Isolating Attosecond Electron Dynamics in Molecules where Nuclei Move Fast.

Capturing electronic dynamics in real time has been the ultimate goal of attosecond science since its beginning. While for atomic targets the existing measurement techniques have been thoroughly validated, in molecules there are open questions due to the inevitable copresence of moving nuclei, which are not always mere spectators of the phototriggered electron dynamics. Previous work has shown that not only can nuclear motion affect the way electrons move in a molecule, but it can also lead to contradictory interpretations depending on the chosen experimental approach. In this Letter we investigate how nuclear motion affects and eventually distorts the electronic dynamics measured by using two of the most popular attosecond techniques, reconstruction of attosecond beating by interference of two-photon transitions and attosecond streaking. Both methods are employed, in combination with ab initio theoretical calculations, to retrieve photoionization delays in the dissociative ionization of H_{2}, H_{2}→H^{+}+H+e^{-}, in the region of the Q_{1} series of autoionizing states, where nuclear motion plays a prominent role. We find that the experimental reconstruction of attosecond beating by interference of two-photon transitions results are very sensitive to bond softening around the Q_{1} threshold (27.8 eV), even at relatively low infrared (IR) intensity (I_{0}∼1.4×10^{11} W/cm^{2}), due to the long duration of the probe pulse that is inherent to this technique. Streaking, on the other hand, seems to be a better choice to isolate attosecond electron dynamics, since shorter pulses can be used, thus reducing the role of bond softening. This conclusion is supported by very good agreement between our streaking measurements and the results of accurate theoretical calculations. Additionally, the streaking technique offers the necessary energy resolution to accurately retrieve the fast-oscillating phase of the photoionization matrix elements, an essential requirement for extending this technique to even more complicated molecular targets.

Iron deficiency anemia in newly diagnosed celiac disease in children.

BACKGROUND: Celiac disease (CD) in children may occur with a wide spectrum of clinical manifestations: anemia is the most frequent extraintestinal manifestation, iron deficiency anemia (IDA) is the common presentation. In our study we aimed to assess IDA condition in a large cohort of pediatric patients with newly diagnosed CD. METHODS: Our study includes a cohort of 518 children (340 females and 178 males), 6 months-18 years old, joined between January 1990 and January 2013. We have analyzed hematological parameters and iron balance: serum iron, serum ferritin and serum transferrin levels. The diagnosis of IDA was considered on the basis of hemoglobin levels below -2SD, associated with serum iron and ferritin reduction, serum transferrin increase; all compared with the normal reference values for age. RESULTS: Of all patients, 156 patients (30.1%) had anemia, including 103 females (19.8%) and 53 males (10.2%); of these, 112 (21.62%) had IDA (in 18 cases associated with α- or ß-thalassemia trait), 22 were thalassemic trait without iron deficiency and the remaining 19 suffered from other forms of anemia. One hundred fifteen patients (22.20%) with low ferritin levels but normal hemoglobin levels were considered as preanemic iron deficient patients. CONCLUSION: Our data confirm that iron depletion and IDA represent a frequent finding at the diagnosis of CD. This significant relation existing between CD and iron deficiency should be considered by pediatricians at the diagnosis of CD in order to treat the patients.

Echo state networks for the recognition of type 1 Brugada syndrome from conventional 12-LEAD ECG.

Artificial Intelligence (AI) applications and Machine Learning (ML) methods have gained much attention in recent years for their ability to automatically detect patterns in data without being explicitly taught rules. Specific features characterise the ECGs of patients with Brugada Syndrome (BrS); however, there is still ambiguity regarding the correct diagnosis of BrS and its differentiation from other pathologies. This work presents an application of Echo State Networks (ESN) in the Recurrent Neural Networks (RNN) class for diagnosing BrS from the ECG time series. 12-lead ECGs were obtained from patients with a definite clinical diagnosis of spontaneous BrS Type 1 pattern (Group A), patients who underwent provocative pharmacological testing to induce BrS type 1 pattern, which resulted in positive (Group B) or negative (Group C), and control subjects (Group D). One extracted beat in the V2 lead was used as input, and the dataset was used to train and evaluate the ESN model using a double cross-validation approach. ESN performance was compared with that of 4 cardiologists trained in electrophysiology. The model performance was assessed in the dataset, with a correct global diagnosis observed in 91.5 % of cases compared to clinicians (88.0 %). High specificity (94.5 %), sensitivity (87.0 %) and AUC (94.7 %) for BrS recognition by ESN were observed in Groups A + B vs. C + D. Our results show that this ML model can discriminate Type 1 BrS ECGs with high accuracy comparable to expert clinicians. Future availability of larger datasets may improve the model performance and increase the potential of the ESN as a clinical support system tool for daily clinical practice.

Hierarchical-Task Reservoir for Online Semantic Analysis From Continuous Speech.

In this article, we propose a novel architecture called hierarchical-task reservoir (HTR) suitable for real-time applications for which different levels of abstraction are available. We apply it to semantic role labeling (SRL) based on continuous speech recognition. Taking inspiration from the brain, this demonstrates the hierarchies of representations from perceptive to integrative areas, and we consider a hierarchy of four subtasks with increasing levels of abstraction (phone, word, part-of-speech (POS), and semantic role tags). These tasks are progressively learned by the layers of the HTR architecture. Interestingly, quantitative and qualitative results show that the hierarchical-task approach provides an advantage to improve the prediction. In particular, the qualitative results show that a shallow or a hierarchical reservoir, considered as baselines, does not produce estimations as good as the HTR model would. Moreover, we show that it is possible to further improve the accuracy of the model by designing skip connections and by considering word embedding (WE) in the internal representations. Overall, the HTR outperformed the other state-of-the-art reservoir-based approaches and it resulted in extremely efficient with respect to typical recurrent neural networks (RNNs) in deep learning (DL) [e.g., long short term memory (LSTMs)]. The HTR architecture is proposed as a step toward the modeling of online and hierarchical processes at work in the brain during language comprehension.

A Robust Method for Assaying the Immunoreactive Fraction in Nonequilibrium Systems.

The immunoreactive fraction r provides important information on the functional purity of radiolabeled proteins. It is traditionally determined by saturating the radioimmunoconjugate with an increasing excess of antigen, followed by linear extrapolation to infinite antigen excess in a double inverse "Lindmo plot". Although several reports have described shortcomings in the Lindmo plot, a systematic examination is lacking. Using an experimental and simulation-based approach, we compared-for accuracy, precision and robustness-the Lindmo plot with the "rectangular hyperbola" extrapolation method based on the Langmuir model. The differences between the theoretical and extrapolated r values demonstrate that nonequilibrium and antigen depletion are important sources of error. The mathematical distortions resulting from the linearization of the data in the Lindmo plot induce fragility towards stochastic errors and make it necessary to exclude low bound fractions. The rectangular hyperbola provides robust and precise r estimates from raw binding data, even for slow kinetics.

Design of deep echo state networks.

In this paper, we provide a novel approach to the architectural design of deep Recurrent Neural Networks using signal frequency analysis. In particular, focusing on the Reservoir Computing framework and inspired by the principles related to the inherent effect of layering, we address a fundamental open issue in deep learning, namely the question of how to establish the number of layers in recurrent architectures in the form of deep echo state networks (DeepESNs). The proposed method is first analyzed and refined on a controlled scenario and then it is experimentally assessed on challenging real-world tasks. The achieved results also show the ability of properly designed DeepESNs to outperform RC approaches on a speech recognition task, and to compete with the state-of-the-art in time-series prediction on polyphonic music tasks.