Wearable Bluetooth Triage Healthcare Monitoring System.

Triage is the first interaction between a patient and a nurse/paramedic. This assessment, usually performed at Emergency departments, is a highly dynamic process and there are international grading systems that according to the patient condition initiate the patient journey. Triage requires an initial rapid assessment followed by routine checks of the patients' vitals, including respiratory rate, temperature, and pulse rate. Ideally, these checks should be performed continuously and remotely to reduce the workload on triage nurses; optimizing tools and monitoring systems can be introduced and include a wearable patient monitoring system that is not at the expense of the patient's comfort and can be remotely monitored through wireless connectivity. In this study, we assessed the suitability of a small ceramic piezoelectric disk submerged in a skin-safe silicone dome that enhances contact with skin, to detect wirelessly both respiration and cardiac events at several positions on the human body. For the purposes of this evaluation, we fitted the sensor with a respiratory belt as well as a single lead ECG, all acquired simultaneously. To complete Triage parameter collection, we also included a medical-grade contact thermometer. Performances of cardiac and respiratory events detection were assessed. The instantaneous heart and respiratory rates provided by the proposed sensor, the ECG and the respiratory belt were compared via statistical analyses. In all considered sensor positions, very high performances were achieved for the detection of both cardiac and respiratory events, except for the wrist, which provided lower performances for respiratory rates. These promising yet preliminary results suggest the proposed wireless sensor could be used as a wearable, hands-free monitoring device for triage assessment within emergency departments. Further tests are foreseen to assess sensor performances in real operating environments.

Real-Time EMG Based Pattern Recognition Control for Hand Prostheses: A Review on Existing Methods, Challenges and Future Implementation.

Upper limb amputation is a condition that significantly restricts the amputees from performing their daily activities. The myoelectric prosthesis, using signals from residual stump muscles, is aimed at restoring the function of such lost limbs seamlessly. Unfortunately, the acquisition and use of such myosignals are cumbersome and complicated. Furthermore, once acquired, it usually requires heavy computational power to turn it into a user control signal. Its transition to a practical prosthesis solution is still being challenged by various factors particularly those related to the fact that each amputee has different mobility, muscle contraction forces, limb positional variations and electrode placements. Thus, a solution that can adapt or otherwise tailor itself to each individual is required for maximum utility across amputees. Modified machine learning schemes for pattern recognition have the potential to significantly reduce the factors (movement of users and contraction of the muscle) affecting the traditional electromyography (EMG)-pattern recognition methods. Although recent developments of intelligent pattern recognition techniques could discriminate multiple degrees of freedom with high-level accuracy, their efficiency level was less accessible and revealed in real-world (amputee) applications. This review paper examined the suitability of upper limb prosthesis (ULP) inventions in the healthcare sector from their technical control perspective. More focus was given to the review of real-world applications and the use of pattern recognition control on amputees. We first reviewed the overall structure of pattern recognition schemes for myo-control prosthetic systems and then discussed their real-time use on amputee upper limbs. Finally, we concluded the paper with a discussion of the existing challenges and future research recommendations.

Opportunities for Electroceuticals in Epilepsy.

Epilepsy is a neurological disorder that affects ∼1% of the world population. Nearly 30% of epilepsy patients suffer from pharmacoresistant epilepsy that cannot be treated with antiepileptic drugs. Depending on seizure type, a diverse range of therapies are available, including surgery, vagus nerve stimulation, and deep brain stimulation. We review the sensing and stimulation technologies most used in neurological disorders, and provide a vision of minimally invasive electroceuticals to enable accurate forecasting of epileptic seizures and therapy. The use of such systems could potentially help patients to prevent injuries and, in combination with an intervention mechanism, could provide a method of suppressing seizures in epileptic patients.

Fully Open-Access Passive Dry Electrodes BIOADC: Open-Electroencephalography (EEG) Re-Invented.

The Open-electroencephalography (EEG) framework is a popular platform to enable EEG measurements and general purposes Brain Computer Interface experimentations. However, the current platform is limited by the number of available channels and electrode compatibility. In this paper we present a fully configurable platform with up to 32 EEG channels and compatibility with virtually any kind of passive electrodes including textile, rubber and contactless electrodes. Together with the full hardware details, results and performance on a single volunteer participant (limited to alpha wave elicitation), we present the brain computer interface (BCI)2000 EEG source driver together with source code as well as the compiled (.exe). In addition, all the necessary device firmware, gerbers and bill of materials for the full reproducibility of the presented hardware is included. Furthermore, the end user can vary the dry-electrode reference circuitry, circuit bandwidth as well as sample rate to adapt the device to other generalized biopotential measurements. Although, not implemented in the tested prototype, the Biomedical Analogue to Digital Converter BIOADC naturally supports SPI communication for an additional 32 channels including the gain controller. In the appendix we describe the necessary modification to the presented hardware to enable this function.

A FPGA Implementation of the CAR-FAC Cochlear Model.

This paper presents a digital implementation of the Cascade of Asymmetric Resonators with Fast-Acting Compression (CAR-FAC) cochlear model. The CAR part simulates the basilar membrane's (BM) response to sound. The FAC part models the outer hair cell (OHC), the inner hair cell (IHC), and the medial olivocochlear efferent system functions. The FAC feeds back to the CAR by moving the poles and zeros of the CAR resonators automatically. We have implemented a 70-section, 44.1 kHz sampling rate CAR-FAC system on an Altera Cyclone V Field Programmable Gate Array (FPGA) with 18% ALM utilization by using time-multiplexing and pipeline parallelizing techniques and present measurement results here. The fully digital reconfigurable CAR-FAC system is stable, scalable, easy to use, and provides an excellent input stage to more complex machine hearing tasks such as sound localization, sound segregation, speech recognition, and so on.

Instantaneous VO2 from a wearable device.

We present a method for calculating instantaneous oxygen uptake (VO2) through the use of a non-invasive and non-obtrusive (i.e. without a face mask) wearable device, together with its clinical evaluation against a standard technique based upon expired gas calorimetry. This method can be integrated with existing wearable devices, we implemented it in the "Device for Reliable Energy Expenditure Monitoring" (DREEM). The DREEM comprises a single lead electrocardiogram (ECG) device combined with a tri-axial accelerometer and is worn around the waist. Our clinical evaluation tests the developed method against a gold standard for VO2, expired gas calorimetry, using an ethically approved protocol comprising active exercise and sedentary periods. The study was performed on 42 participants from a wide sample population including healthy people, athletes and an at-risk health group including persons affected by obesity. We developed an algorithm combining heart rate (HR) and the integral of absolute acceleration (IAA), with results showing a correlation of r = 0.93 for instantaneous VO2, and r = 0.97 for 3 min mean VO2, this is a considerably improved estimation of VO2 in comparison to methods utilising HR and IAA independently.

Investigation of Event-Based Surfaces for High-Speed Detection, Unsupervised Feature Extraction, and Object Recognition.

In this work, we investigate event-based feature extraction through a rigorous framework of testing. We test a hardware efficient variant of Spike Timing Dependent Plasticity (STDP) on a range of spatio-temporal kernels with different surface decaying methods, decay functions, receptive field sizes, feature numbers, and back end classifiers. This detailed investigation can provide helpful insights and rules of thumb for performance vs. complexity trade-offs in more generalized networks, especially in the context of hardware implementation, where design choices can incur significant resource costs. The investigation is performed using a new dataset consisting of model airplanes being dropped free-hand close to the sensor. The target objects exhibit a wide range of relative orientations and velocities. This range of target velocities, analyzed in multiple configurations, allows a rigorous comparison of time-based decaying surfaces (time surfaces) vs. event index-based decaying surface (index surfaces), which are used to perform unsupervised feature extraction, followed by target detection and recognition. We examine each processing stage by comparison to the use of raw events, as well as a range of alternative layer structures, and the use of random features. By comparing results from a linear classifier and an ELM classifier, we evaluate how each element of the system affects accuracy. To generate time and index surfaces, the most commonly used kernels, namely event binning kernels, linearly, and exponentially decaying kernels, are investigated. Index surfaces were found to outperform time surfaces in recognition when invariance to target velocity was made a requirement. In the investigation of network structure, larger networks of neurons with large receptive field sizes were found to perform best. We find that a small number of event-based feature extractors can project the complex spatio-temporal event patterns of the dataset to an almost linearly separable representation in feature space, with best performing linear classifier achieving 98.75% recognition accuracy, using only 25 feature extracting neurons.

Do Lifestyle Activities Protect Against Cognitive Decline in Aging? A Review.

The number of patients suffering from dementia is expected to more than triple by the year 2040, and this represents a major challenge to publicly-funded healthcare systems throughout the world. One of the most effective prevention mechanisms against dementia lies in increasing brain- and cognitive-reserve capacity, which has been found to reduce the behavioral severity of dementia symptoms as neurological degeneration progresses. To date though, most of the factors known to enhance this reserve stem from largely immutable history factors, such as level of education and occupational attainment. Here, we review the potential for basic lifestyle activities, including physical exercise, meditation and musical experience, to contribute to reserve capacity and thus reduce the incidence of dementia in older adults. Relative to other therapies, these activities are low cost, are easily scalable and can be brought to market quickly and easily. Overall, although preliminary evidence is promising at the level of randomized control trials, the state of research on this topic remains underdeveloped. As a result, several important questions remain unanswered, including the amount of training required to receive any cognitive benefit from these activities and the extent to which this benefit continues following cessation. Future research directions are discussed for each lifestyle activity, as well as the potential for these and other lifestyle activities to serve as both a prophylactic and a therapeutic treatment for dementia.

Neuromorphic Hardware Architecture Using the Neural Engineering Framework for Pattern Recognition.

We present a hardware architecture that uses the neural engineering framework (NEF) to implement large-scale neural networks on field programmable gate arrays (FPGAs) for performing massively parallel real-time pattern recognition. NEF is a framework that is capable of synthesising large-scale cognitive systems from subnetworks and we have previously presented an FPGA implementation of the NEF that successfully performs nonlinear mathematical computations. That work was developed based on a compact digital neural core, which consists of 64 neurons that are instantiated by a single physical neuron using a time-multiplexing approach. We have now scaled this approach up to build a pattern recognition system by combining identical neural cores together. As a proof of concept, we have developed a handwritten digit recognition system using the MNIST database and achieved a recognition rate of 96.55%. The system is implemented on a state-of-the-art FPGA and can process 5.12 million digits per second. The architecture and hardware optimisations presented offer high-speed and resource-efficient means for performing high-speed, neuromorphic, and massively parallel pattern recognition and classification tasks.

Bayesian Estimation and Inference Using Stochastic Electronics.

In this paper, we present the implementation of two types of Bayesian inference problems to demonstrate the potential of building probabilistic algorithms in hardware using single set of building blocks with the ability to perform these computations in real time. The first implementation, referred to as the BEAST (Bayesian Estimation and Stochastic Tracker), demonstrates a simple problem where an observer uses an underlying Hidden Markov Model (HMM) to track a target in one dimension. In this implementation, sensors make noisy observations of the target position at discrete time steps. The tracker learns the transition model for target movement, and the observation model for the noisy sensors, and uses these to estimate the target position by solving the Bayesian recursive equation online. We show the tracking performance of the system and demonstrate how it can learn the observation model, the transition model, and the external distractor (noise) probability interfering with the observations. In the second implementation, referred to as the Bayesian INference in DAG (BIND), we show how inference can be performed in a Directed Acyclic Graph (DAG) using stochastic circuits. We show how these building blocks can be easily implemented using simple digital logic gates. An advantage of the stochastic electronic implementation is that it is robust to certain types of noise, which may become an issue in integrated circuit (IC) technology with feature sizes in the order of tens of nanometers due to their low noise margin, the effect of high-energy cosmic rays and the low supply voltage. In our framework, the flipping of random individual bits would not affect the system performance because information is encoded in a bit stream.

A 0.04 mm (2) Buck-Boost DC-DC Converter for Biomedical Implants Using Adaptive Gain and Discrete Frequency Scaling Control.

This paper presents the design of a reconfigurable buck-boost switched-capacitor DC-DC converter suitable for use in a wide range of biomedical implants. The proposed converter has an extremely small footprint and uses a novel control method that allows coarse and fine control of the output voltage. The converter uses adaptive gain control, discrete frequency scaling and pulse-skipping schemes to regulate the power delivered to a range of output voltages and loads. Adaptive gain control is used to implement variable switching gain ratios from a reconfigurable power stage and thereby make coarse steps in output voltage. A discrete frequency scaling controller makes discrete changes in switching frequency to vary the power delivered to the load and perform fine tuning when the output voltage is within 10% of the target output voltage. The control architecture is predominately digital and it has been implemented as part of a fully-integrated switched-capacitor converter design using a standard bulk CMOS 0.18 µm process. Measured results show that the converter has an output voltage range of 1.0 to 2.2 V, can deliver up to 7.5 mW of load power and efficiency up to 75% using an active area of only 0.04 mm (2), which is significantly smaller than that of other designs. This low-area, low-complexity reconfigurable power converter can support low-power circuits in biomedical implant applications.

Concept Design for a 1-Lead Wearable/Implantable ECG Front-End: Power Management.

Power supply quality and stability are critical for wearable and implantable biomedical applications. For this reason we have designed a reconfigurable switched-capacitor DC-DC converter that, aside from having an extremely small footprint (with an active on-chip area of only 0.04 mm²), uses a novel output voltage control method based upon a combination of adaptive gain and discrete frequency scaling control schemes. This novel DC-DC converter achieves a measured output voltage range of 1.0 to 2.2 V with power delivery up to 7.5 mW with 75% efficiency. In this paper, we present the use of this converter as a power supply for a concept design of a wearable (15 mm × 15 mm) 1-lead ECG front-end sensor device that simultaneously harvests power and communicates with external receivers when exposed to a suitable RF field. Due to voltage range limitations of the fabrication process of the current prototype chip, we focus our analysis solely on the power supply of the ECG front-end whose design is also detailed in this paper. Measurement results show not just that the power supplied is regulated, clean and does not infringe upon the ECG bandwidth, but that there is negligible difference between signals acquired using standard linear power-supplies and when the power is regulated by our power management chip.

A new biomarker for subthalamic deep brain stimulation for patients with advanced Parkinson's disease--a pilot study.

OBJECTIVE: Deep brain stimulation (DBS) has become the standard treatment for advanced stages of Parkinson's disease (PD) and other motor disorders. Although the surgical procedure has improved in accuracy over the years thanks to imaging and microelectrode recordings, the underlying principles that render DBS effective are still debated today. The aim of this paper is to present initial findings around a new biomarker that is capable of assessing the efficacy of DBS treatment for PD which could be used both as a research tool, as well as in the context of a closed-loop stimulator. APPROACH: We have used a novel multi-channel stimulator and recording device capable of measuring the response of nervous tissue to stimulation very close to the stimulus site with minimal latency, rejecting most of the stimulus artefact usually found with commercial devices. We have recorded and analyzed the responses obtained intraoperatively in two patients undergoing DBS surgery in the subthalamic nucleus (STN) for advanced PD. MAIN RESULTS: We have identified a biomarker in the responses of the STN to DBS. The responses can be analyzed in two parts, an initial evoked compound action potential arising directly after the stimulus onset, and late responses (LRs), taking the form of positive peaks, that follow the initial response. We have observed a morphological change in the LRs coinciding with a decrease in the rigidity of the patients. SIGNIFICANCE: These initial results could lead to a better characterization of the DBS therapy, and the design of adaptive DBS algorithms that could significantly improve existing therapies and help us gain insights into the functioning of the basal ganglia and DBS.

Sound stream segregation: a neuromorphic approach to solve the "cocktail party problem" in real-time.

The human auditory system has the ability to segregate complex auditory scenes into a foreground component and a background, allowing us to listen to specific speech sounds from a mixture of sounds. Selective attention plays a crucial role in this process, colloquially known as the "cocktail party effect." It has not been possible to build a machine that can emulate this human ability in real-time. Here, we have developed a framework for the implementation of a neuromorphic sound segregation algorithm in a Field Programmable Gate Array (FPGA). This algorithm is based on the principles of temporal coherence and uses an attention signal to separate a target sound stream from background noise. Temporal coherence implies that auditory features belonging to the same sound source are coherently modulated and evoke highly correlated neural response patterns. The basis for this form of sound segregation is that responses from pairs of channels that are strongly positively correlated belong to the same stream, while channels that are uncorrelated or anti-correlated belong to different streams. In our framework, we have used a neuromorphic cochlea as a frontend sound analyser to extract spatial information of the sound input, which then passes through band pass filters that extract the sound envelope at various modulation rates. Further stages include feature extraction and mask generation, which is finally used to reconstruct the targeted sound. Using sample tonal and speech mixtures, we show that our FPGA architecture is able to segregate sound sources in real-time. The accuracy of segregation is indicated by the high signal-to-noise ratio (SNR) of the segregated stream (90, 77, and 55 dB for simple tone, complex tone, and speech, respectively) as compared to the SNR of the mixture waveform (0 dB). This system may be easily extended for the segregation of complex speech signals, and may thus find various applications in electronic devices such as for sound segregation and speech recognition.

A neuromorphic implementation of multiple spike-timing synaptic plasticity rules for large-scale neural networks.

We present a neuromorphic implementation of multiple synaptic plasticity learning rules, which include both Spike Timing Dependent Plasticity (STDP) and Spike Timing Dependent Delay Plasticity (STDDP). We present a fully digital implementation as well as a mixed-signal implementation, both of which use a novel dynamic-assignment time-multiplexing approach and support up to 2(26) (64M) synaptic plasticity elements. Rather than implementing dedicated synapses for particular types of synaptic plasticity, we implemented a more generic synaptic plasticity adaptor array that is separate from the neurons in the neural network. Each adaptor performs synaptic plasticity according to the arrival times of the pre- and post-synaptic spikes assigned to it, and sends out a weighted or delayed pre-synaptic spike to the post-synaptic neuron in the neural network. This strategy provides great flexibility for building complex large-scale neural networks, as a neural network can be configured for multiple synaptic plasticity rules without changing its structure. We validate the proposed neuromorphic implementations with measurement results and illustrate that the circuits are capable of performing both STDP and STDDP. We argue that it is practical to scale the work presented here up to 2(36) (64G) synaptic adaptors on a current high-end FPGA platform.

Turn Down That Noise: Synaptic Encoding of Afferent SNR in a Single Spiking Neuron.

We have added a simplified neuromorphic model of Spike Time Dependent Plasticity (STDP) to the previously described Synapto-dendritic Kernel Adapting Neuron (SKAN), a hardware efficient neuron model capable of learning spatio-temporal spike patterns. The resulting neuron model is the first to perform synaptic encoding of afferent signal-to-noise ratio in addition to the unsupervised learning of spatio-temporal spike patterns. The neuron model is particularly suitable for implementation in digital neuromorphic hardware as it does not use any complex mathematical operations and uses a novel shift-based normalization approach to achieve synaptic homeostasis. The neuron's noise compensation properties are characterized and tested on random spatio-temporal spike patterns as well as a noise corrupted subset of the zero images of the MNIST handwritten digit dataset. Results show the simultaneously learning common patterns in its input data while dynamically weighing individual afferents based on their signal to noise ratio. Despite its simplicity the interesting behaviors of the neuron model and the resulting computational power may also offer insights into biological systems.

A novel method for non-invasive respiration monitoring.

In this paper we present an atypical method for measuring respiration volume. We infer heart rate variability (HRV) from an electrocardiogram (ECG) and present results from a pilot study of 6 participants to validate measuring respiration volume from HRV in comparison to the Cosmed K4b(2). We show a qualitative correlation and trend between the known respiration volume as measured by the Cosmed K4b(2) and the new method for measuring lung volume. From these results, we propose guidelines for an in-depth study of measuring respiration volumes from heart rate variability.

Racing to learn: statistical inference and learning in a single spiking neuron with adaptive kernels.

This paper describes the Synapto-dendritic Kernel Adapting Neuron (SKAN), a simple spiking neuron model that performs statistical inference and unsupervised learning of spatiotemporal spike patterns. SKAN is the first proposed neuron model to investigate the effects of dynamic synapto-dendritic kernels and demonstrate their computational power even at the single neuron scale. The rule-set defining the neuron is simple: there are no complex mathematical operations such as normalization, exponentiation or even multiplication. The functionalities of SKAN emerge from the real-time interaction of simple additive and binary processes. Like a biological neuron, SKAN is robust to signal and parameter noise, and can utilize both in its operations. At the network scale neurons are locked in a race with each other with the fastest neuron to spike effectively "hiding" its learnt pattern from its neighbors. The robustness to noise, high speed, and simple building blocks not only make SKAN an interesting neuron model in computational neuroscience, but also make it ideal for implementation in digital and analog neuromorphic systems which is demonstrated through an implementation in a Field Programmable Gate Array (FPGA). Matlab, Python, and Verilog implementations of SKAN are available at: http://www.uws.edu.au/bioelectronics_neuroscience/bens/reproducible_research.

On the fluid-structure interaction in the cochlea.

The cochlea is known to be a nonlinear system that shows strong fluid-structure coupling. In this work, the monolithic state space approach to cochlear modeling [Rapson et al., J. Acoust. Soc. Am. 131, 3925-3952 (2012)] is used to study the inherent nature of this coupling. Mathematical derivations requiring minimal, widely accepted assumptions about cochlear anatomy provide a clear description of the coupling. In particular, the coupling forces between neighboring cochlear partition segments are demonstrated, with implications for theories of cochlear operation that discount the traveling wave hypothesis. The derivations also reaffirm the importance of selecting a physiologically accurate value for the partition mass in any simulation. Numerical results show that considering the fluid properties in isolation can give a misleading impression of the fluid-structure coupling. Linearization of a nonlinear partition model allows the relationship between the linear and nonlinear fluid-structure interaction to be described. Furthermore, the effect of different classes of nonlinearities on the numerical complexity of a cochlear model is assessed. Cochlear models that assume outer hair cells are able to detect pressure will require implicit solver strategies, should the pressure sensitivity be demonstrated. Classical cochlear models in general do not require implicit solver strategies.

A mixed-signal implementation of a polychronous spiking neural network with delay adaptation.

We present a mixed-signal implementation of a re-configurable polychronous spiking neural network capable of storing and recalling spatio-temporal patterns. The proposed neural network contains one neuron array and one axon array. Spike Timing Dependent Delay Plasticity is used to fine-tune delays and add dynamics to the network. In our mixed-signal implementation, the neurons and axons have been implemented as both analog and digital circuits. The system thus consists of one FPGA, containing the digital neuron array and the digital axon array, and one analog IC containing the analog neuron array and the analog axon array. The system can be easily configured to use different combinations of each. We present and discuss the experimental results of all combinations of the analog and digital axon arrays and the analog and digital neuron arrays. The test results show that the proposed neural network is capable of successfully recalling more than 85% of stored patterns using both analog and digital circuits.