From Seizure Detection to Smart and Fully Embedded Seizure Prediction Engine: A Review

A multi-domain feature fusion epilepsy seizure detection method based on spike matching and PLV functional networks

Objective.The identification of spikes, as a typical characteristic wave of epilepsy, is crucial for diagnosing and locating the epileptogenic region. The traditional seizure detection methods lack spike features and have low sample richness. This paper proposes a seizure detection method with spike-based phase locking value (PLV) functional brain networks and multi-domain fused features.Approach.In the spiking detection part, brain functional networks based on PLV are constructed to explore the changes in brain functional states during spiking discharge, from the perspective of microscopic neuronal activity to macroscopic brain region interactions. Then, in the epilepsy seizure detection task, multi-domain fused feature sequences are constructed using time-domain, frequency-domain, inter-channel correlation, and the spike detection features. Finally, Bi-LSTM and Transformer encoders and their optimized models are used to verify the effectiveness of the proposed method.Main results.Experimental results achieve the best seizure detection metrics on Bi-LSTM-Attention, with accuracy, sensitivity, and specificity reaching 98.40%, 98.94%, and 97.86%, respectively.Significance.The method is significant as it innovatively applies multi channel spike network features to seizure detection. It can potentially improve the diagnosis and location of the epileptogenic region by accurately detecting seizures through the identification of spikes, which is a crucial characteristic wave of epilepsy.

Field-Programmable Gate Array (FPGA)-Based Lock-In Amplifier System with Signal Enhancement: A Comprehensive Review on the Design for Advanced Measurement Applications

Lock-in amplifiers (LIAs) are critical tools in precision measurement, particularly for applications involving weak signals obscured by noise. Advances in signal processing algorithms and hardware synthesis have enabled accurate signal extraction, even in extremely noisy environments, making LIAs indispensable in sensor applications for healthcare, industry, and other services. For instance, the electrical impedance measurement of the human body, organs, tissues, and cells, known as bioelectrical impedance, is commonly used in biomedical and healthcare applications because it is non-invasive and relatively inexpensive. Also, due to its portability and miniaturization capabilities, it has great potential for the development of new point-of-care and portable testing devices. In this document, we highlight existing techniques for high-frequency resolution and precise phase detection in LIA reference signals from field-programmable gate array (FPGA) designs. A comprehensive review is presented under the key requirements and techniques for single- and dual-phase digital LIA architectures, where relevant insights are provided to address the LIAs' digital precision in measurement system configurations. Furthermore, the document highlights a novel method to enhance the spurious-free dynamic range (SFDR), thereby advancing the precision and effectiveness of LIAs in complex measurement environments. Finally, we summarize the diverse applications of impedance measurement, highlighting the wide range of fields that can benefit from the design of high performance in modern measurement technologies.

Local field potential-based brain-machine interface to inhibit epileptic seizures by spinal cord electrical stimulation

Objective.This study proposes a closed-loop brain-machine interface (BMI) based on spinal cord stimulation to inhibit epileptic seizures, applying a semi-supervised machine learning approach that learns from Local Field Potential (LFP) patterns acquired on the pre-ictal (preceding the seizure) condition.Approach.LFP epochs from the hippocampus and motor cortex are band-pass filtered from 1 to 13 Hz, to obtain the time-frequency representation using the continuous Wavelet transform, and successively calculate the phase lock values (PLV). As a novelty, theZ-score-based PLV normalization using both modifiedk-means and Davies-Bouldin's measure for clustering is proposed here. Consequently, a generic seizure's detector is calibrated for detecting seizures on the normalized PLV, and enables the spinal cord stimulation for periods of 30 s in a closed-loop, while the BMI system detects seizure events. To calibrate the proposed BMI, a dataset with LFP signals recorded on five Wistar rats during basal state and epileptic crisis was used. The epileptic crisis was induced by injecting pentylenetetrazol (PTZ). Afterwards, two experiments without/with our BMI were carried out, inducing epileptic crisis by PTZ in Wistar rats.Main results.Stronger seizure events of high LFP amplitudes and long time periods were observed in the rat, when the BMI system was not used. In contrast, short-time seizure events of relative low intensity were observed in the rat, using the proposed BMI. The proposed system detected on unseen data the synchronized seizure activity in the hippocampus and motor cortex, provided stimulation appropriately, and consequently decreased seizure symptoms.Significance.Low-frequency LFP signals from the hippocampus and motor cortex, and cord spinal stimulation can be used to develop accurate closed-loop BMIs for early epileptic seizures inhibition, as an alternative treatment.

A Closed-Loop Ear-Worn Wearable EEG System with Real-Time Passive Electrode Skin Impedance Measurement for Early Autism Detection

Autism spectrum disorder (ASD) is a chronic neurological disorder with the severity directly linked to the diagnosis age. The severity can be reduced if diagnosis and intervention are early (age < 2 years). This work presents a novel ear-worn wearable EEG system designed to aid in the early detection of ASD. Conventional EEG systems often suffer from bulky, wired electrodes, high power consumption, and a lack of real-time electrode-skin interface (ESI) impedance monitoring. To address these limitations, our system incorporates continuous, long-term EEG recording, on-chip machine learning for real-time ASD prediction, and a passive ESI evaluation system. The passive ESI methodology evaluates impedance using the root mean square voltage of the output signal, considering factors like pressure, electrode surface area, material, gel thickness, and duration. The on-chip machine learning processor, implemented in 180 nm CMOS, occupies a minimal 2.52 mm² of active area while consuming only 0.87 µJ of energy per classification. The performance of this ML processor is validated using the Old Dominion University ASD dataset.

A Comprehensive Review of Hardware Acceleration Techniques and Convolutional Neural Networks for EEG Signals

This paper comprehensively reviews hardware acceleration techniques and the deployment of convolutional neural networks (CNNs) for analyzing electroencephalogram (EEG) signals across various application areas, including emotion classification, motor imagery, epilepsy detection, and sleep monitoring. Previous reviews on EEG have mainly focused on software solutions. However, these reviews often overlook key challenges associated with hardware implementation, such as scenarios that require a small size, low power, high security, and high accuracy. This paper discusses the challenges and opportunities of hardware acceleration for wearable EEG devices by focusing on these aspects. Specifically, this review classifies EEG signal features into five groups and discusses hardware implementation solutions for each category in detail, providing insights into the most suitable hardware acceleration strategies for various application scenarios. In addition, it explores the complexity of efficient CNN architectures for EEG signals, including techniques such as pruning, quantization, tensor decomposition, knowledge distillation, and neural architecture search. To the best of our knowledge, this is the first systematic review that combines CNN hardware solutions with EEG signal processing. By providing a comprehensive analysis of current challenges and a roadmap for future research, this paper provides a new perspective on the ongoing development of hardware-accelerated EEG systems.

Calibrating Deep Learning Classifiers for Patient-Independent Electroencephalogram Seizure Forecasting

The recent scientific literature abounds in proposals of seizure forecasting methods that exploit machine learning to automatically analyze electroencephalogram (EEG) signals. Deep learning algorithms seem to achieve a particularly remarkable performance, suggesting that the implementation of clinical devices for seizure prediction might be within reach. However, most of the research evaluated the robustness of automatic forecasting methods through randomized cross-validation techniques, while clinical applications require much more stringent validation based on patient-independent testing. In this study, we show that automatic seizure forecasting can be performed, to some extent, even on independent patients who have never been seen during the training phase, thanks to the implementation of a simple calibration pipeline that can fine-tune deep learning models, even on a single epileptic event recorded from a new patient. We evaluate our calibration procedure using two datasets containing EEG signals recorded from a large cohort of epileptic subjects, demonstrating that the forecast accuracy of deep learning methods can increase on average by more than 20%, and that performance improves systematically in all independent patients. We further show that our calibration procedure works best for deep learning models, but can also be successfully applied to machine learning algorithms based on engineered signal features. Although our method still requires at least one epileptic event per patient to calibrate the forecasting model, we conclude that focusing on realistic validation methods allows to more reliably compare different machine learning approaches for seizure prediction, enabling the implementation of robust and effective forecasting systems that can be used in daily healthcare practice.

A ResNet-LSTM hybrid model for predicting epileptic seizures using a pretrained model with supervised contrastive learning

In this paper, we propose a method for predicting epileptic seizures using a pre-trained model utilizing supervised contrastive learning and a hybrid model combining residual networks (ResNet) and long short-term memory (LSTM). The proposed training approach encompasses three key phases: pre-processing, pre-training as a pretext task, and training as a downstream task. In the pre-processing phase, the data is transformed into a spectrogram image using short time Fourier transform (STFT), which extracts both time and frequency information. This step compensates for the inherent complexity and irregularity of electroencephalography (EEG) data, which often hampers effective data analysis. During the pre-training phase, augmented data is generated from the original dataset using techniques such as band-stop filtering and temporal cutout. Subsequently, a ResNet model is pre-trained alongside a supervised contrastive loss model, learning the representation of the spectrogram image. In the training phase, a hybrid model is constructed by combining ResNet, initialized with weight values from the pre-trained model, and LSTM. This hybrid model extracts image features and time information to enhance prediction accuracy. The proposed method's effectiveness is validated using datasets from CHB-MIT and Seoul National University Hospital (SNUH). The method's generalization ability is confirmed through Leave-one-out cross-validation. From the experimental results measuring accuracy, sensitivity, and false positive rate (FPR), CHB-MIT was 91.90%, 89.64%, 0.058 and SNUH was 83.37%, 79.89%, and 0.131. The experimental results demonstrate that the proposed method outperforms the conventional methods.

An Energy-Efficient Small-Area Configurable Analog Front-End Interface for Diverse Biosignals Recording

We introduce a fully integrated configurable analog front-end (CAFE) sensor intended to accommodate various types of bio-potential signals in this article. The proposed CAFE is composed of an AC-coupled chopper-stabilized amplifier to effectively reduce 1/f noise and an energy- and area-efficient tunable filter to tune this interface to the bandwidth of various specific signals of interest. A tunable active-pseudo-resistor is integrated into the amplifier's feedback to realize a reconfigurable high-pass cutoff frequency and enhance its linearity, while the filter is designed using a subthreshold-source-follower-based pseudo-RC (SSF-PRC) topology to attain the required super-low cutoff frequency without the need for extremely low biasing current sources. Implemented in TSMC 40 nm technology, the chip occupies an active area of 0.048 [Formula: see text] while consuming 2.47 μW DC power from a 1.2-V supply voltage. Measurement results indicate that the proposed design achieved a mid-band gain of 37 dB, with an integrated input-referred noise ( VIRN) of 1.7 μVrms within 1-260 Hz. The total harmonic distortion (THD) of the CAFE is below 1 % with a 2.4 m Vpp input signal. With a wide-range bandwidth adjustment capability, the proposed CAFE can be used in both wearable and implantable recording devices to acquire different bio-potential signals.

Seizure count forecasting to aid diagnostic testing in epilepsy

OBJECTIVE

Epilepsy monitoring unit (EMU) admissions are critical for presurgical evaluation of drug-resistant epilepsy but may be nondiagnostic if an insufficient number of seizures are recorded. Seizure forecasting algorithms have shown promise for estimating the likelihood of seizures as a binary event in individual patients, but methods to predict how many seizures will occur remain elusive. Such methods could increase the diagnostic yield of EMU admissions and help patients mitigate seizure-related morbidity. Here, we evaluated the performance of a state-space method that uses prior seizure count data to predict future counts.

METHODS

A Bayesian negative-binomial dynamic linear model (DLM) was developed to forecast daily electrographic seizure counts in 19 patients implanted with a responsive neurostimulation (RNS) device. Holdout validation was used to evaluate performance in predicting the number of electrographic seizures for forecast horizons ranging 1-7 days ahead.

RESULTS

One-day-ahead prediction of the number of electrographic seizures using a negative-binomial DLM resulted in improvement over chance in 73.1% of time segments compared to a random chance forecaster and remained >50% for forecast horizons of up to 7 days. Superior performance (mean error = .99) was obtained in predicting the number of electrographic seizures in the next day compared to three traditional methods for count forecasting (integer-valued generalized autoregressive conditional heteroskedasticity model or INGARCH, 1.10; Croston, 1.06; generalized linear autoregressive moving average model or GLARMA, 2.00). Number of electrographic seizures in the preceding day and laterality of electrographic pattern detections had highest predictive value, with greater number of electrographic seizures and RNS magnet swipes in the preceding day associated with a higher number of electrographic seizures the next day.

SIGNIFICANCE

This study demonstrates that DLMs can predict the number of electrographic seizures a patient will experience days in advance with above chance accuracy. This study represents an important step toward the translation of seizure forecasting methods into the optimization of EMU admissions.

An Efficient Model-Compressed EEGNet Accelerator for Generalized Brain-Computer Interfaces With Near Sensor Intelligence

Brain-computer interfaces (BCIs) is promising in interacting with machines through electroencephalogram (EEG) signal. The compact end-to-end neural network model for generalized BCIs, EEGNet, has been implemented in hardware to get near sensor intelligence, but without enough efficiency. To utilize EEGNet in low-power wearable device for long-term use, this paper proposes an efficient EEGNet inference accelerator. Firstly, the EEGNet model is compressed by embedded channel selection, normalization merging, and product quantization. The customized accelerator based on the compressed model is then designed. The multilayer convolutions are achieved by reusing multiplying-accumulators and processing elements (PEs) to minimize area of logic circuits, and the weights and intermediate results are quantized to minimize memory sizes. The PEs are clock-gated to save power. Experimental results in FPGA on three datasets show the good generalizing ability of the proposed design across three BCI diagrams, which only consumes 3.31% area and 1.35% power compared to the one-to-one parallel design. The speedup factors of 1.4, 3.5, and 3.7 are achieved by embedded channel selection with negligible loss of accuracy (-0.80%). The presented accelerator is also synthesized in 65 nm CMOS low power (LP) process and consumes 0.23M gates, 24.4 ms/inference, 0.267 mJ/inference, which is 87.22% more efficient than the implementation of EEGNet in a RISC-V MCU realized in 40 nm CMOS LP process in terms of area, and 20.77% more efficient in terms of energy efficiency on BCIC-IV-2a dataset.

Bridging the Gap Between Patient-specific and Patient-independent Seizure Prediction via Knowledge Distillation

Deep neural networks (DNN) have shown unprecedented success in various brain-machine interface (BMI) applications such as epileptic seizure prediction. However, existing approaches typically train models in a patient-specific fashion due to the highly personalized characteristics of epileptic signals. Therefore, only a limited number of labeled recordings from each subject can be used for training. As a consequence, current DNN based methods demonstrate poor generalization ability to some extent due to the insufficiency of training data. On the other hand, patient-independent models attempt to utilize more patient data to train a universal model for all patients by pooling patient data together. Despite different techniques applied, results show that patient-independent models perform worse than patient-specific models due to high individual variation across patients. A substantial gap thus exists between patient-specific and patient-independent models. In this paper, we propose a novel training scheme based on knowledge distillation which makes use of a large amount of data from multiple subjects. It first distills informative features from signals of all available subjects with a pre-trained general model. A patient-specific model can then be obtained with the help of distilled knowledge and additional personalized data. The proposed training scheme significantly improves the performance of patient-specific seizure predictors and bridges the gap between patient-specific and patient-independent predictors. Five state-of-the-art seizure prediction methods are trained on the CHB-MIT sEEG database with our proposed scheme. The resulting accuracy, sensitivity, and false prediction rate show that our proposed training scheme consistently improves the prediction performance of state-of-the-art methods by a large margin.

Power efficient refined seizure prediction algorithm based on an enhanced benchmarking

Deep learning techniques have led to significant advancements in seizure prediction research. However, corresponding used benchmarks are not uniform in published results. Moreover, inappropriate training and evaluation processes used in various work create overfitted models, making prediction performance fluctuate or unreliable. In this study, we analyzed the various data preparation methods, dataset partition methods in related works, and explained the corresponding impacts to the prediction algorithms. Then we applied a robust processing procedure that considers the appropriate sampling parameters and the leave-one-out cross-validation method to avoid possible overfitting and provide prerequisites for ease benchmarking. Moreover, a deep learning architecture takes advantage of a one-dimension convolutional neural network and a bi-directional long short-term memory network is proposed for seizure prediction. The architecture achieves 77.6% accuracy, 82.7% sensitivity, and 72.4% specificity, and it outperforms the indicators of other prior-art works. The proposed model is also hardware friendly; it has 6.274 k parameters and requires only 12.825 M floating-point operations, which is advantageous for memory and power constrained device implementations.

Closed-Loop Neural Prostheses With On-Chip Intelligence: A Review and a Low-Latency Machine Learning Model for Brain State Detection

The application of closed-loop approaches in systems neuroscience and therapeutic stimulation holds great promise for revolutionizing our understanding of the brain and for developing novel neuromodulation therapies to restore lost functions. Neural prostheses capable of multi-channel neural recording, on-site signal processing, rapid symptom detection, and closed-loop stimulation are critical to enabling such novel treatments. However, the existing closed-loop neuromodulation devices are too simplistic and lack sufficient on-chip processing and intelligence. In this paper, we first discuss both commercial and investigational closed-loop neuromodulation devices for brain disorders. Next, we review state-of-the-art neural prostheses with on-chip machine learning, focusing on application-specific integrated circuits (ASIC). System requirements, performance and hardware comparisons, design trade-offs, and hardware optimization techniques are discussed. To facilitate a fair comparison and guide design choices among various on-chip classifiers, we propose a new energy-area (E-A) efficiency figure of merit that evaluates hardware efficiency and multi-channel scalability. Finally, we present several techniques to improve the key design metrics of tree-based on-chip classifiers, both in the context of ensemble methods and oblique structures. A novel Depth-Variant Tree Ensemble (DVTE) is proposed to reduce processing latency (e.g., by 2.5× on seizure detection task). We further develop a cost-aware learning approach to jointly optimize the power and latency metrics. We show that algorithm-hardware co-design enables the energy- and memory-optimized design of tree-based models, while preserving a high accuracy and low latency. Furthermore, we show that our proposed tree-based models feature a highly interpretable decision process that is essential for safety-critical applications such as closed-loop stimulation.

Multi-Level Optimization of an Ultra-Low Power BrainWave System for Non-Convulsive Seizure Detection

We present a systematic evaluation and optimization of a complex bio-medical signal processing application on the BrainWave prototype system, targeted towards ambulatory EEG monitoring within a tiny power budget of 1 mW. The considered BrainWave processor is completely programmable, while maintaining energy-efficiency by means of a Coarse-Grained Reconfigurable Array (CGRA). This is demonstrated through the mapping and evaluation of a state-of-the-art non-convulsive epileptic seizure detection algorithm, while ensuring real-time operation and seizure detection accuracy. Exploiting the CGRA leads to an energy reduction of 73.1%, compared to a highly tuned software implementation (SW-only). A total of 9 complex kernels were benchmarked on the CGRA, resulting in an average 4.7 × speedup and average 4.4 × energy savings over highly tuned SW-only implementations. The BrainWave processor is implemented in 28-nm FDSOI technology with 80 kB of Foundry-provided SRAM. By exploiting near-threshold computing for the logic and voltage-stacking to minimize on-chip voltage-conversion overhead, additional 15.2% and 19.5% energy savings are obtained, respectively. At the Minimum-Energy-Point (MEP) (223 μW, 8 MHz) we report a measured state-of-the-art 90.6% system conversion efficiency, while executing the epileptic seizure detection in real-time.

Searching for a paradigm shift in the research on the epilepsies and associated neuropsychiatric comorbidities. From ancient historical knowledge to the challenge of contemporary systems complexity and emergent functions

In this review, we will discuss in four scenarios our challenges to offer possible solutions for the puzzle associated with the epilepsies and neuropsychiatric comorbidities. We need to recognize that (1) since quite old times, human wisdom was linked to the plural (distinct global places/cultures) perception of the Universe we are in, with deep respect for earth and nature. Plural ancestral knowledge was added with the scientific methods; however, their joint efforts are the ideal scenario; (2) human behavior is not different than animal behavior, in essence the product of Darwinian natural selection; knowledge of animal and human behavior are complementary; (3) the expression of human behavior follows the same rules that complex systems with emergent properties, therefore, we can measure events in human, clinical, neurobiological situations with complexity systems' tools; (4) we can use the semiology of epilepsies and comorbidities, their neural substrates, and potential treatments (including experimental/computational modeling, neurosurgical interventions), as a source and collection of integrated big data to predict with them (e.g.: machine/deep learning) diagnosis/prognosis, individualized solutions (precision medicine), basic underlying mechanisms and molecular targets. Once the group of symptoms/signals (with a myriad of changing definitions and interpretations over time) and their specific sequences are determined, in epileptology research and clinical settings, the use of modern and contemporary techniques such as neuroanatomical maps, surface electroencephalogram and stereoelectroencephalography (SEEG) and imaging (MRI, BOLD, DTI, SPECT/PET), neuropsychological testing, among others, are auxiliary in the determination of the best electroclinical hypothesis, and help design a specific treatment, usually as the first attempt, with available pharmacological resources. On top of ancient knowledge, currently known and potentially new antiepileptic drugs, alternative treatments and mechanisms are usually produced as a consequence of the hard, multidisciplinary, and integrated studies of clinicians, surgeons, and basic scientists, all over the world. The existence of pharmacoresistant patients, calls for search of other solutions, being along the decades the surgeries the most common interventions, such as resective procedures (i.e., selective or standard lobectomy, lesionectomy), callosotomy, hemispherectomy and hemispherotomy, added by vagus nerve stimulation (VNS), deep brain stimulation (DBS), neuromodulation, and more recently focal minimal or noninvasive ablation. What is critical when we consider the pharmacoresistance aspect with the potential solution through surgery, is still the pursuit of localization-dependent regions (e.g.: epileptogenic zone (EZ)), in order to decide, no matter how sophisticated are the brain mapping tools (EEG and MRI), the size and location of the tissue to be removed. Mimicking the semiology and studying potential neural mechanisms and molecular targets - by means of experimental and computational modeling - are fundamental steps of the whole process. Concluding, with the conjunction of ancient knowledge, coupled to critical and creative contemporary, scientific (not dogmatic) clinical/surgical, and experimental/computational contributions, a better world and of improved quality of life can be offered to the people with epilepsy and neuropsychiatric comorbidities, who are still waiting (as well as the scientists) for a paradigm shift in epileptology, both in the Basic Science, Computational, Clinical, and Neurosurgical Arenas. This article is part of the Special Issue "NEWroscience 2018".

Energy-Efficient Neural Network for Epileptic Seizure Prediction

Seizure prediction for drug-refractory epilepsy patients can improve their quality of life, reduce their anxiety, and help them take the necessary precautions. Nowadays, numerous deep learning algorithms have been proposed to predict seizure onset and obtain better performance than traditional machine learning methods. However, these methods require a large set of parameters and large hardware resources; they also have high energy consumption. Therefore, these methods cannot be implemented on compact, low-power wearable, or implantable medical devices. The devices should operate on a real-time basis to continually inform the epileptic patients. In this paper, we describe energy-efficient and hardware-friendly methods to predict the epileptic seizures. A model of only 45 kB was obtained by the neural architecture search and was evaluated across three datasets. The overall accuracy, sensitivity, false prediction rate, and area under receiver operating characteristic curve were 99.53%, 99.81%, 0.005/h, 1 and 93.60%, 93.48%, 0.063/h, 0.977 and 86.86%, 85.19%, 0.116/h, 0.933, respectively, for the CHB-MIT scalp, the AES and Melbourne University intracranial electroencephalography (EEG) datasets. This model was further reduced with network pruning, quantization, and compact neural networks. The performances for the model sizes less than 50 kB for scalp EEG data and less than 10 kB for intracranial EEG data outperformed all the other models of similar model sizes. In particular, the energy consumption estimation was less than 10 mJ per inference for scalp EEG signal and less than 0.5 mJ per inference for intracranial EEG signal, which meet the requirements for low-power wearable and implantable devices, respectively.

Directions of Deep Brain Stimulation for Epilepsy and Parkinson's Disease

BACKGROUND

Deep brain stimulation (DBS) is an effective treatment for movement disorders and neurological/psychiatric disorders. DBS has been approved for the control of Parkinson disease (PD) and epilepsy.

OBJECTIVES

A systematic review and possible future direction of DBS system studies is performed in the open loop and closed-loop configuration on PD and epilepsy.

METHODS

We searched Google Scholar database for DBS system and development. DBS search results were categorized into clinical device and research system from the open-loop and closed-loop perspectives.

RESULTS

We performed literature review for DBS on PD and epilepsy in terms of system development by the open loop and closed-loop configuration. This study described development and trends for DBS in terms of electrode, recording, stimulation, and signal processing. The closed-loop DBS system raised a more attention in recent researches.

CONCLUSION

We overviewed development and progress of DBS. Our results suggest that the closed-loop DBS is important for PD and epilepsy.

Prediction of Seizure Recurrence. A Note of Caution

Great strides have been made recently in documenting that machine-learning programs can predict seizure occurrence in people who have epilepsy. Along with this progress have come claims that appear to us to be a bit premature. We anticipate that many people will benefit from seizure prediction. We also doubt that all will benefit. Although machine learning is a useful tool for aiding discovery, we believe that the greatest progress will come from deeper understanding of seizures, epilepsy, and the EEG features that enable seizure prediction. In this essay, we lay out reasons for optimism and skepticism.

Hardware Implementation of Deep Network Accelerators Towards Healthcare and Biomedical Applications

The advent of dedicated Deep Learning (DL) accelerators and neuromorphic processors has brought on new opportunities for applying both Deep and Spiking Neural Network (SNN) algorithms to healthcare and biomedical applications at the edge. This can facilitate the advancement of medical Internet of Things (IoT) systems and Point of Care (PoC) devices. In this paper, we provide a tutorial describing how various technologies including emerging memristive devices, Field Programmable Gate Arrays (FPGAs), and Complementary Metal Oxide Semiconductor (CMOS) can be used to develop efficient DL accelerators to solve a wide variety of diagnostic, pattern recognition, and signal processing problems in healthcare. Furthermore, we explore how spiking neuromorphic processors can complement their DL counterparts for processing biomedical signals. The tutorial is augmented with case studies of the vast literature on neural network and neuromorphic hardware as applied to the healthcare domain. We benchmark various hardware platforms by performing a sensor fusion signal processing task combining electromyography (EMG) signals with computer vision. Comparisons are made between dedicated neuromorphic processors and embedded AI accelerators in terms of inference latency and energy. Finally, we provide our analysis of the field and share a perspective on the advantages, disadvantages, challenges, and opportunities that various accelerators and neuromorphic processors introduce to healthcare and biomedical domains.