An enzyme-inspired specificity in deep learning model for sleep stage classification using multi-channel PSG signals input: Separating training approach and its performance on cross-dataset validation for generalizability.

Numerous automatic sleep stage classification systems have been developed, but none have become effective assistive tools for sleep technicians due to issues with generalization. Four key factors hinder the generalization of these models are instruments, montage of recording, subject type, and scoring manual factors. This study aimed to develop a deep learning model that addresses generalization problems by integrating enzyme-inspired specificity and employing separating training approaches. Subject type and scoring manual factors were controlled, while the focus was on instruments and montage of recording factors. The proposed model consists of three sets of signal-specific models including EEG-, EOG-, and EMG-specific model. The EEG-specific models further include three sets of channel-specific models. All signal-specific and channel-specific models were established with data manipulation and weighted loss strategies, resulting in three sets of data manipulation models and class-specific models, respectively. These models were CNNs. Additionally, BiLSTM models were applied to EEG- and EOG-specific models to obtain temporal information. Finally, classification task for sleep stage was handled by 'the-last-dense' layer. The optimal sampling frequency for each physiological signal was identified and used during the training process. The proposed model was trained on MGH dataset and evaluated using both within dataset and cross-dataset. For MGH dataset, overall accuracy of 81.05 %, MF1 of 79.05 %, Kappa of 0.7408, and per-class F1-scores: W (84.98 %), N1 (58.06 %), N2 (84.82 %), N3 (79.20 %), and REM (88.17 %) can be achieved. Performances on cross-datasets are as follows: SHHS1 200 records reached 79.54 %, 70.56 %, and 0.7078; SHHS2 200 records achieved 76.77 %, 66.30 %, and 0.6632; Sleep-EDF 153 records gained 78.52 %, 72.13 %, and 0.7031; and BCI-MU (local dataset) 94 records achieved 83.57 %, 82.17 %, and 0.7769 for overall accuracy, MF1, and Kappa respectively. Additionally, the proposed model has approximately 9.3 M trainable parameters and takes around 26 s to process one PSG record. The results indicate that the proposed model demonstrates generalizability in sleep stage classification and shows potential as a feasibility tool for real-world applications. Additionally, enzyme-inspired specificity effectively addresses the challenges posed by varying montage of recording, while the identified optimal frequencies mitigate instrument-related issues.

ZleepAnlystNet: a novel deep learning model for automatic sleep stage scoring based on single-channel raw EEG data using separating training.

Numerous models for sleep stage scoring utilizing single-channel raw EEG signal have typically employed CNN and BiLSTM architectures. While these models, incorporating temporal information for sequence classification, demonstrate superior overall performance, they often exhibit low per-class performance for N1-stage, necessitating an adjustment of loss function. However, the efficacy of such adjustment is constrained by the training process. In this study, a pioneering training approach called separating training is introduced, alongside a novel model, to enhance performance. The developed model comprises 15 CNN models with varying loss function weights for feature extraction and 1 BiLSTM for sequence classification. Due to its architecture, this model cannot be trained using an end-to-end approach, necessitating separate training for each component using the Sleep-EDF dataset. Achieving an overall accuracy of 87.02%, MF1 of 82.09%, Kappa of 0.8221, and per-class F1-socres (W 90.34%, N1 54.23%, N2 89.53%, N3 88.96%, and REM 87.40%), our model demonstrates promising performance. Comparison with sleep technicians reveals a Kappa of 0.7015, indicating alignment with reference sleep stags. Additionally, cross-dataset validation and adaptation through training with the SHHS dataset yield an overall accuracy of 84.40%, MF1 of 74.96% and Kappa of 0.7785 when tested with the Sleep-EDF-13 dataset. These findings underscore the generalization potential in model architecture design facilitated by our novel training approach.

Effect of 1% H2O2 on Three Salivary Stress Biomarkers, Cortisol, Alpha-Amylase, and sIgA.

BACKGROUND: Due to the COVID-19 pandemic, several associations worldwide have been recommending the use of 1% hydrogen peroxide solution as a preprocedural mouth rinse before dental treatments to reduce viral load in saliva. This protocol is also employed in stress studies, especially in the context of dental treatment that uses salivary biomarkers as an indicator. However, the effect of 1% hydrogen peroxide as mouth rinse on salivary biomarkers remains unclear. OBJECTIVE: This study aims to investigate the effects of 1% hydrogen peroxide solution as a preprocedural mouth rinse on 3 salivary stress biomarkers-salivary cortisol, salivary secretory IgA, and salivary α-amylase-both on chemical influence and mechanical irrigation. MATERIALS AND METHODS: Ninety healthy participants with confirmed negative Reverse Transcription Polymerase Chain Reaction results for COVID-19 at most 2 days prior to the experiment were included in this study. All participants were randomly allocated into 3 groups: experimental (1% hydrogen peroxide solution), positive control (distilled water), and negative control (no mouth rinse). Saliva samples were collected before and after mouth rinsing with the respective solutions. Salivary biomarkers were analysed using specific enzyme-linked immunosorbent assay kits. RESULTS: Salivary cortisol and salivary α-amylase did not significantly differ before and after rinsing, whilst salivary sIgA levels decreased in all 3 groups. Nonetheless, there were no significant differences in the changes of these biomarkers across the 3 groups. CONCLUSIONS: This study shows that using 1% hydrogen peroxide solution as a preprocedural mouth rinse for universal precaution does not alter the levels of these 3 salivary biomarkers.

A Novel Insight of Effects of a 3-Hz Binaural Beat on Sleep Stages During Sleep.

The dichotic presentation of two almost equivalent pure tones with slightly different frequencies leads to virtual beat perception by the brain. In this phenomenon, the so-called binaural beat has a frequency equaling the difference of the frequencies of the two pure tones. The binaural beat can entrain neural activities to synchronize with the beat frequency and induce behavioral states related to the neural activities. This study aimed to investigate the effect of a 3-Hz binaural beat on sleep stages, which is considered a behavioral state. Twenty-four participants were allocated to experimental and control groups. The experimental period was three consecutive nights consisting of an adaptation night, a baseline night, and an experimental night. Participants in both groups underwent the same procedures, but only the experimental group was exposed to the 3-Hz binaural beat on the experimental night. The stimulus was initiated when the first epoch of the N2 sleep stage was detected and stopped when the first epoch of the N3 sleep stage detected. For the control group, a silent sham stimulus was used. However, the participants were blinded to their stimulus group. The results showed that the N3 duration of the experimental group was longer than that of the control group, and the N2 duration of the experimental group was shorter than that of the control group. Moreover, the N3 latency of the experimental group was shorter.

Brain Responses to a 6-Hz Binaural Beat: Effects on General Theta Rhythm and Frontal Midline Theta Activity.

A binaural beat is a beat phenomenon that is generated by the dichotic presentation of two almost equivalent pure tones but with slightly different frequencies. The brain responses to binaural beats remain controversial; therefore, the aim of this study was to investigate theta activity responses to a binaural beat by controlling factors affecting localization, including beat frequency, carrier tone frequency, exposure duration, and recording procedure. Exposure to a 6-Hz binaural beat on a 250 Hz carrier tone for 30 min was utilized in this study. Quantitative electroencephalography (QEEG) was utilized as the recording modality. Twenty-eight participants were divided into experimental and control groups. Emotional states were evaluated by Brunel Mood Scale (BRMUS) before and after exposing to the stimulus. The results showed that theta activity was induced in the entire cortex within 10 min of exposure to the stimulus in the experimental group. Compared to the control group, theta activity was also induced at the frontal and parietal-central regions, which included the Fz position, and left hemisphere dominance was presented for other exposure durations. The pattern recorded for 10 min of exposure appeared to be brain functions of a meditative state. Moreover, tension factor of BRUMS was decreased in experimental group compared to control group which resembled the meditation effect. Thus, a 6-Hz binaural beat on a 250 Hz carrier tone was suggested as a stimulus for inducing a meditative state.

Brain responses to 40-Hz binaural beat and effects on emotion and memory.

Gamma oscillation plays a role in binding process or sensory integration, a process by which several brain areas beside primary cortex are activated for higher perception of the received stimulus. Beta oscillation is also involved in interpreting received stimulus and occurs following gamma oscillation, and this process is known as gamma-to-beta transition, a process for neglecting unnecessary stimuli in surrounding environment. Gamma oscillation also associates with cognitive functions, memory and emotion. Therefore, modulation of the brain activity can lead to manipulation of cognitive functions. The stimulus used in this study was 40-Hz binaural beat because binaural beat induces frequency following response. This study aimed to investigate the neural oscillation responding to the 40-Hz binaural beat and to evaluate working memory function and emotional states after listening to that stimulus. Two experiments were developed based on the study aims. In the first experiment, electroencephalograms were recorded while participants listened to the stimulus for 30min. The results suggested that frontal, temporal, and central regions were activated within 15min. In the second experiment, word list recall task was conducted before and after listening to the stimulus for 20min. The results showed that, after listening, the recalled words were increase in the working memory portion of the list. Brunel Mood Scale, a questionnaire to evaluate emotional states, revealed changes in emotional states after listening to the stimulus. The emotional results suggested that these changes were consistent with the induced neural oscillations.

The brain responses to different frequencies of binaural beat sounds on QEEG at cortical level.

Beat phenomenon is occurred when two slightly different frequency waves interfere each other. The beat can also occur in the brain by providing two slightly different frequency waves separately each ear. This is called binaural beat. The brain responses to binaural beat are in discussion process whether the brain side and the brain area. Therefore, this study aims to figure out the brain responses to binaural beat by providing different binaural beat frequencies on 250 carrier tone continuously for 30 minutes to participants and using quantitative electroencephalography (QEEG) to interpret the data. The result shows that different responses appear in different beat frequency. Left hemisphere dominance occur in 3 Hz beat within 15 minutes and 15 Hz beat within 5 minutes. Right hemisphere dominance occurs in 10 Hz beat within 25 minute. 6 Hz beat enhances all area of the brain within 10 minutes. 8 Hz and 25 Hz beats have no clearly responses while 40 Hz beat enhances the responses in frontal lobe. These brain responses can be used for brain modulation application to induce the brain activity in further studies.