Lung Pattern Classification for Interstitial Lung Diseases Using a Deep Convolutional Neural Network

Automated tissue characterization is one of the most crucial components of a computer aided diagnosis (CAD) system for interstitial lung diseases (ILDs). Although much research has been conducted in this field, the problem remains challenging. Deep learning techniques have recently achieved impressive results in a variety of computer vision problems, raising expectations that they might be applied in other domains, such as medical image analysis. In this paper, we propose and evaluate a convolutional neural network (CNN), designed for the classification of ILD patterns. The proposed network consists of 5 convolutional layers with 2×2 kernels and LeakyReLU activations, followed by average pooling with size equal to the size of the final feature maps and three dense layers. The last dense layer has 7 outputs, equivalent to the classes considered: healthy, ground glass opacity (GGO), micronodules, consolidation, reticulation, honeycombing and a combination of GGO/reticulation. To train and evaluate the CNN, we used a dataset of 14696 image patches, derived by 120 CT scans from different scanners and hospitals. To the best of our knowledge, this is the first deep CNN designed for the specific problem. A comparative analysis proved the effectiveness of the proposed CNN against previous methods in a challenging dataset. The classification performance (~85.5%) demonstrated the potential of CNNs in analyzing lung patterns. Future work includes, extending the CNN to three-dimensional data provided by CT volume scans and integrating the proposed method into a CAD system that aims to provide differential diagnosis for ILDs as a supportive tool for radiologists.

The neural network of saccadic foreknowledge.

Foreknowledge about upcoming events may be exploited to optimize behavioural responses. In a previous work, using an eye movement paradigm, we showed that different types of partial foreknowledge have different effects on saccadic efficiency. In the current study, we investigated the neural circuitry involved in processing of partial foreknowledge using functional magnetic resonance imaging. Fourteen subjects performed a mixed antisaccade, prosaccade paradigm with blocks of no foreknowledge, complete foreknowledge or partial foreknowledge about stimulus location, response direction or task. We found that saccadic foreknowledge is processed primarily within the well-known oculomotor network for saccades and antisaccades. Moreover, we found a consistent decrease in BOLD activity in the primary and secondary visual cortex in all foreknowledge conditions compared to the no-foreknowledge conditions. Furthermore we found that the different types of partial foreknowledge are processed in distinct brain areas: response foreknowledge is processed in the frontal eye field, while stimulus foreknowledge is processed in the frontal and parietal eye field. Task foreknowledge, however, revealed no positive BOLD correlate. Our results show different patterns of engagement in the saccade-related neural network depending upon precisely what type of information is known ahead.

Semantic memory involvement in the default mode network: a functional neuroimaging study using independent component analysis

The Default Mode Network (DMN) is a higher order functional neural network that displays activation during passive rest and deactivation during many types of cognitive tasks. Accordingly, the DMN is viewed to represent the neural correlate of internally-generated self-referential cognition. This hypothesis implies that the DMN requires the involvement of cognitive processes, like declarative memory. The present study thus examines the spatial and functional convergence of the DMN and the semantic memory system. Using an active block-design functional Magnetic Resonance Imaging (fMRI) paradigm and Independent Component Analysis (ICA), we trace the DMN and fMRI signal changes evoked by semantic, phonological and perceptual decision tasks upon visually-presented words. Our findings show less deactivation during semantic compared to the two non-semantic tasks for the entire DMN unit and within left-hemispheric DMN regions, i.e., the dorsal medial prefrontal cortex, the anterior cingulate cortex, the retrosplenial cortex, the angular gyrus, the middle temporal gyrus and the anterior temporal region, as well as the right cerebellum. These results demonstrate that well-known semantic regions are spatially and functionally involved in the DMN. The present study further supports the hypothesis of the DMN as an internal mentation system that involves declarative memory functions.

An insulin infusion advisory system based on autotuning nonlinear model-predictive control

This paper aims at the development and evaluation of a personalized insulin infusion advisory system (IIAS), able to provide real-time estimations of the appropriate insulin infusion rate for type 1 diabetes mellitus (T1DM) patients using continuous glucose monitors and insulin pumps. The system is based on a nonlinear model-predictive controller (NMPC) that uses a personalized glucose-insulin metabolism model, consisting of two compartmental models and a recurrent neural network. The model takes as input patient's information regarding meal intake, glucose measurements, and insulin infusion rates, and provides glucose predictions. The predictions are fed to the NMPC, in order for the latter to estimate the optimum insulin infusion rates. An algorithm based on fuzzy logic has been developed for the on-line adaptation of the NMPC control parameters. The IIAS has been in silico evaluated using an appropriate simulation environment (UVa T1DM simulator). The IIAS was able to handle various meal profiles, fasting conditions, interpatient variability, intraday variation in physiological parameters, and errors in meal amount estimations.

An insulin infusion advisory system for type 1 diabetes patients based on non-linear model predictive control methods

In this paper, an Insulin Infusion Advisory System (IIAS) for Type 1 diabetes patients, which use insulin pumps for the Continuous Subcutaneous Insulin Infusion (CSII) is presented. The purpose of the system is to estimate the appropriate insulin infusion rates. The system is based on a Non-Linear Model Predictive Controller (NMPC) which uses a hybrid model. The model comprises a Compartmental Model (CM), which simulates the absorption of the glucose to the blood due to meal intakes, and a Neural Network (NN), which simulates the glucose-insulin kinetics. The NN is a Recurrent NN (RNN) trained with the Real Time Recurrent Learning (RTRL) algorithm. The output of the model consists of short term glucose predictions and provides input to the NMPC, in order for the latter to estimate the optimum insulin infusion rates. For the development and the evaluation of the IIAS, data generated from a Mathematical Model (MM) of a Type 1 diabetes patient have been used. The proposed control strategy is evaluated at multiple meal disturbances, various noise levels and additional time delays. The results indicate that the implemented IIAS is capable of handling multiple meals, which correspond to realistic meal profiles, large noise levels and time delays.

Visuospatial working memory in very preterm and term born children - Impact of age and performance

Working memory is crucial for meeting the challenges of daily life and performing academic tasks, such as reading or arithmetic. Very preterm born children are at risk of low working memory capacity. The aim of this study was to examine the visuospatial working memory network of school-aged preterm children and to determine the effect of age and performance on the neural working memory network. Working memory was assessed in 41 very preterm born children and 36 term born controls (aged 7–12 years) using functional magnetic resonance imaging (fMRI) and neuropsychological assessment. While preterm children and controls showed equal working memory performance, preterm children showed less involvement of the right middle frontal gyrus, but higher fMRI activation in superior frontal regions than controls. The younger and low-performing preterm children presented an atypical working memory network whereas the older high-performing preterm children recruited a working memory network similar to the controls. Results suggest that younger and low-performing preterm children show signs of less neural efficiency in frontal brain areas. With increasing age and performance, compensational mechanisms seem to occur, so that in preterm children, the typical visuospatial working memory network is established by the age of 12 years.

Multi-Model Data Fusion to Improve an Early Warning System for Hypo-/Hyperglycemic Events

Correct predictions of future blood glucose levels in individuals with Type 1 Diabetes (T1D) can be used to provide early warning of upcoming hypo-/hyperglycemic events and thus to improve the patient's safety. To increase prediction accuracy and efficiency, various approaches have been proposed which combine multiple predictors to produce superior results compared to single predictors. Three methods for model fusion are presented and comparatively assessed. Data from 23 T1D subjects under sensor-augmented pump (SAP) therapy were used in two adaptive data-driven models (an autoregressive model with output correction - cARX, and a recurrent neural network - RNN). Data fusion techniques based on i) Dempster-Shafer Evidential Theory (DST), ii) Genetic Algorithms (GA), and iii) Genetic Programming (GP) were used to merge the complimentary performances of the prediction models. The fused output is used in a warning algorithm to issue alarms of upcoming hypo-/hyperglycemic events. The fusion schemes showed improved performance with lower root mean square errors, lower time lags, and higher correlation. In the warning algorithm, median daily false alarms (DFA) of 0.25%, and 100% correct alarms (CA) were obtained for both event types. The detection times (DT) before occurrence of events were 13.0 and 12.1 min respectively for hypo-/hyperglycemic events. Compared to the cARX and RNN models, and a linear fusion of the two, the proposed fusion schemes represents a significant improvement.

An EEG approach to the neurodevelopmental hypothesis of schizophrenia studying schizophrenics, normal controls and adolescents

Based on an integrative brain model which focuses on memory-driven and EEG state-dependent information processing for the organisation of behaviour, we used the developmental changes of the awake EEG to further investigate the hypothesis that neurodevelopmental abnormalities (deviations in organisation and reorganisation of cortico-cortical connectivity during development) are involved in the pathogenesis of schizophrenia. First-episode, neuroleptic-naive schizophrenics and their matched controls and three age groups of normal adolescents were studied (total: 70 subjects). 19-channel EEG delta-theta, alpha and beta spectral band centroid frequencies during resting (baseline) and after verbal stimuli were used as measure of the level of attained complexity and momentary excitability of the neuronal network (working memory). Schizophrenics compared with all control groups showed lower delta-theta activity centroids and higher alpha and beta activity centroids. Reactivity centroids (centroid after stimulus minus centroid during resting) were used as measure of update of working memory. Schizophrenics showed partial similarities in delta-theta and beta reactivity centroids with the 11-year olds and in alpha reactivity centroids with the 13-year olds. Within the framework of our model, the results suggest multifactorially elicited imbalances in the level of excitability of neuronal networks in schizophrenia, resulting in network activation at dissociated complexity levels, partially regressed and partially prematurely developed. It is hypothesised that activation of age- and/or state-inadequate representations for coping with realities becomes manifest as productive schizophrenic symptoms. Thus, the results support some aspects of the neurodevelopmental hypothesis.

A real time simulation model of glucose-insulin metabolism for type 1 diabetes patients

In this paper, a simulation model of glucose-insulin metabolism for Type 1 diabetes patients is presented. The proposed system is based on the combination of Compartmental Models (CMs) and artificial Neural Networks (NNs). This model aims at the development of an accurate system, in order to assist Type 1 diabetes patients to handle their blood glucose profile and recognize dangerous metabolic states. Data from a Type 1 diabetes patient, stored in a database, have been used as input to the hybrid system. The data contain information about measured blood glucose levels, insulin intake, and description of food intake, along with the corresponding time. The data are passed to three separate CMs, which produce estimations about (i) the effect of Short Acting (SA) insulin intake on blood insulin concentration, (ii) the effect of Intermediate Acting (IA) insulin intake on blood insulin concentration, and (iii) the effect of carbohydrate intake on blood glucose absorption from the gut. The outputs of the three CMs are passed to a Recurrent NN (RNN) in order to predict subsequent blood glucose levels. The RNN is trained with the Real Time Recurrent Learning (RTRL) algorithm. The resulted blood glucose predictions are promising for the use of the proposed model for blood glucose level estimation for Type 1 diabetes patients.

Computer-aided diagnosis of carotid atherosclerosis based on ultrasound image statistics, laws' texture and neural networks

Quantitative characterisation of carotid atherosclerosis and classification into symptomatic or asymptomatic is crucial in planning optimal treatment of atheromatous plaque. The computer-aided diagnosis (CAD) system described in this paper can analyse ultrasound (US) images of carotid artery and classify them into symptomatic or asymptomatic based on their echogenicity characteristics. The CAD system consists of three modules: a) the feature extraction module, where first-order statistical (FOS) features and Laws' texture energy can be estimated, b) the dimensionality reduction module, where the number of features can be reduced using analysis of variance (ANOVA), and c) the classifier module consisting of a neural network (NN) trained by a novel hybrid method based on genetic algorithms (GAs) along with the back propagation algorithm. The hybrid method is able to select the most robust features, to adjust automatically the NN architecture and to optimise the classification performance. The performance is measured by the accuracy, sensitivity, specificity and the area under the receiver-operating characteristic (ROC) curve. The CAD design and development is based on images from 54 symptomatic and 54 asymptomatic plaques. This study demonstrates the ability of a CAD system based on US image analysis and a hybrid trained NN to identify atheromatous plaques at high risk of stroke.

Site- and species-specific responses of forest growth to climate across the European continent

Aim To evaluate the climate sensitivity of model-based forest productivity estimates using a continental-scale tree-ring network. Location Europe and North Africa (30–70° N, 10° W–40° E). Methods We compiled close to 1000 annually resolved records of radial tree growth for all major European tree species and quantified changes in growth as a function of historical climatic variation. Sites were grouped using a neural network clustering technique to isolate spatiotemporal and species-specific climate response patterns. The resulting empirical climate sensitivities were compared with the sensitivities of net primary production (NPP) estimates derived from the ORCHIDEE-FM and LPJ-wsl dynamic global vegetation models (DGVMs). Results We found coherent biogeographic patterns in climate response that depend upon (1) phylogenetic controls and (2) ambient environmental conditions delineated by latitudinal/elevational location. Temperature controls dominate forest productivity in high-elevation and high-latitude areas whereas moisture sensitive sites are widespread at low elevation in central and southern Europe. DGVM simulations broadly reproduce the empirical patterns, but show less temperature sensitivity in the boreal zone and stronger precipitation sensitivity towards the mid-latitudes. Main conclusions Large-scale forest productivity is driven by monthly to seasonal climate controls, but our results emphasize species-specific growth patterns under comparable environmental conditions. Furthermore, we demonstrate that carry-over effects from the previous growing season can significantly influence tree growth, particularly in areas with harsh climatic conditions – an element not considered in most current-state DGVMs. Model–data discrepancies suggest that the simulated climate sensitivity of NPP will need refinement before carbon-cycle climate feedbacks can be accurately quantified.