A Flow Sensor-Based Suction-Index Control Strategy for Rotary Left Ventricular Assist Devices.

Rotary left ventricular assist devices (LVAD) have emerged as a long-term treatment option for patients with advanced heart failure. LVADs need to maintain sufficient physiological perfusion while avoiding left ventricular myocardial damage due to suction at the LVAD inlet. To achieve these objectives, a control algorithm that utilizes a calculated suction index from measured pump flow (SIMPF) is proposed. This algorithm maintained a reference, user-defined SIMPF value, and was evaluated using an in silico model of the human circulatory system coupled to an axial or mixed flow LVAD with 5-10% uniformly distributed measurement noise added to flow sensors. Efficacy of the SIMPF algorithm was compared to a constant pump speed control strategy currently used clinically, and control algorithms proposed in the literature including differential pump speed control, left ventricular end-diastolic pressure control, mean aortic pressure control, and differential pressure control during (1) rest and exercise states; (2) rapid, eight-fold augmentation of pulmonary vascular resistance for (1); and (3) rapid change in physiologic states between rest and exercise. Maintaining SIMPF simultaneously provided sufficient physiological perfusion and avoided ventricular suction. Performance of the SIMPF algorithm was superior to the compared control strategies for both types of LVAD, demonstrating pump independence of the SIMPF algorithm.

Ringing Decay of Gamma Oscillations and Transcranial Magnetic Stimulation Therapy in Autism Spectrum Disorder.

Research suggest that in autism spectrum disorder (ASD) a disturbance in the coordinated interactions of neurons within local networks gives rise to abnormal patterns of brainwave activity in the gamma bandwidth. Low frequency transcranial magnetic stimulation (TMS) over the dorsolateral prefrontal cortex (DLPFC) has been proven to normalize gamma oscillation abnormalities, executive functions, and repetitive behaviors in high functioning ASD individuals. In this study, gamma frequency oscillations in response to a visual classification task (Kanizsa figures) were analyzed and compared in 19 ASD (ADI-R diagnosed, 14.2 ± 3.61 years old, 5 girls) and 19 (14.8 ± 3.67 years old, 5 girls) age/gender matched neurotypical individuals. The ASD group was treated with low frequency TMS (1.0 Hz, 90% motor threshold, 18 weekly sessions) targeting the DLPFC. In autistic subjects, as compared to neurotypicals, significant differences in event-related gamma oscillations were evident in amplitude (higher) pre-TMS. In addition, recordings after TMS treatment in our autistic subjects revealed a significant reduction in the time period to reach peak amplitude and an increase in the decay phase (settling time). The use of a novel metric for gamma oscillations. i.e., envelope analysis, and measurements of its ringing decay allowed us to characterize the impedance of the originating neuronal circuit. The ringing decay or dampening of gamma oscillations is dependent on the inhibitory tone generated by networks of interneurons. The results suggest that the ringing decay of gamma oscillations may provide a biomarker reflective of the excitatory/inhibitory balance of the cortex and a putative outcome measure for interventions in autism.

Cloud-Based Monitoring of Thermal Anomalies in Industrial Environments Using AI and the Internet of Robotic Things.

Recent advancements in cloud computing, artificial intelligence, and the internet of things (IoT) create new opportunities for autonomous industrial environments monitoring. Nevertheless, detecting anomalies in harsh industrial settings remains challenging. This paper proposes an edge-fog-cloud architecture with mobile IoT edge nodes carried on autonomous robots for thermal anomalies detection in aluminum factories. We use companion drones as fog nodes to deliver first response services and a cloud back-end for thermal anomalies analysis. We also propose a self-driving deep learning architecture and a thermal anomalies detection and visualization algorithm. Our results show our robot surveyors are low-cost, deliver reduced response time, and more accurately detect anomalies compared to human surveyors or fixed IoT nodes monitoring the same industrial area. Our self-driving architecture has a root mean square error of 0.19 comparable to VGG-19 with a significantly reduced complexity and three times the frame rate at 60 frames per second. Our thermal to visual registration algorithm maximizes mutual information in the image-gradient domain while adapting to different resolutions and camera frame rates.

Behavioral, Cognitive, and Motor Preparation Deficits in a Visual Cued Spatial Attention Task in Autism Spectrum Disorder.

Abnormalities in motor skills have been regarded as part of the symptomatology characterizing autism spectrum disorder (ASD). It has been estimated that 80 % of subjects with autism display "motor dyspraxia" or clumsiness that are not readily identified in a routine neurological examination. In this study we used behavioral measures, event-related potentials (ERP), and lateralized readiness potential (LRP) to study cognitive and motor preparation deficits contributing to the dyspraxia of autism. A modified Posner cueing task was used to analyze motor preparation abnormalities in children with autism and in typically developing children (N = 30/per group). In this task, subjects engage in preparing motor response based on a visual cue, and then execute a motor movement based on the subsequent imperative stimulus. The experimental conditions, such as the validity of the cue and the spatial location of the target stimuli were manipulated to influence motor response selection, preparation, and execution. Reaction time and accuracy benefited from validly cued targets in both groups, while main effects of target spatial position were more obvious in the autism group. The main ERP findings were prolonged and more negative early frontal potentials in the ASD in incongruent trials in both types of spatial location. The LRP amplitude was larger in incongruent trials and had stronger effect in the children with ASD. These effects were better expressed at the earlier stages of LRP, specifically those related to response selection, and showed difficulties at the cognitive phase of stimulus processing rather that at the motor execution stage. The LRP measures at different stages reflect the chronology of cognitive aspects of movement preparation and are sensitive to manipulations of cue correctness, thus representing very useful biomarker in autism dyspraxia research. Future studies may use more advance and diverse manipulations of movement preparation demands in testing more refined specifics of dyspraxia symptoms to investigate functional connectivity abnormalities underlying motor skills deficits in autism.

Neuromodulation integrating rTMS and neurofeedback for the treatment of autism spectrum disorder: an exploratory study.

Autism spectrum disorder (ASD) is a pervasive developmental disorder characterized by deficits in social interaction, language, stereotyped behaviors, and restricted range of interests. In previous studies low frequency repetitive transcranial magnetic stimulation (rTMS) has been used, with positive behavioral and electrophysiological results, for the experimental treatment in ASD. In this study we combined prefrontal rTMS sessions with electroencephalographic (EEG) neurofeedback (NFB) to prolong and reinforce TMS-induced EEG changes. The pilot trial recruited 42 children with ASD (~14.5 years). Outcome measures included behavioral evaluations and reaction time test with event-related potential (ERP) recording. For the main goal of this exploratory study we used rTMS-neurofeedback combination (TMS-NFB, N = 20) and waitlist (WTL, N = 22) groups to examine effects of 18 sessions of integrated rTMS-NFB treatment or wait period) on behavioral responses, stimulus and response-locked ERPs, and other functional and clinical outcomes. The underlying hypothesis was that combined TMS-NFB will improve executive functions in autistic patients as compared to the WTL group. Behavioral and ERP outcomes were collected in pre- and post-treatment tests in both groups. Results of the study supported our hypothesis by demonstration of positive effects of combined TMS-NFB neurotherapy in active treatment group as compared to control WTL group, as the TMS-NFB group showed significant improvements in behavioral and functional outcomes as compared to the WTL group.

Culturing Cardiac Tissue Slices Under Continuous Physiological Mechanical Stretches.

Testing drugs in vivo and in vitro have been essential elements for the discovery of new therapeutics. Due to the recent advances in in vitro cell culture models, such as human-induced pluripotent stem cell-derived cardiomyocytes and 3D multicell type organoid culture methods, the detection of adverse cardiac events prior to human clinical trials has improved. However, there are still numerous therapeutics whose adverse cardiac effects are not detected until human trials due to the inability of these cell cultures to fully model the complex multicellular organization of an intact human myocardium. Cardiac tissue slices are a possible alternative solution. Myocardial slices are a 300-micron thin snapshot of the myocardium, capturing a section of the adult heart in a 1 × 1 cm section. Using a culture method that incorporates essential nutrients and electrical stimulation, tissue slices can be maintained in culture for 6 days with full viability and functionality. With the addition of mechanical stimulation and humoral cues, tissue slices can be cultured for 12 days. Here we provide detailed methods for how to culture cardiac tissue slices under continuous mechanical stimulation in the cardiac tissue culture model (CTCM) device. The CTCM incorporates four essential factors for maintaining tissue slices in culture for 12 days: mechanical stimulation, electrical stimulation, nutrients, and humoral cues. The CTCM can also be used to model disease conditions, such as overstretch-induced cardiac hypertrophy. The versatility of the CTCM illustrates its potential to be a medium-throughput screening platform for personalized drug testing.

A sensorless, physiologic feedback control strategy to increase vascular pulsatility for rotary blood pumps.

Continuous flow rotary blood pumps (RBP) operating clinically at constant rotational speeds cannot match cardiac demand during varying physical activities, are susceptible to suction, diminish vascular pulsatility, and have an increased risk of adverse events. A sensorless, physiologic feedback control strategy for RBP was developed to mitigate these limitations. The proposed algorithm used intrinsic pump speed to obtain differential pump speed (ΔRPM). The proposed gain-scheduled proportional-integral controller, switching of setpoints between a higher pump speed differential setpoint (ΔRPM Hr ) and a lower pump speed differential setpoint (ΔRPM Lr ), generated pulsatility and physiologic perfusion, while avoiding suction. The switching between ΔRPM Hr and ΔRPM Lr setpoints occurred when the measured ΔRPM reached the pump differential reference setpoint. In-silico tests were implemented to assess the proposed algorithm during rest, exercise, a rapid 3-fold pulmonary vascular resistance increase, rapid change from exercise to rest, and compared with maintaining a constant pump speed setpoint. The proposed control algorithm augmented aortic pressure pulsatility to over 35 mmHg during rest and around 30 mmHg during exercise. Significantly, ventricular suction was avoided, and adequate cardiac output was maintained under all simulated conditions. The performance of the sensorless algorithm using estimation was similar to the performance of sensor-based method. This study demonstrated that augmentation of vascular pulsatility was feasible while avoiding ventricular suction and providing physiological pump outflows. Augmentation of vascular pulsatility can minimize adverse events that have been associated with diminished pulsatility. Mock circulation and animal studies would be conducted to validate these results.

Prefrontal neuromodulation using rTMS improves error monitoring and correction function in autism.

One important executive function known to be compromised in autism spectrum disorder (ASD) is related to response error monitoring and post-error response correction. Several reports indicate that children with ASD show reduced error processing and deficient behavioral correction after an error is committed. Error sensitivity can be readily examined by measuring event-related potentials (ERP) associated with responses to errors, the fronto-central error-related negativity (ERN), and the error-related positivity (Pe). The goal of our study was to investigate whether reaction time (RT), error rate, post-error RT change, ERN, and Pe will show positive changes following 12-week long slow frequency repetitive TMS (rTMS) over dorsolateral prefrontal cortex (DLPFC) in high functioning children with ASD. We hypothesized that 12 sessions of 1 Hz rTMS bilaterally applied over the DLPFC will result in improvements reflected in both behavioral and ERP measures. Participants were randomly assigned to either active rTMS treatment or wait-list (WTL) groups. Baseline and post-TMS/or WTL EEG was collected using 128 channel EEG system. The task involved the recognition of a specific illusory shape, in this case a square or triangle, created by three or four inducer disks. ERN in TMS treatment group became significantly more negative. The number of omission errors decreased post-TMS. The RT did not change, but post-error RT became slower. There were no changes in RT, error rate, post-error RT slowing, nor in ERN/Pe measures in the wait-list group. Our results show significant post-TMS differences in the response-locked ERP such as ERN, as well as behavioral response monitoring measures indicative of improved error monitoring and correction function. The ERN and Pe, along with behavioral performance measures, can be used as functional outcome measures to assess the effectiveness of neuromodulation (e.g., rTMS) in children with autism and thus may have important practical implications.

A deep learning framework for automated classification of histopathological kidney whole-slide images.

Background: Renal cell carcinoma is the most common type of malignant kidney tumor and is responsible for 14,830 deaths per year in the United States. Among the four most common subtypes of renal cell carcinoma, clear cell renal cell carcinoma has the worst prognosis and clear cell papillary renal cell carcinoma appears to have no malignant potential. Distinction between these two subtypes can be difficult due to morphologic overlap on examination of histopathological preparation stained with hematoxylin and eosin. Ancillary techniques, such as immunohistochemistry, can be helpful, but they are not universally available. We propose and evaluate a new deep learning framework for tumor classification tasks to distinguish clear cell renal cell carcinoma from papillary renal cell carcinoma. Methods: Our deep learning framework is composed of three convolutional neural networks. We divided whole-slide kidney images into patches with three different sizes where each network processes a specific patch size. Our framework provides patchwise and pixelwise classification. The histopathological kidney data is composed of 64 image slides that belong to 4 categories: fat, parenchyma, clear cell renal cell carcinoma, and clear cell papillary renal cell carcinoma. The final output of our framework is an image map where each pixel is classified into one class. To maintain consistency, we processed the map with Gauss-Markov random field smoothing. Results: Our framework succeeded in classifying the four classes and showed superior performance compared to well-established state-of-the-art methods (pixel accuracy: 0.89 ResNet18, 0.92 proposed). Conclusions: Deep learning techniques have a significant potential for cancer diagnosis.

Thoracolumbar epidural stimulation effects on bladder and bowel function in uninjured and chronic transected anesthetized rats.

Pre-clinical studies have shown that spinal cord epidural stimulation (scES) at the level of pelvic and pudendal nerve inputs/outputs (L5-S1) alters storage and/or emptying functions of both the bladder and bowel. The current mapping experiments were conducted to investigate scES efficacy at the level of hypogastric nerve inputs/outputs (T13-L2) in male and female rats under urethane anesthesia. As found with L5-S1 scES, T13-L2 scES at select frequencies and intensities of stimulation produced an increase in inter-contraction interval (ICI) in non-injured female rats but a short-latency void in chronic T9 transected rats, as well as reduced rectal activity in all groups. However, the detrusor pressure during the lengthened ICI (i.e., urinary hold) remained at a low pressure and was not elevated as seen with L5-S1 scES, an effect that's critical for translation to the clinic as high fill pressures can damage the kidneys. Furthermore, T13-L2 scES was shown to stimulate voiding post-transection by increasing bladder activity while also directly inhibiting the external urethral sphincter, a pattern necessary to overcome detrusor-sphincter dyssynergia. Additionally, select scES parameters at T13-L2 also increased distal colon activity in all groups. Together, the current findings suggest that optimization of scES for bladder and bowel will likely require multiple electrode cohorts at different locations that target circuitries coordinating sympathetic, parasympathetic and somatic outputs.

Biomimetic cardiac tissue culture model (CTCM) to emulate cardiac physiology and pathophysiology ex vivo.

There is need for a reliable in vitro system that can accurately replicate the cardiac physiological environment for drug testing. The limited availability of human heart tissue culture systems has led to inaccurate interpretations of cardiac-related drug effects. Here, we developed a cardiac tissue culture model (CTCM) that can electro-mechanically stimulate heart slices with physiological stretches in systole and diastole during the cardiac cycle. After 12 days in culture, this approach partially improved the viability of heart slices but did not completely maintain their structural integrity. Therefore, following small molecule screening, we found that the incorporation of 100 nM tri-iodothyronine (T3) and 1 µM dexamethasone (Dex) into our culture media preserved the microscopic structure of the slices for 12 days. When combined with T3/Dex treatment, the CTCM system maintained the transcriptional profile, viability, metabolic activity, and structural integrity for 12 days at the same levels as the fresh heart tissue. Furthermore, overstretching the cardiac tissue induced cardiac hypertrophic signaling in culture, which provides a proof of concept for the ability of the CTCM to emulate cardiac stretch-induced hypertrophic conditions. In conclusion, CTCM can emulate cardiac physiology and pathophysiology in culture for an extended time, thereby enabling reliable drug screening.

Bladder and bowel responses to lumbosacral epidural stimulation in uninjured and transected anesthetized rats.

Spinal cord epidural stimulation (scES) mapping at L5-S1 was performed to identify parameters for bladder and bowel inhibition and/or contraction. Using spinally intact and chronic transected rats of both sexes in acute urethane-anesthetized terminal preparations, scES was systematically applied using a modified Specify 5-6-5 (Medtronic) electrode during bladder filling/emptying cycles while recording bladder and colorectal pressures and external urethral and anal sphincter electromyography activity. The results indicate frequency-dependent effects on void volume, micturition, bowel peristalsis, and sphincter activity just above visualized movement threshold intensities that differed depending upon neurological intactness, with some sex-dependent differences. Thereafter, a custom-designed miniature 15-electrode array designed for greater selectivity was tested and exhibited the same frequency-dependent urinary effects over a much smaller surface area without any concurrent movements. Thus, select activation of autonomic nervous system circuitries with scES is a promising neuromodulation approach for expedient translation to individuals with SCI and potentially other neurologic disorders.

Effects of Transcranial Magnetic Stimulation Therapy on Evoked and Induced Gamma Oscillations in Children with Autism Spectrum Disorder.

Autism spectrum disorder (ASD) is a behaviorally diagnosed neurodevelopmental condition of unknown pathology. Research suggests that abnormalities of elecltroencephalogram (EEG) gamma oscillations may provide a biomarker of the condition. In this study, envelope analysis of demodulated waveforms for evoked and induced gamma oscillations in response to Kanizsa figures in an oddball task were analyzed and compared in 19 ASD and 19 age/gender-matched neurotypical children. The ASD group was treated with low frequency transcranial magnetic stimulation (TMS), (1.0 Hz, 90% motor threshold, 18 weekly sessions) targeting the dorsolateral prefrontal cortex. In ASD subjects, as compared to neurotypicals, significant differences in evoked and induced gamma oscillations were evident in higher magnitude of gamma oscillations pre-TMS, especially in response to non-target cues. Recordings post-TMS treatment in ASD revealed a significant reduction of gamma responses to task-irrelevant stimuli. Participants committed fewer errors post-TMS. Behavioral questionnaires showed a decrease in irritability, hyperactivity, and repetitive behavior scores. The use of a novel metric for gamma oscillations. i.e., envelope analysis using wavelet transformation allowed for characterization of the impedance of the originating neuronal circuit. The results suggest that gamma oscillations may provide a biomarker reflective of the excitatory/inhibitory balance of the cortex and a putative outcome measure for interventions in autism.

A convolutional neural network for the screening and staging of diabetic retinopathy.

Diabetic retinopathy (DR) is a serious retinal disease and is considered as a leading cause of blindness in the world. Ophthalmologists use optical coherence tomography (OCT) and fundus photography for the purpose of assessing the retinal thickness, and structure, in addition to detecting edema, hemorrhage, and scars. Deep learning models are mainly used to analyze OCT or fundus images, extract unique features for each stage of DR and therefore classify images and stage the disease. Throughout this paper, a deep Convolutional Neural Network (CNN) with 18 convolutional layers and 3 fully connected layers is proposed to analyze fundus images and automatically distinguish between controls (i.e. no DR), moderate DR (i.e. a combination of mild and moderate Non Proliferative DR (NPDR)) and severe DR (i.e. a group of severe NPDR, and Proliferative DR (PDR)) with a validation accuracy of 88%-89%, a sensitivity of 87%-89%, a specificity of 94%-95%, and a Quadratic Weighted Kappa Score of 0.91-0.92 when both 5-fold, and 10-fold cross validation methods were used respectively. A prior pre-processing stage was deployed where image resizing and a class-specific data augmentation were used. The proposed approach is considerably accurate in objectively diagnosing and grading diabetic retinopathy, which obviates the need for a retina specialist and expands access to retinal care. This technology enables both early diagnosis and objective tracking of disease progression which may help optimize medical therapy to minimize vision loss.

Correction: Novel stochastic framework for automatic segmentation of human thigh MRI volumes and its applications in spinal cord injured individuals.

[This corrects the article DOI: 10.1371/journal.pone.0216487.].

Novel stochastic framework for automatic segmentation of human thigh MRI volumes and its applications in spinal cord injured individuals.

Severe spinal cord injury (SCI) leads to skeletal muscle atrophy and adipose tissue infiltration in the skeletal muscle, which can result in compromised muscle mechanical output and lead to health-related complications. In this study, we developed a novel automatic 3-D approach for volumetric segmentation and quantitative assessment of thigh Magnetic Resonance Imaging (MRI) volumes in individuals with chronic SCI as well as non-disabled individuals. In this framework, subcutaneous adipose tissue, inter-muscular adipose tissue and total muscle tissue are segmented using linear combination of discrete Gaussians algorithm. Also, three thigh muscle groups were segmented utilizing the proposed 3-D Joint Markov Gibbs Random Field model that integrates first order appearance model, spatial information, and shape model to localize the muscle groups. The accuracy of the automatic segmentation method was tested both on SCI (N = 16) and on non-disabled (N = 14) individuals, showing an overall 0.93±0.06 accuracy for adipose tissue and muscle compartments segmentation based on Dice Similarity Coefficient. The proposed framework for muscle compartment segmentation showed an overall higher accuracy compared to ANTs and STAPLE, two previously validated atlas-based segmentation methods. Also, the framework proposed in this study showed similar Dice accuracy and better Hausdorff distance measure to that obtained using DeepMedic Convolutional Neural Network structure, a well-known deep learning network for 3-D medical image segmentation. The automatic segmentation method proposed in this study can provide fast and accurate quantification of adipose and muscle tissues, which have important health and functional implications in the SCI population.

Comparative approaches for classification of diabetes mellitus data: Machine learning paradigm.

BACKGROUND AND OBJECTIVE: Diabetes is a silent killer. The main cause of this disease is the presence of excessive amounts of metabolites such as glucose. There were about 387 million diabetic people all over the world in 2014. The financial burden of this disease has been calculated to be about $13,700 per year. According to the World Health Organization (WHO), these figures will more than double by the year 2030. This cost will be reduced dramatically if someone can predict diabetes statistically on the basis of some covariates. Although several classification techniques are available, it is very difficult to classify diabetes. The main objectives of this paper are as follows: (i) Gaussian process classification (GPC), (ii) comparative classifier for diabetes data classification, (iii) data analysis using the cross-validation approach, (iv) interpretation of the data analysis and (v) benchmarking our method against others. METHODS: To classify diabetes, several classification techniques are used such as linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), and Naive Bayes (NB). However, most of the medical data show non-normality, non-linearity and inherent correlation structure. So in this paper we adapted Gaussian process (GP)-based classification technique using three kernels namely: linear, polynomial and radial basis kernel. We also investigate the performance of a GP-based classification technique in comparison to existing techniques such as LDA, QDA and NB. Performances are evaluated by using the accuracy (ACC), sensitivity (SE), specificity (SP), positive predictive value (PPV), negative predictive value (NPV) and receiver-operating characteristic (ROC) curves. RESULTS: Pima Indian diabetes dataset is taken as part of the study. This consists of 768 patients, of which 268 patients are diabetic and 500 patients are controls. Our machine learning system shows the performance of GP-based model as: ACC 81.97%, SE 91.79%, SP 63.33%, PPV 84.91% and NPV 62.50% which are larger compared to other methods.

A Novel Early Diagnosis System for Mild Cognitive Impairment Based on Local Region Analysis: A Pilot Study.

Alzheimer's disease (AD) is an irreversible neurodegenerative disorder that accounts for 60-70% of cases of dementia in the elderly. An early diagnosis of AD is usually hampered for many reasons including the variable clinical and pathological features exhibited among affected individuals. This paper presents a computer-aided diagnosis (CAD) system with the primary goal of improving the accuracy, specificity, and sensitivity of diagnosis. In this system, PiB-PET scans, which were obtained from the ADNI database, underwent five essential stages. First, the scans were standardized and de-noised. Second, an Automated Anatomical Labeling (AAL) atlas was utilized to partition the brain into 116 regions or labels that served for local (region-based) diagnosis. Third, scale-invariant Laplacian of Gaussian (LoG) was used, per brain label, to detect the discriminant features. Fourth, the regions' features were analyzed using a general linear model in the form of a two-sample t-test. Fifth, the support vector machines (SVM) and their probabilistic variant (pSVM) were constructed to provide local, followed by global diagnosis. The system was evaluated on scans of normal control (NC) vs. mild cognitive impairment (MCI) (19 NC and 65 MCI scans). The proposed system showed superior accuracy, specificity, and sensitivity as compared to other related work.

Precise segmentation of multimodal images.

We propose new techniques for unsupervised segmentation of multimodal grayscale images such that each region-of-interest relates to a single dominant mode of the empirical marginal probability distribution of grey levels. We follow the most conventional approaches in that initial images and desired maps of regions are described by a joint Markov-Gibbs random field (MGRF) model of independent image signals and interdependent region labels. However, our focus is on more accurate model identification. To better specify region borders, each empirical distribution of image signals is precisely approximated by a linear combination of Gaussians (LCG) with positive and negative components. We modify an expectation-maximization (EM) algorithm to deal with the LCGs and also propose a novel EM-based sequential technique to get a close initial LCG approximation with which the modified EM algorithm should start. The proposed technique identifies individual LCG models in a mixed empirical distribution, including the number of positive and negative Gaussians. Initial segmentation based on the LCG models is then iteratively refined by using the MGRF with analytically estimated potentials. The convergence of the overall segmentation algorithm at each stage is discussed. Experiments show that the developed techniques segment different types of complex multimodal medical images more accurately than other known algorithms.

Relative Power of Specific EEG Bands and Their Ratios during Neurofeedback Training in Children with Autism Spectrum Disorder.

Neurofeedback is a mode of treatment that is potentially useful for improving self-regulation skills in persons with autism spectrum disorder. We proposed that operant conditioning of EEG in neurofeedback mode can be accompanied by changes in the relative power of EEG bands. However, the details on the change of the relative power of EEG bands during neurofeedback training course in autism are not yet well explored. In this study, we analyzed the EEG recordings of children diagnosed with autism and enrolled in a prefrontal neurofeedback treatment course. The protocol used in this training was aimed at increasing the ability to focus attention, and the procedure represented the wide band EEG amplitude suppression training along with upregulation of the relative power of gamma activity. Quantitative EEG analysis was completed for each session of neurofeedback using wavelet transform to determine the relative power of gamma and theta/beta ratio, and further to detect the statistical changes within and between sessions. We found a linear decrease of theta/beta ratio and a liner increase of relative power of gamma activity over 18 weekly sessions of neurofeedback in 18 high functioning children with autism. The study indicates that neurofeedback is an effective method for altering EEG characteristics associated with the autism spectrum disorder. Also, it provides information about specific changes of EEG activities and details the correlation between changes of EEG and neurofeedback indexes during the course of neurofeedback. This pilot study contributes to the development of more effective approaches to EEG data analysis during prefrontal neurofeedback training in autism.