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.

Multimodal wearable EEG, EMG and accelerometry measurements improve the accuracy of tonic-clonic seizure detection.

Objective. This paper aims to investigate the possibility of detecting tonic-clonic seizures (TCSs) with behind-the-ear, two-channel wearable electroencephalography (EEG), and to evaluate its added value to non-EEG modalities in TCS detection.Methods. We included 27 participants with a total of 44 TCSs from the European multicenter study SeizeIT2. The wearable Sensor Dot (Byteflies) was used to measure behind-the-ear EEG, electromyography (EMG), electrocardiography, accelerometry (ACC) and gyroscope. We evaluated automatic unimodal detection of TCSs, using sensitivity, precision, false positive rate (FPR) and F1-score. Subsequently, we fused the different modalities and again assessed performance. Algorithm-labeled segments were then provided to two experts, who annotated true positive TCSs, and discarded false positives.Results. Wearable EEG outperformed the other single modalities with a sensitivity of 100% and a FPR of 10.3/24 h. The combination of wearable EEG and EMG proved most clinically useful, delivering a sensitivity of 97.7%, an FPR of 0.4/24 h, a precision of 43%, and an F1-score of 59.7%. The highest overall performance was achieved through the fusion of wearable EEG, EMG, and ACC, yielding a sensitivity of 90.9%, an FPR of 0.1/24 h, a precision of 75.5%, and an F1-score of 82.5%.Conclusions. In TCS detection with a wearable device, combining EEG with EMG, ACC or both resulted in a remarkable reduction of FPR, while retaining a high sensitivity.Significance. Adding wearable EEG could further improve TCS detection, relative to extracerebral-based systems.

Resting state electroencephalographic brain activity in neonates can predict age and is indicative of neurodevelopmental outcome.

OBJECTIVE: Electroencephalography (EEG) can be used to estimate neonates' biological brain age. Discrepancies between postmenstrual age and brain age, termed the brain age gap, can potentially quantify maturational deviation. Existing brain age EEG models are not well suited to clinical cot-side use for estimating neonates' brain age gap due to their dependency on relatively large data and pre-processing requirements. METHODS: We trained a deep learning model on resting state EEG data from preterm neonates with normal neurodevelopmental Bayley Scale of Infant and Toddler Development (BSID) outcomes, using substantially reduced data requirements. We subsequently tested this model in two independent datasets from two clinical sites. RESULTS: In both test datasets, using only 20 min of resting-state EEG activity from a single channel, the model generated accurate age predictions: mean absolute error = 1.03 weeks (p-value = 0.0001) and 0.98 weeks (p-value = 0.0001). In one test dataset, where 9-month follow-up BSID outcomes were available, the average neonatal brain age gap in the severe abnormal outcome group was significantly larger than that of the normal outcome group: difference in mean brain age gap = 0.50 weeks (p-value = 0.04). CONCLUSIONS: These findings demonstrate that the deep learning model generalises to independent datasets from two clinical sites, and that the model's brain age gap magnitudes differ between neonates with normal and severe abnormal follow-up neurodevelopmental outcomes. SIGNIFICANCE: The magnitude of neonates' brain age gap, estimated using only 20 min of resting state EEG data from a single channel, can encode information of clinical neurodevelopmental value.

Interpretable deep learning for the remote characterisation of ambulation in multiple sclerosis using smartphones.

The emergence of digital technologies such as smartphones in healthcare applications have demonstrated the possibility of developing rich, continuous, and objective measures of multiple sclerosis (MS) disability that can be administered remotely and out-of-clinic. Deep Convolutional Neural Networks (DCNN) may capture a richer representation of healthy and MS-related ambulatory characteristics from the raw smartphone-based inertial sensor data than standard feature-based methodologies. To overcome the typical limitations associated with remotely generated health data, such as low subject numbers, sparsity, and heterogeneous data, a transfer learning (TL) model from similar large open-source datasets was proposed. Our TL framework leveraged the ambulatory information learned on human activity recognition (HAR) tasks collected from wearable smartphone sensor data. It was demonstrated that fine-tuning TL DCNN HAR models towards MS disease recognition tasks outperformed previous Support Vector Machine (SVM) feature-based methods, as well as DCNN models trained end-to-end, by upwards of 8-15%. A lack of transparency of "black-box" deep networks remains one of the largest stumbling blocks to the wider acceptance of deep learning for clinical applications. Ensuing work therefore aimed to visualise DCNN decisions attributed by relevance heatmaps using Layer-Wise Relevance Propagation (LRP). Through the LRP framework, the patterns captured from smartphone-based inertial sensor data that were reflective of those who are healthy versus people with MS (PwMS) could begin to be established and understood. Interpretations suggested that cadence-based measures, gait speed, and ambulation-related signal perturbations were distinct characteristics that distinguished MS disability from healthy participants. Robust and interpretable outcomes, generated from high-frequency out-of-clinic assessments, could greatly augment the current in-clinic assessment picture for PwMS, to inform better disease management techniques, and enable the development of better therapeutic interventions.

Neonatal EEG sleep stage classification based on deep learning and HMM.

OBJECTIVE: Automatic sleep stage scoring is of great importance for investigating sleep architecture during infancy. In this work, we introduce a novel multichannel approach based on deep learning networks and hidden Markov models (HMM) to improve the accuracy of sleep stage classification in term neonates. APPROACH: The classification performance was evaluated on quiet sleep (QS) and active sleep (AS) stages, each with two sub-states, using multichannel EEG data recorded from sixteen neonates with postmenstrual age of 38-40 weeks. A comprehensive set of linear and nonlinear features were extracted from thirty-second EEG segments. The feature space dimensionality was then reduced by using an evolutionary feature selection method called MGCACO (Modified Graph Clustering Ant Colony Optimization) based on the relevance and redundancy analysis. A bi-directional long-short time memory (BiLSTM) network was trained for sleep stage classification. The number of channels was optimized using the sequential forward selection method to reduce the spatial space. Finally, an HMM-based postprocessing stage was used to reduce false positives by incorporating the knowledge of transition probabilities between stages into the classification process. The method performance was evaluated using the K-fold (KFCV) and leave-one-out cross-validation (LOOCV) strategies. MAIN RESULTS: Using six-bipolar channels, our method achieved a mean kappa and an overall accuracy of 0.71-0.76 and 78.9%-82.4% using the KFCV and LOOCV strategies, respectively. SIGNIFICANCE: The presented automatic sleep stage scoring method can be used to study the neurodevelopmental process and to diagnose brain abnormalities in term neonates.

Building a Machine-Learning Framework to Remotely Assess Parkinson's Disease Using Smartphones.

OBJECTIVE: Parkinson's disease (PD) is a neurodegenerative disorder that affects multiple neurological systems. Traditional PD assessment is conducted by a physician during infrequent clinic visits. Using smartphones, remote patient monitoring has the potential to obtain objective behavioral data semi-continuously, track disease fluctuations, and avoid rater dependency. METHODS: Smartphones collect sensor data during various active tests and passive monitoring, including balance (postural instability), dexterity (skill in performing tasks using hands), gait (the pattern of walking), tremor (involuntary muscle contraction and relaxation), and voice. Some of the features extracted from smartphone data are potentially associated with specific PD symptoms identified by physicians. To leverage large-scale cross-modality smartphone features, we propose a machine-learning framework for performing automated disease assessment. The framework consists of a two-step feature selection procedure and a generic model based on the elastic-net regularization. RESULTS: Using this framework, we map the PD-specific architecture of behaviors using data obtained from both PD participants and healthy controls (HCs). Utilizing these atlases of features, the framework shows promises to (a) discriminate PD participants from HCs, and (b) estimate the disease severity of individuals with PD. SIGNIFICANCE: Data analysis results from 437 behavioral features obtained from 72 subjects (37 PD and 35 HC) sampled from 17 separate days during a period of up to six months suggest that this framework is potentially useful for the analysis of remotely collected smartphone sensor data in individuals with PD.

A Bayesian parametric model for quantifying brain maturation from sleep-EEG in the vulnerable newborn baby.

Newborn babies, particularly preterms, can exhibit early deviations in sleep maturation as seen by Electroencephalogram (EEG) recordings. This may be indicative of cognitive problems by school-age. The current 'clinically-driven' approach uses separate algorithms to first extract sleep states and then predict EEG 'brain-age'. Maturational deviations are identified when the brain-age no longer matches the Postmenstrual Age (PMA, the age since the last menstrual cycle of the mother). However, the PMA range where existing sleep staging algorithms perform optimally, is limited, which subsequently limits the PMA range for brain-age prediction. We introduce a Bayesian Parametric Model (BPM) as a single end-to-end solution to directly estimate brain-age, modelling for sleep state maturation without requiring a separately optimized sleep staging algorithm. Comparison of this model with a traditional multi-stage approach, yields a similar Krippendorff's $\alpha = 0.92$ (a performance measure ranging from 0 (chance agreement) to 1 (perfect agreement)) with the BPM performing better at younger ages <30 weeks PMA. The BPM's potential to detect maturational deviations is also explored on a few preterm babies who were abnormal at 9 months follow-up.

Automatic Sleep Stage Classification Using Single-Channel EEG: Learning Sequential Features with Attention-Based Recurrent Neural Networks.

We propose in this work a feature learning approach using deep bidirectional recurrent neural networks (RNNs) with attention mechanism for single-channel automatic sleep stage classification. We firstly decompose an EEG epoch into multiple small frames and subsequently transform them into a sequence of frame-wise feature vectors. Given the training sequences, the attention-based RNN is trained in a sequence-to-label fashion for sleep stage classification. Due to discriminative training, the network is expected to encode information of an input sequence into a high-level feature vector after the attention layer. We, therefore, treat the trained network as a feature extractor and extract these feature vectors for classification which is accomplished by a linear SVM classifier. We also propose a discriminative method to learn a filter bank with a DNN for preprocessing purpose. Filtering the frame-wise feature vectors with the learned filter bank beforehand leads to further improvement on the classification performance. The proposed approach demonstrates good performance on the Sleep-EDF dataset.

Mobile EEG on the bike: disentangling attentional and physical contributions to auditory attention tasks.

OBJECTIVE: In the past few years there has been a growing interest in studying brain functioning in natural, real-life situations. Mobile EEG allows to study the brain in real unconstrained environments but it faces the intrinsic challenge that it is impossible to disentangle observed changes in brain activity due to increase in cognitive demands by the complex natural environment or due to the physical involvement. In this work we aim to disentangle the influence of cognitive demands and distractions that arise from such outdoor unconstrained recordings. APPROACH: We evaluate the ERP and single trial characteristics of a three-class auditory oddball paradigm recorded in outdoor scenario's while peddling on a fixed bike or biking freely around. In addition we also carefully evaluate the trial specific motion artifacts through independent gyro measurements and control for muscle artifacts. MAIN RESULTS: A decrease in P300 amplitude was observed in the free biking condition as compared to the fixed bike conditions. Above chance P300 single-trial classification in highly dynamic real life environments while biking outdoors was achieved. Certain significant artifact patterns were identified in the free biking condition, but neither these nor the increase in movement (as derived from continuous gyrometer measurements) can explain the differences in classification accuracy and P300 waveform differences with full clarity. The increased cognitive load in real-life scenarios is shown to play a major role in the observed differences. SIGNIFICANCE: Our findings suggest that auditory oddball results measured in natural real-life scenarios are influenced mainly by increased cognitive load due to being in an unconstrained environment.

Tensor-based classification of an auditory mobile BCI without a subject-specific calibration phase.

OBJECTIVE: One of the major drawbacks in EEG brain-computer interfaces (BCI) is the need for subject-specific training of the classifier. By removing the need for a supervised calibration phase, new users could potentially explore a BCI faster. In this work we aim to remove this subject-specific calibration phase and allow direct classification. APPROACH: We explore canonical polyadic decompositions and block term decompositions of the EEG. These methods exploit structure in higher dimensional data arrays called tensors. The BCI tensors are constructed by concatenating ERP templates from other subjects to a target and non-target trial and the inherent structure guides a decomposition that allows accurate classification. We illustrate the new method on data from a three-class auditory oddball paradigm. MAIN RESULTS: The presented approach leads to a fast and intuitive classification with accuracies competitive with a supervised and cross-validated LDA approach. SIGNIFICANCE: The described methods are a promising new way of classifying BCI data with a forthright link to the original P300 ERP signal over the conventional and widely used supervised approaches.

The "why" and "how" of JointICA: results from a visual detection task.

Since several years, neuroscience research started to focus on multimodal approaches. One such multimodal approach is the combination of electroencephalography (EEG) and functional magnetic resonance imaging (fMRI). However, no standard integration procedure has been established so far. One promising data-driven approach consists of a joint decomposition of event-related potentials (ERPs) and fMRI maps derived from the response to a particular stimulus. Such an algorithm (joint independent component analysis or JointICA) has recently been proposed by Calhoun et al. (2006). This method provides sources with both a fine spatial and temporal resolution, and has shown to provide meaningful results. However, the algorithm's performance has not been fully characterized yet, and no procedure has been proposed to assess the quality of the decomposition. In this paper, we therefore try to answer why and how JointICA works. We show the performance of the algorithm on data obtained in a visual detection task, and compare the performance for EEG recorded simultaneously with fMRI data and for EEG recorded in a separate session (outside the scanner room). We perform several analyses in order to set the necessary conditions that lead to a sound decomposition, and to give additional insights for exploration in future studies. In that respect, we show how the algorithm behaves when different EEG electrodes are used and we test the robustness with respect to the number of subjects in the study. The performance of the algorithm in all the experiments is validated based on results from previous studies.