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.

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.

Innate and adaptive immunity in host-microbiota mutualism

Healthy individuals live in peaceful co-existence with an immense load of intestinal bacteria. This symbiosis is advantageous for both the host and the bacteria. For the host it provides access to otherwise undigestible nutrients and colonization resistance against pathogens. In return the bacteria receive an excellent nutrient habitat. The mucosal immune adaptations to the presence of this commensal intestinal microflora are manifold. Although bacterial colonization has clear systemic consequences, such as maturation of the immune system, it is striking that the mutualistic adaptive (T and B cells) and innate immune responses are precisely compartmentalized to the mucosal immune system. Here we summarize the mechanisms of mucosal immune compartmentalization and its importance for a healthy host-microbiota mutualism.

[How the bovine viral diarrhea virus outwits the immune system]

The interaction of bovine viral diarrhea virus (BVD virus) with its host has several unique features, most notably the capacity to infect its host either transiently or persistently. The transient infection stimulates an antiviral immune reaction similar to that seen in other transient viral infections. In contrast, being associated with immunotolerance specific for the infecting BVD viral strain, the persistent infection differs fundamentally from other persistent infections like those caused by lentiviruses. Whereas the latter are characterized by complex viral evasion of the host's adaptive immune response by mechanisms such as antigenic drift and interference with presentation of T cell epitopes, BVD virus avoids the immune response altogether by inducing both humoral and cellular immune tolerance. This is made possible by invasion of the fetus at an early stage of development. In addition to adaptive immunity, BVD virus also manipulates key elements of the host's innate immune response. The non-cytopathic biotype of BVD virus, which is capable of persistently infecting its host, fails to induce type I interferon. In addition, persistently infected cells are resistant to the induction of apoptosis by double-stranded RNA and do not produce interferon when treated with this pathogen-associated molecular pattern (PAMP) that signals viral infection. Moreover, when treated with interferon, cells persistently infected with non-cytopathic BVD virus do not clear the virus. Surprisingly, however, despite this lack of effect on persistent infection, interferon readily induces an antiviral state in these cells, as shown by the protection against infection by unrelated viruses. Overall, BVD virus manipulates the host's interferon defense in a manner that optimises its chances of maintaining the persistent infection as well as decreasing the risks that heterologous viral infections may carry for the host. Thus, since not all potential host cells are infected in animals persistently infected with BVD virus, heterologous viruses replicating in cells uninfected with BVD virus will still trigger production of interferon. Interferon produced by such cells will curtail the replication of heterologous viruses only, be that in cells already infected with BVD virus, or in cells in which the heterologous virus may replicate alone. From an evolutionary viewpoint, this strategy clearly enhances the chances of transmission of BVD virus to new hosts, as it attenuates the negative effects that a global immunosuppression would have on the survival of persistently infected animals.

The mucosal immune system at the gastrointestinal barrier

The immune system faces a considerable challenge in its efforts to maintain tissue homeostasis in the intestinal mucosa. It is constantly confronted with a large array of antigens, and has to prevent the dissemination and proliferation of potentially harmful agents while sparing the vital structures of the intestine from immune-mediated destruction. Complex interactions between the highly adapted effector cells and mechanisms of the innate and adaptive immune system generally prevent the luminal microflora from penetrating the intestinal mucosa and from spreading systemically. Non-haematopoietic cells critically contribute to the maintenance of local tissue homeostasis in an antigen-rich environment by producing protective factors (e.g. production of mucus by goblet cells, or secretion of microbicidal defensins by Paneth cells) and also through interactions with the adaptive and innate immune system (such as the production of chemotactic factors that lead to the selective recruitment of immune cell subsets). The complexity of the regulatory mechanisms that control the local immune response to luminal antigens is also reflected in the observation that mutations in immunologically relevant genes often lead to the development of uncontrolled inflammatory reactions in the microbially colonized intestine of experimental animals.

A two-microphone noise reduction system for cochlear implant users with nearby microphones. Part I: Signal processing algorithm design and development

Users of cochlear implant systems, that is, of auditory aids which stimulate the auditory nerve at the cochlea electrically, often complain about poor speech understanding in noisy environments. Despite the proven advantages of multimicrophone directional noise reduction systems for conventional hearing aids, only one major manufacturer has so far implemented such a system in a product, presumably because of the added power consumption and size. We present a physically small (intermicrophone distance 7 mm) and computationally inexpensive adaptive noise reduction system suitable for behind-the-ear cochlear implant speech processors. Supporting algorithms, which allow the adjustment of the opening angle and the maximum noise suppression, are proposed and evaluated. A portable real-time device for test in real acoustic environments is presented.

A two-microphone noise reduction system for cochlear implant users with nearby microphones. Part II: Performance Evaluation

Users of cochlear implants (auditory aids, which stimulate the auditory nerve electrically at the inner ear) often suffer from poor speech understanding in noise. We evaluate a small (intermicrophone distance 7 mm) and computationally inexpensive adaptive noise reduction system suitable for behind-the-ear cochlear implant speech processors. The system is evaluated in simulated and real, anechoic and reverberant environments. Results from simulations show improvements of 3.4 to 9.3 dB in signal to noise ratio for rooms with realistic reverberation and more than 18 dB under anechoic conditions. Speech understanding in noise is measured in 6 adult cochlear implant users in a reverberant room, showing average improvements of 7.9–9.6 dB, when compared to a single omnidirectional microphone or 1.3–5.6 dB, when compared to a simple directional two-microphone device. Subjective evaluation in a cafeteria at lunchtime shows a preference of the cochlear implant users for the evaluated device in terms of speech understanding and sound quality.

The Immune System of Spiders

Spiders, as all other arthropods, have an open circulatory system, and their body fluid, the hemolymph, freely moves between lymphatic vessels and the body cavities (see Wirkner and Huckstorf 2013). The hemolymph can be considered as a multifunctional organ, central for locomotion (Kropf 2013), respiration (Burmester 2013) and nutrition, and it amounts to approximately 20 % of a spider’s body weight. Any injury includes not only immediate hemolymph loss but also pathogen attacks and subsequent infections. Therefore spiders have to react to injuries in a combined manner to stop fluid loss and to defend against microbial invaders. This is achieved by an innate immune system which involves several host defence systems such as hemolymph coagulation and the production of a variety of defensive substances (Fukuzawa et al.2008). In spiders, the immune system is localised in hemocytes which are derived from the myocardium cells of the heart wall where they are produced as prohemocytes and from where they are released as different cell types into the hemolymph (Seitz 1972). They contribute to the defence against pathogens by phagocytosis, nodulation and encapsulation of invaders. The humoral response includes mechanisms which induce melanin production to destroy pathogens, a clotting cascade to stop hemolymph loss and the constitutive production of several types of antimicrobial peptides, which are stored in hemocyte granules and released into the hemolymph (Fukuzawa et al.2008) (Fig.7.1). The immune system of spiders is an innate immune system. It is hemolymph-based and characterised by a broad but not very particular specificity. Its advantage is a fast response within minutes to a few hours. This is in contrast to the adaptive immune system of vertebrates which can react to very specific pathogens, thus resulting in much more specific responses. Moreover, it creates an immunological memory during the lifetime of the species. The disadvantage is that it needs more time to react with antibody production, usually many hours to a few days, and needs to be built up during early ontogenesis.

Adaptive Algorithms for Personalized Diabetes Treatment

Dynamic systems, especially in real-life applications, are often determined by inter-/intra-variability, uncertainties and time-varying components. Physiological systems are probably the most representative example in which population variability, vital signal measurement noise and uncertain dynamics render their explicit representation and optimization a rather difficult task. Systems characterized by such challenges often require the use of adaptive algorithmic solutions able to perform an iterative structural and/or parametrical update process towards optimized behavior. Adaptive optimization presents the advantages of (i) individualization through learning of basic system characteristics, (ii) ability to follow time-varying dynamics and (iii) low computational cost. In this chapter, the use of online adaptive algorithms is investigated in two basic research areas related to diabetes management: (i) real-time glucose regulation and (ii) real-time prediction of hypo-/hyperglycemia. The applicability of these methods is illustrated through the design and development of an adaptive glucose control algorithm based on reinforcement learning and optimal control and an adaptive, personalized early-warning system for the recognition and alarm generation against hypo- and hyperglycemic events.

An adaptive Newton-method based on a dynamical systems approach

The traditional Newton method for solving nonlinear operator equations in Banach spaces is discussed within the context of the continuous Newton method. This setting makes it possible to interpret the Newton method as a discrete dynamical system and thereby to cast it in the framework of an adaptive step size control procedure. In so doing, our goal is to reduce the chaotic behavior of the original method without losing its quadratic convergence property close to the roots. The performance of the modified scheme is illustrated with various examples from algebraic and differential equations.

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.

Contemporary ecotypic divergence during a recent range expansion was facilitated by adaptive introgression

Although rapid phenotypic evolution during range expansion associated with colonization of contrasting habitats has been documented in several taxa, the evolutionary mechanisms that underlie such phenotypic divergence have less often been investigated. A strong candidate for rapid ecotype formation within an invaded range is the three-spine stickleback in the Lake Geneva region of central Europe. Since its introduction only about 140 years ago, it has undergone a significant expansion of its range and its niche, now forming phenotypically differentiated parapatric ecotypes that occupy either the pelagic zone of the large lake or small inlet streams, respectively. By comparing museum collections from different times with contemporary population samples, we here reconstruct the evolution of parapatric phenotypic divergence through time. Using genetic data from modern samples, we infer the underlying invasion history. We find that parapatric habitat-dependent phenotypic divergence between the lake and stream was already present in the first half of the twentieth century, but the magnitude of differentiation increased through time, particularly in antipredator defence traits. This suggests that divergent selection between the habitats occurred and was stable through much of the time since colonization. Recently, increased phenotypic differentiation in antipredator defence traits likely results from habitat-dependent selection on alleles that arrived through introgression from a distantly related lineage from outside the Lake Geneva region. This illustrates how hybridization can quickly promote phenotypic divergence in a system where adaptation from standing genetic variation was constrained.

Ecosystem regimes and responses in a coupled ancient lake system from MIS 5b to present: the diatom record of lakes Ohrid and Prespa

We reconstruct the aquatic ecosystem interactions since the last interglacial period in the oldest, most diverse, hydrologically connected European lake system, by using palaeolimnological diatom and selected geochemistry data from Lake Ohrid “DEEP site” core and equivalent data from Lake Prespa core, Co1215. Driven by climate forcing, the lakes experienced two adaptive cycles during the last 92 ka: "interglacial and interstadial" and "glacial" cycle. The short-term ecosystems reorganizations, e.g. regime shifts within these cycles substantially differ between the lakes, as evident from the inferred amplitudes of variation. The deeper Lake Ohrid shifted between ultra oligo- and oligotrophic regimes in contrast to the much shallower Lake Prespa, which shifted from a deeper, (oligo-) mesotrophic to a shallower, eutrophic lake and vice versa. Due to the high level of ecosystem stability (e.g. trophic state, lake level), Lake Ohrid appears relatively resistant to external forcing, such as climate and environmental change. Recovering in a relatively short time from major climate change, Lake Prespa is a resilient ecosystem. At the DEEP site, the decoupling between the lakes' response to climate change is marked in the prolonged and gradual changes during the MIS 5/4 and 2/1 transitions. These response differences and the lakes' different physical and chemical properties may limit the influence of Lake Prespa on Lake Ohrid. Regime shifts of Lake Ohrid due to potential hydrological change in Lake Prespa are not evident in the data presented here. Moreover, a complete collapse of the ecosystems functionality and loss of their diatom communities did not happen in either lake for the period presented in the study.