Learning real-world stimuli in a neural network with spike-driven synaptic dynamics

We present a model of spike-driven synaptic plasticity inspired by experimental observations and motivated by the desire to build an electronic hardware device that can learn to classify complex stimuli in a semisupervised fashion. During training, patterns of activity are sequentially imposed on the input neurons, and an additional instructor signal drives the output neurons toward the desired activity. The network is made of integrate-and-fire neurons with constant leak and a floor. The synapses are bistable, and they are modified by the arrival of presynaptic spikes. The sign of the change is determined by both the depolarization and the state of a variable that integrates the postsynaptic action potentials. Following the training phase, the instructor signal is removed, and the output neurons are driven purely by the activity of the input neurons weighted by the plastic synapses. In the absence of stimulation, the synapses preserve their internal state indefinitely. Memories are also very robust to the disruptive action of spontaneous activity. A network of 2000 input neurons is shown to be able to classify correctly a large number (thousands) of highly overlapping patterns (300 classes of preprocessed Latex characters, 30 patterns per class, and a subset of the NIST characters data set) and to generalize with performances that are better than or comparable to those of artificial neural networks. Finally we show that the synaptic dynamics is compatible with many of the experimental observations on the induction of long-term modifications (spike-timing-dependent plasticity and its dependence on both the postsynaptic depolarization and the frequency of pre- and postsynaptic neurons).

Unprotected left main stenting in the real world: two-year outcomes of the French left main taxus registry

BACKGROUND: Cardiac surgery is the reference treatment for patients with left main (LM) disease, although percutaneous coronary intervention with drug-eluting stents is emerging as a possible alternative. The objective of this registry was to evaluate the 2-year outcome of elective percutaneous coronary intervention for unprotected LM disease with paclitaxel-eluting stents. METHODS AND RESULTS: A total of 291 patients were prospectively included from 4 centers. Acute myocardial infarction and cardiogenic shock were the only exclusion criteria. Patients were 69+/-11 years old, 29% were diabetic, and 25% had 3-vessel disease. For distal LM lesions (78%), the provisional side-branch T-stenting approach was used in 92% of cases and final kissing balloon inflation in 97%. Angiographic success was obtained in 99.7% of cases. At 2-year follow-up, the total cardiac death rate was 5.4% (1 EuroSCORE point was associated with a 15% [95% confidence interval 2.9% to 28.2%, P=0.013] higher risk of cardiac death), target-lesion revascularization was 8.7%, and incidence of Q-wave or non-Q-wave myocardial infarction was 0.9% and 3.1%, respectively. The combined end point occurred in 15.8% of cases and stroke in 0.7%. The incidence of definite and probable LM stent thrombosis was 0.7%, whereas the incidence of any stent thrombosis was 3.8%, with a higher risk in patients with side-branch stenting in the presence of LM bifurcation lesions (hazard ratio 9.6, 95% confidence interval 1.2 to 77.7, P=0.035). CONCLUSIONS: Unprotected LM stenting with paclitaxel-eluting stents, with a strategy of provisional side-branch T-stenting for distal lesions, provides excellent acute angiographic results and good mid-term clinical outcomes, with a 15.8% rate of major adverse cardiac events at 2-year follow-up.

Calibrating Wireless Sensor Network Simulation Models with Real-World Experiments

This paper studies the energy-efficiency and service characteristics of a recently developed energy-efficient MAC protocol for wireless sensor networks in simulation and on a real sensor hardware testbed. This opportunity is seized to illustrate how simulation models can be verified by cross-comparing simulation results with real-world experiment results. The paper demonstrates that by careful calibration of simulation model parameters, the inevitable gap between simulation models and real-world conditions can be reduced. It concludes with guidelines for a methodology for model calibration and validation of sensor network simulation models.

Real-World Energy Measurements of a Wireless Mesh Network

Over the past several years the topics of energy consumption and energy harvesting have gained significant importance as a means for improved operation of wireless sensor and mesh networks. Energy-awareness of operation is especially relevant for application scenarios from the domain of environmental monitoring in hard to access areas. In this work we reflect upon our experiences with a real-world deployment of a wireless mesh network. In particular, a comprehensive study on energy measurements collected over several weeks during the summer and the winter period in a network deployment in the Swiss Alps is presented. Energy performance is monitored and analysed for three system components, namely, mesh node, battery and solar panel module. Our findings cover a number of aspects of energy consumption, including the amount of load consumed by a mesh node, the amount of load harvested by a solar panel module, and the dependencies between these two. With our work we aim to shed some light on energy-aware network operation and to help both users and developers in the planning and deployment of a new wireless (mesh) network for environmental research.

Real-World Evaluation of Sensor Context-aware Adaptive Duty-cycled Opportunistic Routing

Energy is of primary concern in wireless sensor networks (WSNs). Low power transmission makes the wireless links unreliable, which leads to frequent topology changes. Resulting packet retransmissions aggravate the energy consumption. Beaconless routing approaches, such as opportunistic routing (OR) choose packet forwarders after data transmissions, and are promising to support dynamic features of WSNs. This paper proposes SCAD - Sensor Context-aware Adaptive Duty-cycled beaconless OR for WSNs. SCAD is a cross-layer routing solution and it brings the concept of beaconless OR into WSNs. SCAD selects packet forwarders based on multiple types of network contexts. To achieve a balance between performance and energy efficiency, SCAD adapts duty-cycles of sensors based on real-time traffic loads and energy drain rates. We implemented SCAD in TinyOS running on top of Tmote Sky sensor motes. Real-world evaluations show that SCAD outperforms other protocols in terms of both throughput and network lifetime.

Evidence from the real world: N-15 natural abundances reveal enhanced nitrogen use at high plant diversity in Central European grasslands

Complementarity that leads to more efficient resource use is presumed to be a key mechanism explaining positive biodiversity–productivity relationships but has been described solely for experimental set-ups with controlled environmental settings or for very short gradients of abiotic conditions, land-use intensity and biodiversity. Therefore, we analysed plant diversity effects on nitrogen dynamics across a broad range of Central European grasslands. The 15N natural abundance in soil and plant biomass reflects the net effect of processes affecting ecosystem N dynamics. This includes the mechanism of complementary resource utilization that causes a decrease in the 15N isotopic signal. We measured plant species richness, natural abundance of 15N in soil and plants, above-ground biomass of the community and three single species (an herb, grass and legume) and a variety of additional environmental variables in 150 grassland plots in three regions of Germany. To explore the drivers of the nitrogen dynamics, we performed several analyses of covariance treating the 15N isotopic signals as a function of plant diversity and a large set of covariates. Increasing plant diversity was consistently linked to decreased δ15N isotopic signals in soil, above-ground community biomass and the three single species. Even after accounting for multiple covariates, plant diversity remained the strongest predictor of δ15N isotopic signals suggesting that higher plant diversity leads to a more closed nitrogen cycle due to more efficient nitrogen use. Factors linked to increased δ15N values included the amount of nitrogen taken up, soil moisture and land-use intensity (particularly fertilization), all indicators of the openness of the nitrogen cycle due to enhanced N-turnover and subsequent losses. Study region was significantly related to the δ15N isotopic signals indicating that regional peculiarities such as former intensive land use could strongly affect nitrogen dynamics. Synthesis. Our results provide strong evidence that the mechanism of complementary resource utilization operates in real-world grasslands where multiple external factors affect nitrogen dynamics. Although single species may differ in effect size, actively increasing total plant diversity in grasslands could be an option to more effectively use nitrogen resources and to reduce the negative environmental impacts of nitrogen losses.

The effects of qualitatively different acute physical activity interventions in real-world settings on executive functions in preadolescent children

There is growing evidence indicating a positive effect of acute physical activity on cognitive performance in children. Most of the evidence originates, however, from studies in highly controlled laboratory settings. The aim of the present study was to investigate whether the same effects can be found in more real-world settings. We examined the effects of qualitatively different acute physical activity interventions on the three core dimensions of executive functions (updating, inhibition, shifting). In an experimental between-subject design, 219 ten to twelve year-olds were assigned to one of four conditions which varied systematically in physical activation and cognitive engagement. Executive functions were measured before and immediately after the intervention. Contrary to the hypothesis, no effects of acute physical activity with and without cognitive engagement were found on executive functions in the overall sample. Only children with higher fitness and/or higher academic achievement benefitted from the interventions in terms of their updating performance. Thus, the results indicate that it may be more difficult to attain positive effects through acute physical activity in real-world settings than in laboratory settings and that physiological and cognitive requirements may have to be adjusted to individual capacity to make an intervention effective.

Plant diversity moderates drought stress in grasslands: Implications from a large real-world study on 13C natural abundances

Land-use change and intensification play a key role in the current biodiversity crisis. The resulting species loss can have severe effects on ecosystem functions and services, thereby increasing ecosystem vulnerability to climate change. We explored whether land-use intensification (i.e. fertilization intensity), plant diversity and other potentially confounding environmental factors may be significantly related to water use (i.e. drought stress) of grassland plants. Drought stress was assessed using δ13C abundances in aboveground plant biomass of 150 grassland plots across a gradient of land-use intensity. Under water shortage, plants are forced to increasingly take up the heavier 13C due to closing stomata leading to an enrichment of 13C in biomass. Plants were sampled at the community level and for single species, which belong to three different functional groups (one grass, one herb, two legumes). Results show that plant diversity was significantly related to the δ13C signal in community, grass and legume biomass indicating that drought stress was lower under higher diversity, although this relation was not significant for the herb species under study. Fertilization, in turn, mostly increased drought stress as indicated by more positive δ13C values. This effect was mostly indirect by decreasing plant diversity. In line with these results, we found similar patterns in the δ13C signal of the organic matter in the topsoil, indicating a long history of these processes. Our study provided strong indication for a positive biodiversity-ecosystem functioning relationship with reduced drought stress at higher plant diversity. However, it also underlined a negative reinforcing situation: as land-use intensification decreases plant diversity in grasslands, this might subsequently increases drought sensitivity. Vice-versa, enhancing plant diversity in species-poor agricultural grasslands may moderate negative effects of future climate change.

Modeling the role of fixational eye movements in real-world scenes

Our eyes never remain still. Even when we stare at a fixed point, small involuntary movements take place in our eyes in an imperceptible manner. Researchers agree on the presence of three main contributions to eye movements when we fix the gaze: microsaccades, drifts and tremor. These small movements carry the image across the retina stimulating the photoreceptors and thus avoiding fading. Nowadays it is commonly accepted that these movements can improve the discrimination performance of the retina. In this paper, several retina models with and without fixational eye movements were implemented by mean of RetinaStudio tool to test the feasibility of these models to be incorporated in future neuroprostheses. For this purpose each retina model has been stimulated with natural scene images in two experiments. Results are discussed from the point of view of a neuroprosthesis development.

Patient-reported outcomes and visual acuity after 12 months of anti-VEGF-treatment for sight-threatening diabetic macular edema in a real world setting

Aims To examine objective visual acuity measured with ETDRS, retinal thickness (OCT), patient reported outcome and describe levels of glycated hemoglobin and its association with the effects on visual acuity in patients treated with anti-VEGF for visual impairment due to diabetic macular edema (DME) during 12 months in a real world setting. Methods In this cross-sectional study, 58 patients (29 females and 29 males; mean age, 68 years) with type 1 and type 2 diabetes diagnosed with DME were included. Medical data and two questionnaires were collected; an eye-specific (NEI VFQ-25) and a generic health-related quality of life questionnaire (SF-36) were used. Results The total patient group had significantly improved visual acuity and reduced retinal thickness at 4 months and remains at 12 months follow up. Thirty patients had significantly improved visual acuity, and 27 patients had no improved visual acuity at 12 months. The patients with improved visual acuity had significantly improved scores for NEI VFQ-25 subscales including general health, general vision, near activities, distance activities, and composite score, but no significant changes in scores were found in the group without improvements in visual acuity. Conclusions Our study revealed that anti-VEGF treatment improved visual acuity and central retinal thickness as well as patient-reported outcome in real world 12 months after treatment start.

The blue side of diabetes: assessing the impact of depression in people with type 2 diabetes using real-world data

The aim of the thesis is to assess the impact of depression in people with type 2 diabetes. Using Healthcare Utilization Databases, I estimated in a large population-based cohort with type 2 diabetes the incidence of depression over 10 year-period, identified the demographic and clinical predictors of depression, and determined the extent to which depression is a risk factor for acute and long-term complications and mortality. In the context of COVID-19 pandemic, I evaluated whether the presence of a history of depression in type 2 diabetes increased the Emergency Department (ED) access rate for diabetes-related complications, and I investigated changes in the incidence of depression during the first year of the pandemic. Findings from the first study indicated that developing depression was associated with being a woman, being over 65 years, living in rural areas, having insulin as initial diabetes medication and having comorbid conditions; the study also confirmed that depression was associated with an increased risk for acute and long-term diabetes complications and all-cause mortality. The second observational study showed a higher rate of ED access for diabetes-related complications during the pandemic in people with type 2 diabetes and a history of depression than in those without a history of depression, similar to what was observed in a pre-pandemic period. As shown in the third population-based study, the incidence of depression decreased in 2020 compared to 2019, mainly during the first and the second waves of the COVID-19 pandemic, when people probably had difficulty reaching healthcare services. This new real-world evidence will help healthcare professionals identify timely patients at high risk of developing depression. Lastly, policymakers and physicians will benefit from new evidence of the effects of the COVID-19 pandemic on depression in people with type 2 diabetes to ensure a high level of care during crisis periods.

Models and algorithms for real-world optimization problems

This thesis deals with efficient solution of optimization problems of practical interest. The first part of the thesis deals with bin packing problems. The bin packing problem (BPP) is one of the oldest and most fundamental combinatorial optimiza- tion problems. The bin packing problem and its generalizations arise often in real-world ap- plications, from manufacturing industry, logistics and transportation of goods, and scheduling. After an introductory chapter, I will present two applications of two of the most natural extensions of the bin packing: Chapter 2 will be dedicated to an application of bin packing in two dimension to a problem of scheduling a set of computational tasks on a computer cluster, while Chapter 3 deals with the generalization of BPP in three dimensions that arise frequently in logistic and transportation, often com- plemented with additional constraints on the placement of items and characteristics of the solution, like, for example, guarantees on the stability of the items, to avoid potential damage to the transported goods, on the distribution of the total weight of the bins, and on compatibility with loading and unloading operations. The second part of the thesis, and in particular Chapter 4 considers the Trans- mission Expansion Problem (TEP), where an electrical transmission grid must be expanded so as to satisfy future energy demand at the minimum cost, while main- taining some guarantees of robustness to potential line failures. These problems are gaining importance in a world where a shift towards renewable energy can impose a significant geographical reallocation of generation capacities, resulting in the ne- cessity of expanding current power transmission grids.

ICE Emissions Estimation through Machine-learning techniques to support real-world R&D experiments

The emissions estimation, both during homologation and standard driving, is one of the new challenges that automotive industries have to face. The new European and American regulation will allow a lower and lower quantity of Carbon Monoxide emission and will require that all the vehicles have to be able to monitor their own pollutants production. Since numerical models are too computationally expensive and approximated, new solutions based on Machine Learning are replacing standard techniques. In this project we considered a real V12 Internal Combustion Engine to propose a novel approach pushing Random Forests to generate meaningful prediction also in extreme cases (extrapolation, very high frequency peaks, noisy instrumentation etc.). The present work proposes also a data preprocessing pipeline for strongly unbalanced datasets and a reinterpretation of the regression problem as a classification problem in a logarithmic quantized domain. Results have been evaluated for two different models representing a pure interpolation scenario (more standard) and an extrapolation scenario, to test the out of bounds robustness of the model. The employed metrics take into account different aspects which can affect the homologation procedure, so the final analysis will focus on combining all the specific performances together to obtain the overall conclusions.