How representative are air transport functional complex networks? A quantitative validation.

Functional networks have emerged as powerful instruments to characterize the propagation of information in complex systems, with applications ranging from neuroscience to climate and air transport. In spite of their success, reliable methods for validating the resulting structures are still missing, forcing the community to resort to expert knowledge or simplified models of the system's dynamics. We here propose the use of a real-world problem, involving the reconstruction of the structure of flights in the US air transport system from the activity of individual airports, as a way to explore the limits of such an approach. While the true connectivity is known and is, therefore, possible to provide a quantitative benchmark, this problem presents challenges commonly found in other fields, including the presence of non-stationarities and observational noise, and the limitedness of available time series. We explore the impact of elements like the specific functional metric employed, the way of detrending the time series, or the size of the reconstructed system and discuss how the conclusions here drawn could have implications for similar analyses in neuroscience.

Normal tissue content impact on the GBM molecular classification.

Molecular classification of glioblastoma has enabled a deeper understanding of the disease. The four-subtype model (including Proneural, Classical, Mesenchymal and Neural) has been replaced by a model that discards the Neural subtype, found to be associated with samples with a high content of normal tissue. These samples can be misclassified preventing biological and clinical insights into the different tumor subtypes from coming to light. In this work, we present a model that tackles both the molecular classification of samples and discrimination of those with a high content of normal cells. We performed a transcriptomic in silico analysis on glioblastoma (GBM) samples (n = 810) and tested different criteria to optimize the number of genes needed for molecular classification. We used gene expression of normal brain samples (n = 555) to design an additional gene signature to detect samples with a high normal tissue content. Microdissection samples of different structures within GBM (n = 122) have been used to validate the final model. Finally, the model was tested in a cohort of 43 patients and confirmed by histology. Based on the expression of 20 genes, our model is able to discriminate samples with a high content of normal tissue and to classify the remaining ones. We have shown that taking into consideration normal cells can prevent errors in the classification and the subsequent misinterpretation of the results. Moreover, considering only samples with a low content of normal cells, we found an association between the complexity of the samples and survival for the three molecular subtypes.

Deep learning in systems medicine.

Systems medicine (SM) has emerged as a powerful tool for studying the human body at the systems level with the aim of improving our understanding, prevention and treatment of complex diseases. Being able to automatically extract relevant features needed for a given task from high-dimensional, heterogeneous data, deep learning (DL) holds great promise in this endeavour. This review paper addresses the main developments of DL algorithms and a set of general topics where DL is decisive, namely, within the SM landscape. It discusses how DL can be applied to SM with an emphasis on the applications to predictive, preventive and precision medicine. Several key challenges have been highlighted including delivering clinical impact and improving interpretability. We used some prototypical examples to highlight the relevance and significance of the adoption of DL in SM, one of them is involving the creation of a model for personalized Parkinson's disease. The review offers valuable insights and informs the research in DL and SM.

Continuous ordinal patterns: Creating a bridge between ordinal analysis and deep learning.

We introduce a generalization of the celebrated ordinal pattern approach for the analysis of time series, in which these are evaluated in terms of their distance to ordinal patterns defined in a continuous way. This allows us to naturally incorporate information about the local amplitude of the data and to optimize the ordinal pattern(s) to the problem under study. This last element represents a novel bridge between standard ordinal analysis and deep learning, allowing the achievement of results comparable to the latter in real-world classification problems while also retaining the conceptual simplicity, computational efficiency, and easy interpretability of the former. We test this through the use of synthetic time series, generated by standard chaotic maps and dynamical models, data sets representing brain activity in health and schizophrenia, and the dynamics of delays in the European air transport system. We further show how the continuous ordinal patterns can be used to assess other aspects of the dynamics, like time irreversibility.

On the complementarity of ordinal patterns-based entropy and time asymmetry metrics.

Entropy and time asymmetry are two intertwined aspects of a system's dynamics, with the production of entropy marking a clear direction in the temporal dimension. In the last few years, metrics to quantify both properties in time series have been designed around the same concept, i.e., the use of ordinal patterns. In spite of this, the relationship between these two families of metrics is yet not well understood. In this contribution, we study this problem by constructing an entropy-time asymmetry plane and evaluating it on a large set of synthetic and real-world time series. We show how the two metrics can at times behave independently, the main reason being the presence of patterns with turning points; due to this, they yield complementary information about the underlying systems, and they have different discriminating performance.

Principles and open questions in functional brain network reconstruction.

Graph theory is now becoming a standard tool in system-level neuroscience. However, endowing observed brain anatomy and dynamics with a complex network representation involves often covert theoretical assumptions and methodological choices which affect the way networks are reconstructed from experimental data, and ultimately the resulting network properties and their interpretation. Here, we review some fundamental conceptual underpinnings and technical issues associated with brain network reconstruction, and discuss how their mutual influence concurs in clarifying the organization of brain function.

Community effort endorsing multiscale modelling, multiscale data science and multiscale computing for systems medicine.

Systems medicine holds many promises, but has so far provided only a limited number of proofs of principle. To address this road block, possible barriers and challenges of translating systems medicine into clinical practice need to be identified and addressed. The members of the European Cooperation in Science and Technology (COST) Action CA15120 Open Multiscale Systems Medicine (OpenMultiMed) wish to engage the scientific community of systems medicine and multiscale modelling, data science and computing, to provide their feedback in a structured manner. This will result in follow-up white papers and open access resources to accelerate the clinical translation of systems medicine.

Uncertainty in Functional Network Representations of Brain Activity of Alcoholic Patients.

In spite of the large attention received by brain activity analyses through functional networks, the effects of uncertainty on such representations have mostly been neglected. We here elaborate the hypothesis that such uncertainty is not just a nuisance, but that on the contrary is condition-dependent. We test this hypothesis by analysing a large set of EEG brain recordings corresponding to control subjects and patients suffering from alcoholism, through the reconstruction of the corresponding Maximum Spanning Trees (MSTs), the assessment of their topological differences, and the comparison of two frequentist and Bayesian reconstruction approaches. A machine learning model demonstrates that the Bayesian reconstruction encodes more information than the frequentist one, and that such additional information is related to the uncertainty of the topological structures. We finally show how the Bayesian approach is more effective in the validation of generative models, over and above the frequentist one, by proposing and disproving two models based on additive noise.

Assessing time series irreversibility through micro-scale trends.

Time irreversibility, defined as the lack of invariance of the statistical properties of a system or time series under the operation of time reversal, has received increasing attention during the last few decades, thanks to the information it provides about the mechanisms underlying the observed dynamics. Following the need of analyzing real-world time series, many irreversibility metrics and tests have been proposed, each one associated with different requirements in terms of, e.g., minimum time series length or computational cost. We here build upon previously proposed tests based on the concept of permutation patterns but deviating from them through the inclusion of information about the amplitude of the signal and how this evolves over time. We show, by means of synthetic time series, that the results yielded by this method are complementary to the ones obtained by using permutation patterns alone, thus suggesting that "one irreversibility metric does not fit all." We further apply the proposed metric to the analysis of two real-world data sets.

Algorithmic Approaches for Assessing Irreversibility in Time Series: Review and Comparison.

The assessment of time irreversibility, i.e., of the lack of invariance of the statistical properties of a system under the operation of time reversal, is a topic steadily gaining attention within the research community. Irreversible dynamics have been found in many real-world systems, with alterations being connected to, for instance, pathologies in the human brain, heart and gait, or to inefficiencies in financial markets. Assessing irreversibility in time series is not an easy task, due to its many aetiologies and to the different ways it manifests in data. It is thus not surprising that several numerical methods have been proposed in the last decades, based on different principles and with different applications in mind. In this contribution we review the most important algorithmic solutions that have been proposed to test the irreversibility of time series, their underlying hypotheses, computational and practical limitations, and their comparative performance. We further provide an open-source software library that includes all tests here considered. As a final point, we show that "one size does not fit all", as tests yield complementary, and sometimes conflicting views to the problem; and discuss some future research avenues.

Travel restrictions during pandemics: A useful strategy?

Though carrying considerable economic and societal costs, restricting individuals' traveling freedom appears as a logical way to curb the spreading of an epidemic. However, whether, under what conditions, and to what extent travel restrictions actually exert a mitigating effect on epidemic spreading are poorly understood issues. Recent studies have actually suggested the opposite, i.e., that allowing some movements can hinder the propagation of a disease. Here, we explore this topic by modeling the spreading of a generic contagious disease where susceptible, infected, or recovered point-wise individuals are uncorrelated random-walkers evolving within a space comprising two equally sized separated compartments. We evaluate the spreading process under different separation conditions between the two spatial compartments and a forced relocation schedule. Our results confirm that, under certain conditions, allowing individuals to move from regions of high to low infection rates may turn out to have a positive effect on aggregate; such positive effect is nevertheless reduced if a directional flow is allowed. This highlights the importance of considering travel restriction policies alternative to classical ones.

Assessing functional propagation patterns in COVID-19.

Among the many efforts done by the scientific community to help coping with the COVID-19 pandemic, one of the most important has been the creation of models to describe its propagation, as these are expected to guide the deployment of containment and health policies. These models are commonly based on exogenous information, as e.g. mobility data, whose limitedness always compromise the reliability of obtained results. In this contribution we propose a different approach, based on extracting relationships between the evolution of the disease in different regions through information theoretical metrics. In a way similar to what is commonly done in neuroscience, propagation is understood as information transfer, and the resulting propagation patterns are represented and studied as functional networks. By applying this methodology to the dynamics of COVID-19 in several countries and regions thereof, we were able to reconstruct static and time-varying propagation graphs. We further discuss the advantages, promises and open research questions associated with this functional approach.

A minimal model of hospital patients' dynamics in COVID-19.

Italy has been one of the countries hardest hit by the coronavirus disease (COVID-19) pandemic. While the overall policy in response to the epidemic was to a large degree centralised, the regional basis of the healthcare system represented an important factor affecting the natural dynamics of the disease induced geographic specificities. Here, we characterise the region-specific modulation of COVID dynamics with a reduced exponential model leveraging available data on sub-intensive and intensive care unit patients made available by all regional councils from the very onset of the disease. This simple model provides a rather good fit of regional patient dynamics, particularly for regions where the affected population was large, highlighting important region-specific patterns of epidemic dynamics.

Disease networks and their contribution to disease understanding: A review of their evolution, techniques and data sources.

Over a decade ago, a new discipline called network medicine emerged as an approach to understand human diseases from a network theory point-of-view. Disease networks proved to be an intuitive and powerful way to reveal hidden connections among apparently unconnected biomedical entities such as diseases, physiological processes, signaling pathways, and genes. One of the fields that has benefited most from this improvement is the identification of new opportunities for the use of old drugs, known as drug repurposing. The importance of drug repurposing lies in the high costs and the prolonged time from target selection to regulatory approval of traditional drug development. In this document we analyze the evolution of disease network concept during the last decade and apply a data science pipeline approach to evaluate their functional units. As a result of this analysis, we obtain a list of the most commonly used functional units and the challenges that remain to be solved. This information can be very valuable for the generation of new prediction models based on disease networks.

Indoor Temperature Prediction in an IoT Scenario.

One of the hottest topics being researched in the field of IoT relates to making connected devices smarter, by locally computing relevant information and integrating data coming from other sensors through a local network. Such works are still in their early stages either by lack of access to data or, on the other hand, by the lack of simple test cases with a clear added value. This contribution aims at shading some light on how knowledge can be obtained, using a simple use case. It focuses on the feasibility of having a home refrigerator performing temperature forecasts, using information provided by both internal and external sensors. The problem is reviewed for both its potential applications and to compare the use of different algorithms, from simple linear correlations to ARIMA models. We analyse the precision and computational cost using real data from a refrigerator. Results indicate that small average errors, down to ≈0.09 ∘ C, can be obtained. Lastly, it is devised how can the scenario be improved, and, most importantly, how this work can be extended in the future.

Profiling Lung Cancer Patients Using Electronic Health Records.

If Electronic Health Records contain a large amount of information about the patient's condition and response to treatment, which can potentially revolutionize the clinical practice, such information is seldom considered due to the complexity of its extraction and analysis. We here report on a first integration of an NLP framework for the analysis of clinical records of lung cancer patients making use of a telephone assistance service of a major Spanish hospital. We specifically show how some relevant data, about patient demographics and health condition, can be extracted; and how some relevant analyses can be performed, aimed at improving the usefulness of the service. We thus demonstrate that the use of EHR texts, and their integration inside a data analysis framework, is technically feasible and worth of further study.

Assessing Time Series Reversibility through Permutation Patterns.

Time irreversibility, i.e., the lack of invariance of the statistical properties of a system under time reversal, is a fundamental property of all systems operating out of equilibrium. Time reversal symmetry is associated with important statistical and physical properties and is related to the predictability of the system generating the time series. Over the past fifteen years, various methods to quantify time irreversibility in time series have been proposed, but these can be computationally expensive. Here, we propose a new method, based on permutation entropy, which is essentially parameter-free, temporally local, yields straightforward statistical tests, and has fast convergence properties. We apply this method to the study of financial time series, showing that stocks and indices present a rich irreversibility dynamics. We illustrate the comparative methodological advantages of our method with respect to a recently proposed method based on visibility graphs, and discuss the implications of our results for financial data analysis and interpretation.

Characterizing Normal and Pathological Gait through Permutation Entropy.

Cerebral palsy is a physical impairment stemming from a brain lesion at perinatal time, most of the time resulting in gait abnormalities: the first cause of severe disability in childhood. Gait study, and instrumental gait analysis in particular, has been receiving increasing attention in the last few years, for being the complex result of the interactions between different brain motor areas and thus a proxy in the understanding of the underlying neural dynamics. Yet, and in spite of its importance, little is still known about how the brain adapts to cerebral palsy and to its impaired gait and, consequently, about the best strategies for mitigating the disability. In this contribution, we present the hitherto first analysis of joint kinematics data using permutation entropy, comparing cerebral palsy children with a set of matched control subjects. We find a significant increase in the permutation entropy for the former group, thus indicating a more complex and erratic neural control of joints and a non-trivial relationship between the permutation entropy and the gait speed. We further show how this information theory measure can be used to train a data mining model able to forecast the child's condition. We finally discuss the relevance of these results in clinical applications and specifically in the design of personalized medicine interventions.

Detecting switching and intermittent causalities in time series.

During the last decade, complex network representations have emerged as a powerful instrument for describing the cross-talk between different brain regions both at rest and as subjects are carrying out cognitive tasks, in healthy brains and neurological pathologies. The transient nature of such cross-talk has nevertheless by and large been neglected, mainly due to the inherent limitations of some metrics, e.g., causality ones, which require a long time series in order to yield statistically significant results. Here, we present a methodology to account for intermittent causal coupling in neural activity, based on the identification of non-overlapping windows within the original time series in which the causality is strongest. The result is a less coarse-grained assessment of the time-varying properties of brain interactions, which can be used to create a high temporal resolution time-varying network. We apply the proposed methodology to the analysis of the brain activity of control subjects and alcoholic patients performing an image recognition task. Our results show that short-lived, intermittent, local-scale causality is better at discriminating both groups than global network metrics. These results highlight the importance of the transient nature of brain activity, at least under some pathological conditions.