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.

Evaluation of Three State-of-the-Art Classifiers for Recognition of Activities of Daily Living from Smart Home Ambient Data

Smart homes for the aging population have recently started attracting the attention of the research community. The "health state" of smart homes is comprised of many different levels; starting with the physical health of citizens, it also includes longer-term health norms and outcomes, as well as the arena of positive behavior changes. One of the problems of interest is to monitor the activities of daily living (ADL) of the elderly, aiming at their protection and well-being. For this purpose, we installed passive infrared (PIR) sensors to detect motion in a specific area inside a smart apartment and used them to collect a set of ADL. In a novel approach, we describe a technology that allows the ground truth collected in one smart home to train activity recognition systems for other smart homes. We asked the users to label all instances of all ADL only once and subsequently applied data mining techniques to cluster in-home sensor firings. Each cluster would therefore represent the instances of the same activity. Once the clusters were associated to their corresponding activities, our system was able to recognize future activities. To improve the activity recognition accuracy, our system preprocessed raw sensor data by identifying overlapping activities. To evaluate the recognition performance from a 200-day dataset, we implemented three different active learning classification algorithms and compared their performance: naive Bayesian (NB), support vector machine (SVM) and random forest (RF). Based on our results, the RF classifier recognized activities with an average specificity of 96.53%, a sensitivity of 68.49%, a precision of 74.41% and an F-measure of 71.33%, outperforming both the NB and SVM classifiers. Further clustering markedly improved the results of the RF classifier. An activity recognition system based on PIR sensors in conjunction with a clustering classification approach was able to detect ADL from datasets collected from different homes. Thus, our PIR-based smart home technology could improve care and provide valuable information to better understand the functioning of our societies, as well as to inform both individual and collective action in a smart city scenario.

Recognition of activities of daily living in healthy subjects using two ad-hoc classifiers

Activities of daily living (ADL) are important for quality of life. They are indicators of cognitive health status and their assessment is a measure of independence in everyday living. ADL are difficult to reliably assess using questionnaires due to self-reporting biases. Various sensor-based (wearable, in-home, intrusive) systems have been proposed to successfully recognize and quantify ADL without relying on self-reporting. New classifiers required to classify sensor data are on the rise. We propose two ad-hoc classifiers that are based only on non-intrusive sensor data. METHODS: A wireless sensor system with ten sensor boxes was installed in the home of ten healthy subjects to collect ambient data over a duration of 20 consecutive days. A handheld protocol device and a paper logbook were also provided to the subjects. Eight ADL were selected for recognition. We developed two ad-hoc ADL classifiers, namely the rule based forward chaining inference engine (RBI) classifier and the circadian activity rhythm (CAR) classifier. The RBI classifier finds facts in data and matches them against the rules. The CAR classifier works within a framework to automatically rate routine activities to detect regular repeating patterns of behavior. For comparison, two state-of-the-art [Naïves Bayes (NB), Random Forest (RF)] classifiers have also been used. All classifiers were validated with the collected data sets for classification and recognition of the eight specific ADL. RESULTS: Out of a total of 1,373 ADL, the RBI classifier correctly determined 1,264, while missing 109 and the CAR determined 1,305 while missing 68 ADL. The RBI and CAR classifier recognized activities with an average sensitivity of 91.27 and 94.36%, respectively, outperforming both RF and NB. CONCLUSIONS: The performance of the classifiers varied significantly and shows that the classifier plays an important role in ADL recognition. Both RBI and CAR classifier performed better than existing state-of-the-art (NB, RF) on all ADL. Of the two ad-hoc classifiers, the CAR classifier was more accurate and is likely to be better suited than the RBI for distinguishing and recognizing complex ADL.

An Early-Warning System for Hypo-/Hyperglycemic Events Based on Fusion of Adaptive Prediction Models

Introduction: Early warning of future hypoglycemic and hyperglycemic events can improve the safety of type 1 diabetes mellitus (T1DM) patients. The aim of this study is to design and evaluate a hypoglycemia / hyperglycemia early warning system (EWS) for T1DM patients under sensor-augmented pump (SAP) therapy. Methods: The EWS is based on the combination of data-driven online adaptive prediction models and a warning algorithm. Three modeling approaches have been investigated: (i) autoregressive (ARX) models, (ii) auto-regressive with an output correction module (cARX) models, and (iii) recurrent neural network (RNN) models. The warning algorithm performs postprocessing of the models′ outputs and issues alerts if upcoming hypoglycemic/hyperglycemic events are detected. Fusion of the cARX and RNN models, due to their complementary prediction performances, resulted in the hybrid autoregressive with an output correction module/recurrent neural network (cARN)-based EWS. Results: The EWS was evaluated on 23 T1DM patients under SAP therapy. The ARX-based system achieved hypoglycemic (hyperglycemic) event prediction with median values of accuracy of 100.0% (100.0%), detection time of 10.0 (8.0) min, and daily false alarms of 0.7 (0.5). The respective values for the cARX-based system were 100.0% (100.0%), 17.5 (14.8) min, and 1.5 (1.3) and, for the RNN-based system, were 100.0% (92.0%), 8.4 (7.0) min, and 0.1 (0.2). The hybrid cARN-based EWS presented outperforming results with 100.0% (100.0%) prediction accuracy, detection 16.7 (14.7) min in advance, and 0.8 (0.8) daily false alarms. Conclusion: Combined use of cARX and RNN models for the development of an EWS outperformed the single use of each model, achieving accurate and prompt event prediction with few false alarms, thus providing increased safety and comfort.

Fusion of fundus images and MRI data of the human eye

Purpose Ophthalmologists are confronted with a set of different image modalities to diagnose eye tumors e.g., fundus photography, CT and MRI. However, these images are often complementary and represent pathologies differently. Some aspects of tumors can only be seen in a particular modality. A fusion of modalities would improve the contextual information for diagnosis. The presented work attempts to register color fundus photography with MRI volumes. This would complement the low resolution 3D information in the MRI with high resolution 2D fundus images. Methods MRI volumes were acquired from 12 infants under the age of 5 with unilateral retinoblastoma. The contrast-enhanced T1-FLAIR sequence was performed with an isotropic resolution of less than 0.5mm. Fundus images were acquired with a RetCam camera. For healthy eyes, two landmarks were used: the optic disk and the fovea. The eyes were detected and extracted from the MRI volume using a 3D adaption of the Fast Radial Symmetry Transform (FRST). The cropped volume was automatically segmented using the Split Bregman algorithm. The optic nerve was enhanced by a Frangi vessel filter. By intersection the nerve with the retina the optic disk was found. The fovea position was estimated by constraining the position with the angle between the optic and the visual axis as well as the distance from the optic disk. The optical axis was detected automatically by fitting a parable on to the lens surface. On the fundus, the optic disk and the fovea were detected by using the method of Budai et al. Finally, the image was projected on to the segmented surface using the lens position as the camera center. In tumor affected eyes, the manually segmented tumors were used instead of the optic disk and macula for the registration. Results In all of the 12 MRI volumes that were tested the 24 eyes were found correctly, including healthy and pathological cases. In healthy eyes the optic nerve head was found in all of the tested eyes with an error of 1.08 +/- 0.37mm. A successful registration can be seen in figure 1. Conclusions The presented method is a step toward automatic fusion of modalities in ophthalmology. The combination enhances the MRI volume with higher resolution from the color fundus on the retina. Tumor treatment planning is improved by avoiding critical structures and disease progression monitoring is made easier.

Toward optimal display of physiologic status in critical care: I. Recreating bedside displays from archived physiologic data

BACKGROUND: Physiologic data display is essential to decision making in critical care. Current displays echo first-generation hemodynamic monitors dating to the 1970s and have not kept pace with new insights into physiology or the needs of clinicians who must make progressively more complex decisions about their patients. The effectiveness of any redesign must be tested before deployment. Tools that compare current displays with novel presentations of processed physiologic data are required. Regenerating conventional physiologic displays from archived physiologic data is an essential first step. OBJECTIVES: The purposes of the study were to (1) describe the SSSI (single sensor single indicator) paradigm that is currently used for physiologic signal displays, (2) identify and discuss possible extensions and enhancements of the SSSI paradigm, and (3) develop a general approach and a software prototype to construct such "extended SSSI displays" from raw data. RESULTS: We present Multi Wave Animator (MWA) framework-a set of open source MATLAB (MathWorks, Inc., Natick, MA, USA) scripts aimed to create dynamic visualizations (eg, video files in AVI format) of patient vital signs recorded from bedside (intensive care unit or operating room) monitors. Multi Wave Animator creates animations in which vital signs are displayed to mimic their appearance on current bedside monitors. The source code of MWA is freely available online together with a detailed tutorial and sample data sets.

Rapid fusion and syncytium formation of heterologous cells upon expression of the FGFRL1 receptor

The fusion of mammalian cells into syncytia is a developmental process that is tightly restricted to a limited subset of cells. Besides gamete and placental trophoblast fusion, only macrophages and myogenic stem cells fuse into multinucleated syncytia. In contrast to viral cell fusion, which is mediated by fusogenic glycoproteins that actively merge membranes, mammalian cell fusion is poorly understood at the molecular level. A variety of mammalian transmembrane proteins, among them many of the immunoglobulin superfamily, have been implicated in cell-cell fusion, but none has been shown to actively fuse cells in vitro. Here we report that the FGFRL1 receptor, which is up-regulated during the differentiation of myoblasts into myotubes, fuses cultured cells into large, multinucleated syncytia. We used luciferase and GFP-based reporter assays to confirm cytoplasmic mixing and to identify the fusion inducing domain of FGFRL1. These assays revealed that Ig-like domain III and the transmembrane domain are both necessary and sufficient to rapidly fuse CHO cells into multinucleated syncytia comprising several hundred nuclei. Moreover, FGFRL1 also fused HEK293 and HeLa cells with untransfected CHO cells. Our data show that FGFRL1 is the first mammalian protein that is capable of inducing syncytium formation of heterologous cells in vitro.

MaxMAC: a Maximally Traffic-Adaptive MAC Protocol for Wireless Sensor Networks

Energy efficiency is a major concern in the design of Wireless Sensor Networks (WSNs) and their communication protocols. As the radio transceiver typically accounts for a major portion of a WSN node’s power consumption, researchers have proposed Energy-Efficient Medium Access (E2-MAC) protocols that switch the radio transceiver off for a major part of the time. Such protocols typically trade off energy-efficiency versus classical quality of service parameters (throughput, latency, reliability). Today’s E2-MAC protocols are able to deliver little amounts of data with a low energy footprint, but introduce severe restrictions with respect to throughput and latency. Regrettably, they yet fail to adapt to varying traffic load at run-time. This paper presents MaxMAC, an E2-MAC protocol that targets at achieving maximal adaptivity with respect to throughput and latency. By adaptively tuning essential parameters at run-time, the protocol reaches the throughput and latency of energy-unconstrained CSMA in high-traffic phases, while still exhibiting a high energy-efficiency in periods of sparse traffic. The paper compares the protocol against a selection of today’s E2-MAC protocols and evaluates its advantages and drawbacks.

TARWIS - A Testbed Management Architecture for Wireless Sensor Network Testbeds

With research on Wireless Sensor Networks (WSNs) becoming more and more mature in the past five years, researchers from universities all over the world have set up testbeds of wireless sensor networks, in most cases to test and evaluate the real-world behavior of developed WSN protocol mechanisms. Although these testbeds differ heavily in the employed sensor node types and the general architectural set up, they all have similar requirements with respect to management and scheduling functionalities: as every shared resource, a testbed requires a notion of users, resource reservation features, support for reprogramming and reconfiguration of the nodes, provisions to debug and remotely reset sensor nodes in case of node failures, as well as a solution for collecting and storing experimental data. The TARWIS management architecture presented in this paper targets at providing these functionalities independent from node type and node operating system. TARWIS has been designed as a re-usable management solution for research and/or educational oriented research testbeds of wireless sensor networks, relieving researchers intending to deploy a testbed from the burden to implement their own scheduling and testbed management solutions from scratch.

Intraoperative determination of the load-displacement behavior of scoliotic spinal motion segments: preliminary clinical results

Introduction: Spinal fusion is a widely and successfully performed strategy for the treatment of spinal deformities and degenerative diseases. The general approach has been to stabilize the spine with implants so that a solid bony fusion between the vertebrae can develop. However, new implant designs have emerged that aim at preservation or restoration of the motion of the spinal segment. In addition to static, load sharing principles, these designs also require a profound knowledge of kinematic and dynamic properties to properly characterise the in vivo performance of the implants. Methods: To address this, an apparatus was developed that enables the intraoperative determination of the load–displacement behavior of spinal motion segments. The apparatus consists of a sensor-equipped distractor to measure the applied force between the transverse processes, and an optoelectronic camera to track the motion of vertebrae and the distractor. In this intraoperative trial, measurements from two patients with adolescent idiopathic scoliosis with right thoracic curves were made at four motion segments each. Results: At a lateral bending moment of 5 N m, the mean flexibility of all eight motion segments was 0.18 ± 0.08°/N m on the convex side and 0.24 ± 0.11°/N m on the concave side. Discussion: The results agree with published data obtained from cadaver studies with and without axial preload. Intraoperatively acquired data with this method may serve as an input for mathematical models and contribute to the development of new implants and treatment strategies.

An early warning system for hypoglycemic/hyperglycemic events based on fusion of adaptive prediction models

Early warning of future hypoglycemic and hyperglycemic events can improve the safety of type 1 diabetes mellitus (T1DM) patients. The aim of this study is to design and evaluate a hypoglycemia/hyperglycemia early warning system (EWS) for T1DM patients under sensor-augmented pump (SAP) therapy.

A smart multi-hop hierarchical routing protocol for efficient video communication over wireless multimedia sensor networks

For smart applications, nodes in wireless multimedia sensor networks (MWSNs) have to take decisions based on sensed scalar physical measurements. A routing protocol must provide the multimedia delivery with quality level support and be energy-efficient for large-scale networks. With this goal in mind, this paper proposes a smart Multi-hop hierarchical routing protocol for Efficient VIdeo communication (MEVI). MEVI combines an opportunistic scheme to create clusters, a cross-layer solution to select routes based on network conditions, and a smart solution to trigger multimedia transmission according to sensed data. Simulations were conducted to show the benefits of MEVI compared with the well-known Low-Energy Adaptive Clustering Hierarchy (LEACH) protocol. This paper includes an analysis of the signaling overhead, energy-efficiency, and video quality.

SNOMC: An overlay multicast protocol for Wireless Sensor Networks

Using multicast communication in Wireless Sensor Networks (WSNs) is an efficient way to disseminate the same data (from one sender) to multiple receivers, e.g., transmitting code updates to a group of sensor nodes. Due to the nature of code update traffic a multicast protocol has to support bulky traffic and end-to-end reliability. We are interested in an energy-efficient multicast protocol due to the limited resources of wireless sensor nodes. Current data dissemination schemes do not fulfill the above requirements. In order to close the gap, we designed and implemented the SNOMC (Sensor Node Overlay Multicast) protocol. It is an overlay multicast protocol, which supports reliable, time-efficient, and energy-efficient data dissemination of bulky data from one sender to many receivers. To ensure end-to-end reliability, SNOMC uses a NACK-based reliability mechanism with different caching strategies.