Berenike Lisa Blaser1*, Mathias Weymar1,2 & Julia Wendt11 University of Potsdam, Department of Biological Psychology and Affective Science, Faculty of Human Sciences, Karl-Liebknecht Strasse 24/25, 14476 Potsdam, Germany2 Faculty of Health Sciences Brandenburg, University of Potsdam, Potsdam, Germany * Correspondence:Berenike Lisa BlaserUniversity of PotsdamKarl-Liebknecht Strasse 24/2514476 Potsdam, Germany Email: berenike.blaser@uni-potsdam.de Running head: Heart Rate Variability and premenstrual symptoms Keywords: heart rate variability, vagal tone, premenstrual syndrome, premenstrual symptoms, menstrual cycle, ambulatory assessmentAbstractIntroductionPremenstrual syndrome (PMS) affects up to 90% of individuals with an active menstrual cycle. Several studies have observed reduced vagally mediated heart rate variability in a single assessment during the luteal phase compared to an assessment during the follicular phase, especially in participants experiencing strong PMS symptoms. The aim of this investigation was to initially assess the relationship between premenstrual symptoms and vagally mediated heart rate variability throughout the menstrual cycle, as well as to examine the feasibility of conducting a large-scale study to verify this association.MethodsThree participants completed daily ambulatory assessments of resting vagally mediated heart rate variability using mobile electrocardiographs and typical PMS symptoms. We calculated correlations between these measurements for each participant.ResultsPMS symptoms and vagally mediated heart rate variability showed medium to high correlations in each of the participants throughout the cycle. These associations were primarily driven by the relationship between vagally mediated heart rate variability and psychological symptoms rather than physiological symptoms. Visual inspection of the fluctuations confirmed the concurrent occurrence of a phasic reduction in vagally mediated heart rate variability parallel to the increases in PMS symptoms experienced during the mid to late luteal phase in each participant.DiscussionThe results support the notion of an association between PMS symptoms and vagally mediated heart rate variability. An ambulatory daily assessment paradigm proves to be feasible. Studies with larger samples are necessary to provide deeper insights into inter- and intra-individual differences as well as stronger knowledge on the mechanistic pathways of PMS. 1 IntroductionPremenstrual syndrome (PMS) encompasses a heterogeneous collection of symptoms that typically manifest in the week preceding menstruation, during the luteal phase of the menstrual cycle, and dissipate within a few days after menstruation begins. These symptoms can be of a physiological nature, such as bloating and water retention, or psychological, including feelings of stress, anxiety, or irritability. It is noteworthy that as many as 90% of menstruating individuals regularly experience at least one symptom associated with PMS (Tschudin et al., 2010).The etiology of PMS remains unclear. A common hypothesis is that varying sensitivities to fluctuations in gonadal hormones throughout the menstrual cycle play a role (Rapkin & Akopians, 2012). This differential sensitivity may involve several systems, including the Gamma-Aminobutyric Acid (GABA) and serotonin systems (Nappi et al., 2022; Rapkin & Akopians, 2012). Vagally mediated heart rate variability (vmHRV) serves as a potential physiological marker that could contribute to our understanding of PMS. VmHRV is regarded as a marker for cardiac vagal control (Laborde et al., 2023), and research has linked it to a wide range of psychopathological states (Heiss et al., 2021) and cognitive outcomes (Holzman & Bridgett, 2017; Zahn et al., 2016). The associations are so consistent, that Beauchaine and Thayer (2015) have proposed vmHRV as a transdiagnostic biomarker for psychopathology.In a meta-analysis conducted by Schmalenberger et al. (2019), consistent reductions in vmHRV of medium effect size were identified during the luteal phase when compared to measurements during the follicular phase. A limited number of studies, however, have explored the relationship between this vmHRV reduction and PMS. The fluctuations in vmHRV are found to be moderated by the extend of premenstrual symptomatology (Schmalenberger et al., 2023) . The observed effect indicates that high PMS groups experience more substantial reductions in vmHRV during the luteal phase, whereas control groups show either smaller fluctuations (Zambotti et al., 2013) or no discernible difference (Baker et al., 2008; Matsumoto et al., 2007) in vmHRV between the follicular and luteal phases.The consistent direction of the association between vmHRV and PMS is noteworthy, but it is important to acknowledge that all previous studies on this topic have involved only a single measurement during each cycle phase. Consequently, it is challenging to ascertain whether symptoms and vmHRV fluctuate in parallel throughout the menstrual cycle. In an effort to shed more light on the relationship between vmHRV and PMS and to assess feasibility, we therefore initiated a pilot diary study involving three participants. This study aimed to gather daily assessments of premenstrual symptoms alongside measurements of resting vmHRV to follow the course of their association.2 Methods and Materials2.1 ParticipantsWe tested three participants of different ages (ageparticipant 1 = 44, ageparticipant 2 = 27, ageparticipant 3 = 20) who were recruited within our department. These participants provided informed consent and received either course credit or no compensation for their participation. In line with the guidelines suggested by Laborde et al. (2017), the participants did not take medication that could affect vmHRV, had no chronic diseases, and were not pregnant.2.2 Testing protocolEach participant received an introduction on how to use the mobile electrocardiography (ECG) device and a document with written instructions outlining the procedure. Measurements were taken each day at the same time, between 7 and 8 pm. The participants began by completing an online questionnaire assessing premenstrual symptoms, recording their last menstrual period, and responding to a number of control variables. Following the questionnaire, the ECG measurement was conducted. Participants attached the ECG device to their chest, set a timer for 6 minutes, initiated the ECG recording, and closed their eyes while the resting vmHRV measurement was taken. This measurement was performed with participants in a sitting position.Participants were requested to complete the assessments daily over 1.5 menstrual cycles to ensure the inclusion of one complete cycle. We employed the backward- and forward-counting method to assess the cycle phase (Schmalenberger et al., 2021). We included two weeks before a reported menstruation onset (luteal phase) and two to three weeks (depending on reported average cycle length) after a reported menstruation onset (follicular phase) for the analysis.2.3 Heart rate variabilityVagally mediated heart rate variability was assessed using a mobile 1-channel ECG device, the Bittium Faros 180. The device electrodes were attached to the chest, and data were collected at a sampling rate of 1000 Hz. Data preprocessing was performed using the Faros Software, which generated R-R interval and R peak timestamp series. The first and last 30 seconds of each measurement were removed, resulting in a 5-minute interval, to avoid artifacts caused by participant movement, as participants initiated and concluded the measurement themselves.The root mean square of successive differences (RMSSD) was derived from the time series as measurement of vmHRV. The time series were analyzed in R (version 4.2.2) using the RHRV package (https://rhrv.r-forge.r-project.org/), following the package documentation. We chose this measurement over the high-frequency component of power spectral analysis due to its robustness to breathing rate and its clearer indication of parasympathetic activity (Chapleau & Sabharwal, 2011).2.4 Premenstrual symptomsPremenstrual symptoms were assessed with the German version of the shortened Premenstrual Assessment Form (PAF20) (Allen et al., 1991; Blaser et al., 2023). The questionnaire comprises the 20 most endorsed items from the long form PAF, which includes nearly 100 items in total. Each item represents a specific symptom, and participants are asked to rate how strongly they experienced each symptom during the last premenstrual phase using a 6-point Likert scale, ranging from “not at all/no change from normal” to “extreme change from normal”. The German version of this questionnaire has demonstrated good validity and reliability and loads onto two distinct factors, creating psychological and physiological symptom scales. For this study, we adapted the questionnaire to a diary format, where participants reported how strongly they experienced each symptom in the previous 24 hours.2.5 Control variablesThe daily online questionnaire included several control variables that are known to influence vmHRV or PMS. Participants were asked to provide retrospective assessments of these variables for the last 24 hours. The control variables encompassed substance intake (alcohol, caffeine, nicotine), a one-item rating of the level of stress experienced that day on a Likert scale ranging from 1 to 9, a one-item rating of sleep quality on a 1-9 Likert scale, and reports of any physical health symptoms related to acute diseases, such as respiratory symptoms. 2.6 AnalysisAll statistical analyses were conducted with R (version 4.2.2). To assess the association between premenstrual symptoms and RMSSD over the menstrual cycle, we conducted Pearson correlations between the two measurements for each participant individually. Furthermore, separate correlations were calculated for the physiological and the psychological subscale of the PAF20 with the RMSSD.To test the association between PAF20 and RMSSD for all three participants, independently of the control variables, we conducted a linear mixed model predicting PAF20 sum scores. Participant intercepts were modeled as random effects to account for nesting of the data. The RMSSD and control variables were introduced as fixed effects.3 ResultsThe RMSSD values were subjected to a log transformation to approximate a normal distribution, aligning with the conventions of other vmHRV research (Laborde et al., 2017). A visual representation of the symptom course and RMSSD for each of the three participants is presented in Figure 1. Pearson correlations between log-transformed RMSSD and daily symptom scores were moderate to high, rparticipant1(25) = -.41, p < .05, rparticipant 2(35) = -.48, p < .01, rparticipant 3(29) = -.43, p < .05. The correlations were consistently higher in the psychological subscale than the physiological subscale (see Table 1). The associations between RMSSD and physiological symptoms were not significant in all three participants.Table 1. Pearson correlations of vagally mediated heart rate variability and premenstrual symptom scores Psychological symptoms Physiological symptoms r df p r df p Participant 1 -.50 25 .008 -.25 25 .21 log(RMSSD) Participant 2 -.57 35 <.001 -.07 35 .68 Participant 3 -.46 29 .010 -.31 29 .093 Notes. The symptom scores are the sum scores of the psychological and physiological subscale of the daily ratings of the short form of the premenstrual assessment form (PAF20). RMSSD – root mean square of successive differences; df – degrees of freedom.