Note. Colored lines indicate individual participants. Black lines indicate the predicted interaction effect based on the mixed effect model depicted in Table 5. The waitlist period occurred between sessions W1 and T1, while the heart rate variability biofeedback intervention period took place between T1 and T5. a) Orienting Score – (reaction time incongruent spatial cues - reaction time congruent spatial cues); b) Executive Score – (reaction time invalid flankers - reaction time valid flankers). *** – p<.001; ns – not significant.
4 Discussion
In this study, we investigated the impact of a 4-week photoplethysmography smartphone-based HRVB intervention on premenstrual and depressive symptoms using a waitlist-control design. Additionally, we assessed its effects on stress symptoms, vmHRV, and attentional control. Our findings revealed improvements in premenstrual, depressive, and stress symptoms following the intervention, with no significant changes observed during the waitlist period. However, we did not observe any effects on vmHRV, and the results for attentional control were mixed.
The positive effects on premenstrual, depressive, anxiety, and stress symptoms found in our study highlight the feasibility and effectiveness of a smartphone-based HRVB intervention without the need for external devices. Laborde, Allen, Borges, Iskra, et al. (2022) reported similar effects of a slow-paced breathing intervention, whether or not visual feedback on current HRV was provided. This suggests that slow-paced breathing is the primary driver of the intervention's effectiveness. However, learning the technique of abdominal slow-paced breathing, which allows for a relaxed resonance frequency activation, can be challenging. The drastically slowed-down breathing can be uncomfortable and may even lead to hyperventilation (Szulczewski & Rynkiewicz, 2018). Providing visual feedback on respiratory sinus arrhythmia can assist users in correctly employing the method at home without extensive training. Moreover, this biofeedback feature can enhance user engagement with the app, as it offers immediate visual results of the breathing intervention, providing a sense of immediate gratification. Essentially, this form of feedback incorporates a simplified version of gamification into the intervention, which can increase motivation and engagement (Johnson et al., 2016).
Our primary objective was to investigate whether HRVB could help alleviate PMS symptoms. We observed a significant improvement in the premenstrual phase following 2-4 weeks of HRVB practice compared to the preceding cycle in which no HRVB was practiced. The effect size we observed is approximately 0.3, with a slightly larger effect of 0.4 in the psychological symptom scale. In this context, the effect size is similar to the moderate effect size found in meta-analyses of depressive symptoms (Pizzoli et al., 2021). This finding supports the notion that PMS may result from an altered reactivity within the ALLO/GABA system (Hantsoo & Epperson, 2020), affecting the CAN as proposed by Thayer and Lane (2000). Consequently, interventions designed to target the interconnectivity and functional capacity of the CAN may hold promise in improving PMS symptoms.
However, further studies with expanded paradigms are needed to determine whether this effect is mediated by a reduction in stress throughout the menstrual cycle or by an increase in the capacity of the CAN, which may buffer modulatory fluctuations throughout the cycle. It also remains unclear whether there is a critical phase in the menstrual cycle when HRVB has a more significant impact on subsequent premenstrual symptoms. It is possible that practicing HRVB throughout the preceding follicular phase reduces chronic stress (Goessl et al., 2017) and, consequently, chronic ALLO exposure. This prolonged exposure has been suggested to contribute to atypical GABA receptor reactivity to ALLO fluctuations later in the cycle, which is assumed to cause PMS (Hantsoo & Epperson, 2020). Another possibility is that during the premenstrual phase, the acute stress-relieving effects (see e.g. Meier & Welch, 2016; Prinsloo et al., 2013) of HRVB directly alleviate the symptom burden.
Independently of mechanisms of action, HRVB is an easily learned intervention with minimal to no side effects, offering PMS sufferers an opportunity to enhance their symptom management self-efficacy. Furthermore, it can lower the threshold for receiving treatment. When delivered through a smartphone application, it becomes even more accessible and, if its effectiveness is established, can be seamlessly integrated into widely used menstrual cycle-related health apps.
Our results also demonstrated improvements in depressive symptoms, as well as anxiety and stress symptoms, through the smartphone-based HRVB intervention. This aligns with prior research using other HRVB intervention methods (Goessl et al., 2017; Pizzoli et al., 2021), although the effects we observed are somewhat smaller than the meta-analytic effects on depression and anxiety/stress. This discrepancy may be due to ceiling effects. The majority of studies investigating HRVB effects typically focus on clinical populations with very high values in the respective outcome measures. Our inclusion criteria, on the other hand, involved individuals with above-average PMS symptoms or slightly elevated depressive symptoms. 30% of the participants were included in our study due to only PMS symptoms. While clinically significant affective premenstrual symptoms (premenstrual dysphoric disorder) are often comorbid with anxiety and depression (Yonkers & McCunn, 2007), our subclinical sample likely had lower baseline scores in both BDI-II and DASS values compared to the samples in most studies included in the meta-analysis. This could account for the slightly smaller effect sizes observed in our study.
We did not observe improvements in vmHRV (RMSSD) following the HRVB intervention. Contrary to our findings, Laborde, Allen, Borges, Dosseville, et al. (2022) reported consistent elevations in vmHRV parameters after HRVB and slow-paced breathing interventions in their meta-analysis. The effect size for this improvement was approximately Hedges' g = 0.3. Notably, this effect size is smaller than the meta-analytic effect sizes for depressive symptoms (g = 0.4; Pizzoli et al., 2021) or anxiety/stress (g = 0.8; Goessl et al., 2017). These findings suggest that improvements in vmHRV alone do not fully account for the observed affective and cognitive effects. This is consistent with other findings that demonstrate physiological and clinical outcomes do not always change simultaneously in HRVB (Wheat & Larkin, 2010). Critically, our study may not have detected this effect due to its relatively small sample size. A G*Power analysis indicates that a sample size of more than 70 would be required to replicate the effect of 0.3 with a power of .8. Our sample size was significantly below this threshold. The reason why effects were found in other outcome measures lies in the advantage of mixed model analysis, which allows us to include each item/trial individually without the need to condense the information into a composite score. This significantly increases the number of data points included, by 20 to 144 times, depending on the measurement. However, in the case of vmHRV, it is necessary compute a single value per measurement, which drastically reduces statistical power. This suggests the possibility that HRVB may have had a beneficial impact on vmHRV, but our study may have been underpowered to detect this effect.
The influence of the HRVB on the attentional control domains yielded mixed results. We chose to employ the Executive and Orienting Scores of the ANT-R as outcome measures, given their documented correlations with vmHRV (Blaser et al., 2023a; Quintana et al., 2017; Ramírez et al., 2015; Sørensen et al., 2019). Interestingly, despite previous studies (Quintana et al., 2017; Sørensen et al., 2019) indicating a stronger association between the Orienting Score and vmHRV (compared to the Executive Score vmHRV association), we did not observe any improvements in this measure. However, we did find enhancements in the Executive Score following the intervention period. These varying outcomes align with the findings of a meta-analysis conducted by Tinello et al. (2022), who reported positive effects of HRVB on Executive Functions in approximately half of the studies they reviewed, with a slightly higher likelihood of effects in the attention domain. This highlights the need for further research to elucidate the specific impact of HRVB on cognitive outcomes.
Although we found significant beneficial effects of smartphone-based HRVB on several mental health and some cognitive outcomes, our study has several limitations that need to be considered. Firstly, the most notable limitation is our small sample size. Although linear mixed model analyses help to overcome the limited statistical power due to small samples by including individual items/trials in the analysis, it is important that the interpretation of the results and their applicability to larger populations must be considered with caution. Another limitation is related to the passive control group, which has methodological weaknesses. In future studies, it would be beneficial to implement active control groups to better distinguish the intervention’s effects from any potential placebo effects. Additionally, we did not conduct clinical assessments of PMS or depression. However, it is essential to note that our study was not aimed at identifying treatments for a clinically relevant premenstrual dysphoric disorder. Instead, our primary focus was on investigating a user-friendly intervention to help individuals self-manage premenstrual symptoms, regardless of their severity levels. Furthermore, our study included participants regardless of their use of hormonal contraceptives and their current menstrual cycle phase. For future studies, it is advisable to standardize these criteria to ensure a more consistent assessment.
5 Conclusion
In summary, smartphone-based HRVB has proven to be effective in enhancing both emotional and cognitive well-being, without the need for external devices. This intervention holds promise as a novel approach for self-managing premenstrual symptoms and provides a more accessible solution for harnessing the known benefits of HRVB for depressive, stress, and anxiety symptoms, as well as certain aspects of attention.
Acknowledgements
The project was partly funded by the Funding of Knowledge and Technology Transfer (FöWiTec) of the University of Potsdam. The first author was partly supported by a PhD scholarship from the University of Potsdam (Potsdam Graduate School) with fundings from the Graduate Fund of the State of Brandenburg (Germany).
During the preparation of this work, the first author utilized ChatGPT 3.5 to enhance the text's readability. Prompts used were aimed at tasks such as “check grammar and spelling in this paragraph.” After employing this tool/service, the author reviewed and edited the content as necessary and assumes full responsibility for the publication's content.
We would like to thank Dr. Michael Vesker and Kenkou GmbH for providing the biofeedback App as well as Kenkou GmbH for partly providing the participant compensation.
6 References