Machine learning based hybrid anomaly detection technique for automatic diagnosis of cardiovascular diseases using cardiac sympathetic nerve activity and electrocardiogram.

OBJECTIVES: Coronary artery diseases (CADs) are the leading cause of death worldwide and early diagnosis is crucial for timely treatment. To address this, our study presents a novel automated Artificial Intelligence (AI)-based Hybrid Anomaly Detection (AIHAD) technique that combines various signal processing, feature extraction, supervised, and unsupervised machine learning methods. By jointly and simultaneously analyzing 12-lead cardiac sympathetic nerve activity (CSNA) and electrocardiogram (ECG) data, the automated AIHAD technique performs fast, early, and accurate diagnosis of CADs. METHODS: In order to develop and evaluate the proposed automated AIHAD technique, we utilized the fully labeled STAFF III and PTBD databases, which contain the 12-lead wideband raw recordings non-invasively acquired from 260 subjects. Using these wideband raw recordings, we developed a signal processing technique that simultaneously detects the 12-lead CSNA and ECG signals of all subjects. Using the pre-processed 12-lead CSNA and ECG signals, we developed a time-domain feature extraction technique that extracts the statistical CSNA and ECG features critical for the reliable diagnosis of CADs. Using the extracted discriminative features, we developed a supervised classification technique based on Artificial Neural Networks (ANNs) that simultaneously detects anomalies in the 12-lead CSNA and ECG data. Furthermore, we developed an unsupervised clustering technique based on Gaussian mixture models (GMMs) and Neyman-Pearson criterion, which robustly detects outliers corresponding to CADs. RESULTS: Using the automated AIHAD technique, we have, for the first time, demonstrated a significant association between the increase in CSNA signals and anomalies in ECG signals during CADs. The AIHAD technique achieved highly reliable detection of CADs with a sensitivity of 98.48 %, specificity of 97.73 %, accuracy of 98.11 %, positive predictive value of 97.74 %, negative predictive value of 98.47 %, and F1-score of 98.11 %. Hence, the automated AIHAD technique demonstrates superior performance compared to the gold standard diagnostic test ECG in the diagnosis of CADs. Additionally, it outperforms other techniques developed in this study that separately utilize either only CSNA data or only ECG data. Therefore, it significantly increases the detection performance of CADs by taking advantage of the diversity in different data types and leveraging their strengths. Furthermore, its performance is comparatively better than that of most previously proposed machine and deep learning methods that exclusively used ECG data to diagnose or classify CADs. Additionally, it has a very low implementation time, which is highly desirable for real-time detection of CADs. CONCLUSIONS: The proposed automated AIHAD technique may serve as an efficient decision-support system to increase physicians' success in fast, early, and accurate diagnosis of CADs. It may be highly beneficial and valuable, particularly for asymptomatic patients, for whom the diagnostic information provided by ECG alone is not sufficient to reliably diagnose the disease. Hence, it may significantly improve patient outcomes by enabling timely treatments and considerably reducing the mortality of cardiovascular diseases (CVDs).

Machine-based classification of ADHD and nonADHD participants using time/frequency features of event-related neuroelectric activity.

OBJECTIVE: Attention-deficit/hyperactivity disorder (ADHD) is the most frequent diagnosis among children who are referred to psychiatry departments. Although ADHD was discovered at the beginning of the 20th century, its diagnosis is still confronted with many problems. METHOD: A novel classification approach that discriminates ADHD and nonADHD groups over the time-frequency domain features of event-related potential (ERP) recordings that are taken during Stroop task is presented. Time-Frequency Hermite-Atomizer (TFHA) technique is used for the extraction of high resolution time-frequency domain features that are highly localized in time-frequency domain. Based on an extensive investigation, Support Vector Machine-Recursive Feature Elimination (SVM-RFE) was used to obtain the best discriminating features. RESULTS: When the best three features were used, the classification accuracy for the training dataset reached 98%, and the use of five features further improved the accuracy to 99.5%. The accuracy was 100% for the testing dataset. Based on extensive experiments, the delta band emerged as the most contributing frequency band and statistical parameters emerged as the most contributing feature group. CONCLUSION: The classification performance of this study suggests that TFHA can be employed as an auxiliary component of the diagnostic and prognostic procedures for ADHD. SIGNIFICANCE: The features obtained in this study can potentially contribute to the neuroelectrical understanding and clinical diagnosis of ADHD.

Idler-efficiency-enhanced long-wave infrared beam generation using aperiodic orientation-patterned GaAs gratings.

In this paper, we design aperiodic gratings based on orientation-patterned gallium arsenide (OP-GaAs) for converting 2.1 µm pump laser radiation into long-wave infrared (8-12 µm) in an idler-efficiency-enhanced scheme. These single OP-GaAs gratings placed in an optical parametric oscillator (OPO) or an optical parametric generator (OPG) can simultaneously phase match two optical parametric amplification (OPA) processes, OPA 1 and OPA 2. We use two design methods that allow simultaneous phase matching of two arbitrary χ(2) processes and also free adjustment of their relative strength. The first aperiodic grating design method (Method 1) relies on generating a grating structure that has domain walls located at the zeros of the summation of two cosine functions, each of which has a spatial frequency that equals one of the phase-mismatch terms of the two processes. Some of the domain walls are discarded considering the minimum domain length that is achievable in the production process. In this paper, we propose a second design method (Method 2) that relies on discretizing the crystal length with sample lengths that are much smaller than the minimum domain length and testing each sample's contribution in such a way that the sign of the nonlinearity maximizes the magnitude sum of the real and imaginary parts of the Fourier transform of the grating function at the relevant phase mismatches. Method 2 produces a similar performance as Method 1 in terms of the maximization of the height of either Fourier peak located at the relevant phase mismatch while allowing an adjustable relative height for the two peaks. To our knowledge, this is the first time that aperiodic OP-GaAs gratings have been proposed for efficient long-wave infrared beam generation based on simultaneous phase matching.

Fourier transform magnetic resonance current density imaging (FT-MRCDI) from one component of magnetic flux density.

Fourier transform (FT)-based algorithms for magnetic resonance current density imaging (MRCDI) from one component of magnetic flux density have been developed for 2D and 3D problems. For 2D problems, where current is confined to the xy-plane and z-component of the magnetic flux density is measured also on the xy-plane inside the object, an iterative FT-MRCDI algorithm is developed by which both the current distribution inside the object and the z-component of the magnetic flux density on the xy-plane outside the object are reconstructed. The method is applied to simulated as well as actual data from phantoms. The effect of measurement error on the spatial resolution of the current density reconstruction is also investigated. For 3D objects an iterative FT-based algorithm is developed whereby the projected current is reconstructed on any slice using as data the Laplacian of the z-component of magnetic flux density measured for that slice. In an injected current MRCDI scenario, the current is not divergence free on the boundary of the object. The method developed in this study also handles this situation.

Efficient computation of joint fractional Fourier domain signal representation.

A joint fractional domain signal representation is proposed based on an intuitive understanding from a time-frequency distribution of signals that designates the joint time and frequency energy content. The joint fractional signal representation (JFSR) of a signal is so designed that its projections onto the defining joint fractional Fourier domains give the modulus square of the fractional Fourier transform of the signal at the corresponding orders. We derive properties of the JFSR, including its relations to quadratic time-frequency representations and fractional Fourier transformations, which include the oblique projections of the JFSR. We present a fast algorithm to compute radial slices of the JFSR and the results are shown for various signals at different fractionally ordered domains.

Early gamma response of sleep is sensory/perceptual in origin.

The goal of the study was to investigate the gamma response of the brain and its functional correlates in rapid eye movements (REM) sleep and the three stages of non-REM sleep. Data on overnight sleep were acquired from 16 healthy, young adult, volunteer males. Neuroelectric activity was recorded from seven recording sites (Fz, Cz, Pz, F3, F4, P3, P4) in response to auditory stimuli (2000 Hz deviant and 1000 Hz standard stimuli: 65 dB, 10 ms r/f time, 50 ms duration) under passive oddball paradigm. Data were analyzed with the Fourier transform and digital filtering and also the recently developed technique of time-frequency component analysis (TFCA). TFCA displayed the gamma response under all stages of sleep. Statistical analysis did not reveal a significant effect of stimulus type, recording site or sleep stage on the three parameters of TFCA, which included maximum value of the time-frequency representation of the extracted gamma component, maximum magnitude of the time-domain representation of the component and the energy of this component. The gamma period included N1 and the early theta response, both of which are related to sensory-perceptual processing in the literature. According to these findings, the gamma response is possibly related, as in wakefulness, to early stimulus processing that also includes sensory/perceptual operations.

Early time-locked gamma response and gender specificity.

The aim was to investigate whether gender is a causative factor in the gamma status according to which some individuals respond with time-locked, early gamma response, G+, while the others do not show this response, G-. The sample consisted of 42 volunteer participants (between 19 and 37 years of age with at least 9 years of education). There were 22 females and 20 males. Data were collected under the oddball paradigm. Auditory stimulation (10 ms r/f time, 50 ms duration, 65 dB SPL) consisted of target (2000 Hz; p = .20) stimuli that occurred randomly within a series of standard stimuli (1000 Hz; p = .80). Gamma responses were studied in the amplitude frequency characteristics, in the digitally filtered event-related potentials (f-ERPs) and in the distributions which were obtained using the recently developed time-frequency component analysis (TFCA) technique. Participants were classified into G+ and G- groups with a criterion of full agreement between the results of an automated gamma detection technique and expert opinion. The 2 x 2 x 2 ANOVA on f-ERPs and 2 x 2 x 2 multivariate ANOVA on TFCA distributions showed the main effect of gamma status and gender as significant, and the interaction between gamma status and gender as nonsignificant. Accordingly, individual difference in gamma status is a reliable phenomenon, but this does not depend on gender. There are conflicting findings in the literature concerning the effect of gender on ERP components (N100, P300). The present study showed that if the gamma status is not included in research designs, it may produce a confounding effect on ERP parameters.

Time-frequency component analyser and its application to brain oscillatory activity.

Currently, event-related potential (ERP) signals are analysed in the time domain (ERP technique) or in the frequency domain (Fourier analysis and variants). In techniques of time-domain and frequency-domain analysis (short-time Fourier transform, wavelet transform) assumptions concerning linearity, stationarity, and templates are made about the brain signals. In the time-frequency component analyser (TFCA), the assumption is that the signal has one or more components with non-overlapping supports in the time-frequency plane. In this study, the TFCA technique was applied to ERPs. TFCA determined and extracted the oscillatory components from the signal and, simultaneously, localized them in the time-frequency plane with high resolution and negligible cross-term contamination. The results obtained by means of TFCA were compared with those obtained by means of other commonly used techniques of ERP analysis, such as bilinear time-frequency distributions and wavelet analysis. It is suggested that TFCA may serve as an appropriate tool for capturing the localized ERP components in the time-frequency domain and for studying the intricate, frequency-based dynamics of the human brain.