Four State Sleep Staging From a Multilayered Algorithm Using Electrocardiographic and Actigraphic Data

PURPOSE

Sleep studies are important to evaluate sleep and sleep-related disorders. The standard test for evaluating sleep is polysomnography, during which several physiological signals are recorded separately and simultaneously with specialized equipment that requires a technologist. Simpler recordings that can model the results of a polysomnography would provide the benefit of expanding the possibilities of sleep recordings.

METHODS

Using the publicly available sleep data set from the multiethnic study of atherosclerosis and 1769 nights of sleep, we extracted a distinct data subset with engineered features of the biomarkers collected by actigraphic, oxygenation, and electrocardiographic sensors. We then applied scalable models with recurrent neural network and Extreme Gradient Boosting (XGBoost) with a layered approach to produce an algorithm that we then validated with a separate data set of 177 nights.

RESULTS

The algorithm achieved an overall performance of 0.833 accuracy and 0.736 kappa in classifying into four states: wake, light sleep, deep sleep, and rapid eye movement (REM). Using feature analysis, we demonstrated that heart rate variability is the most salient feature, which is similar to prior reports.

CONCLUSIONS

Our results demonstrate the potential benefit of a multilayered algorithm and achieved higher accuracy and kappa than previously described approaches for staging sleep. The results further the possibility of simple, wearable devices for sleep staging. Code is available at https://github.com/NovelaNeuro/nEureka-SleepStaging.

Classification of directionally specific vagus nerve activity using an upper airway obstruction model in anesthetized rodents

Electrical signals from the peripheral nervous system have the potential to provide the necessary motor, sensory or autonomic information for implementing closed-loop control of neuroprosthetic or neuromodulatory systems. However, developing methods to recover information encoded in these signals is a significant challenge. Our goal was to test the feasibility of measuring physiologically generated nerve action potentials that can be classified as sensory or motor signals. A tetrapolar recording nerve cuff electrode was used to measure vagal nerve (VN) activity in a rodent model of upper airway obstruction. The effect of upper airway occlusions on VN activity related to respiration (RnP) was calculated and compared for 4 different cases: (1) intact VN, (2) VN transection only proximal to recording electrode, (3) VN transection only distal to the recording electrode, and (4) transection of VN proximal and distal to electrode. We employed a Support Vector Machine (SVM) model with Gaussian Kernel to learn a model capable of classifying efferent and afferent waveforms obtained from the tetrapolar electrode. In vivo results showed that the RnP values decreased significantly during obstruction by 91.7% ± 3.1%, and 78.2% ± 3.4% for cases of intact VN or proximal transection, respectively. In contrast, there were no significant changes for cases of VN transection at the distal end or both ends of the electrode. The SVM model yielded an 85.8% accuracy in distinguishing motor and sensory signals. The feasibility of measuring low-noise directionally-sensitive neural activity using a tetrapolar nerve cuff electrode along with the use of an SVM classifier was shown. Future experimental work in chronic implant studies is needed to support clinical translatability.

Reduced genioglossus muscle activity caused by fluid overload in anesthetized rats

INTRODUCTION

Although the precise cause of obstructive sleep apnea (OSA) remains unknown, various anatomical or structural factors are thought to influence upper airway patency. Recent clinical studies show that OSA is frequently observed among patients with fluid-retaining states, such as heart/renal failure and postsurgery. It is important to note that a cause-effect relationship is not yet established, and our understanding of the effects of fluid overload is limited. The goal of this study was to investigate an animal model that can characterize the physiological changes that occur in response to fluid overload.

METHOD

Acute nonsurvival experiments were conducted in 16 Sprague-Dawley rats. Rats were initially anesthetized by inhaled isoflurane, while the femoral vein was cannulated and urethane (1.2-1.5 g/Kg body weight) was gradually delivered intravenously to induce anesthesia. Additional doses of urethane were delivered as necessary to maintain a surgical plane of anesthesia. A surgical incision was made on the cervical area to catheterize carotid artery to measure blood pressure. A pair of stainless-steel wires was injected into the tongue to measure genioglossus muscle activity (GGEMG). All physiological measurements were recorded as intravenous infusion of saline was provided to the rat (infusion rate = 22 ml/kg over 30 min).

RESULTS

Acute saline overloading resulted in a 33% decrease in GGEMG, when compared to baseline. There was also a gradual drop in the respiratory rate (13% decrease) that reached statistical significance at 10 min after infusion was stopped. The blood pressure exhibited a 14% increase which subsequently returned to baseline within 40 min stopping infusion. There were no significant changes in the heart rate.

CONCLUSION

The results of this study indicate that systemic fluid overload can affect significant changes in different physiological systems including reduction in genioglossus muscle activity, increase in blood pressure, and change autonomic nervous system function.

Feasibility of differentially measuring afferent and efferent neural activity with a single nerve cuff electrode

OBJECTIVE

Advances in electrode technology have facilitated the development of neuroprostheses for restoring motor/sensory function in disabled individuals. Information extracted from a whole nerve, recorded using cuffs, can provide signals that control the operation of neuroprostheses. However, the amount of information that can be extracted from a tripolar cuff-which provides the highest signal-to-noise ratio (SNR)-is limited. The physical symmetry of the tripolar cuff results in neural recordings that cannot differentiate afferent versus efferent signals. In this study, we introduced a tetrapolar cuff to achieve low-noise and directionally sensitive recording.

APPROACH

The tetrapolar cuff was initially designed using a computational approach. A finite element model was used to solve the electric potential generated at the electrode contacts by active electrical sources, such as the nodes of Ranvier and an artifact noise source. The resulting single fiber action potentials (SFAPs) and artifact noise signals (ANS) were used to characterize the performance of the tetrapolar configuration of the electrode length (EL) and electrode edge length (EEL) on simulated SFAP and ANS. The feasibility of using a tetrapolar cuff to differentiate afferent/efferent action potentials by applying potassium chloride in anesthetized rats was also investigated.

MAIN RESULTS

Both the computational and experimental results of this study indicated that directional recording can be achieved using a tetrapolar cuff. Testing different design criteria (e.g. EL and EEL) showed that at EL values above 15 mm and EEL  ⩾  2 mm, the tetrapolar cuffs can yield larger SNRs than equally-sized tripolar cuffs.

SIGNIFICANCE

This study indicated that low-noise directionally sensitive measurement of neural activity can be achieved with a tetrapolar cuff. The experimental results confirmed the feasibility of using a tetrapolar cuff to differentiate afferent/efferent signals by applying potassium chloride. Further work is needed to determine whether the tetrapolar cuff can differentiate afferent/efferent physiologically elicited neural activities.

Optimizing the design of bipolar nerve cuff electrodes for improved recording of peripheral nerve activity

OBJECTIVE

Differential measurement of efferent and afferent peripheral nerve activity offers a promising means of improving the clinical utility of implantable neuroprostheses. The tripolar nerve cuff electrode has historically served as the gold standard for achieving high signal-to-noise ratios (SNRs) of the recordings. However, the symmetrical geometry of this electrode array (i.e. electrically-shorted side contacts) precludes it from measuring electrical signals that can be used to obtain directional information. In this study, we investigated the feasibility of using a bipolar nerve cuff electrode to achieve high-SNR of peripheral nerve activity.

APPROACH

A finite element model was implemented to investigate the effects of electrode design parameters-electrode length, electrode edge length (EEL), and a conductive shielding layer (CSL)-on simulated single fiber action potentials (SFAP) and also artifact noise signals (ANS).

MAIN RESULTS

Our model revealed that the EEL was particularly effective in increasing the peak-to-peak amplitude of the SFAP (319%) and reducing the common mode ANS (67%) of the bipolar cuff electrode. By adding a CSL to the bipolar cuff electrode, the SNR was found to be 65.2% greater than that of a conventional tripolar cuff electrode. In vivo experiments in anesthetized rats confirmed that a bipolar cuff electrode can achieve a SNR that is 38% greater than that achieved by a conventional tripolar cuff electrode (p  <  0.05).

SIGNIFICANCE

The current study showed that bipolar nerve cuff electrodes can be designed to achieve SNR levels that are comparable to that of tripolar configuration. Further work is needed to confirm that these bipolar design parameters can be used to record bi-directional neural activity in a physiological setting.

Directionally-sensitive peripheral nerve recording: bipolar nerve cuff design

Although bipolar nerve cuff electrodes (NCEs) are used in clinical neurostimulation therapies (e.g., epilepsy, obstructive sleep apnea), the use of this electrode as part of a closed-up controlled system is very limited. The literature points to high noise content (i.e., poor signal-to-noise ratio, SNR) as the primary reason for not using bipolar NCEs to record neural activity. In fact, over the past several decades, the pseudo-tripolar NCEs, the symmetric configuration of this electrode array (i.e., electrically-shorted side contacts) makes it very difficult to obtain directional information from the recorded neural activity (efferent vs. afferent). In this study, we investigated the feasibility of using a bipolar NCE to achieve low-noise, directionally-sensitive recording of peripheral nerve activity.

Survey on the adequacy of depth of anaesthesia with bispectral index and isolated forearm technique in elective Caesarean section under general anaesthesia with sevoflurane

BACKGROUND

Awareness during general anaesthesia for Caesarean section (C/S), although uncommon, remains a concern for anaesthesiologists. We examined the relationship between the bispectral index (BIS) and responses to the isolated forearm technique (IFT) to evaluate the adequacy of general anaesthesia in C/S and determine a suitable cut-off point for BIS values based on IFT results.

METHODS

In 61 parturients, a standardized anaesthetic technique was applied. It included sodium thiopental and succinylcholine for induction, and O2, N2O, and sevoflurane for maintenance of anaesthesia. BIS values and IFT response were recorded at 16 predetermined events during anaesthesia.

RESULTS

Positive IFT responses were seen in 41%, 46%, and 23% of the parturients at laryngoscopy, intubation, and skin incision, respectively. BIS could not reliably differentiate between IFT responders and non-responders during these three stages. The receiver operating characteristic curve cut-off points for BIS to predict IFT responders with 100% sensitivity were 34, 37, and 27, respectively, for these stages. In all stages of the operation after skin incision, more than 90% of parturients had no IFT test response, and BIS values between 40 and 63 were associated with negative IFT results. During a structured interview within 12-24 h after the operation, no patient had evidence of explicit recall of intraoperative events.

CONCLUSIONS

The BIS is not reliable for monitoring anaesthesia depth in C/S. Lower than previously recommended values are needed to avoid IFT test responses during laryngoscopy, intubation, and skin incision.

Parametric modelling of cardiac system multiple measurement signals: an open-source computer framework for performance evaluation of ECG, PCG and ABP event detectors

The major focus of this study is to present a performance accuracy assessment framework based on mathematical modelling of cardiac system multiple measurement signals. Three mathematical algebraic subroutines with simple structural functions for synthetic generation of the synchronously triggered electrocardiogram (ECG), phonocardiogram (PCG) and arterial blood pressure (ABP) signals are described. In the case of ECG signals, normal and abnormal PQRST cycles in complicated conditions such as fascicular ventricular tachycardia, rate dependent conduction block and acute Q-wave infarctions of inferior and anterolateral walls can be simulated. Also, continuous ABP waveform with corresponding individual events such as systolic, diastolic and dicrotic pressures with normal or abnormal morphologies can be generated by another part of the model. In addition, the mathematical synthetic PCG framework is able to generate the S4-S1-S2-S3 cycles in normal and in cardiac disorder conditions such as stenosis, insufficiency, regurgitation and gallop. In the PCG model, the amplitude and frequency content (5-700 Hz) of each sound and variation patterns can be specified. The three proposed models were implemented to generate artificial signals with varies abnormality types and signal-to-noise ratios (SNR), for quantitative detection-delineation performance assessment of several ECG, PCG and ABP individual event detectors designed based on the Hilbert transform, discrete wavelet transform, geometric features such as area curve length (ACLM), the multiple higher order moments (MHOM) metric, and the principal components analysed geometric index (PCAGI). For each method the detection-delineation operating characteristics were obtained automatically in terms of sensitivity, positive predictivity and delineation (segmentation) error rms and checked by the cardiologist. The Matlab m-file script of the synthetic ECG, ABP and PCG signal generators are available in the Appendix.