at the beginning of the experimental session.
Participants took part in an interactive task presented as a word association game with an Embodied Conversational Agent. The between-participant manipulation was induced state rumination. The within-participant factor in the task was the Embodied Conversational Agent’s vocal pitch (High/Low). The dependent variables measured were self-reported MW, the polarity of the words that participants generated during the word association game (positive, negative, neutral), pitch imitation parameters, and eye movement parameters (blinks, saccades, fixations, and average fixation duration).
Trait ruminationWe used the Rumination Inventory (, ), which was designed to measure a tendency toward distractability, a tendency to engage in mental rehearsal of future and past events, and a tendency toward repetitive thought, including increased frequency and decreased controllability of thoughts. Participants had to indicate whether the statements (e.g., “I often get distracted from what I am doing with thoughts about something else” and “I have no trouble focusing all of my attention on one thing”) describe them well or not on a scale from 1-7. The scale contained 10 items and was found to be reliable ( α =.67).
Participants were asked to complete the RI for a second time at the end of the experiment ( α =.75). State ruminationIn the rumination induction procedure, treated as a proxy for state rumination, participants were asked to think about a variety of (45) recurring self, emotion, and symptom focused thoughts (e.g., “your character and who you strive to be” and “the way you feel inside”). In the control condition, participants thought about 45 items that were not related to the self, emotions, or symptoms (e.g., “the shape of the continent of Africa” and “a group of polar bears fishing in a stream”; ). Mind wanderingIn order to measure MW, we used a subjective measure of task-unrelated thought (, Thinking component of the Dundee Stress Questionnaire; ). This scale assesses what participants are thinking about during a recently completed task. It contains two 8-item parts: (1) one measuring task-related interferences (TRI; e.g., “I thought about how I should work more carefully”) and (2) one referring to task-unrelated thoughts (TUT; e.g., “I thought about personal worries”).
Participants were asked to indicate on a 5-point Likert scale how well each of the statements described them, ranging from 1(never) to 5 (very often). The scale contained 16 items and was found to be highly reliable ( α =.89).
Reliability coefficients for each subscale are reported in. ProcedureParticipants arrived in the lab and were asked to sit in the sound-proof booth in front of a computer screen where they first filled out the Rumination Inventory. In order to ensure that participants in both conditions were in a comparable mood, participants watched a ninety second long relaxation video consisting of underwater scenes and accompanied by soothing music (, ). Then they were randomly assigned to either the rumination induction or to the control group. After the rumination/ control task, they played a word association game with the Embodied Conversational Agent in which their voice and eye movements were recorded.
Participants followed the instructions on the screen and were asked to sit as still as possible in order for the device to capture their eye movements accurately. Before having their eye movements recorded, they underwent a calibration and validation procedure in order to ensure that the eye tracker was measuring correctly. Next, the Embodied Conversational Agent began the word association task, pronouncing either high- or low-pitched words. Each stimulus was preceded by a centralised fixation cross that was on the screen for one second. Participants then had to pronounce the first association that came to mind, within four seconds. After the word association task, participants had to fill out two post-questionnaires; one about their MW episodes during the task (Thought Component of the Dundee Stress State Questionnaire (DSSQ) and the Rumination Inventory for the second time, in order to verify the effectiveness of the rumination induction. A full version of the materials is available in the supplementary files.
Pitch data analysisThe aim of the auditory data analysis was to determine whether there was a significant change in the vocal pitch of participants in response to a low or high-pitched utterance by the Embodied Conversational Agent, and whether this change was higher for participants in the rumination condition compared to the control. The recordings were analysed with Praat 5.3.04. They were first visually inspected in order to establish a pitch floor and ceiling for the speakers to prevent errors.
![]()
As a result, the range was set to 40 Hz—400 Hz for both the male and female voices. Prior to the analysis of the word recordings, extraneous noises and non-speech sounds (pauses, hesitations, clearing of the throat, and background noises) were edited from the recordings and octave jumps were manually corrected. Creaky voices and octave jumps frequently result in errors of pitch determination, and hence had to be manually corrected –. This required visual and auditory inspection of each of the 1233 recordings in order to remove pauses, hesitations, stuttering, creaky voices and octave jumps. The mean pitch for each segment was determined with the standard autocorrelation-based pitch detection in Praat in semitones. A total of 1233 pitch measurements were obtained for 1342 (22 × 61) of the experimental trials. Output was missing for 109 of the trials (8.12%) where participants were unable to come up with a word in the association task within the allotted time.
Pitch measurements were averaged per participant and per condition (high or low pitch uttered by the Embodied Conversational Agent). Lexical data analysisPrior to the lexical analysis several words had to be adjusted in order to be recognised by the semantic analysis tools. Words in the past tense had to be modified to present tense and plural words had to be changed to their singular form. When the response of the participants contained more than one word (e.g., “hard work”), the words were analysed separately, and both words were included in the final analysis.
When the participant was undecided about an answer and uttered multiple words, the first word was always analysed (e.g., “words/limit/word limit”).The polarity of the words (positive/negative) was analysed with LIWC (Linguistic Inquiry and Word Count, ), a text analysis software that is widely used in a broad range of experimental settings to evaluate emotionality. Since LIWC may not have always captured the meaning of compounds (e.g., ‘give up’ would be coded as two separate items with positive polarity, while the expression as a whole has negative polarity), the lexical associations were also analysed by SentiWordNet , a publicly available lexical tool for opinion mining. SentiWordNet determines the polarity of a word by assigning three numbers to it, a positivity, a negativity and an objectivity value that always add up to one. In SentiWordNet, contrary to the LIWC, compound words are recognised as single entries. Eye movement data analysisParticipants’ eye movements were recorded with an SMI RED 250 eye-tracking device, with a sampling rate of 250 Hz positioned below a Dell computer (22-inch monitor, 1680x1050 resolution). This system uses infrared tracking technology which measures pupil center and size of both eyes.
Blinks, saccades, fixations, and average fixation duration were processed in MATLAB R2015a (8.5.0) in order to prepare the data for statistical testing. Often in eye tracking experiments measuring mind wandering, if the quality of data is low (e.g. Due to loss of signal or a participant not facing the eye tracker), data points or entire participants are excluded from the analysis ,. Following such practices, eye movement data for 2 participants were excluded, as there were insufficient eye movements recorded for over 50% of these participants’ trials. ResultsData were analysed for 61 participants. Means, Standard Deviations, and Cronbach’s α’s are displayed in. Pitch, lexical associations and eye movements were averaged across trials and aggregated to the participant level.
We first performed a Kolmogorov-Smirnov test of normality for rumination measured at the beginning and end of the experiment, MW(TRI/TUT), high and low pitch, each eye movement parameter, and lexical associations in order to evaluate the distribution of values in comparison to the standard normal distribution. Rumination measured prior to the experiment, MW, saccades, and fixations were normally distributed. For blinks, lexical parameters and pitch, the test indicated a non-normal distribution of values ( p. Effect of ruminationA mixed ANOVA with the first and second measures of rumination as within-participant variables and condition (rumination/control) as the between-participant variable, indicated that there was a significant effect of the experimental manipulation on participants’ rumination tendencies, with main effects found for condition F(1, 59) = 4.50, p =.038, η p 2 =. No main effects were found for trait rumination, F(1, 59) = 2.21, p =.143, η p 2 =.
04 and there were no interaction effects between trait rumination and condition F(1, 59) = 1.20, p =.278, η p 2 =. Means and standard deviations reveal that participants in both the rumination induction condition and in the control group were already highly ruminative prior to the manipulation. There were no significant differences between both groups at T1.
However, at T2, means for trait rumination in the control group were significantly lower than means for the rumination group, which on average, stayed the same from T1 to T2. Although we expected the rumination induction to have led to a significant increase in rumination scores, the reverse occurred, and instead, there was a significant decrease in rumination scores in the control group. The fact that the rumination induction did not lead to an increase in rumination scores was a caveat in this study; therefore, to pry out the effects of the rumination manipulation from trait rumination, we introduced trait rumination as a control variable and treated induced rumination as a proxy for state rumination for all subsequent analyses. Mind wanderingIn order to test whether rumination increased MW, a two-way ANOVA was performed on answers to the Thinking Component of the Dundee Stress Questionnaire with both trait rumination and state rumination as covariates. There was a marginally statistically significant effect of both trait rumination F(1, 58) = 3.54, p =.065, η p 2 =. 06 and state rumination F(1, 58) = 3.22, p =.078.
05 on TRI scores, but no effect of either trait rumination F(1, 58) = 4, p =.409, η p 2 =. 01 or state rumination F(1, 58) =, p =.102, η p 2 =. 05 on TUT scores. Lexical associationsAs the negative associations were not normally distributed, a Mann-Whitney test was performed in order to test if there was a significant effect of state rumination on participants’ generation of negative lexical associations. Descriptive statistics are shown in. When comparing the scores provided by the LIWC, the results of the test indicated that participants came up with marginally significant more negative words in the rumination ( Mdn = 9.31) than in the control condition ( Mdn = 5.26), U = 331.00, z = −1.94, p =.052, while the negativity scores from SentiWordNet were significantly higher in the rumination condition ( Mdn = 0.08) than in the control condition ( Mdn =.05), U = 303.50, z = −2.33, p =.020.
As there were no nonparametric alternatives which would also account for the effect of trait rumination on lexical associations, we also performed a two-way ANOVA with state and trait rumination as covariates. There was no effect of either state rumination F(1, 58) = 2.40, p =.127, η p 2 =.
04, or trait rumination on the LIWC scores F(1, 58) = 2.11, p =.152, η p 2 =. Moreover, there was no effect of either state rumination F(1, 58) = 1.72, p =.195, η p 2 =. 029, or of trait rumination on the SentiWordNet scores, F(1, 58) = 0.15, p =.700, η p 2. Pitch imitationIn order to test whether there was a significant effect of rumination on pitch accommodation, a mixed ANOVA was performed. The within-participant factors were high and low pitch, while the between-participant factor was state rumination, with trait rumination as a covariate.
Pitch imitation was operationalised as adaptation to the Embodied Conversational Agent’s pitch, so that high-pitched utterances would have induced higher pitch responses, while lower-pitched utterances would have induced lower pitch responses. There was a statistically significant interaction between state rumination and high and low pitch, F(1, 58) = 8.23, p =.006, η p 2 =. 12, as well as between trait rumination and high and low pitch, F(1, 58) = 5.22, p =.026, η p 2 =. In the control condition, mean pitch was lower following vocalizations by a low-pitch prime ( M = 5.57 semitones, SD = 4.55 semitones), and mean pitch was higher following vocalizations by a high pitch prime ( M = 6.38 semitones, SD = 4.78 semitones). In the rumination condition, this was not the case, as there were no significant differences in pitch after either a low prime ( M = 6.22 semitones, SD = 4.95 semitones) or a high prime ( M = 6.19 semitones, SD = 5.20 semitones). Follow-up comparisons indicate a significant effect ( d = 0.50) of the high and low pitch experimental manipulation in the control condition ( M = 0.81, SD = 1.61), t(30) = 2.79, p =.009, 95% CI 0.22, 1.40, indicating that participants in the control condition accommodated their pitch significantly more to the Embodied Conversational Agent’s pitch, while those in a ruminative state did not ( M = 0.03, SD = 1.15), t(29) = 0.14, p =.891, 95% CI -0.45, 0.40. Rumination and mind wandering as predictive of eye movementsSeparate regressions were conducted in order to verify the effect of state and trait rumination, as well as mind wandering on eye movements.
In this analysis, we included mind wandering as combination of the TUT and TRI scales a predictor, as our hypotheses for eye movements derived from studies investigating either rumination or mind wandering in general in relation to eye movements. The regression model in Step 1 predicted the number of blinks when MW and the condition were not included in the analysis.
The results of the regression indicated that trait rumination explained 7% of the variance in the number of blinks, R =.27, adjusted- R 2 =.05, Δ R 2 =.07, F(1, 57) = 4.32, p =.042. Regression coefficients of the predictors of the number of blinks are shown in. DiscussionBuilding on previous research , , we found both trait and state rumination to have a marginally significant effect on TRIs but not on TUTs, that is, on thoughts about the task which actually interfere with performance of the task itself. It is necessary to note that in this study, participants scored highly on trait rumination, leading to a bias in our sample. When using both state and trait rumination to predict TRIs, it may be that one cannot dissociate one from the other, especially not in such a highly ruminative group. Previous studies investigating the relationship between rumination and task-related MW seem to have either only used a trait or a state measure of rumination.
In our study, however, we included both measures so that we could assess the success of a validated ruminative state induction procedure and discovered how volatile trait rumination really can be. Importantly, it may be that keeping high trait ruminators in a ruminative state keeps them in that rigid form of thinking. Meanwhile, distracting high trait ruminators from their sticky thoughts enables them to momentarily detach from their rigid patterns of thought.
As rumination at time 2 was measured only at the end of the experiment, it is not that surprising that trait rumination scores reduced significantly only in the control group, but that there was no significant interaction between the induction conditions and rumination at times 1 and 2. Any differences found may be a result of a combination between the induction procedures and the word association game. In the control condition, participants were further distracted by the word association game after being asked to think about random thoughts such as “the structure of a long bridge” and “a row of shampoo bottles on display”. In the rumination condition, any increases caused by the rumination induction were likely counteracted by the distracting effects of the word association game. In order to find out if any differences in trait rumination were a result of the manipulation alone, it would have been necessary to measure trait rumination directly after the induction.As expected, participants in the ruminative state condition generated more negative associations than in the control condition. However, what remains unclear, is why only SWN yielded significant results.
This may be due to the fact that it rated words as having positive, objective, and negative scores, while LIWC only rated words for their positivity and negativity. As SWN rated words for these three categories, it seemed to be able to pry apart the objectivity from the positivity and negativity scores of words. In LIWC, on the other hand, a word’s objectivity was likely conflated with either its positivity or negativity scores. Regardless, the finding that ruminative state is generally associated with negative word associations is in line with previous studies that emphasised the effect of rumination not only on negative affect , , but also on negative cognition and demonstrates the value of analyzing language usage in the exploration of psychological processes. However, when controlling for trait rumination, we did not find any effects of state rumination on the polarity of lexical associations. Indeed, the design of our current study did not include measures for the possible effects of negative mood on negative cognitions.
Although experimental studies indicate that rumination increases negative mood relative to distraction, the effect of rumination alone is less clear. In future experiments, it would be interesting to pry apart the effect of rumination, negative mood, and negative cognitions in a more complex experimental design.The results of our experimental investigation suggest that rumination may play an important role in disrupting the establishment of connection between interaction partners.
Previous research concerning pitch and MW yielded mixed findings; Drummond and Litman (2010) found that minimum pitch is a powerful predictor of MW when reading texts aloud. Franklin, Mooneyham, Baird, and Schooler (2014) , on the other hand, found no evidence for differences in pitch or pitch variability during MW; however, they did find MW to be was associated with less variability in volume when reading aloud. Rumination, as a type of MW which tends to be fixed on a single theme or topic and is marked by a high degree of automatic constraints , attenuates the normal variation in speaker’s pitch in relation to a conversation partner’s pitch. This falls in line with research that has associated rumination with reduced variability in physiological cues, including heart rate variability, which suggests rumination is associated with higher levels of cognitive inflexibility , –. Our study is novel in that it demonstrates that rumination not only predicts reduced variability in physiological cues, it also attenuates pitch imitation which occurs during interactions.
Although we framed our experiment in a social, interactive setting, a word association game with an Embodied Conversational Agent only taps into particular aspects of social interactions, and does not take into account their full complexity. Despite this limitation, automatic imitation is one of the most-basic nonverbal components of successful human interactions, and plays an essential role in creating rapport, empathy, and social bonding.
In our study, both state and trait rumination were related to a lack of accommodation to the Embodied Conversational Agent’s high and low pitches. This is in line with a previous finding regarding the effect of self-focus –an important feature of rumination– on reduced gesture imitation. While pitch convergence indicates rapport and desire for social approval, pitch divergence may be interpreted as speakers’ desire to be seen as dissimilar and wish to increase the social distance between themselves and their communication partners.Our study demonstrated that trait rumination is related to an increased amount of blinks during an interactive game, which indicates a decoupling of attention from the external environment and a focus on internal thoughts. Smilek and colleagues (2010) found that participants blinked more whenever mind wandering than when focused on reading. Beyond previous findings relating mind wandering as a state to more blinks, our study suggests that rumination, as a stable trait, is also related to more blinks.
Although rumination (as both a state and trait) may serve to exacerbate task-related mind wandering, the two cannot be equated, and hence it would be valuable to pry apart actual ruminative MW from from task-related MW episodes during a task. Furthermore, contrary to our predictions that ruminative MW would be related to less saccades, we instead found mind wandering in general to be related to more saccades, while finding no effects for fixation count and fixation duration. It is important to note that the methodology of previous studies upon which we based our predictions differed considerably from ours. Previous studies investigated eye movements in relation to rumination during either an emotional dot-probe task or during abstract animations. With regards to mind wandering, Reichle and colleagues (2010) , Smilek and colleagues (2010) , Uzzaman and Joordens (2011) , and Faber and colleagues (2017) , used self-caught or probe-caught measures, to distinguish between periods of MW and focus during reading and only analysed eye movement features from a short period of time preceding each auditory probe. In our study, we used a retrospective measure of MW, and accordingly, measured average eye movement parameters during the entire word association task. Indeed, online thought probes are arguably a more reliable method for measuring mind wandering , reducing the probability of confabulation as a result of having to retrospectively assess the content of one’s thoughts over a long period of time.
Intermittent thought probes would more accurately pinpoint the moments of time during which participants were mind wandering; however, they would disrupt the natural and automatic flow as well as the covert nature of the word association game. ConclusionIn our study, we examined the relationship between rumination and mind wandering and their impact on social cognitive mechanisms that support successful interactions with others.
First, we found rumination to marginally predict task-related interferences, suggesting a possible directionality in the relation between rumination and task-related mind wandering. Mind wandering is a broad term, which encompasses a wide variety of self-generated thoughts; hence, rumination elicits a particular type of mind wandering—rigid, self-focused and repetitive—and overlaps with thoughts that are related to a particular task, but interfere with performance of the task itself.Second, we observed that the emotional valence of lexical associations generated by participants in a condition where a ruminative state was induced was more frequently negative when compared to a control group. Next to that, both trait rumination and a ruminative state led to a decrease in pitch imitation, a more or less automatic process that is used to signal rapport and group membership. It also resulted in an increased number of blinks during the interaction, suggesting that participants were not engaging with their interaction partner. Taken together, these results suggest that ruminative MW may lead to an increased social distance and have the potential of disrupting our social relations. Considering that a substantial portion of our lives is social and interactive, our data highlight the need for further studies of MW in interactive environments.
![]() Comments are closed.
|
AuthorWrite something about yourself. No need to be fancy, just an overview. Archives
March 2023
Categories |