Machine learning-based prediction of acute coronary syndrome using only the pre-hospital 12-lead electrocardiogram

Prompt identification of acute coronary syndrome is a challenge in clinical practice. The 12-lead electrocardiogram (ECG) is readily available during initial patient evaluation, but current rule-based interpretation approaches lack sufficient accuracy. Here we report machine learning-based methods for the prediction of underlying acute myocardial ischemia in patients with chest pain. Using 554 temporal-spatial features of the 12-lead ECG, we train and test multiple classifiers on two independent prospective patient cohorts (n = 1244). While maintaining higher negative predictive value, our final fusion model achieves 52% gain in sensitivity compared to commercial interpretation software and 37% gain in sensitivity compared to experienced clinicians. Such an ultra-early, ECG-based clinical decision support tool, when combined with the judgment of trained emergency personnel, would help to improve clinical outcomes and reduce unnecessary costs in patients with chest pain.

Radiological images and machine learning: Trends, perspectives, and prospects

The application of machine learning to radiological images is an increasingly active research area that is expected to grow in the next five to ten years. Recent advances in machine learning have the potential to recognize and classify complex patterns from different radiological imaging modalities such as x-rays, computed tomography, magnetic resonance imaging and positron emission tomography imaging. In many applications, machine learning based systems have shown comparable performance to human decision-making. The applications of machine learning are the key ingredients of future clinical decision making and monitoring systems. This review covers the fundamental concepts behind various machine learning techniques and their applications in several radiological imaging areas, such as medical image segmentation, brain function studies and neurological disease diagnosis, as well as computer-aided systems, image registration, and content-based image retrieval systems. Synchronistically, we will briefly discuss current challenges and future directions regarding the application of machine learning in radiological imaging. By giving insight on how take advantage of machine learning powered applications, we expect that clinicians can prevent and diagnose diseases more accurately and efficiently.

A comprehensive assessment of gait accelerometry signals in time, frequency and time-frequency domains

Gait accelerometry is a promising tool to assess human walking and reveal deteriorating gait characteristics in patients and can be a rich source of clinically relevant information about functional declines in older adults. Therefore, in this paper, we present a comprehensive set of signal features that may be used to extract clinically valuable information from gait accelerometry signals. To achieve our goal, we collected tri-axial gait accelerometry signals from 35 adults 65 years of age and older. Fourteen subjects were healthy controls, 10 participants were diagnosed with Parkinson's disease, and 11 participants were diagnosed with peripheral neuropathy. The data were collected while the participants walked on a treadmill at a preferred walking speed. Accelerometer signal features in time, frequency and time-frequency domains were extracted. The results of our analysis showed that some of the extracted features were able to differentiate between healthy and clinical populations. Signal features in all three domains were able to emphasize variability among different groups, and also revealed valuable information about variability of the signals between anterior-posterior, mediolateral, and vertical directions within subjects. The current results imply that the proposed signal features can be valuable tools for the analysis of gait accelerometry data and should be utilized in future studies.

An empirical examination of detrended fluctuation analysis for gait data

Stride interval series exhibit statistical persistence, and detrended fluctuation analysis (DFA) is a routinely employed technique for describing this behavior. However, the implementation of DFA to gait data varies considerably between studies. We empirically examine two practical aspects of DFA which significantly affect the analysis outcome: the box size range and the stride interval series length. We conduct an analysis of their effect using stride intervals from 16 able-bodied adults, for overground walking, treadmill walking while holding a handrail, and treadmill walking without using a handrail. Our goal is to provide general guidelines for these two choices, with the aim of standardizing the application of DFA and facilitating inter-study comparisons. Based on the results of our analysis, we propose the use of box sizes from 16 to N/9, where N is the number of stride intervals. Moreover, for differentiating between normal and pathological walking with reasonable accuracy, we recommend a minimum of 600 stride intervals.

Necessity of noise in physiology and medicine

Noise is omnipresent in biomedical systems and signals. Conventional views assume that its presence is detrimental to systems' performance and accuracy. Hence, various analytic approaches and instrumentation have been designed to remove noise. On the contrary, recent contributions have shown that noise can play a beneficial role in biomedical systems. The results of this literature review indicate that noise is an essential part of biomedical systems and often plays a fundamental role in the performance of these systems. Furthermore, in preliminary work, noise has demonstrated therapeutic potential to alleviate the effects of various diseases. Further research into the role of noise and its applications in medicine is likely to lead to novel approaches to the treatment of diseases and prevention of disability.

The effects of rhythmic sensory cues on the temporal dynamics of human gait

Walking is a complex, rhythmic task performed by the locomotor system. However, natural gait rhythms can be influenced by metronomic auditory stimuli, a phenomenon of particular interest in neurological rehabilitation. In this paper, we examined the effects of aural, visual and tactile rhythmic cues on the temporal dynamics associated with human gait. Data were collected from fifteen healthy adults in two sessions. Each session consisted of five 15-minute trials. In the first trial of each session, participants walked at their preferred walking speed. In subsequent trials, participants were asked to walk to a metronomic beat, provided through visually, aurally, tactile or all three cues (simultaneously and in sync), the pace of which was set to the preferred walking speed of the first trial. Using the collected data, we extracted several parameters including: gait speed, mean stride interval, stride interval variability, scaling exponent and maximum Lyapunov exponent. The extracted parameters showed that rhythmic sensory cues affect the temporal dynamics of human gait. The auditory rhythmic cue had the greatest influence on the gait parameters, while the visual cue had no statistically significant effect on the scaling exponent. These results demonstrate that visual rhythmic cues could be considered as an alternative cueing modality in rehabilitation without concern of adversely altering the statistical persistence of walking.

Dysphagia Screening: Contributions of Cervical Auscultation Signals and Modern Signal-Processing Techniques

Cervical auscultation is the recording of sounds and vibrations caused by the human body from the throat during swallowing. While traditionally done by a trained clinician with a stethoscope, much work has been put towards developing more sensitive and clinically useful methods to characterize the data obtained with this technique. The eventual goal of the field is to improve the effectiveness of screening algorithms designed to predict the risk that swallowing disorders pose to individual patients' health and safety. This paper provides an overview of these signal processing techniques and summarizes recent advances made with digital transducers in hopes of organizing the highly varied research on cervical auscultation. It investigates where on the body these transducers are placed in order to record a signal as well as the collection of analog and digital filtering techniques used to further improve the signal quality. It also presents the wide array of methods and features used to characterize these signals, ranging from simply counting the number of swallows that occur over a period of time to calculating various descriptive features in the time, frequency, and phase space domains. Finally, this paper presents the algorithms that have been used to classify this data into 'normal' and 'abnormal' categories. Both linear as well as non-linear techniques are presented in this regard.

Segmentation of dual-axis swallowing accelerometry signals in healthy subjects with analysis of anthropometric effects on duration of swallowing activities

Dysphagia (swallowing difficulty) is a serious and debilitating condition that often accompanies stroke, acquired brain injury, and neurodegenerative illnesses. Individuals with dysphagia are prone to aspiration (the entry of foreign material into the airway), which directly increases the risk of serious respiratory consequences such as pneumonia. Swallowing accelerometry is a promising noninvasive tool for the detection of aspiration and the evaluation of swallowing. In this paper, dual-axis accelerometry was implemented since the motion of the hyolaryngeal complex occurs in both anterior-posterior and superior-inferior directions during swallowing. Dual-axis cervical accelerometry signals were acquired from 408 healthy subjects during dry, wet, and wet chin tuck swallowing tasks. The proposed segmentation algorithm is based on the idea of sequential fuzzy partitioning of the signal and is well suited for long signals with nonstationary variance. The algorithm was validated with simulated signals with known swallowing locations and a subset of 295 real swallows manually segmented by an experienced speech language pathologist. In both cases, the algorithm extracted individual swallows with over 90% accuracy. The time duration analysis was carried out with respect to gender, body mass index (BMI), and age. Demographic and anthropometric variables influenced the duration of these segmented signals. Male participants exhibited longer swallows than female participants (p=0.05). Older participants and participants with higher BMIs exhibited swallows with significantly longer (p=0.05) duration than younger participants and those with lower BMIs, respectively.

Effects of liquid stimuli on dual-axis swallowing accelerometry signals in a healthy population

Background

Dual-axis swallowing accelerometry has recently been proposed as a tool for non-invasive analysis of swallowing function. Although swallowing is known to be physiologically modifiable by the type of food or liquid (i.e., stimuli), the effects of stimuli on dual-axis accelerometry signals have never been thoroughly investigated. Thus, the objective of this study was to investigate stimulus effects on dual-axis accelerometry signal characteristics. Signals were acquired from 17 healthy participants while swallowing 4 different stimuli: water, nectar-thick and honey-thick apple juices, and a thin-liquid barium suspension. Two swallowing tasks were examined: discrete and sequential. A variety of features were extracted in the time and time-frequency domains after swallow segmentation and pre-processing. A separate Friedman test was conducted for each feature and for each swallowing task.

Results

Significant main stimulus effects were found on 6 out of 30 features for the discrete task and on 5 out of 30 features for the sequential task. Analysis of the features with significant stimulus effects suggested that the changes in the signals revealed slower and more pronounced swallowing patterns with increasing bolus viscosity.

Conclusions

We conclude that stimulus type does affect specific characteristics of dual-axis swallowing accelerometry signals, suggesting that associated clinical screening protocols may need to be stimulus specific.

Carbon nanotube chemiresistor for wireless pH sensing

The ability to accurately measure real-time pH fluctuations in-vivo could be highly advantageous. Early detection and potential prevention of bacteria colonization of surgical implants can be accomplished by monitoring associated acidosis. However, conventional glass membrane or ion-selective field-effect transistor (ISFET) pH sensing technologies both require a reference electrode which may suffer from leakage of electrolytes and potential contamination. Herein, we describe a solid-state sensor based on oxidized single-walled carbon nanotubes (ox-SWNTs) functionalized with the conductive polymer poly(1-aminoanthracene) (PAA). This device had a Nernstian response over a wide pH range (2–12) and retained sensitivity over 120 days. The sensor was also attached to a passively-powered radio-frequency identification (RFID) tag which transmits pH data through simulated skin. This battery-less, reference electrode free, wirelessly transmitting sensor platform shows potential for biomedical applications as an implantable sensor, adjacent to surgical implants detecting for infection.

A comparative analysis of DBSCAN, K-means, and quadratic variation algorithms for automatic identification of swallows from swallowing accelerometry signals

BACKGROUND

Cervical auscultation with high resolution sensors is currently under consideration as a method of automatically screening for specific swallowing abnormalities. To be clinically useful without human involvement, any devices based on cervical auscultation should be able to detect specified swallowing events in an automatic manner.

METHODS

In this paper, we comparatively analyze the density-based spatial clustering of applications with noise algorithm (DBSCAN), a k-means based algorithm, and an algorithm based on quadratic variation as methods of differentiating periods of swallowing activity from periods of time without swallows. These algorithms utilized swallowing vibration data exclusively and compared the results to a gold standard measure of swallowing duration. Data was collected from 23 subjects that were actively suffering from swallowing difficulties.

RESULTS

Comparing the performance of the DBSCAN algorithm with a proven segmentation algorithm that utilizes k-means clustering demonstrated that the DBSCAN algorithm had a higher sensitivity and correctly segmented more swallows. Comparing its performance with a threshold-based algorithm that utilized the quadratic variation of the signal showed that the DBSCAN algorithm offered no direct increase in performance. However, it offered several other benefits including a faster run time and more consistent performance between patients. All algorithms showed noticeable differentiation from the endpoints provided by a videofluoroscopy examination as well as reduced sensitivity.

CONCLUSIONS

In summary, we showed that the DBSCAN algorithm is a viable method for detecting the occurrence of a swallowing event using cervical auscultation signals, but significant work must be done to improve its performance before it can be implemented in an unsupervised manner.

A comparative analysis of swallowing accelerometry and sounds during saliva swallows

BACKGROUND

Accelerometry (the measurement of vibrations) and auscultation (the measurement of sounds) are both non-invasive techniques that have been explored for their potential to detect abnormalities in swallowing. The differences between these techniques and the information they capture about swallowing have not previously been explored in a direct comparison.

METHODS

In this study, we investigated the differences between dual-axis swallowing accelerometry and swallowing sounds by recording data from adult participants and calculating a number of time and frequency domain features. During the experiment, 55 participants (ages 18-65) were asked to complete five saliva swallows in a neutral head position. The resulting data was processed using previously designed techniques including wavelet denoising, spline filtering, and fuzzy means segmentation. The pre-processed signals were then used to calculate 9 time, frequency, and time-frequency domain features for each independent signal. Wilcoxon signed-rank and Wilcoxon rank-sum tests were utilized to compare feature values across transducers and patient demographics, respectively.

RESULTS

In addition to finding a number of features that varied between male and female participants, our statistical analysis determined that the majority of our chosen features were statistically significantly different across the two sensor methods and that the dependence on within-subject factors varied with the transducer type. However, a regression analysis showed that age accounted for an insignificant amount of variation in our signals.

CONCLUSIONS

We conclude that swallowing accelerometry and swallowing sounds provide different information about deglutition despite utilizing similar transduction methods. This contradicts past assumptions in the field and necessitates the development of separate analysis and processing techniques for swallowing sounds and vibrations.

A brain-computer interface based on bilateral transcranial Doppler ultrasound

In this study, we investigate the feasibility of a BCI based on transcranial Doppler ultrasound (TCD), a medical imaging technique used to monitor cerebral blood flow velocity. We classified the cerebral blood flow velocity changes associated with two mental tasks - a word generation task, and a mental rotation task. Cerebral blood flow velocity was measured simultaneously within the left and right middle cerebral arteries while nine able-bodied adults alternated between mental activity (i.e. word generation or mental rotation) and relaxation. Using linear discriminant analysis and a set of time-domain features, word generation and mental rotation were classified with respective average accuracies of 82.9%10.5 and 85.7%10.0 across all participants. Accuracies for all participants significantly exceeded chance. These results indicate that TCD is a promising measurement modality for BCI research.

Adaptive Transcutaneous Power Transfer to Implantable Devices: A State of the Art Review

Wireless energy transfer is a broad research area that has recently become applicable to implantable medical devices. Wireless powering of and communication with implanted devices is possible through wireless transcutaneous energy transfer. However, designing wireless transcutaneous systems is complicated due to the variability of the environment. The focus of this review is on strategies to sense and adapt to environmental variations in wireless transcutaneous systems. Adaptive systems provide the ability to maintain performance in the face of both unpredictability (variation from expected parameters) and variability (changes over time). Current strategies in adaptive (or tunable) systems include sensing relevant metrics to evaluate the function of the system in its environment and adjusting control parameters according to sensed values through the use of tunable components. Some challenges of applying adaptive designs to implantable devices are challenges common to all implantable devices, including size and power reduction on the implant, efficiency of power transfer and safety related to energy absorption in tissue. Challenges specifically associated with adaptation include choosing relevant and accessible parameters to sense and adjust, minimizing the tuning time and complexity of control, utilizing feedback from the implanted device and coordinating adaptation at the transmitter and receiver.

Automatic hyoid bone detection in fluoroscopic images using deep learning

The displacement of the hyoid bone is one of the key components evaluated in the swallow study, as its motion during swallowing is related to overall swallowing integrity. In daily research settings, experts visually detect the hyoid bone in the video frames and manually plot hyoid bone position frame by frame. This study aims to develop an automatic method to localize the location of the hyoid bone in the video sequence. To automatically detect the location of the hyoid bone in a frame, we proposed a single shot multibox detector, a deep convolutional neural network, which is employed to detect and classify the location of the hyoid bone. We also evaluated the performance of two other state-of-art detection methods for comparison. The experimental results clearly showed that the single shot multibox detector can detect the hyoid bone with an average precision of 89.14% and outperform other auto-detection algorithms. We conclude that this automatic hyoid bone tracking system is accurate enough to be widely applied as a pre-processing step for image processing in dysphagia research, as well as a promising development that may be useful in the diagnosis of dysphagia.

Noninvasive detection of thin-liquid aspiration using dual-axis swallowing accelerometry

Aspiration (the entry of foreign contents into the upper airway) is a serious concern for individuals with dysphagia and can lead to pneumonia. However, overt signs of aspiration, such as cough, are not always present, making noninstrumental diagnosis challenging. Valid, reliable tools for detecting aspiration during clinical screening and assessment are needed. In this study we investigated the validity of a noninvasive accelerometry signal-processing classifier for detecting aspiration. Dual-axis cervical accelerometry signals were collected from 40 adults on thin-liquid swallowing tasks during videofluoroscopic swallowing examinations. Signal-processing algorithms were used to remove known sources of artifact and a classifier was trained to identify signals associated with penetration-aspiration. Validity was measured in comparison to blinded ratings of penetration-aspiration from the concurrently recorded videofluoroscopies. On a bolus-by-bolus basis, the accelerometry classifier had a 10 % false-negative rate (90 % sensitivity) and a 23 % false-positive rate (77 % specificity) for detecting penetration-aspiration. We conclude that accelerometry can be used to support valid, reliable, and efficient detection of aspiration risk in patients with suspected dysphagia.

Decoding human swallowing via electroencephalography: a state-of-the-art review

Swallowing and swallowing disorders have garnered continuing interest over the past several decades. Electroencephalography (EEG) is an inexpensive and non-invasive procedure with very high temporal resolution which enables analysis of short and fast swallowing events, as well as an analysis of the organizational and behavioral aspects of cortical motor preparation, swallowing execution and swallowing regulation. EEG is a powerful technique which can be used alone or in combination with other techniques for monitoring swallowing, detection of swallowing motor imagery for diagnostic or biofeedback purposes, or to modulate and measure the effects of swallowing rehabilitation. This paper provides a review of the existing literature which has deployed EEG in the investigation of oropharyngeal swallowing, smell, taste and texture related to swallowing, cortical pre-motor activation in swallowing, and swallowing motor imagery detection. Furthermore, this paper provides a brief review of the different modalities of brain imaging techniques used to study swallowing brain activities, as well as the EEG components of interest for studies on swallowing and on swallowing motor imagery. Lastly, this paper provides directions for future swallowing investigations using EEG.

Extraction of stride events from gait accelerometry during treadmill walking

Objective: evaluating stride events can be valuable for understanding the changes in walking due to aging and neurological diseases. However, creating the time series necessary for this analysis can be cumbersome. In particular, finding heel contact and toe-off events which define the gait cycles accurately are difficult. Method: we proposed a method to extract stride cycle events from tri-axial accelerometry signals. We validated our method via data collected from 14 healthy controls, 10 participants with Parkinson’s disease, and 11 participants with peripheral neuropathy. All participants walked at self-selected comfortable and reduced speeds on a computer-controlled treadmill. Gait accelerometry signals were captured via a tri-axial accelerometer positioned over the L3 segment of the lumbar spine. Motion capture data were also collected and served as the comparison method. Results: our analysis of the accelerometry data showed that the proposed methodology was able to accurately extract heel and toe-contact events from both feet. We used t-tests, analysis of variance (ANOVA) and mixed models to summarize results and make comparisons. Mean gait cycle intervals were the same as those derived from motion capture, and cycle-to-cycle variability measures were within 1.5%. Subject group differences could be similarly identified using measures with the two methods. Conclusions: a simple tri-axial acceleromter accompanied by a signal processing algorithm can be used to capture stride events. Clinical impact: the proposed algorithm enables the assessment of stride events during treadmill walking, and is the first step toward the assessment of stride events using tri-axial accelerometers in real-life settings.

Non-invasive identification of swallows via deep learning in high resolution cervical auscultation recordings

High resolution cervical auscultation is a very promising noninvasive method for dysphagia screening and aspiration detection, as it does not involve the use of harmful ionizing radiation approaches. Automatic extraction of swallowing events in cervical auscultation is a key step for swallowing analysis to be clinically effective. Using time-varying spectral estimation of swallowing signals and deep feed forward neural networks, we propose an automatic segmentation algorithm for swallowing accelerometry and sounds that works directly on the raw swallowing signals in an online fashion. The algorithm was validated qualitatively and quantitatively using the swallowing data collected from 248 patients, yielding over 3000 swallows manually labeled by experienced speech language pathologists. With a detection accuracy that exceeded 95%, the algorithm has shown superior performance in comparison to the existing algorithms and demonstrated its generalizability when tested over 76 completely unseen swallows from a different population. The proposed method is not only of great importance to any subsequent swallowing signal analysis steps, but also provides an evidence that such signals can capture the physiological signature of the swallowing process.

Motor sequence learning-induced neural efficiency in functional brain connectivity

Previous studies have shown the functional neural circuitry differences before and after an explicitly learned motor sequence task, but have not assessed these changes during the process of motor skill learning. Functional magnetic resonance imaging activity was measured while participants (n=13) were asked to tap their fingers to visually presented sequences in blocks that were either the same sequence repeated (learning block) or random sequences (control block). Motor learning was associated with a decrease in brain activity during learning compared to control. Lower brain activation was noted in the posterior parietal association area and bilateral thalamus during the later periods of learning (not during the control). Compared to the control condition, we found the task-related motor learning was associated with decreased connectivity between the putamen and left inferior frontal gyrus and left middle cingulate brain regions. Motor learning was associated with changes in network activity, spatial extent, and connectivity.

Classification of penetration--aspiration versus healthy swallows using dual-axis swallowing accelerometry signals in dysphagic subjects

Swallowing accelerometry is a promising noninvasive approach for the detection of swallowing difficulties. In this paper, we propose an approach for classification of swallowing accelerometry recordings containing either healthy swallows or penetration-aspiration (entry of material into the airway) in dysphagic patients. The proposed algorithm is based on the wavelet packet decomposition of swallowing accelerometry signals in combination with linear discriminant analysis as a feature reduction method and Bayes classification. The proposed algorithm was tested using swallowing accelerometry signals collected from 40 patients during the regularly scheduled videoflouroscopy exam. The participants were instructed to swallow several 5-mL sips of thin liquid barium in a head neutral position. The results of our numerical analysis showed that the proposed algorithm can differentiate healthy swallows from aspiration swallows with an accuracy greater than 90%. These results position swallowing accelerometry as a valid approach for the detection of swallowing difficulties, particularly penetration-aspiration in patients suspected of dysphagia.