Personalization of Automatic Sleep Scoring: How Best to Adapt Models to Personal Domains in Wearable EEG.

Wearable EEG enables us to capture large amounts of high-quality sleep data for diagnostic purposes. To make full use of this capacity we need high-performance automatic sleep scoring models. To this end, it has been noted that domain mismatch between recording equipment can be considerable, e.g. PSG to wearable EEG, but a previously observed benefit from personalizing models to individual subjects further indicates a personal domain in sleep EEG. In this work, we have investigated the extent of such a personal domain in wearable EEG, and review supervised and unsupervised approaches to personalization as found in the literature. We investigated the personalization effect of the unsupervised Adversarial Domain Adaptation and implemented an unsupervised method based on statistics alignment. No beneficial personalization effect was observed using these unsupervised methods. We find that supervised personalization leads to a substantial performance improvement on the target subject ranging from 15% Cohen's Kappa for subjects with poor performance ( ) to roughly 2% on subjects with high performance ( ). This improvement was present for models trained on both small and large data sets, indicating that even high-performance models benefit from supervised personalization. We found that this personalization can be beneficially regularized using Kullback-Leibler regularization, leading to lower variance with negligible cost to improvement. Based on the experiments, we recommend model personalization using Kullback-Leibler regularization.

Automated sleep staging on reduced channels in children with epilepsy.

Objectives: This study aimed to validate a sleep staging algorithm using in-hospital video-electroencephalogram (EEG) in children without epilepsy, with well-controlled epilepsy (WCE), and with drug-resistant epilepsy (DRE). Methods: Overnight video-EEG, along with electrooculogram (EOG) and chin electromyogram (EMG), was recorded in children between 4 and 18 years of age. Classical sleep staging was performed manually as a ground truth. An end-to-end hierarchical recurrent neural network for sequence-to-sequence automatic sleep staging (SeqSleepNet) was used to perform automated sleep staging using three channels: C4-A1, EOG, and chin EMG. Results: In 176 children sleep stages were manually scored: 47 children without epilepsy, 74 with WCE, and 55 with DRE. The 5-class sleep staging accuracy of the automatic sleep staging algorithm was 84.7% for the children without epilepsy, 83.5% for those with WCE, and 80.8% for those with DRE (Kappa of 0.79, 0.77, and 0.73 respectively). Performance per sleep stage was assessed with an F1 score of 0.91 for wake, 0.50 for N1, 0.83 for N2, 0.84 for N3, and 0.86 for rapid eye movement (REM) sleep. Conclusion: We concluded that the tested algorithm has a high accuracy in children without epilepsy and with WCE. Performance in children with DRE was acceptable, but significantly lower, which could be explained by a tendency of more time spent in N1, and by abundant interictal epileptiform discharges and intellectual disability leading to less recognizable sleep stages. REM sleep time, however, significantly affected in children with DRE, can be detected reliably by the algorithm.Clinical trial registration: ClinicalTrials.gov, identifier NCT04584385.

U-PASS: An uncertainty-guided deep learning pipeline for automated sleep staging.

With the increasing prevalence of machine learning in critical fields like healthcare, ensuring the safety and reliability of these systems is crucial. Estimating uncertainty plays a vital role in enhancing reliability by identifying areas of high and low confidence and reducing the risk of errors. This study introduces U-PASS, a specialized human-centered machine learning pipeline tailored for clinical applications, which effectively communicates uncertainty to clinical experts and collaborates with them to improve predictions. U-PASS incorporates uncertainty estimation at every stage of the process, including data acquisition, training, and model deployment. Training is divided into a supervised pre-training step and a semi-supervised recording-wise finetuning step. We apply U-PASS to the challenging task of sleep staging and demonstrate that it systematically improves performance at every stage. By optimizing the training dataset, actively seeking feedback from domain experts for informative samples, and deferring the most uncertain samples to experts, U-PASS achieves an impressive expert-level accuracy of 85% on a challenging clinical dataset of elderly sleep apnea patients. This represents a significant improvement over the starting point at 75% accuracy. The largest improvement gain is due to the deferral of uncertain epochs to a sleep expert. U-PASS presents a promising AI approach to incorporating uncertainty estimation in machine learning pipelines, improving their reliability and unlocking their potential in clinical settings.

Acoustic stimulation as a promising technique to enhance slow-wave sleep in Alzheimer's disease: results of a pilot study.

STUDY OBJECTIVES: Sleep disturbances are common in people with Alzheimer's disease (AD), and a reduction in slow-wave activity is the most striking underlying change. Acoustic stimulation has emerged as a promising approach to enhance slow-wave activity in healthy adults and people with amnestic mild cognitive impairment. In this phase 1 study we investigated, for the first time, the feasibility of acoustic stimulation in AD and piloted the effect on slow-wave sleep (SWS). METHODS: Eleven adults with mild to moderate AD first wore the DREEM 2 headband for 2 nights to establish a baseline registration. Using machine learning, the DREEM 2 headband automatically scores sleep stages in real time. Subsequently, the participants wore the headband for 14 consecutive "stimulation nights" at home. During these nights, the device applied phase-locked acoustic stimulation of 40-dB pink noise delivered over 2 bone-conductance transducers targeted to the up-phase of the delta wave or SHAM, if it detected SWS in sufficiently high-quality data. RESULTS: Results of the DREEM 2 headband algorithm show a significant average increase in SWS (minutes) [t(3.17) = 33.57, P = .019] between the beginning and end of the intervention, almost twice as much time was spent in SWS. Consensus scoring of electroencephalography data confirmed this trend of more time spent in SWS [t(2.4) = 26.07, P = .053]. CONCLUSIONS: Our phase 1 study provided the first evidence that targeted acoustic stimuli is feasible and could increase SWS in AD significantly. Future studies should further test and optimize the effect of stimulation on SWS in AD in a large randomized controlled trial. CITATION: Van den Bulcke L, Peeters A-M, Heremans E, et al. Acoustic stimulation as a promising technique to enhance slow-wave sleep in Alzheimer's disease: results of a pilot study. J Clin Sleep Med. 2023;19(12):2107-2112.

L-SeqSleepNet: Whole-cycle Long Sequence Modeling for Automatic Sleep Staging.

Human sleep is cyclical with a period of approximately 90 minutes, implying long temporal dependency in the sleep data. Yet, exploring this long-term dependency when developing sleep staging models has remained untouched. In this work, we show that while encoding the logic of a whole sleep cycle is crucial to improve sleep staging performance, the sequential modelling approach in existing state-of-the-art deep learning models are inefficient for that purpose. We thus introduce a method for efficient long sequence modelling and propose a new deep learning model, L-SeqSleepNet, which takes into account whole-cycle sleep information for sleep staging. Evaluating L-SeqSleepNet on four distinct databases of various sizes, we demonstrate state-of-the-art performance obtained by the model over three different EEG setups, including scalp EEG in conventional Polysomnography (PSG), in-ear EEG, and around-the-ear EEG (cEEGrid), even with a single EEG channel input. Our analyses also show that L-SeqSleepNet is able to alleviate the predominance of N2 sleep (the major class in terms of classification) to bring down errors in other sleep stages. Moreover the network becomes much more robust, meaning that for all subjects where the baseline method had exceptionally poor performance, their performance are improved significantly. Finally, the computation time only grows at a sub-linear rate when the sequence length increases.

From unsupervised to semi-supervised adversarial domain adaptation in electroencephalography-based sleep staging.

Objective.The recent breakthrough of wearable sleep monitoring devices has resulted in large amounts of sleep data. However, as limited labels are available, interpreting these data requires automated sleep stage classification methods with a small need for labeled training data. Transfer learning and domain adaptation offer possible solutions by enabling models to learn on a source dataset and adapt to a target dataset.Approach.In this paper, we investigate adversarial domain adaptation applied to real use cases with wearable sleep datasets acquired from diseased patient populations. Different practical aspects of the adversarial domain adaptation framework are examined, including the added value of (pseudo-)labels from the target dataset and the influence of domain mismatch between the source and target data. The method is also implemented for personalization to specific patients.Main results.The results show that adversarial domain adaptation is effective in the application of sleep staging on wearable data. When compared to a model applied on a target dataset without any adaptation, the domain adaptation method in its simplest form achieves relative gains of 7%-27% in accuracy. The performance in the target domain is further boosted by adding pseudo-labels and real target domain labels when available, and by choosing an appropriate source dataset. Furthermore, unsupervised adversarial domain adaptation can also personalize a model, improving the performance by 1%-2% compared to a non-personalized model.Significance.In conclusion, adversarial domain adaptation provides a flexible framework for semi-supervised and unsupervised transfer learning. This is particularly useful in sleep staging and other wearable electroencephalography applications. (Clinical trial registration number: S64190.).

Artificial Neural Network-Based Automatic Detection of Food Intake for Neuromodulation in Treating Obesity and Diabetes.

PURPOSE: Neuromodulation, such as vagal nerve stimulation and intestinal electrical stimulation, has been introduced for the treatment of obesity and diabetes. Ideally, neuromodulation should be applied automatically after food intake. The purpose of this study was to develop a method of automatic food intake detection through dynamic analysis of heart rate variability (HRV). MATERIALS AND METHODS: Two experiments were conducted: (1) a small sample series with a standard test meal and (2) a large sample series with varying meal size. Electrocardiograms (ECGs) were collected in the fasting and postprandial states. Each ECG was processed to compute the HRV. For each HRV segment, time- and frequency-domain features were derived and used as inputs to train and test an artificial neural network (ANN). The ANN was trained and tested with different cross-validation methods. RESULTS: The highest classification accuracy reached with leave-one-subject-out-leave-one-sample-out cross-validation was 0.93 in experiment 1 and 0.88 in experiment 2. Retraining the ANN on recordings of a subject drastically increased the achieved accuracy for that subject to values of 0.995 and 0.95 in experiments 1 and 2, respectively. CONCLUSIONS: Automatic food intake detection by ANNs, using features from the HRV, is feasible and may have a great potential for neuromodulation-based treatments of meal-related disorders.

Maturation of the Autonomic Nervous System in Premature Infants: Estimating Development Based on Heart-Rate Variability Analysis.

This study aims at investigating the development of premature infants' autonomic nervous system (ANS) based on a quantitative analysis of the heart-rate variability (HRV) with a variety of novel features. Additionally, the role of heart-rate drops, known as bradycardias, has been studied in relation to both clinical and novel sympathovagal indices. ECG data were measured for at least 3 h in 25 preterm infants (gestational age ≤32 weeks) for a total number of 74 recordings. The post-menstrual age (PMA) of each patient was estimated from the RR interval time-series by means of multivariate linear-mixed effects regression. The tachograms were segmented based on bradycardias in periods after, between and during bradycardias. For each of those epochs, a set of temporal, spectral and fractal indices were included in the regression model. The best performing model has R 2 = 0.75 and mean absolute error MAE = 1.56 weeks. Three main novelties can be reported. First, the obtained maturation models based on HRV have comparable performance to other development models. Second, the selected features for age estimation show a predominance of power and fractal features in the very-low- and low-frequency bands in explaining the infants' sympathovagal development from 27 PMA weeks until 40 PMA weeks. Third, bradycardias might disrupt the relationship between common temporal indices of the tachogram and the age of the infant and the interpretation of sympathovagal indices. This approach might provide a novel overview of post-natal autonomic maturation and an alternative development index to other electrophysiological data analysis.