Ecological eHealth Platforms

Text

Sports

Text

DigiMarkers

Text

Hospitals and Nursing Homes Ward Reduced Stressors Designs

Text

Objective indexing of neruro-pathologies

Our approach is based on the assumption, confirmed by our prior research, that indexing of behavioral “micro-patterns” provides objective and quantitative measurements of diseases as well as response to drugs and other treatments. The concept itself stems from and relies on mathematical frameworks yielding macroscopic characterizations of physical systems based on their microscopic properties. This type of coarse-graining approach has been used many times in fluids or smart materials with remarkable success throughout the past century. We adopt an approach in which the coarse-graining is provided by macroscopic surrogate behavioral patterns in place of the direct EEG of fMRI scanning of brain activities.

The microscopic level is represented by tiny changes in physiological readings that we map onto the Hurst index, equivalently onto the Hausdorff-Besicovitch dimension, that can be thought of as a macroscopic system parameter. In analogy, this macroscopic system parameter would correspond to the diffusion parameter in fluids. There are fundamental differences though compared to fluids. While the Brownian motion and its macroscopic model given by diffusion is driven by a single mechanism, namely the molecular collisions, giving raise to, e.g., free-path or cardinality, the behavioral complexity indices are macroscopic pronunciations due to a variety of actions sustained by different neural subnets driven by both internal and external parameters. There are similarities, too. The indices we introduce to objectively measure various physiological effects can be thought of as being subsumable and describable through one single measure. Similar to, say, the entropy or temperature. Consider temperature. Temperature is a macroscopic statistical quantifier and so are the indices we consider. Temperature itself cannot indicate a specific disease or etiology, but it is routinely used as a forefront diagnostic tool to judge physiological states. The indices we introduce are also macroscopic statistical quantities that can be thought of a generalization of the temperature analogy to gauge physiological states.

Athletes' performance, improvement and health risk indexing

We introduce an approach to monitor athletic training patterns using Neural Networks, a part of the Deep Learning and Artificial Intelligence frameworks, applied to subjective yet consistent monitoring, in order to optimize training effectiveness, and to preserve health. One of the most common problems within a suboptimally designed training program is the occurrence of the overtraining syndrome. The definition of this syndrome is currently accepted by both the European College of Sport Science and the American College of Sports Medicine.

The current recommendations concerning training monitoring and early detection of the overtraining syndrome advise to carry out psychological monitoring using validated questionnaires (such as the Profile of Mood State (POMS)), a monitoring of the training load perceived by the athlete (e.g. session rating of perceived exertion (RPE), a health problem monitoring tool, and a performance monitoring tool. The blood parameters and hormonal follow-up is not advised because of cost and feasability reasons, and also, because the current scientific literature has not reached consensus about the usefulness of specific biomarkers. In order to achieve such type of training monitoring, the development of easy-to-use electronic tools allows the collection of athlete’s data in a standarised manner.

Our approach allows for identification of recovery and training parameters patterns among young swimmers which are linked with performance improvement or otherwise. We acheive these goals by considering six different combinations of different subjective data that are projected on three dimensional Euclidian space. We select one of the three reading to serve a role of binary parameter by measuring a particular datum distance from the selected data mean. We then use Neural Networks based analysis to construct separation hyperlanes. These planes indicate two predictive distinct areas of performance. We use our propritery Geometric Activity Performance Index (GAPI) and Spectral Performance Ranking (SPR). These indices provide significant and both quantitative and qualitative inside into the performance and improvements assessments.

Objective indexing of stress

The Human Stress is a system responce to various perceived external and/or internal stimuli. The initial response is governed by the sympathetic nervous system while the return to system equilibrium might be controlled by a convex feedback loop provided by homestasis or non-convex control due to time-local energy expenditure yielding allometric feedback responce.

The reason to consider the allometry is that it correlates amount and availability of resources in fight-or-flight responce. More importantly allometric feedback implies self-similarity of the underlying microscopic behavioural time-series. The self-similarity has a fundamental role in the presented approach. The more complex nonlinear control mechanisms can be characterised by existence of the allometric power law. The existence of multiple stable behavioural states, including both health and diseased states, is not anticipated by homeostatic linear descriptions of behavioural control mechanisms. Multiple stable states are, however, predicted by allometric based models in which the resistance to perturbation derives from feedback loop interactions among controlling subnetworks, that is just an another pronunciation of self-similarity. From the phenomenological prospective, we view human stress as a state when either external or internal demands exceed the regulatory capabilities of the autonomous part of the Central Nervous System. We use our proprietary mathematical constitutive model that can be used in clinical settings to objectively quantify a level of stress a human subject would be experiencing. We rely on body-attached sensors recording certain behavioral surrogate data. In other words, we compute an objective measure of stress severity using a combination of sensing, posteriorly mathematical analysis, and constitutive assessments to arrive at both the Compound Spectral Stress Index and Geometric Stress Index.

Analysis of stressors

The hospitalisation-associated disability might be due to several reasons, e.g. immobilisation, reduced activity, stress, poor sleep, nutritional problems, medication, etc. Of course, the prehospitalisation functional capacity reserve is also important: a patient with a larger reserve can decrease more in his function before becoming disabled. Older adults may lose about one kilogram of lean muscle mass per ten day of bed rest. Muscle protein catabolism is increased by muscular inactivity (e.g. immobilisation, bed-rest, increased bed-time) and the stress response with an increase in counterregulatory hormone concentrations. Inflammation is another factor associated with muscle weakness and also nutrition plays a role during hospitalisation.

There is an urgent need for health care owing to a high number of hospitalized older patients presenting several symptoms of frailty and other geriatric syndromes to early monitor hospitalized adverse events. Some studies have observed a strong relationship between cognitive and physical decline, as manifested by an elevated number of hospital-related falls. Above all, hospitalization itself should be considered as an important delirium and frailty risk factor. Interdisciplinary preventive care strategies focusing on limiting physical and cognitive decline during hospitalization by early monitoring using technology may be effective in averting delirium and frailty and other poor posthospitalization outcomes. Health care reforms focusing on building more effective and efficient care models for older inpatients, integrating newly developed ehealth devices, are urgently needed to develop “senior-friendly” hospitals, including specialized geriatric care units.

Unique identification of humans using behavioural complexity patterns

We provide an analytic technology using complexity of behavioural variables to identify human subjects. The identification is made by projection of self-similar and normally distributed behavioural data onto Euclidian product spaces based on a non-disruptive sensing. We introduce platform combining data acquisitions using body attached sensing devices, mathematical analysis of complexity of the acquired signals, and subsequent projections onto Euclidian spaces representing behavioural states.

Externally applied wearable sensors al-low for collecting self-similar time-series of behavioural or physiological parame-ters such as heart rate, blood oxygena-tion, skin temperature, steps frequency, and others. The measurements of the behavioural parameters represent surro-gate data that characterize patterns through segmentation and a subsequent complexity analysis using data pertaining to expended time spans. Behavioural surrogate data are segmented and analysed to extract meaningful patterns hidden in the measured time se-ries of behavioural data. The word segmentation is used to indicate a connection between behavioural temporal segmentation and image spatial segmentation. This approach is similar to the concepts of the Gestalt school and psycho-physiologists analysing perceptions of images in the retina in the first instants of their arrival. Segmentation of captured images is aimed at and allows for recognition of well-defined objects from discrete patterns.

We use complexity indexing to recognize well-defined behaviours that can differentiate various subjects using time-discrete behavioural time-series. We build geometrical images of behavioural states evolving over time that can be measured and analysed using mathematical tools based on complexity indexing. Our approach shares some fundamental underpinnings with image segmentations, in particular the use of a measure generating dimension.

Cassiopée Computational Ecosystem

Cassiopée is unique and the worlds first complete analytical diagnosis tool providing complexity analysis of behavioral and physiological data. It detects scale invariance and it provides a measure of complexity of behaviors and the underlying micro-patterns as well as many other physiological sensory recordings.

The complexity analysis of the behavioral and physiological data is the key objective measurements of effects of drugs, detection and/or analysis of progression of a disease and, consequently, an objective evaluation of applicability of a particular drug for treatment. Cassiopée is fully compatible with a number of behavioral and physiological data acquisition systems so that it can produce complex analyses of both objective and subjective human surrogate data.

The three basic structures of the system are shown below.

Principles of complexity analysis

Characterization of behavioral patterns and in particular the locomotion and its micro-patterns using either the non-integer Hausdorff-Besicovitch dimension or the Hurst exponent pertain to the analysis of the complexity. The higher these two measures are the less complex are the behavioral patterns and vise versa. There are two underlying assumptions that have to be valid to apply our approach: (1) the data representing both the locomotion itself and its micro-patterns have to be stochastic processes themselves exhibiting self-similarity, (2) and that such patterns can be modeled by fractional Brownian motions/functions.Technically, the first of the requirements is verified by computing the joint probability density function that have to be nearly identical over a range of the coarse-grained discrete recordings. The identification of the self-similarity scaling index, the Hurst exponent, is obtained using maximum likelihood estimate of the process identification procedure.Our overall approach is based on the overhelming scientiic evidence that indexing of behavioral micro-patterns provides objective and quantitative measurements of drugs, treatments or injury pronunciations. The basic scheme is shown below. It is important to realize that the direction of deviation from the Base Line indices in either of the two directions shown must be interpreted with respect to the expected outcome of a drug application, treatment or injury type.

Founders