Combining detrended cross-correlation analysis with Riemannian geometry-based classification for improved brain-computer interface performance.

Riemannian geometry-based classification (RGBC) gained popularity in the field of brain-computer interfaces (BCIs) lately, due to its ability to deal with non-stationarities arising in electroencephalography (EEG) data. Domain adaptation, however, is most often performed on sample covariance matrices (SCMs) obtained from EEG data, and thus might not fully account for components affecting covariance estimation itself, such as regional trends. Detrended cross-correlation analysis (DCCA) can be utilized to estimate the covariance structure of such signals, yet it is computationally expensive in its original form. A recently proposed online implementation of DCCA, however, allows for its fast computation and thus makes it possible to employ DCCA in real-time applications. In this study we propose to replace the SCM with the DCCA matrix as input to RGBC and assess its effect on offline and online BCI performance. First we evaluated the proposed decoding pipeline offline on previously recorded EEG data from 18 individuals performing left and right hand motor imagery (MI), and benchmarked it against vanilla RGBC and popular MI-detection approaches. Subsequently, we recruited eight participants (with previous BCI experience) who operated an MI-based BCI (MI-BCI) online using the DCCA-enhanced Riemannian decoder. Finally, we tested the proposed method on a public, multi-class MI-BCI dataset. During offline evaluations the DCCA-based decoder consistently and significantly outperformed the other approaches. Online evaluation confirmed that the DCCA matrix could be computed in real-time even for 22-channel EEG, as well as subjects could control the MI-BCI with high command delivery (normalized Cohen's κ: 0.7409 ± 0.1515) and sample-wise MI detection (normalized Cohen's κ: 0.5200 ± 0.1610). Post-hoc analysis indicated characteristic connectivity patterns under both MI conditions, with stronger connectivity in the hemisphere contralateral to the MI task. Additionally, fractal scaling exponent of neural activity was found increased in the contralateral compared to the ipsilateral motor cortices (C4 and C3 for left and right MI, respectively) in both classes. Combining DCCA with Riemannian geometry-based decoding yields a robust and effective decoder, that not only improves upon the SCM-based approach but can also provide relevant information on the neurophysiological processes behind MI.

Transfer learning promotes acquisition of individual BCI skills.

Subject training is crucial for acquiring brain-computer interface (BCI) control. Typically, this requires collecting user-specific calibration data due to high inter-subject neural variability that limits the usability of generic decoders. However, calibration is cumbersome and may produce inadequate data for building decoders, especially with naïve subjects. Here, we show that a decoder trained on the data of a single expert is readily transferrable to inexperienced users via domain adaptation techniques allowing calibration-free BCI training. We introduce two real-time frameworks, (i) Generic Recentering (GR) through unsupervised adaptation and (ii) Personally Assisted Recentering (PAR) that extends GR by employing supervised recalibration of the decoder parameters. We evaluated our frameworks on 18 healthy naïve subjects over five online sessions, who operated a customary synchronous bar task with continuous feedback and a more challenging car racing game with asynchronous control and discrete feedback. We show that along with improved task-oriented BCI performance in both tasks, our frameworks promoted subjects' ability to acquire individual BCI skills, as the initial neurophysiological control features of an expert subject evolved and became subject specific. Furthermore, those features were task-specific and were learned in parallel as participants practiced the two tasks in every session. Contrary to previous findings implying that supervised methods lead to improved online BCI control, we observed that longitudinal training coupled with unsupervised domain matching (GR) achieved similar performance to supervised recalibration (PAR). Therefore, our presented frameworks facilitate calibration-free BCIs and have immediate implications for broader populations-such as patients with neurological pathologies-who might struggle to provide suitable initial calibration data.

A Novel Form Factor For PPG-based Blood Pressure Monitoring.

Blood pressure (BP) is one of the four main vital signs in medicine and may be a useful signal for wellness tracking and for user-aware interfaces in human-computer interaction. The current standard for BP measurement uses cuff-based devices that block an artery temporarily to get a single, discrete measurement of BP. Recently, there have been significant efforts to measure correlates of BP continuously and non-invasively from relevant signals like photoplethysmography (PPG), which responds to volumetric changes in arteries due to blood pulsations. In this paper, we explore a novel setup with two points of instrumentation, one on the head and a second on the wrist, for recording PPG and estimating the pulse wave velocity, which is a major correlate of BP, along with other waveform-related features. We prospectively tested the device on 10 subjects who followed a protocol for the deliberate variation of BP while ground truth measurements were taken using a reference cuff-device. Generic absolute BP models, which use the collected data for leave-one-subject-out cross-validation, yielded an error of -0.14 ± 7.3 mmHg for systolic BP (SBP) and -0.21±6.7 mmHg for diastolic BP (DBP), which are within the regulatory limits of 5 ± 8 mmHg. Notably, the correlation between the predicted BPs and the ground truth BPs was higher for SBP (r = 0.74, p < 0.001) than for DBP (r = 0.34, p < 0.001). The results show that the proposed form factor can extract BP-related features that could be used for continuous, cuff-less BP monitoring.

A highly stable electrode with low electrode-skin impedance for wearable brain-computer interface.

To date, brain-computer interfaces (BCIs) have proved to play a key role in many medical applications, for example, the rehabilitation of stroke patients. For post-stroke rehabilitation, the BCIs require the EEG electrodes to precisely translate the brain signals of patients into intended movements of the paralyzed limb for months. However, the gold standard silver/silver-chloride electrodes cannot satisfy the requirements for long-term stability and preparation-free recording capability in wearable EEG devices, thus limiting the versatility of EEG in wearable BCI applications over time outside the rehabilitation center. Here, we design a long-term stable and low electrode-skin interfacial impedance conductive polymer-hydrogel EEG electrode that maintains a lower impedance value than gel-based electrodes for 29 days. With this technology, EEG-based long-term and wearable BCIs could be realized in the near future. To demonstrate this, our designed electrode is applied for a wireless single-channel EEG device that detects changes in alpha rhythms in eye-open/eye-close conditions. In addition, we validate that the designed electrodes could capture oscillatory rhythms in motor imagery protocols as well as low-frequency time-locked event-related potentials from healthy subjects, with similar or better performance than gel-based electrodes. Finally, we demonstrate the use of the designed electrode in online BCI-based functional electrical stimulation, which could be used for post-stroke rehabilitation.

Towards Point-of-Care Heart Failure Diagnostic Platforms: BNP and NT-proBNP Biosensors.

Heart failure is a class of cardiovascular diseases that remains the number one cause of death worldwide with a substantial economic burden of around $18 billion incurred by the healthcare sector in 2017 due to heart failure hospitalization and disease management. Although several laboratory tests have been used for early detection of heart failure, these traditional diagnostic methods still fail to effectively guide clinical decisions, prognosis, and therapy in a timely and cost-effective manner. Recent advances in the design and development of biosensors coupled with the discovery of new clinically relevant cardiac biomarkers are paving the way for breakthroughs in heart failure management. Natriuretic neurohormone peptides, B-type natriuretic peptide (BNP) and N-terminal prohormone of BNP (NT-proBNP), are among the most promising biomarkers for clinical use. Remarkably, they result in an increased diagnostic accuracy of around 80% owing to the strong correlation between their circulating concentrations and different heart failure events. The latter has encouraged research towards developing and optimizing BNP biosensors for rapid and highly sensitive detection in the scope of point-of-care testing. This review sheds light on the advances in BNP and NT-proBNP sensing technologies for point-of-care (POC) applications and highlights the challenges of potential integration of these technologies in the clinic. Optical and electrochemical immunosensors are currently used for BNP sensing. The performance metrics of these biosensors-expressed in terms of sensitivity, selectivity, reproducibility, and other criteria-are compared to those of traditional diagnostic techniques, and the clinical applicability of these biosensors is assessed for their potential integration in point-of-care diagnostic platforms.

A Real-Time ECG Feature Extraction Algorithm for Detecting Meditation Levels within a General Measurement Setup.

This paper presents a setup for the real-time extraction of Electroencephalography (EEG) and Electrocardiogram (ECG) features indicating the level of focus, relaxation, or meditation of a given subject. An algorithm for detecting meditation in real-time using the extracted ECG features is designed and shown to lead to accurate results using an online ECG measurement dataset. Similar methods can be used for EEG data, such that the proposed measurement setup can be used, for example, for investigating the effect of virtual reality based EEG training, with and without neurofeedback, on the capability of subjects to focus, relax, or meditate.