Subject-specific Characteristics Research Articles

#2898 Exposure response model describing hematologic changes with chronic treatment of CKD anemia with vadadustat

Abstract Background and Aims Vadadustat (VADA) is an oral hypoxia-inducible factor (HIF) prolyl hydroxylase inhibitor, a class of drugs that stabilize HIF and stimulate endogenous erythropoietin production, increasing iron mobilization and red blood cell production. VADA is approved in 36 countries globally and is indicated in the European Union for the treatment of symptomatic anaemia associated with chronic kidney disease (CKD) in adults on chronic maintenance dialysis and is an alternative to treatment with erythropoiesis-stimulating agents (ESAs). The aim of this work was to develop a predictive, semi-mechanistic model to understand the effects of VADA and erythropoiesis-stimulating agents (ESAs) on hemoglobin response and to evaluate the impact of intrinsic and extrinsic patient characteristics on that response for non-dialysis dependent (NDD) and dialysis-dependent (DD) patient populations. Method A predictive semi-mechanistic pharmacokinetic/pharmacodynamic (PK/PD) model for VADA and ESA treatment was developed using nonlinear mixed effects methods. The model describes changes in reticulocyte count (RETICs), mature erythrocytes (RBCs) and hemoglobin (Hb) as a function of drug exposure and subject-specific characteristics. The dataset was based on data from 7 Phase 1 to Phase 3 studies (Table 1; N = 894 for VADA; N = 560 for ESA modeling) in healthy volunteers and subjects with NDD- or DD-CKD who were either ESA-naïve or ESA experienced but not receiving treatment at screening. Prior population PK modeling of VADA provided estimated drug exposures for VADA-treated subjects, while dose normalized to standard units (IU/kg/wk) was used as the exposure metric for subjects on ESA. Each of these exposure metrics could vary with time, as the dose was adjusted according to observed Hb response. This work evaluated alternative model structures, as well as covariate effects affecting Hb individual response. Results A predictive semi-mechanistic pharmacokinetic/pharmacodynamic (PK/PD) model for VADA and ESA treatment was developed using nonlinear mixed effects methods. The model describes changes in reticulocyte count (RETICs), mature erythrocytes (RBCs) and hemoglobin (Hb) as a function of drug exposure and subject-specific characteristics. The dataset was based on data from 7 Phase 1 to Phase 3 studies (Table 1; N = 894 for VADA; N = 560 for ESA modeling) in healthy volunteers and subjects with NDD- or DD-CKD who were either ESA-naïve or ESA experienced but not receiving treatment at screening. Prior population PK modeling of VADA provided estimated drug exposures for VADA-treated subjects, while dose normalized to standard units (IU/kg/wk) was used as the exposure metric for subjects on ESA. Each of these exposure metrics could vary with time, as the dose was adjusted according to observed Hb response. This work evaluated alternative model structures, as well as covariate effects affecting Hb individual response. A similar model structure was found to describe early changes in RETICs and longer-term changes in Hb for both VADA and ESAs. In each case, the magnitude of response increases with increased drug exposure up to a maximum value (Emax), and the Emax values were similar between VADA and ESAs. Subject intrinsic predictors of response included eGFR, subject age, BMI, and serum albumin, while extrinsic factors affecting response included geographic region, limits on acceptable Hb values (which also differ by geographic region) and the presence of previous ESA treatment. Interestingly, Hb responses were also correlated with changes in serum hepcidin, a key regulator of iron homeostasis, following both VADA and ESA treatment. Differences between the response to VADA and ESA were also described, including differences in temporal changes in RETICs and Hb. Conclusion This model showed an exposure-response relationship between VADA or ESA treatment and changes in RETICs and Hb for ESA naïve or ESA-experienced CKD subjects. The model will be used to inform dosage recommendations for VADA based on patient characteristics.

Elucidating vaccine efficacy using a correlate of protection, demographics, and logistic regression

BackgroundVaccine efficacy (VE) assessed in a randomized controlled clinical trial can be affected by demographic, clinical, and other subject-specific characteristics evaluated as baseline covariates. Understanding the effect of covariates on efficacy is key to decisions by vaccine developers and public health authorities.MethodsThis work evaluates the impact of including correlate of protection (CoP) data in logistic regression on its performance in identifying statistically and clinically significant covariates in settings typical for a vaccine phase 3 trial. The proposed approach uses CoP data and covariate data as predictors of clinical outcome (diseased versus non-diseased) and is compared to logistic regression (without CoP data) to relate vaccination status and covariate data to clinical outcome.ResultsClinical trial simulations, in which the true relationship between CoP data and clinical outcome probability is a sigmoid function, show that use of CoP data increases the positive predictive value for detection of a covariate effect. If the true relationship is characterized by a decreasing convex function, use of CoP data does not substantially change positive or negative predictive value. In either scenario, vaccine efficacy is estimated more precisely (i.e., confidence intervals are narrower) in covariate-defined subgroups if CoP data are used, implying that using CoP data increases the ability to determine clinical significance of baseline covariate effects on efficacy.ConclusionsThis study proposes and evaluates a novel approach for assessing baseline demographic covariates potentially affecting VE. Results show that the proposed approach can sensitively and specifically identify potentially important covariates and provides a method for evaluating their likely clinical significance in terms of predicted impact on vaccine efficacy. It shows further that inclusion of CoP data can enable more precise VE estimation, thus enhancing study power and/or efficiency and providing even better information to support health policy and development decisions.

Resting state EEG assisted imagined vowel phonemes recognition by native and non-native speakers using brain connectivity measures.

Communication is challenging for disabled individuals, but with advancement of brain-computer interface (BCI) systems, alternative communication systems can be developed. Current BCI spellers, such as P300, SSVEP, and MI, have drawbacks like reliance on external stimuli or conversation irrelevant mental tasks. In contrast to these systems, Imagined speech based BCI systems rely on directly decoding the vowels/words user is thinking, making them more intuitive, user friendly and highly popular among Brain-Computer-Interface (BCI) researchers. However, more research needs to be conducted on how subject-specific characteristics such as mental state, age, handedness, nativeness and resting state activity affects the brain's output during imagined speech. In an overt speech, it is evident that native and non-native speakers' brains function differently. Therefore, this paper explores how nativeness to language affects EEG signals while imagining vowel phonemes, using brain-map analysis and scalogram and also investigates the inclusion of features extracted from resting state EEG with imagined state EEG. The Fourteen-channel EEG for Imagined Speech (FEIS) dataset was used to analyse the EEG signals recorded while imagining vowel phonemes for 16 subjects (nine native English and seven non-native Chinese). For the classification of vowel phonemes, different connectivity measures such as covariance, coherence, and Phase Synchronous Index-PSI were extracted and analysed using statistics based Multivariate Analysis of Variance (MANOVA) approach. Different fusion strategies (difference, concatenation, Common Spatial Pattern-CSP and Canonical Correlation Analysis-CCA) were carried out to incorporate resting state EEG connectivity measures with imagined state connectivity measures for enhancing the accuracy of imagined vowel phoneme recognition. Simulation results revealed that concatenating imagined state and rest state covariance and PSI features provided the maximum accuracy of 92.78% for native speakers and 94.07% for non-native speakers.

Effects of observation mode on ratings of teaching quality in secondary mathematics classrooms

AbstractIn educational research, teaching quality is extensively studied because of its role of a mediator between teacher characteristics and student learning. However, empirical evidence on differences between video and live scoring of teaching quality is rare. In the present study, thirty lessons from 15 secondary mathematics classrooms in a German metropolitan area were observed. Lessons were scored both live in the classroom and using video recordings. Live and video scoring was conducted by (different) trained observers. Ratings were obtained with a “hybrid” observational instrument that covers generic and subject-specific characteristics of teaching quality in mathematics classrooms. Generalizability analysis and paired t tests were performed to investigate mode effects. The findings showed that in live scoring, classroom management was rated lower, and cognitive activation was rated higher. Rankings of lessons or classrooms were very similar across modes, and reliabilities did not differ to a meaningful extent either, except for classroom management reaching better results for live ratings. This suggests that based on the present findings, classroom observation performed with our hybrid framework of teaching quality generalizes across observation mode only under certain circumstances. Further research is necessary to better understand the relation between observation mode and teaching quality ratings, as well as the impact of the scoring procedures. We discuss the implications of our findings for educational research and practice.

Individuals with rotator cuff tears unsuccessfully treated with exercise therapy have less inferiorly oriented net muscle forces during scapular plane abduction

Exercise therapy for individuals with rotator cuff tears fails in approximately 25.0 % of cases. One reason for failure of exercise therapy may be the inability to strengthen and balance the muscle forces crossing the glenohumeral joint that act to center the humeral head on the glenoid. The objective of the current study was to compare the magnitude and orientation of the net muscle force pre- and post-exercise therapy between subjects successfully and unsuccessfully (e.g. eventually underwent surgery) treated with a 12-week individualized exercise therapy program. Twelve computational musculoskeletal models (n = 6 successful, n = 6 unsuccessful) were developed in OpenSim (v4.0) that incorporated subject specific tear characteristics, muscle peak isometric force, in-vivo kinematics and bony morphology. The models were driven with experimental kinematics and the magnitude and orientation of the net muscle force was determined during scapular plane abduction at pre- and post-exercise therapy timepoints. Subjects unsuccessfully treated had less inferiorly oriented net muscle forces pre- and post-exercise therapy compared to subjects successfully treated (p = 0.039 & 0.045, respectively). No differences were observed in the magnitude of the net muscle force (p > 0.05). The current study developed novel computational musculoskeletal models with subject specific inputs capable of distinguishing between subjects successfully and unsuccessfully treated with exercise therapy. A less inferiorly oriented net muscle force in subjects unsuccessfully treated may increase the risk of superior migration leading to impingement. Adjustments to exercise therapy programs may be warranted to avoid surgery in subjects at risk of unsuccessful treatment.

Echocardiographic Reference Ranges of Global Longitudinal Strain for All Cardiac Chambers Using Guideline-Directed Dedicated Views

BackgroundMyocardial deformation by echocardiographic strain imaging is a key measurement in cardiology, providing valuable diagnostic and prognostic information. Reference ranges for strain should be established from large healthy populations with minimal methodologic biases and variability. ObjectivesThe aim of this study was to establish echocardiographic reference ranges, including lower normal limits of global strains for all 4 cardiac chambers, by guideline-directed dedicated views from a large healthy population and to evaluate the influence of subject-specific characteristics on strain. MethodsIn total, 1,329 healthy participants from HUNT4Echo, the echocardiographic substudy of the 4th wave of the Trøndelag Health Study, were included. Echocardiographic recordings specific for each chamber were optimized according to current recommendations. Two experienced sonographers recorded all echocardiograms using GE HealthCare Vivid E95 scanners. Analyses were performed by experts using GE HealthCare EchoPAC. ResultsThe reference ranges for left ventricular (LV) global longitudinal strain and right ventricular free-wall strain were −24% to −16% and −35% to −17%, respectively. Correspondingly, left atrial (LA) and right atrial (RA) reservoir strains were 17% to 49% and 17% to 59%. All strains showed lower absolute values with higher age, except for LA and RA contractile strains, which were higher. The feasibility for strain was overall good (LV 96%, right ventricular 83%, LA 94%, and RA 87%). All chamber-specific strains were associated with age, and LV strain was associated with sex. ConclusionsReference ranges of strain for all cardiac chambers were established based on guideline-directed chamber-specific recordings. Age and sex were the most important factors influencing reference ranges and should be considered when using strain echocardiography.

Exploring the Effects of Computer and Smart Device-Assisted Learning on Students’ Achievements: Empirical Evidence from Korea

Computer and Smart Device-assisted Learning (CSDL) has gained increasing attention from educational researchers and practitioners in recent years. However, it remains controversial whether students can benefit from CSDL and what moderators could affect the impact of CSDL. Within the specific context of Korea, where the interest in digital education is steadily increasing, the number of empirical studies exploring the causal effect of CSDL remains relatively scarce. The primary objective of this empirical study was to investigate the impact of CSDL on students’ academic achievements in Korea. To achieve this objective, a two-way fixed effect model was employed, utilizing a panel dataset spanning three years derived from the “Korean Education Longitudinal Study 2013”. The findings revealed a significant positive impact of CSDL on students’ mathematics achievements. Notably, higher income levels, increased availability of computer resources provided by schools, and the implementation of more individualized education were identified as factors that moderate the effect of CSDL on students’ achievement levels in Korean and English subjects. These findings underscore the need for an approach that optimizes the educational benefits of CSDL by considering subject-specific characteristics. Furthermore, this study highlights the importance of allocating educational resources, such as computers and smart devices, and integrating individualized educational activities within the classroom environment.

An Incremental Learning Approach to Detect Muscular Fatigue in Human– Robot Collaboration

Human–robot collaboration aims to join the distinctive strengths of humans and robots to compensate for the weaknesses associated with each party and, thus, to enable synergetic effects. Robots are characteristically considered fatigue-proof. Hence, they are utilized to assist human operators during heavy pushing and pulling activities. To detect physical fatigue or high payloads held by a human operator, wearable sensors, such as electromyographys (EMGs), are deployed. The EMG data are typically processed via machine learning, which includes training models offline before an application in an online system. However, these approaches often demonstrate varying performances between offline and online applications due to subject-specific characteristics within the data. An opportunity to tackle this challenge can be found in incremental learning, as these models purely learn online and constantly fine-tune the model's structure. In this article, a Mondrian Forest is applied to predict payloads and physical fatigue of human operators during an assistance scenario with a collaborative robot. An experiment was conducted with a total of 12 participants, where the payload was increased until participants initiated an assistance request from a Universal Robots model 10 cobot. This allowed for testing whether the Mondrian Forest can accurately predict the payload and fatigue levels from the acquired EMG signals. Overall, the approach demonstrates a promising potential toward higher awareness when an operator might require assistance from a robot and ultimately toward a more effective human–robot collaboration.

Assessment of the sensitivity of thoracic injury criteria to subject-specific characteristics using human body models.

Introduction: Chest deformation has been proposed as the best predictor of thoracic injury risk in frontal impacts. Finite Element Human Body Models (FE-HBM) can enhance the results obtained in physical crash tests with Anthropometric Test Devices (ATD) since they can be exposed to omnidirectional impacts and their geometry can be modified to reflect specific population groups. This study aims to assess the sensitivity of two thoracic injury risk criteria (PC Score and Cmax) to several personalization techniques of FE-HBMs. Methods: Three 30° nearside oblique sled tests were reproduced using the SAFER HBM v8 and three personalization techniques were applied to this model to evaluate the influence on the risk of thoracic injuries. First, the overall mass of the model was adjusted to represent the weight of the subjects. Second, the model anthropometry and mass were modified to represent the characteristics of the post-mortem human subjects (PMHS). Finally, the spine alignment of the model was adapted to the PMHS posture at t = 0ms, to conform to the angles between spinal landmarks measured in the PMHS. The following two metrics were used to predict three or more fractured ribs (AIS3+) of the SAFER HBM v8 and the effect of personalization techniques: the maximum posterior displacement of any studied chest point (Cmax), and the sum of the upper and lower deformation of selected rib points (PC score). Results: Despite having led to statistically significant differences in the probability of AIS3+ calculations, the mass-scaled and morphed version provided, in general, lower values for injury risk than the baseline model and the postured version being the latter, which exhibited the better approximation to the PMHS tests in terms of probability of injury. Additionally, this study found that the prediction of AIS3+ chest injuries based on PC Score resulted in higher probability values than the prediction based on Cmax for the loading conditions and personalization techniques analyzed within this study. Discussion: This study could demonstrate that the personalization techniques do not lead to linear trends when they are used in combination. Furthermore, the results included here suggest that these two criteria will result in significantly different predictions if the chest is loaded more asymmetrically.

Musculoskeletal-Modeling-Based, Full-Body Load-Assessment Tool for Ergonomists (MATE): Method Development and Proof of Concept Case Studies.

A new ergonomic-risk-assessment tool was developed that combines musculoskeletal-model-based loading estimates with insights from fatigue failure theory to evaluate full-body musculoskeletal loading during dynamic tasks. Musculoskeletal-modeling output parameters, i.e., joint contact forces and muscle forces, were combined with tissue-specific injury thresholds that account for loading frequency to determine the injury risk for muscles, lower back, and hip cartilage. The potential of this new risk-assessment tool is demonstrated for defining ergonomic interventions in terms of lifting characteristics, back and shoulder exoskeleton assistance, box transferring, stoop lifting, and an overhead wiring task, respectively. The MATE identifies the risk of WMSDs in different anatomical regions during occupational tasks and allows for the evaluation of the impact of interventions that modify specific lifting characteristics, i.e., load weight versus task repetition. Furthermore, and in clear contrast to currently available ergonomic assessment scores, the effects of the exoskeleton assistance level on the risk of WMSDs of full-body musculoskeletal loading (in particular, the muscles, lower back, and hips) can be evaluated and shows small reductions in musculoskeletal loading but not in injury risk. Therefore, the MATE is a risk-assessment tool based on a full-body, musculoskeletal-modeling approach combined with insights from the fatigue failure theory that shows the proof of concept of a shoulder and back exoskeleton. Furthermore, it accounts for subject-specific characteristics (age and BMI), further enhancing individualized ergonomic-risk assessment.

Person-specific biophysical modelling of alpha-motoneuron pools driven by in vivo decoded neural synaptic input.

Interfacing with alpha-motoneurons (MNs) is key to understand and control motor impairment and neurorehabilitation technologies. Depending on the neurophysiological condition of each individual, MN pools exhibit distinct neuro-anatomical properties and firing behaviors. Hence, the ability to assess subject-specific characteristics of MN pools is essential for unravelling the neural mechanisms and adaptations underlying motor control, both in healthy and impaired individuals. However, measuring in vivo the properties of complete human MN pools remains an open challenge. Therefore, this work proposes a novel approach based on decoding neural discharges from human MNs in vivo for driving the metaheuristic optimization of biophysically realistic MN models. First, we show that this framework provides subject-specific estimates of MN pool properties from the tibialis anterior muscle on five healthy individuals. Second, we propose a methodology to create complete pools of in silico MNs for each subject. Lastly, we show that neural-data driven complete in silico MN pools reproduce in vivo MN firing characteristics and muscle activation profiles during force-tracking tasks involving isometric ankle dorsi-flexion, at different levels of amplitude. This approach can open new avenues for understanding human neuro-mechanics and, particularly, MN pool dynamics, in a person-specific way. Thereby enabling the development of personalized neurorehabilitation and motor restoring technologies.

Population pharmacokinetic modeling of daridorexant, a novel dual orexin receptor antagonist.

The analysis aimed at identifying subject-specific characteristics (covariates) influencing exposure to daridorexant and quantification of covariate effects to determine clinical relevance. Data from 13 phase I, two phase II, and two phase III studies were pooled to develop a population pharmacokinetic model describing daridorexant concentration over time. Covariate effects were quantified based on model predictions. A two-compartment model with dose-dependent bioavailability, absorption lag time, linear absorption, and nonlinear elimination described the data best. Statistically significant covariates were food status on absorption (lag time and rate constant), time of drug administration (morning, bedtime) on absorption rate constant, lean body weight on central volume of distribution and elimination, fat mass on peripheral volume of distribution and intercompartmental drug transfer, and age and alkaline phosphatase on elimination. Age, lean body weight, fat mass, and alkaline phosphatase influence exposure (area under the curve, time of maximum concentration after dose administration, maximum plasma concentration, and next-morning concentration) to a limited extent, that is, less than 20% difference from a typical subject. Morning administration is not relevant for daridorexant use by insomnia patients. The food effect with simultaneous intake of a high-fat, high-calorie food is an extreme-case scenario unlikely to occur in clinical practice. Body composition, alkaline phosphatase, and age showed clinically negligible effects on exposure to daridorexant. Lean body weight and fat mass described the pharmacokinetics of daridorexant better than other body size descriptors (body weight, height, body mass index), suggesting a convenient physiological alternative to reduce the number of covariates in population pharmacokinetic models. The results indicate that differences between subjects do not require dose adjustments.

The Covariate-Adjusted ROC Curve: The Concept and Its Importance, Review of Inferential Methods, and a New Bayesian Estimator

Accurate diagnosis of disease is of fundamental importance in clinical practice and medical research. Before a medical diagnostic test is routinely used in practice, its ability to distinguish between diseased and nondiseased states must be rigorously assessed. The receiver operating characteristic (ROC) curve is the most popular used tool for evaluating the diagnostic accuracy of continuous-outcome tests. It has been acknowledged that several factors (e.g., subject-specific characteristics such as age and/or gender) can affect the test outcomes and accuracy beyond disease status. Recently, the covariate-adjusted ROC curve has been proposed and successfully applied as a global summary measure of diagnostic accuracy that takes covariate information into account. The aim of this paper is three-fold. First, we motivate the importance of including covariate-information, whenever available, in ROC analysis and, in particular, how the covariate-adjusted ROC curve is an important tool in this context. Second, we review and provide insight on the existing approaches for estimating the covariate-adjusted ROC curve. Third, we develop a highly flexible Bayesian method, based on the combination of a Dirichlet process mixture of additive normal models and the Bayesian bootstrap, for conducting inference about the covariate-adjusted ROC curve. A simulation study is conducted to assess the performance of the different methods and it also demonstrates the ability of our proposed Bayesian model to successfully recover the true covariate-adjusted ROC curve and to produce valid inferences in a variety of complex scenarios. The methods are applied to an endocrine study where the goal is to assess the accuracy of the body mass index, adjusted for age and gender, for detecting clusters of cardiovascular disease risk factors.

Framework of sampling the subject-specific static loads from the gait cycle of interindividual variation

Background and ObjectiveNumerous techniques for bone remodeling simulation have been developed based on Wolff's law. However, most studies have been conducted with empirically determined static loads, which cannot reflect subject-specific characteristics. We recently proposed a new concept of representative static loads (RSLs) to efficiently consider the effect of cyclically repeated dynamic loads on bone remodeling simulation. Based on this concept, the goal of this study is to sample the subject-specific static loads (SSL) from a general gait cycle of interindividual variation. MethodsA total of 15 gait cycles (ten normal and five abnormal cycles) obtained from the public database were used in this study. Each gait cycle was applied to a femur FE model constructed from the clinical CT scan data to evaluate the strain energy distribution as a reference. Then, a natural coordinate was introduced to maintain the predefined locations of extreme points (i.e., two peaks and one valley) for both normal and abnormal gait cycles. To determine the RSLs in the natural coordinate, five out of ten normal gait cycles were used. Through an inverse transformation for each gait cycle, the RSLs in the natural coordinate were converted to the SSLs in the original coordinate. Topology optimization results with the proposed SSLs were compared with those with a single full gait cycle (reference). For comparison, topology optimization was also conducted with empirically determined loads (EDLs) which have been widely used in the literature. ResultsFor normal gait cycles, the proposed SSLs reduced the average computing cost by 95.86% while suppressing the errors of bone mass distribution and apparent stiffness below maximum 4.24% and 1.72%, respectively. Even for abnormal gait cycles, the errors of bone mass distribution and apparent stiffness were suppressed below maximum 9.49% and 2.12%, respectively. Conversely, the conventional EDLs (peak loads selected in this study) showed significantly larger errors of maximum 47.28% and 30.31% in bone mass distribution and apparent stiffness for normal gait cycles. ConclusionBy virtue of using the coordinate transformation for each gait cycle, the proposed SSLs achieved a higher accuracy in the bone mass distribution and apparent stiffness than the previous RSLs and EDLs. Furthermore, this approach can be used for abnormal gait cycles which have higher interindividual variation.

MPose: Environment- and subject-agnostic 3D skeleton posture reconstruction leveraging a single mmWave device

Human skeleton posture reconstruction is an essential component for human–computer interactions (HCI) in various application domains. Traditional approaches usually rely on either cameras or on-body sensors, which induce privacy concerns or inconvenient practical setups. To address these practical concerns, this paper proposes a low-cost contactless skeleton posture reconstruction system, mPose , which can reconstruct a user’s 3D skeleton postures using a single mmWave device. mPose does not require the user to wear any sensors and can enable a broad range of emerging mobile applications (e.g., VR gaming and pervasive user input) via mmWave-5G ready Internet of Things (IoT) devices. Particularly, the system extracts multi-dimensional spatial information from mmWave signals which characterizes the skeleton postures in a 3D space. To mitigate the impacts of environmental changes, mPose dynamically detects the user location and extracts spatial features from the mmWave signals reflected only from the user. Furthermore, we develop a deep regression method with a domain discriminator to learn a mapping between the spatial features and the joint coordinates of human body while removing subject-specific characteristics, realizing robust posture reconstruction across users. Extensive experiments, involving 17 representative body postures, 7 subjects, and 3 indoor environments, show that mPose outperforms contemporary state-of-the-art RF-based solutions with a lower average joint error of only ∼ 30 mm, while achieving transferability across environments and subjects at the same time.

Geometric Deep Learning for Subject Independent Epileptic Seizure Prediction Using Scalp EEG Signals.

Recently, researchers in the biomedical community have introduced deep learning-based epileptic seizure prediction models using electroencephalograms (EEGs) that can anticipate an epileptic seizure by differentiating between the pre-ictal and interictal stages of the subject's brain. Despite having the appearance of a typical anomaly detection task, this problem is complicated by subject-specific characteristics in EEG data. Therefore, studies that investigate seizure prediction widely employ subject-specific models. However, this approach is not suitable in situations where a target subject has limited (or no) data for training. Subject-independent models can address this issue by learning to predict seizures from multiple subjects, and therefore are of greater value in practice. In this study, we propose a subject-independent seizure predictor using Geometric Deep Learning (GDL). In the first stage of our GDL-based method we use graphs derived from physical connections in the EEG grid. We subsequently seek to synthesize subject-specific graphs using deep learning. The models proposed in both stages achieve state-of-the-art performance using a one-hour early seizure prediction window on two benchmark datasets (CHB-MIT-EEG: 95.38% with 23 subjects and Siena-EEG: 96.05% with 15 subjects). To the best of our knowledge, this is the first study that proposes synthesizing subject-specific graphs for seizure prediction. Furthermore, through model interpretation we outline how this method can potentially contribute towards Scalp EEG-based seizure localization.

A Pattern Mining Method for Teaching Practices

When integrating digital technology into teaching, many teachers experience similar challenges. Nevertheless, sharing experiences is difficult as it is usually not possible to transfer teaching scenarios directly from one subject to another because subject-specific characteristics make it difficult to reuse them. To address this problem, instructional scenarios can be described as patterns, which has already been applied in educational contexts. Patterns capture proven teaching strategies and describe teaching scenarios in a unified structure that can be reused. Since priorities for content, methods, and tools are different in each subject, we show an approach to develop a domain-independent graph database to collect digital teaching practices from a taxonomic structure via the intermediate step of an ontology. Furthermore, we outline a method to identify effective teaching practices from interdisciplinary data as patterns from the graph database using an association rule algorithm. The results show that an association-based analysis approach can derive initial indications of effective teaching scenarios.

An Analysis of Glucose Effectiveness in Subjects With or Without Type 2 Diabetes via Hierarchical Modeling.

Glucose effectiveness, defined as the ability of glucose itself to increase glucose utilization and inhibit hepatic glucose production, is an important mechanism maintaining normoglycemia. We conducted a minimal modeling analysis of glucose effectiveness at zero insulin (GEZI) using intravenous glucose tolerance test data from subjects with type 2 diabetes (T2D, n=154) and non-diabetic (ND) subjects (n=343). A hierarchical statistical analysis was performed, which provided a formal mechanism for pooling the data from all study subjects, to yield a single composite population model that quantifies the role of subject specific characteristics such as weight, height, age, sex, and glucose tolerance. Based on the resulting composite population model, GEZI was reduced from 0.021 min–1 (standard error – 0.00078 min–1) in the ND population to 0.011 min–1 (standard error – 0.00045 min–1) in T2D. The resulting model was also employed to calculate the proportion of the non–insulin-dependent net glucose uptake in each subject receiving an intravenous glucose load. Based on individual parameter estimates, the fraction of total glucose disposal independent of insulin was 72.8% ± 12.0% in the 238 ND subjects over the course of the experiment, indicating the major contribution to the whole-body glucose clearance under non-diabetic conditions. This fraction was significantly reduced to 48.8% ± 16.9% in the 30 T2D subjects, although still accounting for approximately half of the total in the T2D population based on our modeling analysis. Given the potential application of glucose effectiveness as a predictor of glucose intolerance and as a potential therapeutic target for treating diabetes, more investigations of glucose effectiveness in other disease conditions can be conducted using the hierarchical modeling framework reported herein.

Learning Continuous Face Age Progression: A Pyramid of GANs.

The two underlying requirements of face age progression, i.e., aging accuracy and identity permanence, are not well studied in the literature. This paper presents a novel generative adversarial network based approach to address the issues in a coupled manner. It separately models the constraints for the intrinsic subject-specific characteristics and the age-specific facial changes with respect to the elapsed time, ensuring that the generated faces present desired aging effects while keeping personalized properties stable. To render photo-realistic facial details, high-level age-specific features conveyed by the synthesized face are estimated by a pyramidal adversarial discriminator at multiple scales, which simulates the aging effects in a finer way. Further, an adversarial learning scheme is introduced to simultaneously train a single generator and multiple parallel discriminators, resulting in smooth continuous face aging sequences. The proposed method is applicable even in the presence of variations in pose, expression, makeup, etc., achieving remarkably vivid aging effects. Quantitative evaluations by a COTS face recognition system demonstrate that the target age distributions are accurately recovered, and 99.88 and 99.98 percent age progressed faces can be correctly verified at 0.001 percent FAR after age transformations of approximately 28 and 23 years elapsed time on the MORPH and CACD databases, respectively. Both visual and quantitative assessments show that the approach advances the state-of-the-art.

EVALUATION OF SPECIFIC ABSORPTION RATE IN THE FAR-FIELD, NEAR-TO-FAR FIELD AND NEAR-FIELD REGIONS FOR INTEGRATIVE RADIOFREQUENCY EXPOSURE ASSESSMENT.

The specific absorption rate (SAR) induced by wireless radiofrequency (RF) systems depends on different parameters. Previously, SAR was mainly assessed under conditions of a single frequency and technology and for a limited number of localized RF sources. The current and emerging mobile systems involve a wider range of usage scenarios and are frequently used simultaneously, leading to combined exposures for which almost no exposure evaluation exists. The aim and novelty of this study is to close this gap of knowledge by developing new methods to rapidly evaluate the SAR induced by RF systems in such scenarios at frequencies from 50MHz to 5.5GHz. To this aim, analytical methods for SAR estimation in several usage scenarios were derived through a large-scale numerical study. These include subject-specific characteristics, properties of the RF systems and provide an estimation of the SAR in the whole body, tissues and organs, and different brain regions.