Respiration Rate (RR) is an indicator of numerous conditions and its importance in condition evaluation and monitoring has been well established. This paper deals with RR estimation based on the photoplethysmogram (PPG), a signal that can be easily acquired by wearables and is already captured by pulse oximeters. PPG is modulated by respiration and thus statistical methods can be used to extract the respiratory information encoded in the signal. A critical preprocessing step of RR estimation is that of reliable PPG peak detection, which is a form of signal quality evaluation when incorporated in PPG beat segmentation. To demonstrate the importance of this step we compare three peak detection algorithms against three RR estimation methods and show that peak extraction based on beat morphological characteristics leads to better RR estimation. Furthermore, we propose a new PPG-based method fusion that leads to higher estimation accuracy. All methods are evaluated on CapnoBase, a benchmark database on respiration, the best method achieving a RMSE of 3.4 bpm for 60-second intervals.

Drug-drug interactions (DDIs) pose a significant issue in modern healthcare, potentially compromising treatment efficacy and patient safety. DDIs arise when significant alterations occur in the pharmacological action of a drug due to its co-administration with another drug, leading to potential adverse drug reactions (ADRs), toxicity or diminished therapeutic efficacy. Apart from the obvious cases of drug combinations that should be avoided, there are instances where risk-benefit analysis may allow co-administration. Hence, DDIs may represent clinically significant cases depending on the clinical outcome, time point of administration, etc. The issue is especially critical in cases of patients with multimorbidity and complex therapeutic regimens with different time points in drug administrations. This work employs a graph-based approach aimed at optimizing therapeutic regiments while considering the minimization of DDIs potential. In this approach each drug is represented as a node, and edges represent the clinical significance of DDIs. We aim to identify sets of drugs that either have no DDIs or exhibit minor to moderate clinical significance (referred to as Maximal Independent Sets), indicating that they can be taken together. In practice, we solved the complementary problem, which is finding Maximal Cliques. Both problems are NP-hard, but for small graphs, they can be solved exactly. From all the cliques we identify, those selected to be a part of each proposed therapeutic regimen must consist of nodes that appear only once. This problem is once again reduced to clique finding. The above approach is demonstrated using two clinical scenarios involving two patients who are experiencing polypharmacy and are at risk for ADRs due to potential DDIs of varying clinical significance. By applying our approach, the therapeutic schemes are optimized towards minimizing the risk of ADRs.

Electroencephalography-based brain computer interfaces (BCIs) have been widely used in assistive applications for patients suffering from quadriplegia, or even the locked-in syndrome, to promote autonomy and control. Steady-state visual evoked potentials (SSVEPs) is a BCI stimulation protocol that has been employed widely in navigation applications, due to their efficiency and fast response time. In the current study, we use a previously developed SSVEP-based BCI for robotic car navigation with a low-cost EEG recording device, to compare the performance of the easier-to-use dry EEG-electrodes and the commonly-used wet electrodes. We also employ two different stimulus scenarios, a fixed one and an adaptive one optimized for each participant. We tested our system on 23 healthy participants in both an offline session and an online navigation task on a predefined route. The results indicate that our system with the combination of the low-cost EEG device and the dry electrodes achieves comparable performance to the same system using wet electrodes. Moreover, it ensures a more user-friendly and affordable alternative that could be adopted by a larger part of potential users.

During sleep., breathing-related sleep disorders (BSD) are very probable to cause distortions on human health and even be life-threatening. Among the different types of BSD., apnea accounts for one of the most common. Many detection algorithms have been proposed for spotting and classifying apneas, using one feature or being designed for binary classification. Also, many proposed clinical setups for respiratory data acquisition are invasive, making the application to patients a non-trial task. In this study, we aim to propose an easy-to-apply and patient-friendly clinical setup with a BSD detection that utilizes a multi-feature classification scheme for binary (apnea, healthy), as well as multiple classes (healthy, central, mixed, and obstructive apneas and hypopneas). Our clinical setup includes a high-resolution microphone attached to the bed at a very close distance to the patient. Our multi-feature approach contains spectral, statistical, and symbolic-based characteristics of respiratory signals of five patients admitted for a first BSD diagnosis and assesses the performance of different classification algorithms iteratively. The results show a high classification performance ($>$98% for binary and$>$84% for multi-class classification) for either classification scheme. A robust classification scheme is thus proposed, utilizing the entire content of the recorded respiratory signal. Such a classification scheme leads to a promising result towards the design of portable devices with multi-features for real-time detection of BSD.

Electrocardiogram (ECG) signal constitutes a valuable technique that provides considerable information towards the early diagnosis of several cardiovascular diseases, especially regarding the detection of abnormal heart rate, namely arrhythmias. In this paper, innovative methodologies that allow for the efficient classification of cardiac rhythm are presented. The proposed methods are based on ECG signal analysis, extraction of significant features, as well as classification algorithms. Several clinical, time- and frequency-domain features are either calculated, or automatically extracted by means of a Convolutional Neural Network, while traditional machine learning algorithms, such as k-Nearest Neighbors and Random Forests are employed in order to classify the ECG signals among 7 different cases of abnormal and normal heart rate. The learning methods are carried out within the JADBio software tool, that also performs feature selection prior to classification. The experimental results demonstrate high performance of the deployed methods in terms of relevant statistical metrics, while they yielded an average validation Area Under the Curve (AUC) of 99.9%.

Brain computer interfaces (BCIs) that are focused on navigation applications have been developed for patients suffering from severe paralysis to offer a means of autonomy. In SSVEP-based BCIs, users focus their gaze on flickering targets, which correspond to specific commands. Besides the accuracy of the target identification, several additional aspects are important for the development of a practical and useful BCI, such as low cost, ease of use and robustness to everyday-life conditions. In a previous paper, we presented an SSVEP-based BCI for remote robotic car navigation offering live camera feedback. In this paper, we further improve our implementation by adding a fourth direction (backwards), while redeveloping the software in a free-license environment (Python) using a smaller and lighter robotic car, easier to maneuver in interior spaces. Additionally, we study the possibility of using a single channel EEG and test the performance of our system in an offline session, as well as in an online realistic navigation in a predefined remote route. A total of 14 participants achieved an average offline accuracy of 81%, an average offline ITR of 117.1 bits/min and an average online completion time ratio (BCI completion time against optimal button completion time) of 2.27. All of the participants managed to finish the route under realistic conditions which indicates that our system has the potential to be integrated in the everyday life of immobilized patients.

Metabolic reprogramming is a hallmark of cancer. The main aim of this paper is to integrate a genome-scale metabolic description of tumor cells into a tumor growth model that accounts for the spatiotemporally heterogeneous tumor microenvironment, in order to study the effects of microscopic characteristics on tumor evolution. A lactate maximization metabolic strategy that allows near-optimal growth solution, while maximizing lactate secretion, is assumed. The proposed sub-cellular metabolic model is then incorporated into a hybrid discrete-continuous model of tumor growth. We produced several phenotypes by applying different constraints and optimization criteria in the metabolic model and explored the tumor evolution of the various phenotypes in different vasculature conditions and extracellular matrix densities. At first, we showed that the metabolic capabilities of phenotypes depending on resource availability can vary in a counter-intuitive manner. We then showed that: first, tumor population, morphology, and spread are affected differently in different conditions, allowing thus phenotypes to be superior than others in different conditions; and second, polyclonal tumors consisting of different phenotypes can exploit their different metabolic capabilities to enhance further tumor evolution. The proposed framework comprises a proof-of-concept demonstration showing the importance of considering the metabolic capabilities of phenotypes on predicting tumor evolution. The proposed framework allows the incorporation of context-specific and patient-specific data for the study of personalized tumor evolution and therapy efficacy, linking genome to metabolic capabilities and tumor dynamics.

This study investigates the discrimination between calm, exciting positive and exciting negative emotional states using EEG signals. Towards this direction, a publicly available dataset from eNTERFACE Workshop 2006 was used having as stimuli emotionally evocative images. At first, EEG features were extracted based on literature review. Then, a computational framework is proposed using machine learning techniques, performing feature selection and classification into two at a time emotional states. The procedure described in this paper investigates and assess the effectiveness of selection and classification techniques providing improved classification accuracy. The proposed methodology is able to obtain accuracy of 75.12% in classifying the two emotional states comparing with similar studies using the same dataset.

Significant Virtual Physiological Human efforts and projects have been concerned with cancer modeling, especially in the European Commission Seventh Framework research program, with the ambitious goal to approach personalized cancer simulation based on patient-specific data and thereby optimize therapy decisions in the clinical setting. However, building realistic in silico predictive models targeting the clinical practice requires interactive, synergetic approaches to integrate the currently fragmented efforts emanating from the systems biology and computational oncology communities all around the globe. To further this goal, we propose an intelligent graphical workflow planning system that exploits the multiscale and modular nature of cancer and allows building complex cancer models by intuitively linking/interchanging highly specialized models. The system adopts and extends current standardization efforts, key tools, and infrastructure in view of building a pool of reliable and reproducible models capable of improving current therapies and demonstrating the potential for clinical translation of these technologies.

The generation of “virtual twins” of patients (Avatars) through integration of multiscale data gained from both the clinical profile of the patient and - omics tools, could create an appropriate environment for stratification of patients in fitting cohorts of “virtual populations”. Physiologically based pharmacokinetic & pharmacodynamic (PB/PK/PD) models as in silico clinical trial tools can estimate the PK/PD profiles in specific populations. In this work we discuss examples of how patient Avatars could be exploited in the context of in silico clinical trials and help in identifying novel biomarkers for personalized diagnosis. The PB/PK/PD models, neuroimaging and - omics data, may be fused together to further advance current decision making processes in clinical practice.

Cancer cells inefficiently produce energy through glycolysis even in ample oxygen, a phenomenon known as “aerobic glycolysis”. A characteristic of the rapid and incomplete catabolism of glucose is the secretion of lactate. Genome-scale metabolic models have been recently employed to describe the glycolytic phenotype of highly proliferating human cancer cells. Genome-scale models describe genotype-phenotype relations revealing the full extent of metabolic capabilities of genotypes under various environmental conditions. The importance of these approaches in understanding some aspects of cancer complexity, as well as in cancer diagnostics and individualized therapeutic schemes related to metabolism is evident. Based on previous metabolic models, we explore the metabolic capabilities and rerouting that occur in cancer metabolism when we apply a strategy that allows near optimal growth solution while maximizing lactate secretion. The simulations show that slight deviations around the optimal growth are sufficient for adequate lactate release and that glucose uptake and lactate secretion are correlated at high proliferation rates as it has been observed. Inhibition of lactate dehydrogenase-A, an enzyme involved in the conversion of pyruvate to lactate, substantially reduces lactate release. We also observe that activating specific reactions associated with the migration-related PLCγ enzyme, the proliferation rate decreases. Furthermore, we incorporate flux constraints related to differentially expressed genes in Glioblastoma Multiforme in an attempt to construct a Glioblastoma-specific metabolic model and investigate its metabolic capabilities across different glucose uptake bounds.

Improving the initial diagnosis and the assessment of response to treatment in malignant gliomas, while avoiding invasive methods as much as justifiable, is one major aspect actual research is focusing on. Imaging studies are used to calculate tumor volume and define vital, necrotic and cystic areas within a tumor. Though the visual interpretation of magnetic resonance (MR) images is based on qualitative observation of variation in signal intensity, a correlation of signal intensities with histological features of a tumor is not possible. Better methods are needed for a reliable interpretation of follow-up studies in single patients. Histograms of signal intensities might serve as a method adding quantitative data to the description of a tumor. Using DoctorEye software, tumors can be easily rendered and histograms of the signal intensities within a tumor as well as mean and median signal intensities are possible to calculate. Our results in glioblastoma suggest that these histograms are an innovative method of gaining new tumor-specific information without performing additional investigations in a patient. It can be an additional diagnostic tool in differentiating various intracranial lesions from each other, as well as in assessing response to treatment or progression of malignant glioma.

Glioma, especially glioblastoma, is a leading cause of brain cancer fatality involving highly invasive and neoplastic growth. Diffusive models of glioma growth use variations of the diffusion-reaction equation in order to simulate the invasive patterns of glioma cells by approximating the spatiotemporal change of glioma cell concentration. The most advanced diffusive models take into consideration the heterogeneous velocity of glioma in gray and white matter, by using two different discrete diffusion coefficients in these areas. Moreover, by using diffusion tensor imaging (DTI), they simulate the anisotropic migration of glioma cells, which is facilitated along white fibers, assuming diffusion tensors with different diffusion coefficients along each candidate direction of growth. Our study extends this concept by fully exploiting the proportions of white and gray matter extracted by normal brain atlases, rather than discretizing diffusion coefficients. Moreover, the proportions of white and gray matter, as well as the diffusion tensors, are extracted by the respective atlases; thus, no DTI processing is needed. Finally, we applied this novel glioma growth model on real data and the results indicate that prognostication rates can be improved.

Glioma is one of the most aggressive types of brain tumor. Several mathematical models have been developed during the past two decades, toward simulating the mechanisms that govern the development of glioma. The most common models use the diffusion-reaction equation (DRE) for simulating the spatiotemporal variation of tumor cell concentration. Nevertheless, despite the applications presented, there has been little work on studying the details of the mathematical solution and implementation of the 3-D diffusion model and presenting a qualitative analysis of the algorithmic results. This paper presents a complete mathematical framework on the solution of the DRE using different numerical schemes. This framework takes into account all characteristics of the latest models, such as brain tissue heterogeneity, anisotropic tumor cell migration, chemotherapy, and resection modeling. The different numerical schemes presented have been evaluated based upon the degree to which the DRE exact solution is approximated. Experiments have been conducted both on real datasets and a test case for which there is a known algebraic expression of the solution. Thus, it is possible to calculate the accuracy of the different models.

This paper presents the work of our group concerning cancer image analysis and modeling. The adopted strategy aims to build a complete system for analysis and visualization of DICOM tomographic data, offering a variety of annotation or automatic segmentation tools as well as tools for tumor growth simulation and visualization.

Epilepsy is one of the most common brain disorders and may result in brain dysfunction and cognitive disturbances. Epileptic seizures usually begin in childhood without being accommodated by brain damage and are tolerated by drugs that produce no brain dysfunction. In this study, cognitive function is evaluated in children with mild epileptic seizures controlled with common antiepileptic drugs. Under this prism, we propose a concise technical framework of combining and validating both linear and nonlinear methods to efficiently evaluate (in terms of synchronization) neurophysiological activity during a visual cognitive task consisting of fractal pattern observation. We investigate six measures of quantifying synchronous oscillatory activity based on different underlying assumptions. These measures include the coherence computed with the traditional formula and an alternative evaluation of it that relies on autoregressive models, an information theoretic measure known as minimum description length, a robust phase coupling measure known as phase-locking value, a reliable way of assessing generalized synchronization in state-space and an unbiased alternative called synchronization likelihood. Assessment is performed in three stages; initially, the nonlinear methods are validated on coupled nonlinear oscillators under increasing noise interference; second, surrogate data testing is performed to assess the possible nonlinear channel interdependencies of the acquired EEGs by comparing the synchronization indexes under the null hypothesis of stationary, linear dynamics; and finally, synchronization on the actual data is measured. The results on the actual data suggest that there is a significant difference between normal controls and epileptics, mostly apparent in occipital-parietal lobes during fractal observation tests.