Combining general and personal models for epilepsy detection with hyperdimensional computing.

Epilepsy is a highly prevalent chronic neurological disorder with great negative impact on patients' daily lives. Despite this there is still no adequate technological support to enable epilepsy detection and continuous outpatient monitoring in everyday life. Hyperdimensional (HD) computing is a promising method for epilepsy detection via wearable devices, characterized by a simpler learning process and lower memory requirements compared to other methods. In this work, we demonstrate additional avenues in which HD computing and the manner in which its models are built and stored can be used to better understand, compare and create more advanced machine learning models for epilepsy detection. These possibilities are not feasible with other state-of-the-art models, such as random forests or neural networks. We compare inter-subject model similarity of different classes (seizure and non-seizure), study the process of creating general models from personal ones, and finally posit a method of combining personal and general models to create hybrid models. This results in an improved epilepsy detection performance. We also tested knowledge transfer between models trained on two different datasets. The attained insights are highly interesting not only from an engineering perspective, to create better models for wearables, but also from a neurological perspective, to better understand individual epilepsy patterns.

Importance of methodological choices in data manipulation for validating epileptic seizure detection models.

Epilepsy is a chronic neurological disorder that affects a significant portion of the human population and imposes serious risks in the daily life. Despite advances in machine learning and IoT, small, non-stigmatizing wearable devices for continuous monitoring and detection in outpatient environments are not yet widely available. Part of the reason is the complexity of epilepsy itself, including highly imbalanced data, multimodal nature, and very subject-specific signatures. However, another problem is the heterogeneity of methodological approaches in research, leading to slower progress, difficulty in comparing results, and low reproducibility. Therefore, this article identifies a wide range of methodological decisions that must be made and reported when training and evaluating the performance of epilepsy detection systems. We characterize the influence of individual choices using a typical ensemble random-forest model and the publicly available CHB-MIT database, providing a broader picture of each decision and giving good-practice recommendations, based on our experience, where possible.

A Multimodal Dataset for Automatic Edge-AI Cough Detection.

Counting the number of times a patient coughs per day is an essential biomarker in determining treatment efficacy for novel antitussive therapies and personalizing patient care. Automatic cough counting tools must provide accurate information, while running on a lightweight, portable device that protects the patient's privacy. Several devices and algorithms have been developed for cough counting, but many use only error-prone audio signals, rely on offline processing that compromises data privacy, or utilize processing and memory-intensive neural networks that require more hardware resources than can fit on a wearable device. Therefore, there is a need for wearable devices that employ multimodal sensors to perform accurate, privacy-preserving, automatic cough counting algorithms directly on the device in an edge Artificial Intelligence (edge-AI) fashion. To advance this research field, we contribute the first publicly accessible cough counting dataset of multimodal biosignals. The database contains nearly 4 hours of biosignal data, with both acoustic and kinematic modalities, covering 4,300 annotated cough events from 15 subjects. Furthermore, a variety of non-cough sounds and motion scenarios mimicking daily life activities are also present, which the research community can use to accelerate machine learning (ML) algorithm development. A technical validation of the dataset reveals that it represents a wide variety of signal-to-noise ratios, which can be expected in a real-life use case, as well as consistency across experimental trials. Finally, to demonstrate the usability of the dataset, we train a simple cough vs non-cough signal classifier that obtains a 91% sensitivity, 92% specificity, and 80% precision on unseen test subject data. Such edge-friendly AI algorithms have the potential to provide continuous ambulatory monitoring of the numerous chronic cough patients.

A semi-supervised algorithm for improving the consistency of crowdsourced datasets: The COVID-19 case study on respiratory disorder classification.

BACKGROUND AND OBJECTIVE: Cough audio signal classification is a potentially useful tool in screening for respiratory disorders, such as COVID-19. Since it is dangerous to collect data from patients with contagious diseases, many research teams have turned to crowdsourcing to quickly gather cough sound data. The COUGHVID dataset enlisted expert physicians to annotate and diagnose the underlying diseases present in a limited number of recordings. However, this approach suffers from potential cough mislabeling, as well as disagreement between experts. METHODS: In this work, we use a semi-supervised learning (SSL) approach - based on audio signal processing tools and interpretable machine learning models - to improve the labeling consistency of the COUGHVID dataset for 1) COVID-19 versus healthy cough sound classification 2) distinguishing wet from dry coughs, and 3) assessing cough severity. First, we leverage SSL expert knowledge aggregation techniques to overcome the labeling inconsistencies and label sparsity in the dataset. Next, our SSL approach is used to identify a subsample of re-labeled COUGHVID audio samples that can be used to train or augment future cough classifiers. RESULTS: The consistency of the re-labeled COVID-19 and healthy data is demonstrated in that it exhibits a high degree of inter-class feature separability: 3x higher than that of the user-labeled data. Similarly, the SSL method increases this separability by 11.3x for cough type and 5.1x for severity classifications. Furthermore, the spectral differences in the user-labeled audio segments are amplified in the re-labeled data, resulting in significantly different power spectral densities between healthy and COVID-19 coughs in the 1-1.5 kHz range (p=1.2×10-64), which demonstrates both the increased consistency of the new dataset and its explainability from an acoustic perspective. Finally, we demonstrate how the re-labeled dataset can be used to train a COVID-19 classifier, achieving an AUC of 0.797. CONCLUSIONS: We propose a SSL expert knowledge aggregation technique for the field of cough sound classification for the first time, and demonstrate how it can be used to combine the medical knowledge of multiple experts in an explainable fashion, thus providing abundant, consistent data for cough classification tasks.

Event-based sampled ECG morphology reconstruction through self-similarity.

BACKGROUND AND OBJECTIVE: Event-based analog-to-digital converters allow for sparse bio-signal acquisition, enabling local sub-Nyquist sampling frequency. However, aggressive event selection can cause the loss of important bio-markers, not recoverable with standard interpolation techniques. In this work, we leverage the self-similarity of the electrocardiogram (ECG) signal to recover missing features in event-based sampled ECG signals, dynamically selecting patient-representative templates together with a novel dynamic time warping algorithm to infer the morphology of event-based sampled heartbeats. METHODS: We acquire a set of uniformly sampled heartbeats and use a graph-based clustering algorithm to define representative templates for the patient. Then, for each event-based sampled heartbeat, we select the morphologically nearest template, and we then reconstruct the heartbeat with piece-wise linear deformations of the selected template, according to a novel dynamic time warping algorithm that matches events to template segments. RESULTS: Synthetic tests on a standard normal sinus rhythm dataset, composed of approximately 1.8 million normal heartbeats, show a big leap in performance with respect to standard resampling techniques. In particular (when compared to classic linear resampling), we show an improvement in P-wave detection of up to 10 times, an improvement in T-wave detection of up to three times, and a 30% improvement in the dynamic time warping morphological distance. CONCLUSION: In this work, we have developed an event-based processing pipeline that leverages signal self-similarity to reconstruct event-based sampled ECG signals. Synthetic tests show clear advantages over classical resampling techniques.

Adaptive R-Peak Detection on Wearable ECG Sensors for High-Intensity Exercise.

OBJECTIVE: Continuous monitoring of biosignals via wearable sensors has quickly expanded in the medical and wellness fields. At rest, automatic detection of vital parameters is generally accurate. However, in conditions such as high-intensity exercise, sudden physiological changes occur to the signals, compromising the robustness of standard algorithms. METHODS: Our method, called BayeSlope, is based on unsupervised learning, Bayesian filtering, and non-linear normalization to enhance and correctly detect the R peaks according to their expected positions in the ECG. Furthermore, as BayeSlope is computationally heavy and can drain the device battery quickly, we propose an online design that adapts its robustness to sudden physiological changes, and its complexity to the heterogeneous resources of modern embedded platforms. This method combines BayeSlope with a lightweight algorithm, executed in cores with different capabilities, to reduce the energy consumption while preserving the accuracy. RESULTS: BayeSlope achieves an F1 score of 99.3% in experiments during intense cycling exercise with 20 subjects. Additionally, the online adaptive process achieves an F1 score of 99% across five different exercise intensities, with a total energy consumption of 1.55±0.54 mJ. CONCLUSION: We propose a highly accurate and robust method, and a complete energy-efficient implementation in a modern ultra-low-power embedded platform to improve R peak detection in challenging conditions, such as during high-intensity exercise. SIGNIFICANCE: The experiments show that BayeSlope outperforms state-of-the-art QRS detectors up to 8.4% in F1 score, while our online adaptive method can reach energy savings up to 38.7% on modern heterogeneous wearable platforms.

Exploration of Hyperdimensional Computing Strategies for Enhanced Learning on Epileptic Seizure Detection.

Wearable and unobtrusive monitoring and prediction of epileptic seizures has the potential to significantly increase the life quality of patients, but is still an unreached goal due to challenges of real-time detection and wearable devices design. Hyperdimensional (HD) computing has evolved in recent years as a new promising machine learning approach, especially when talking about wearable applications. But in the case of epilepsy detection, standard HD computing is not performing at the level of other state-of-the-art algorithms. This could be due to the inherent complexity of the seizures and their signatures in different biosignals, such as the electroencephalogram (EEG), the highly personalized nature, and the disbalance of seizure and non-seizure instances. In the literature, different strategies for improved learning of HD computing have been proposed, such as iterative (multi-pass) learning, multi-centroid learning and learning with sample weight ("OnlineHD"). Yet, most of them have not been tested on the challenging task of epileptic seizure detection, and it stays unclear whether they can increase the HD computing performance to the level of the current state-of-the-art algorithms for wearable devices, such as random forests. Thus, in this paper, we implement different learning strategies and assess their performance on an individual basis, or in combination, regarding detection performance and memory and computational requirements. Results show that the best-performing algorithm, which is a combination of multi-centroid and multi-pass, can indeed reach the performance of the random forest model on a highly unbalanced dataset imitating a real-life epileptic seizure detection application.

Multi-Centroid Hyperdimensional Computing Approach for Epileptic Seizure Detection.

Long-term monitoring of patients with epilepsy presents a challenging problem from the engineering perspective of real-time detection and wearable devices design. It requires new solutions that allow continuous unobstructed monitoring and reliable detection and prediction of seizures. A high variability in the electroencephalogram (EEG) patterns exists among people, brain states, and time instances during seizures, but also during non-seizure periods. This makes epileptic seizure detection very challenging, especially if data is grouped under only seizure (ictal) and non-seizure (inter-ictal) labels. Hyperdimensional (HD) computing, a novel machine learning approach, comes in as a promising tool. However, it has certain limitations when the data shows a high intra-class variability. Therefore, in this work, we propose a novel semi-supervised learning approach based on a multi-centroid HD computing. The multi-centroid approach allows to have several prototype vectors representing seizure and non-seizure states, which leads to significantly improved performance when compared to a simple single-centroid HD model. Further, real-life data imbalance poses an additional challenge and the performance reported on balanced subsets of data is likely to be overestimated. Thus, we test our multi-centroid approach with three different dataset balancing scenarios, showing that performance improvement is higher for the less balanced dataset. More specifically, up to 14% improvement is achieved on an unbalanced test set with 10 times more non-seizure than seizure data. At the same time, the total number of sub-classes is not significantly increased compared to the balanced dataset. Thus, the proposed multi-centroid approach can be an important element in achieving a high performance of epilepsy detection with real-life data balance or during online learning, where seizures are infrequent.

Personalized seizure signature: An interpretable approach to false alarm reduction for long-term epileptic seizure detection.

OBJECTIVE: Long-term automatic detection of focal seizures remains one of the major challenges in epilepsy due to the unacceptably high number of false alarms from state-of-the-art methods. Our aim was to investigate to what extent a new patient-specific approach based on similarly occurring morphological electroencephalographic (EEG) signal patterns could be used to distinguish seizures from nonseizure events, as well as to estimate its maximum performance. METHODS: We evaluated our approach on >5500 h of long-term EEG recordings using two public datasets: the PhysioNet.org Children's Hospital Boston-Massachusetts Institute of Technology (CHB-MIT) Scalp EEG database and the EPILEPSIAE European epilepsy database. We visually identified a set of similarly occurring morphological patterns (seizure signature) seen simultaneously over two different EEG channels, and within two randomly selected seizures from each individual. The same seizure signature was then searched for in the entire recording from the same patient using dynamic time warping (DTW) as a similarity metric, with a threshold set to reflect the maximum sensitivity our algorithm could achieve without false alarm. RESULTS: At a DTW threshold providing no false alarm during the entire recordings, the mean seizure detection sensitivity across patients was 84%, including 96% for the CHB-MIT database and 74% for the European epilepsy database. A 100% sensitivity was reached in 50% of patients, including 79% from the CHB-MIT database and 27% from the European epilepsy database. The median latency from seizure onset to its detection was 17 ± 10 s, with 84% of seizures being detected within 40 s. SIGNIFICANCE: Personalized EEG signature combined with DTW appears to be a promising method to detect ictal events from a limited number of EEG channels with high sensitivity despite low rate of false alarms, high degree of interpretability, and low computational complexity, compatible with its future use in wearable devices.

Personalized Real-Time Federated Learning for Epileptic Seizure Detection.

Epilepsy is one of the most prevalent paroxystic neurological disorders. It is characterized by the occurrence of spontaneous seizures. About 1 out of 3 patients have drug-resistant epilepsy, thus their seizures cannot be controlled by medication. Automatic detection of epileptic seizures can substantially improve the patient's quality of life. To achieve a high-quality model, we have to collect data from various patients in a central server. However, sending the patient's raw data to this central server puts patient privacy at risk and consumes a significant amount of energy. To address these challenges, in this work, we have designed and evaluated a standard federated learning framework in the context of epileptic seizure detection using a deep learning-based approach, which operates across a cluster of machines. We evaluated the accuracy and performance of our proposed approach on the NVIDIA Jetson Nano Developer Kit based on the EPILEPSIAE database, which is one of the largest public epilepsy datasets for seizure detection. Our proposed framework achieved a sensitivity of 81.25%, a specificity of 82.00%, and a geometric mean of 81.62%. It can be implemented on embedded platforms that complete the entire training process in 1.86 hours using 344.34 mAh energy on a single battery charge. We also studied a personalized variant of the federated learning, where each machine is responsible for training a deep neural network (DNN) to learn the discriminative electrocardiography (ECG) features of the epileptic seizures of the specific person monitored based on its local data. In this context, the DNN benefitted from a well-trained model without sharing the patient's raw data with a server or a central cloud repository. We observe in our results that personalized federated learning provides an increase in all the performance metric, with a sensitivity of 90.24%, a specificity of 91.58%, and a geometric mean of 90.90%.

Systematic Assessment of Hyperdimensional Computing for Epileptic Seizure Detection.

Hyperdimensional computing is a promising novel paradigm for low-power embedded machine learning. It has been applied on different biomedical applications, and particularly on epileptic seizure detection. Unfortunately, due to differences in data preparation, segmentation, encoding strategies, and performance metrics, results are hard to compare, which makes building upon that knowledge difficult. Thus, the main goal of this work is to perform a systematic assessment of the HD computing framework for the detection of epileptic seizures, comparing different feature approaches mapped to HD vectors. More precisely, we test two previously implemented features as well as several novel approaches with HD computing on epileptic seizure detection. We evaluate them in a comparable way, i.e., with the same preprocessing setup and with identical performance measures. We use two different datasets in order to assess the generalizability of our conclusions. The systematic assessment involved three primary aspects relevant for potential wearable implementations: 1) detection performance, 2) memory requirements, and 3) computational complexity. Our analysis shows a significant difference in detection performance between approaches, but also that the ones with the highest performance might not be ideal for wearable applications due to their high memory or computational requirements. Furthermore, we evaluate a post-processing strategy to adjust the predictions to the dynamics of epileptic seizures, showing that performance is significantly improved in all the approaches and also that after post-processing, differences in performance are much smaller between approaches.

The COUGHVID crowdsourcing dataset, a corpus for the study of large-scale cough analysis algorithms.

Cough audio signal classification has been successfully used to diagnose a variety of respiratory conditions, and there has been significant interest in leveraging Machine Learning (ML) to provide widespread COVID-19 screening. The COUGHVID dataset provides over 25,000 crowdsourced cough recordings representing a wide range of participant ages, genders, geographic locations, and COVID-19 statuses. First, we contribute our open-sourced cough detection algorithm to the research community to assist in data robustness assessment. Second, four experienced physicians labeled more than 2,800 recordings to diagnose medical abnormalities present in the coughs, thereby contributing one of the largest expert-labeled cough datasets in existence that can be used for a plethora of cough audio classification tasks. Finally, we ensured that coughs labeled as symptomatic and COVID-19 originate from countries with high infection rates. As a result, the COUGHVID dataset contributes a wealth of cough recordings for training ML models to address the world's most urgent health crises.

Interpreting deep learning models for epileptic seizure detection on EEG signals.

While Deep Learning (DL) is often considered the state-of-the art for Artificial Intel-ligence-based medical decision support, it remains sparsely implemented in clinical practice and poorly trusted by clinicians due to insufficient interpretability of neural network models. We have approached this issue in the context of online detection of epileptic seizures by developing a DL model from EEG signals, and associating certain properties of the model behavior with the expert medical knowledge. This has conditioned the preparation of the input signals, the network architecture, and the post-processing of the output in line with the domain knowledge. Specifically, we focused the discussion on three main aspects: (1) how to aggregate the classification results on signal segments provided by the DL model into a larger time scale, at the seizure-level; (2) what are the relevant frequency patterns learned in the first convolutional layer of different models, and their relation with the delta, theta, alpha, beta and gamma frequency bands on which the visual interpretation of EEG is based; and (3) the identification of the signal waveforms with larger contribution towards the ictal class, according to the activation differences highlighted using the DeepLIFT method. Results show that the kernel size in the first layer determines the interpretability of the extracted features and the sensitivity of the trained models, even though the final performance is very similar after post-processing. Also, we found that amplitude is the main feature leading to an ictal prediction, suggesting that a larger patient population would be required to learn more complex frequency patterns. Still, our methodology was successfully able to generalize patient inter-variability for the majority of the studied population with a classification F1-score of 0.873 and detecting 90% of the seizures.

Heartbeat Classification Using Abstract Features From the Abductive Interpretation of the ECG.

OBJECTIVE: This paper aims to prove that automatic beat classification on ECG signals can be effectively solved with a pure knowledge-based approach, using an appropriate set of abstract features obtained from the interpretation of the physiological processes underlying the signal. METHODS: A set of qualitative morphological and rhythm features are obtained for each heartbeat as a result of the abductive interpretation of the ECG. Then, a QRS clustering algorithm is applied in order to reduce the effect of possible errors in the interpretation. Finally, a rule-based classifier assigns a tag to each cluster. RESULTS: The method has been tested with the MIT-BIH Arrhythmia Database records, showing a significantly better performance than any other automatic approach in the state-of-the-art, and even improving most of the assisted approaches that require the intervention of an expert in the process. CONCLUSION: The most relevant issues in ECG classification, related to a large extent to the variability of the signal patterns between different subjects and even in the same subject over time, will be overcome by changing the reasoning paradigm. SIGNIFICANCE: This paper demonstrates the power of an abductive framework for time-series interpretation to make a qualitative leap in the significance of the information extracted from the ECG by automatic methods.

Heart rate variability in patients with severe chronic obstructive pulmonary disease in a home care program.

BACKGROUND: Chronic obstructive pulmonary disease (COPD) patients present functional and structural changes of the respiratory system that have a profound influence on cardiac autonomic dysfunction. OBJETIVE: To analyse heart rate variability in COPD patients under stable condition and during acute exacerbation episodes (AECOPD). METHODS: Twenty three severe COPD male patients, 69.6 ± 7.3 years, in stable condition were followed up for two years. Home visits were carried out by a nurse every month, and home or hospital visits were arranged on demand. Every three months an ECG, oxygen saturation and spirometric recording was obtained for each patient. If the patient presented AECOPD compatible clinical data the same measurements were performed before any change of treatment. Spectral parameters of heart rate variability in time and frequency domains were obtained from ECG. The time evolution of power in low frequency (LF) and high frequency (HF) bands were obtained from the spectrogram. In addition, we calculated the LF/HF ratio and total heart rate variability power (POW). RESULTS: We analysed 154 patient-visit records during the follow up, pertaining to 23 patients and 8 controls; 19 of the patients had experienced at least one AECOPD. Stable COPD patients had higher HF values than control subjects. No significant differences were found in LF, LF/HF ratio or POW variables. AECOPD patients had higher LF, HF and POW than the stable COPD and control groups. CONCLUSION: AECOPD patients exhibited signs of increased autonomic activity compared with stable COPD.

A home-care PDA-based program for the management of COPD patients.

This paper proposes a PDA-based system, SERVANDO, for the home-care management of patients with COPD. In addition to making the supervision of such patients easier, the system permits the collection and structuring of large quantities of information on the evolution of COPD, allowing the application of data-mining techniques, which could open up new lines of research on this disease.