https://orcid.org/0000-0001-6512-0167
Abstract
Binaural beat (BB) has been investigated as a potential modality to enhance sleep quality. In this study, we introduce a new form of BB, referred to as dynamic BB (DBB), which incorporates dynamically changing carrier frequency differences between the left and right ears. Specifically, the carrier frequency of the right ear varied between 100 and 103 Hz over a period, while the left ear remained fixed at 100 Hz, yielding a frequency difference range of 0 to 3 Hz. The objective of this study was to examine the effect of DBB on sleep quality. Ten healthy participants were included in a cross-over design, where they experienced both DBB and a SHAM (absence of sound) condition across two consecutive nights, with polysomnography evaluation. DBB was administrated during pre-sleep initiation, sleep onset, and transition from rapid eye movement (REM) to non-REM stage. DBB significantly reduced sleep latency compared to the SHAM condition. Electrocardiogram analysis revealed that exposure to DBB led to diminished heart rate variability during the pre-sleep initiation and sleep onset periods, accompanied by a decrease in low-frequency power of heart rate during the sleep onset period. DBB might be effective in improving sleep quality, suggesting its possible application in insomnia treatments.
This study introduces dynamic binaural beats (DBB) as a novel method for improving sleep quality. The effectiveness of DBB is validated by questionnaire results and biosignals analysis, demonstrating significant improvement in overall sleep quality, especially in reducing sleep onset time. This presents a viable, non-pharmacological alternative in the treatment of insomnia.
Sleep disorder is highly prevalent and has a substantial implication on general health. Sleep deficiency caused by sleep disorder can lead to impaired cognitive function, decreased immunity, and, in chronic conditions, increased risk of obesity, diabetes, cardiovascular diseases, and psychiatric disorders [1–5]. There has been a rapid increase in the number of individuals affected by sleep disorders and the demand for their treatment.
Insomnia is a prevalent sleep disorder, and its key pathomechanistic mechanism involves hyperarousal, which increases anxiety, stress, and alertness during the pre-sleep initiation period and interferes with the ability to initiate or maintain sleep. Hyperarousal is also associated with disrupted autonomic nervous system (ANS) regulation during the process of sleep initiation. Therefore, various investigations have focused on improving the balance of the ANS during the pre-sleep initiation period by decreasing the activity of the sympathetic nervous system and increasing the activity of the parasympathetic nervous system. These interventions aim to facilitate the onset of sleep and ameliorate symptoms of insomnia [6–8].
Previous studies have attempted to address this hyperarousal and ANS dysregulation by introducing neuromodulation methods. For example, electrical stimulation has been shown to increase total sleep time and sleep efficiency when applied before bedtime, and to improve memory functions when applied during sleep [9–13]. Auditory stimuli, such as white noise, autonomous sensory meridian response (ASMR), and binaural beats (BB) have also been used to relieve sleep disorders and improve sleep quality [6, 14–29]. For example, listening to white noise at bedtime has been shown to significantly decrease time to fall asleep, reduce wakefulness after sleep onset, and improve overall sleep quality [14–22]. Similarly, ASMR has been shown to improve sleep quality by promoting faster sleep onset and quicker transitions to deep sleep stages [23–25].
BB are sounds subjectively perceived by listeners when each ear is presented with a small frequency difference, meaning that the brain perceives a pulsation with a frequency equal to the frequency difference between the two sounds presented to each ear [30]. Interestingly, the difference between two carrier frequencies perceived by the brain consists of a single note with a single frequency. Numerous studies have demonstrated the positive impact of BB on improving sleep quality [6, 26–29], e.g. shortening the transition from light to deep sleep, where BB was presented before bedtime [6, 27–29] or during sleep [26]. Although the physiological mechanism by which BB enhances sleep is not fully elucidated, hypotheses such as the entrainment of brain rhythm to the beat frequency, modulation of cortical gamma-range activity through synaptic plasticity, and regulation of ANS activity have been suggested to explain the possible mechanism of effect [30-36]. However, limited research has been performed to examine the effects of BB during the pre-sleep initiation period, specifically its potential to shorten sleep onset latency, which directly affects sleep quality.
Traditionally, BB sounds have been created with a constant frequency difference between the left and right ears over time. However, in this study, we introduce a new type of BB called dynamic BB (DBB), where the frequency difference between the ears changes over time. Given that the effect of BB in the cortex varies according to the beat frequency [35], we hypothesized that DBB might offer advantages in inducing sleep by covering the entire range of the delta frequency band. The objective of this study was to examine the impact of DBB on sleep by administering DBB during three specific periods that have rarely been investigated so far: (1) pre-sleep initiation, (2) sleep onset, and (3) transition from rapid eye movement (REM) to non-REM stage during sleep (non-REM transition). We evaluated the effects of DBB on sleep quality using polysomnography (PSG) parameters, changes in brain activity patterns measured using electroencephalography (EEG), and changes in ANS function regulation measured using heart rate variability (HRV). The current study was carried out with individuals who have normal sleep patterns as an initial proof of concept study.
Materials and Methods
Participants
We recruited a total of 10 healthy men and women (male: 4, female: 6, mean age: 22.0 ± 1.9 years, range 19‒24 years). At baseline, participants’ medical histories were reviewed and a battery of sleep questionnaire, including Insomnia Severity Index for insomnia, Berline Questionnaire for the risk of obstructive sleep apnea, Cambridge-Hopkins questionnaire for restless leg syndrome, Morningness–Eveningness Questionnaire for circadian rhythm sleep–wake disorders, Epworth Sleepiness Scale for daytime sleepiness, Pittsburgh Sleep Quality Index for sleep quality, Beck Depression Inventory for depressive symptoms, and the Korean version of Hearing Handicap Inventory for the Elderly for a hearing disorder were performed. We excluded participants who (1) are screened as positive for sleep disorders such as insomnia, obstructive sleep apnea, restless leg syndrome, or circadian rhythm sleep–wake disorder from the sleep questionnaire; (2) are being treated for major mental illnesses such as mood disorders; (3) have been diagnosed with a fixed neurological disorder; (4) are taking medications for insomnia, or other dietary supplements that may affect sleep structure or quality within 4 weeks from the time of inclusion; (5) addicted to alcohol or other drugs; or (6) have contraindications for the application of auditory stimulation such as fixed hearing disorder. The study procedures were approved by the Institutional Review Board of the Seoul National University Bundang Hospital (B-2205-754-303). Written informed consent was obtained from all study participants.
Dynamic BB
In this study, we developed a new type of BB called dynamic BB (DBB), in which the difference in carrier frequencies between the ears varies over time. The carrier frequency of the left ear was fixed at 100 Hz, while the carrier frequency of the right ear continuously fluctuated between 100 and 103 Hz over a duration of one minute. Consequently, the frequency difference between the ears ranged from 0 to 3 Hz, falling within the delta frequency band of EEG (Figure 1). By using the delta frequency band for DBB, we expected the neuromodulation impact of DBB on delta power increase because delta frequency power significantly increases during deep sleep (non-REM 2 and non-REM 3 stage) [31]. DBB was generated using MATLAB R2020b and was played through earphones (ataw 008; ATECH INVENTION, Inc., USA) during this experiment. The volume of DBB was adjusted to a comfortable level for the participants to listen to while sleeping on the first day of the experiment (50.50 ± 5.93 dB) and it was used on the second and third days of the experiment. The specific volume levels for each participant can be found in Supplementary Table S1.
Scheme of dynamic binaural beat. BDBB: Dynamic Binaural Beats.
Polysomnography
Overnight video polysomnography (PSG) was performed in specialized facilities at a tertiary hospital (Seoul National University Bundang Hospital), specifically designed for PSG execution. According to the latest version of the ASSM accreditation standards, the PSG bedroom was single occupancy, private and comfortable, and shielded from noise with triple-glazed windows, double doors, and solid floor-to-ceiling walls. The PSG was performed using a 64-channel Embla RemLogic PSG system (NATUS, Inc., USA). The recording system included six-channel EEG (F3-A2, F4-A1, C3-A2, C4-A1, O1-A2, and O2-A1), bilateral electrooculography (EOG), electromyography (EMG) leads for the submentalis, and electrocardiography (ECG). In order to evaluate respiratory events, a nasal thermistor, nasal airflow pressure transducer, thoracic and abdominal strain gauges, position sensor, and finger pulse oximetry were applied. Synchronized audio and video recordings were also performed during the entire PSG procedure. Sleep stages during 30-second epochs, arousals, and respiratory events (apnea, hypopnea, and respiratory effort-related arousal), were scored according to the American Academy of Sleep Medicine manual [21].
Experiment procedure
All experiments were conducted in the sleep laboratory at Seoul National University Bundang Hospital. Each participant was tested for three consecutive nights: Friday, Saturday, and Sunday. On the first night, the participants were adapted to sleeping in a hospital laboratory environment, which was different from their usual sleep environment (adaptation night). On the second and third nights, the participants were randomly assigned to either sleep with DBB (DBB condition) or without DBB (SHAM condition) in the same environment as the first night. For each night, the participants completed a questionnaire (rated on a scale of 1–9) about five types of their physical condition before and after sleep, including drowsiness, fatigue, anxiety, depression, and stress. In the DBB condition, DBB was played during three specific periods (Figure 2). The first period was from the time the participant was lying comfortably in bed until they fell asleep (pre-sleep initiation). The second period was after the participant had fallen asleep, as determined by a sleep technician using PSG results (non-REM stage identified three consecutive times), where DBB was played for 20 minutes after a 3-minute stabilization delay (sleep onset). The third period occurred after the participant transitioned from the REM stage to the non-REM stage (three consecutive confirmations of the non-REM stage), where DBB was played for 20 minutes after a 3-minute stabilization delay (non-REM transition). The occurrence number of the third period varied between the participants and it was 3.80 ± 0.79 in DBB condition and 3.20 ± 0.63 in SHAM condition. In the SHAM condition, the timing of the DBB presentation was also recorded, but no actual DBB stimulus was presented to the participants. Participants wore same earphones throughout the three consecutive PSG nights to mitigate any potential impact of wearing earphones on sleep quality. They were blinded to the condition—DBB or SHAM—assigned on the second night. Nonetheless, it is possible that they might have recognized the DBB condition by detecting the sound during the evaluation. Seven participants acknowledged hearing a DBB sound at sleep onset, but all participants reported that the sleep environment during the examination was overall comfortable, with no significant discomfort stemming from the DBB.
Presentation periods of DBB.
Questionnaire analysis
To examine the effect of DBB on subjective physical condition, the questionnaire results from the experiment were compared between the DBB and SHAM conditions by calculating the mean and variance across all participants. This was accomplished by subtracting the pre-sleep quantitative value from the post-sleep quantitative value for each of the five categories, i.e. drowsiness, fatigue, anxiety, depression, and stress.
Sleep macrostructure analysis
Sleep macrostructure indices, which included total sleep time, sleep latency, sleep efficiency, WASO/SPT (ratio of wakefulness after sleep onset), and the ratio of each sleep stage to the total sleep time, were compared between the DBB and SHAM conditions. We also compared the proportion of each sleep stage between the DBB and SHAM conditions for three stimulus presentations, i.e. pre-sleep initiation, sleep onset, and non-REM transition.
EEG analysis
EEG data was first bandpass-filtered between 0.1 and 55 Hz and independent component analysis was applied to remove the noise introduced into the EEG. Following preprocessing, the EEG data were epoched into a 2-second window with 50% overlap, and if the EEG amplitude of each 2-second epoch was larger than ± 70 μV, the segment was regarded as a noise epoch and was excluded from analysis. A fast Fourier transform was applied to each epoch to calculate power spectral density (PSD) for five frequency bands: delta (0.1–3 Hz), theta (4–7 Hz), alpha (8–13 Hz), beta (14–30 Hz), and gamma (31–55 Hz). The PSD analysis was performed and averaged separately for the three periods of pre-sleep initiation, sleep onset, and non-REM transition.
ECG analysis
HRV was extracted from ECG data, and the low-frequency (LF: 0.04–0.12 Hz) and high-frequency (HF: 0.15–0.35 Hz) bands of the ECG were also calculated using FFT. In order to account for the relatively slower response of the ECG data compared to the EEG, changes in HRV and frequency powers were assessed for a duration of 20 minutes before listening to the music (pre), during the music listening period (during) and 20 minutes after listening (post) for the two periods of sleep onset and non-REM transition. However, we calculated and compared both HRV and frequency power only during the listening period (during) between the DBB and SHAM conditions for the pre-sleep initiation period because there is no time period preceding the pre-sleep initiation period (pre) and the post-period following the pre-sleep initiation period (post) overlapped with the pre-period of sleep onset.
Statistical analysis
The data are presented as means ± standard deviation, median (interquartile range [IQR]), or number (percentage). Statistical analyses were performed using MATLAB (ver. R2020b, MathWorks, MA, USA), where non-parametric statistical methods were used because the data did not follow normal distribution. The Wilcoxon signed-rank test was used to compare the questionnaire survey, sleep macrostructure, EEG, and HRV data between the DBB and SHAM conditions. The Kruskal–Wallis test with post hoc Tukey’s honest significant difference was used to compare the three conditions (pre, during, and post) within each DBB or SHAM condition. A p-value of < .05 was considered statistically significant for all analyses.
Data availability
The datasets generated for the current study are available from the corresponding author on request.
Results
Questionnaire results
Both drowsiness and fatigue tended to decrease after sleep in the DBB condition compared to the SHAM condition, implying that the DBB condition considerably reduced the degree of drowsiness and fatigue. However, only drowsiness showed a statistically significant decrease (Figure 3; DBB: −2.1 ± 2.1 vs. SHAM: −0.3 ± 1.7, p = .004). Supplementary Table S2 provides all questionnaire results for five categories, i.e. drowsiness, fatigue, anxiety, depression, and stress.
Mean changes in “drowsiness” and “fatigue” across all participants for the DBB and SHAM conditions. An asterisk indicates statistical significance (p < .05).
Sleep macrostructure results
Figure 4A shows mean sleep latency (the amount of time spent sleeping) and it was significantly shorter for the DBB condition than for the SHAM condition (DBB: 6.1 ± 4.3 minutes vs. SHAM: 12.5 ± 7.9 minutes, p = .027). Figure 4B shows the mean values of three sleep indices, i.e. sleep efficiency, WASO/SPT, and sleep stages (N1, N2, N3, and REM). Even though no statistical difference was found for all sleep indices, most sleep index results also indicated the positive impact of DBB on sleep quality; (1) the sleep efficiency of the DBB condition was marginally higher than that of the SHAM condition (DBB: 89.6% ± 3.4% vs. SHAM: 85.8% ± 5.2%, p = .064), (2) the occurrence of wakefulness after sleep onset was reduced for the DBB condition than for the SHAM condition (DBB: 9.1% ± 4.0% vs. SHAM: 12.3% ± 5.5%, p = .10), and (3) the total sleep time of the DBB condition was higher than that of the SHAM condition by about 18 minutes (DBB: 438.1 ± 17.6 vs. SHAM: 420.1 ± 29.0, p = .13; Supplementary Table S3). The quantitative values of all sleep indices can be found in Supplementary Table S3.
(A) Mean sleep latency time across all participants for the DBB and SHAM condition. (B) Mean values of the three sleep indices: (1) sleep efficiency, (2) WASO (Wakefulness After Sleep Onset)/SPT (Sleep Period Time), and (3) sleep stages (N1, N2, N3, and REM) for the DBB and SHAM condition. An asterisk and triangle indicate statistical significance (p < .05) and marginal significance, respectively.
Figure 5 presents a comparison of the mean proportion of each sleep stage between the DBB and SHAM conditions for the three periods of pre-sleep initiation, sleep onset, and non-REM transition, respectively. During the pre-sleep initiation period, the proportion of the wake stage was significantly lower in the DBB condition compared to the SHAM condition (DBB: 53.1% ± 20.4% vs. SHAM: 70.2% ± 15.4%, p = .006) while the proportion of the N1 stage was significantly higher (DBB: 40.6% ± 20.2% vs. SHAM: 23.0% ± 12.8%, p = .013). These results align with the sleep latency results presented in Figure 4. No statistical significance was observed for any of the other sleep stages, except for the pre-sleep initiation period.
Mean proportion of each sleep stage for the DBB and SHAM conditions in the three periods of (A) pre-sleep initiation (B) sleep onset (C) non-REM transition. An asterisk indicates statistical significance (p < .05).
EEG results
Figure 6 shows the average PSD proportion of each frequency band for the DBB and SHAM conditions in pre-sleep initiation, sleep onset, and non-REM transition. The delta PSD proportion in the DBB condition was slightly higher than that of the SAHM condition for all three periods, but no statistical significance was found.
Mean PSD proportion of each frequency band for the DBB and SHAM conditions in the three periods of (A) pre-sleep initiation (B) sleep onset (C) non-REM transition.
ECG results
Figure 7 shows the mean HRV for the DBB and SHAM conditions across the three periods: (1) pre-sleep initiation, (2) sleep onset, and (3) non-REM transition, respectively. During the pre-sleep initiation, the HRV was significantly lower in the DBB condition compared to the SHAM condition (DBB: 19.6 ± 9.0 vs. SHAM: 27.1 ± 13.2, p = .037; Figure 7A). In addition, the HRV significantly decreased in the DBB condition only (Pre: 19.6 ± 9.0 vs. During: 13.8 ± 9.0, p = .026) during listening to DBB (During) compared to before listening DBB (Pre), and the HRV in the DBB condition after listening to DBB was significantly lower than the SHAM condition (DBB: 14.6 ± 9.6 vs. SHAM: 19.0 ± 13.1, p = .049; Figure 7B). However, no statistical significance was observed for any of the conditions during the non-REM transition, as depicted in Figure 7C. Note that the R-R intervals of the two conditions were not statistically different (Supplementary Figure S1). A diminished HRV in the DBB condition indicates heart rate stabilization, which subsequently contributes to reduced sleep latency, as depicted in Figure 4A [32].
Mean HRV for the DBB and SHAM conditions in the (A) pre-sleep initiation (B) sleep onset (C) non-REM transition. “Pre,” “During,” and “Post” mean the time periods before, during, and after the presentation of DBB for 20 minutes, respectively. An asterisk indicates statistical significance (p < .05).
A significant decrease in LF PSD was observed during sleep (during) in the sleep onset phase compared to before sleep (pre; pre: 0.04 ± 0.02 vs. during: 0.01 ± 0.01, p = .002; Figure 8B). This reduction in LF PSD implies a modulation in the ANS, characterized by inhibition of the sympathetic nervous system and enhancement of the parasympathetic nervous system, which subsequently exerts a beneficial influence on sleep [33]. When examining the mean LF PSD for the DBB and SHAM conditions in the pre-sleep initiation and non-REM transition stages, no statistically significant differences were observed (Figure 8, A and C). Similarly, the HF PSD remained statistically invariant across all three intervals.
Mean LF PSD for the DBB and SHAM conditions in the (A) pre-sleep initiation (B) sleep onset (C) non-REM transition. “Pre,” “During,” and “Post” mean the time periods before, during, and after the presentation of DBB for 20 minutes, respectively. An asterisk indicates statistical significance (p < .05).
Discussion
In this study, DBB was presented at pre-sleep initiation, after sleep onset, and non-REM transition to investigate the effect of DBB on overall sleep quality. We found that DBB can significantly reduce sleep latency (DBB: 6.1 ± 4.3 minutes vs. SHAM: 12.5 ± 7.9 minutes) as well as marginally increase sleep efficiency compared to the SHAM condition. DBB was also associated with decreased proportion of wake stage and increased proportion of N1 stage during the pre-sleep initiation period, as well as a significant decrease in drowsiness after sleep. Moreover, HRV was significantly lower for the DBB condition than the SHAM condition in both pre-sleep initiation and sleep onset periods, and LF power did in sleep onset period. Our results are consistent with several previous studies, most notably Roya Dabiri et al.’s study in which listening to a conventional BB before going to bed calmed participants and resulted in a significantly shorter sleep latency [27]. However, to the best of our knowledge, this is the first study to introduce DBB and demonstrate its positive impact on overall sleep quality based on overnight PSG and HRV analyses. Given that there was no significant discomfort associated with DBB in the study population, DBB might have potential as a safe, simple, and effective strategy for improving sleep quality and treating insomnia with sleep onset difficulties.
The effect of BB on ANS modulation has been reported in previous studies. BB provided at post-exercise status increased parasympathetic activation and sympathetic withdrawal [30], and BB also elevated the high-frequency power of HRV, indicating parasympathetic activity, during daytime sleep [34], as well as other HRV parameters for parasympathetic activity at baseline when provided long-term [35]. The potential effect of BB on ANS modulation, especially parasympathetic activation and sympathetic withdrawal, is consistent with our observation that DBB reduced HRV during both the pre-sleep initiation and sleep onset periods and decreased LF power during the sleep onset compared to before sleep onset in the DBB condition. Decreased HRV under the DBB condition can be interpreted as HR stabilization due to the relaxation effect induced by DBB. LF power, indicative of sympathetic activities or the modulation of cardiac autonomic outflows by baroreflexes, was significantly reduced exclusively in the DBB condition during the sleep onset period compared to before sleep onset. This observation aligns with a previous study by Trinder, J., et al. in [36], where LF power decreased rapidly in the pre-sleep initiation and sleep onset periods, particularly during the transition from wakefulness to a state of unconsciousness. Even though these results might directly indicate BB’s role in modulating ANS to promote sleep by diminishing sympathetic activity related to arousal and boosting parasympathetic response, the precise mechanism by which BB can modulate the ANS into a more sleep-conductive status has not been fully demonstrated. One of the underlying mechanisms might be that BB can modulate the cortex-level brain activity estimated based on source imaging, such as gamma activity, which involves post-synaptic regulation of cortical neuronal firing [37]. Gamma oscillation in the cortex and hippocampus is implicated in the modulation of autonomic tone during sleep [38], indicating the possible effect of BB on ANS modulation to enhance sleep initiation.
BB might modulate the EEG activity through entrainment, a phenomenon where the frequency of EEG activity synchronizes with the external rhythmic stimulus. A recent study reported that a 3-Hz BB, initiated at the first epoch of N2 sleep, shortened the latency to N3 sleep, increased the proportion of N3 sleep, along with the increment in delta frequency power in EEG during the night, which supports its possible effect of EEG entrainment [26]. However, the effect of BB on entraining EEG is not consistent across studies [39] and varies according to the beat frequency [40]. Furthermore, the optimal beat frequency for inducing entrainment has yet to be demonstrated. In this study, unfortunately, we did not observe any significant entrainment effect between the two conditions (DBB vs. SHAM) which could be due to various DBB parameters, such as intensity, pattern, and form, not being optimized in this proof-of-concept study. Note that the entrainment of EEG activity might not be the major effect of BB in promoting sleep initiation, as the augmentation of delta EEG activity is not the process associated with sleep initiation. Therefore, further research involving a larger number of participants is essential to precisely validate the effect of DBB on brain entrainment as well as the enhancement of sleep initiation.
The advantages of DBB over SBB are far less investigated. In this study, we chose DBB which encompasses the delta frequency band range over SBB with a fixed beat frequency. The cycler duration of DBB was 60 s, and the slow fluctuation of beat frequency (0.017 Hz = 1/60) in DBB stimulation is more aligned to the brain’s intrinsic rhythm in the very low frequency (VLF) band (< 0.1 Hz). This VLF rhythm, specifically augmented during sleep, was observed across EEG, hemodynamic, and CSF recirculation systems in the brain, and is associated with parasympathetic activation, glymphatic system activation, and functional recovery of the brain [41, 42]. By adjusting to this VLF rhythm, DBB may facilitate a more pronounced enhancement of the parasympathetic tone and a more efficient induction of sleep.
Various types of sound manipulations have been introduced to improve sleep quality, such as ASMR, white noise, and BB (Table 1). White noise was effective for sleep efficiency [20] and WASO [15] while BB was beneficial for EEG entrainment [6, 26]. In this study, we demonstrated the positive impact of DBB on sleep latency and HRV, indicating that each sound manipulation has unique advantages for different sleep quality metrics. Consequently, it is not definitive that DBB is more effective overall than other auditory stimuli in enhancing sleep quality. However, the uniqueness of this study lies in its first-time exploration of DBB’s comprehensive impact on multiple sleep stages (pre-sleep initiation, sleep onset, NREM transition) and its simultaneous assessment of both sleep metrics and biomarkers. On the other hand, regarding DBB sound parameters such as intensity, pattern, and form, our study utilized a carrier frequency of 100 Hz, notably lower than the 250 Hz commonly employed in other studies [6, 26]. This selection was motivated by evidence suggesting that LF sound waves can promote relaxation and tension release [43]. Additionally, the DBB cycle duration in our study was set to one minute without optimization, and the DBB volume was adjusted to a comfortable listening level for participants during sleep on the initial day of the experiment (50.50 ± 5.93 dB). Given the existing research gap on the specific effects of different DBB sound parameters on sleep quality, further research into the optimal parameters for DBB, including aspects such as intensity, pattern, and form, is indeed a promising and valuable direction for future studies.
Comparison of Sleep and Biosignal Index Values in Sleep Research Based on Auditory Stimulation
| Auditory stimulation | Sleep quality index | EEG | ECG | ||||
|---|---|---|---|---|---|---|---|
| Sleep latency | Sleep efficiency | WASO | Entrainment | HRV | LF PSD | ||
| Ours | DBB | −51.2% | +4.3%∆ | ns | ns | −27.7% | ns |
| Jirakittayakorn N, et al. [26] | BB | ns | ns | ns | +3.2% | — | — |
| Lee M, et al. [6] | ASMR + BB | — | — | — | +18.0% | — | — |
| Ebben MR, et al. [15] | White noise | ns | ns | −10.4% | — | — | — |
| Stanchina ML, et al. [17] | White noise | ns | — | — | — | — | — |
| Cho ME, et al. [20] | White noise | — | +4.8% | ns | — | — | — |
| Auditory stimulation | Sleep quality index | EEG | ECG | ||||
|---|---|---|---|---|---|---|---|
| Sleep latency | Sleep efficiency | WASO | Entrainment | HRV | LF PSD | ||
| Ours | DBB | −51.2% | +4.3%∆ | ns | ns | −27.7% | ns |
| Jirakittayakorn N, et al. [26] | BB | ns | ns | ns | +3.2% | — | — |
| Lee M, et al. [6] | ASMR + BB | — | — | — | +18.0% | — | — |
| Ebben MR, et al. [15] | White noise | ns | ns | −10.4% | — | — | — |
| Stanchina ML, et al. [17] | White noise | ns | — | — | — | — | — |
| Cho ME, et al. [20] | White noise | — | +4.8% | ns | — | — | — |
Bold values and ∆ indicate statistically significant and marginal significance, respectively; ns: no significance, —: not investigated.
Comparison of Sleep and Biosignal Index Values in Sleep Research Based on Auditory Stimulation
| Auditory stimulation | Sleep quality index | EEG | ECG | ||||
|---|---|---|---|---|---|---|---|
| Sleep latency | Sleep efficiency | WASO | Entrainment | HRV | LF PSD | ||
| Ours | DBB | −51.2% | +4.3%∆ | ns | ns | −27.7% | ns |
| Jirakittayakorn N, et al. [26] | BB | ns | ns | ns | +3.2% | — | — |
| Lee M, et al. [6] | ASMR + BB | — | — | — | +18.0% | — | — |
| Ebben MR, et al. [15] | White noise | ns | ns | −10.4% | — | — | — |
| Stanchina ML, et al. [17] | White noise | ns | — | — | — | — | — |
| Cho ME, et al. [20] | White noise | — | +4.8% | ns | — | — | — |
| Auditory stimulation | Sleep quality index | EEG | ECG | ||||
|---|---|---|---|---|---|---|---|
| Sleep latency | Sleep efficiency | WASO | Entrainment | HRV | LF PSD | ||
| Ours | DBB | −51.2% | +4.3%∆ | ns | ns | −27.7% | ns |
| Jirakittayakorn N, et al. [26] | BB | ns | ns | ns | +3.2% | — | — |
| Lee M, et al. [6] | ASMR + BB | — | — | — | +18.0% | — | — |
| Ebben MR, et al. [15] | White noise | ns | ns | −10.4% | — | — | — |
| Stanchina ML, et al. [17] | White noise | ns | — | — | — | — | — |
| Cho ME, et al. [20] | White noise | — | +4.8% | ns | — | — | — |
Bold values and ∆ indicate statistically significant and marginal significance, respectively; ns: no significance, —: not investigated.
This study has several limitations to be addressed. As a preliminary study, our participant sample size was limited, which might have underpowered the statistical robustness in assessing the effect of DBB. We attempted to reduce the occurrence of return to wake-like REM sleep stages by introducing DBB during each transition from REM to non-REM sleep stages, with the intent to elevate the non-REM sleep stage, which is associated with physical recovery and memory consolidation. However, due to the phenomenon known as REM rebound—a natural homeostatic mechanism to maintain a constant proportion of REM sleep in the sleep cycle, we were unable to identify a significant difference between the two conditions at any DBB presentation junctures [44]. Additionally, EEG analysis did not reveal any significant differences between the two conditions at any of the DBB presentation junctures. Such an observation aligns with studies using external stimuli for EEG entrainment, which are typically confined to singular session studies and those conducted on individuals without sleep abnormalities [45–48]. As mentioned, because various DBB parameters, such as intensity, pattern, and form, were not optimized in this proof-of-concept study, a meaningful effect of EEG entrainment might not have been observed. So, further studies with a larger number of participants including individuals with insomnia and extended application of DBB might reinforce and validate the findings presented in this study.
In conclusion, this study proposed a DBB characterized by a time-varying carrier frequency and demonstrated the enhancement of overall sleep quality following exposure to DBB across varied sleep phases. In particular, sleep onset time was significantly reduced with the decrease in both HRV and LF power. These results validate the efficacy of DBB in ameliorating sleep-related disturbances and underscore its prospective utility as a therapeutic intervention for insomnia.
Supplementary material
Supplementary material is available at SLEEP online.
Acknowledgment
This work was mainly supported by LG Electronics (Q2204181), and partially supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) [No. RS- 2023-00302489] and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. IITP-2024-RS-2023-00258971).
Disclosure Statement
Financial disclosure: This research was supported & funded by LG Electronics (Q2204181). Nonfinancial disclosure: none.
References
Author notes
Hwa-Ah-Ni Lee and Woo-Jin Lee share the first authorship.
Comments