A wrist-worn biosensor system for assessment of neurological status

Nonlinear Features from Multi-Modal Signals for Continuous Stress Monitoring

Continuous monitoring of stress in individuals during their daily activities has become an inevitable need in present times. Unattended stress is a silent killer and may lead to fatal physical and mental disorders if left unidentified. Stress identification based on individual judgement often leads to under-diagnosis and delayed treatment possibilities. EEG-based stress monitoring is quite popular in this context, but impractical to use for continuous remote monitoring.Continuous remote monitoring of stress using signals acquired from everyday wearables like smart watches is the best alternative here. Non-EEG data such as heart rate and ectodermal activity can also act as indicators of physiological stress. In this work, we have explored the possibility of using nonlinear features from non-EEG data such as (a) heart rate, (b) ectodermal activity, (c) body temperature (d) SpO2 and (e) acceleration in detecting four different types of neurological states; namely (1) Relaxed state, (2) State of Physical stress, (3) State of Cognitive stress and (4) State of Emotional stress. Physiological data of 20 healthy adults have been used from the noneeg database of PhysioNet.Results: We used two machine learning models; a linear logistic regression and a nonlinear random forest to detect (a) stress from relaxed state and (4) the four different neurological states. We trained the models using linear and nonlinear features separately. For the 2-class and 4-class problems, using nonlinear features increased the accuracy of the models. Moreover, it is also proved in this study that by using nonlinear features, we can avoid the use of complex machine learning models.

Universal Physiological Representation Learning With Soft-Disentangled Rateless Autoencoders

Human computer interaction (HCI) involves a multidisciplinary fusion of technologies, through which the control of external devices could be achieved by monitoring physiological status of users. However, physiological biosignals often vary across users and recording sessions due to unstable physical/mental conditions and task-irrelevant activities. To deal with this challenge, we propose a method of adversarial feature encoding with the concept of a Rateless Autoencoder (RAE), in order to exploit disentangled, nuisance-robust, and universal representations. We achieve a good trade-off between user-specific and task-relevant features by making use of the stochastic disentanglement of the latent representations by adopting additional adversarial networks. The proposed model is applicable to a wider range of unknown users and tasks as well as different classifiers. Results on cross-subject transfer evaluations show the advantages of the proposed framework, with up to an 11.6% improvement in the average subject-transfer classification accuracy.

Disentangled Adversarial Transfer Learning for Physiological Biosignals

Recent developments in wearable sensors demonstrate promising results for monitoring physiological status in effective and comfortable ways. One major challenge of physiological status assessment is the problem of transfer learning caused by the domain inconsistency of biosignals across users or different recording sessions from the same user. We propose an adversarial inference approach for transfer learning to extract disentangled nuisance-robust representations from physiological biosignal data in stress status level assessment. We exploit the trade-off between task-related features and person-discriminative information by using both an adversary network and a nuisance network to jointly manipulate and disentangle the learned latent representations by the encoder, which are then input to a discriminative classifier. Results on cross-subjects transfer evaluations demonstrate the benefits of the proposed adversarial framework, and thus show its capabilities to adapt to a broader range of subjects. Finally we highlight that our proposed adversarial transfer learning approach is also applicable to other deep feature learning frameworks.

Disentangled Adversarial Autoencoder for Subject-Invariant Physiological Feature Extraction

Recent developments in biosignal processing have enabled users to exploit their physiological status for manipulating devices in a reliable and safe manner. One major challenge of physiological sensing lies in the variability of biosignals across different users and tasks. To address this issue, we propose an adversarial feature extractor for transfer learning to exploit disentangled universal representations. We consider the trade-off between task-relevant features and user-discriminative information by introducing additional adversary and nuisance networks in order to manipulate the latent representations such that the learned feature extractor is applicable to unknown users and various tasks. Results on cross-subject transfer evaluations exhibit the benefits of the proposed framework, with up to 8.8% improvement in average accuracy of classification, and demonstrate adaptability to a broader range of subjects.

Epileptic seizure detection using wristworn biosensors

Single signal seizure detection algorithms suffer from high false positive rates. We have found a set of signals which can be easily monitored by a wristworn device and which produce a distinctive pattern during seizure for patients in an epilepsy monitoring unit (EMU). This pattern is much less likely to be reproduced by nonseizure events in the patient's daily life than are changes in heart rate alone. We collected 108 hours of data from three EMU patients who suffered a combined total of seven seizures, then developed a time series analysis/pattern recognition based algorithm which distinguishes the seizures from nonseizure events with 100% accuracy.

Sleep state classification using pressure sensor mats

Sleep state detection is valuable in assessing patient's sleep quality and in-bed general behavior. In this paper, a novel classification approach of sleep states (sleep, pre-wake, wake) is proposed that uses only surface pressure sensors. In our method, a mobility metric is defined based on successive pressure body maps. Then, suitable statistical features are computed based on the mobility metric. Finally, a customized random forest classifier is employed to identify various classes including a new class for pre-wake state. Our algorithm achieves 96.1% and 88% accuracies for two (sleep, wake) and three (sleep, pre-wake, wake) class identification, respectively.