D2.1: Functional and Non-Functional Requirements of the ARMOR Services

 

D2.1: Functional and Non-Functional Requirements of the ARMOR ServicesThis document is part of the WP2, whose main objective is to evaluate the most efficient body sensors available and adapt them in order to be able to collect the required physiological data and store them in a proper way.
The main goal of this document, according to the objective of Task 2.1, is to ensure practical applicability of the ARMOR system with respect to the added-value services that will be implemented. A comprehensive analysis of the clinical scenarios and the rationale that led to their selection is presented in order to ensure system usability, reliability and responsiveness.
Furthermore, the gaps in the current state-of-the-art of epilepsy treatment/ monitoring are mentioned. It is then defined which of these gaps will be addressed by the ARMOR Consortium to create new benefits for the epileptic patient and healthcare professional. In this framework, the benefits for the patient are distinguished from those expected for the researcher (multiparametric analysis model) and the health professional are distinguished (doctors, support staff, other that will use the ARMOR services to improve the support of epilepsy diagnosis and treatment).
In accordance with the scenarios defined, a detailed presentation of the related Use Cases is displayed. The ARMOR services have been classified into two main software packages, the "ARMOR Monitoring Local System" and the "ARMOR Data Management Centre", for the needs of the project. Therefore, separate applications are considered in the home environment and the remote environment for offline analysis and ARMOR services synthesis.

 

D2.2: Real time data Requirements

 

D2.2: Real time data RequirementsIn this deliverable  the  real time data  requirements are analysed  from three perspectives: (1) from the theoretical perspective, where the ideal case for the real time analysis is shaped, (2) from the perspective of maximum clinical utility, where the ideal clinical solution is explored, and (3) from the practical perspective, where  the "ideal" requirements are contrasted with the  real-life practical limitations. These conflicting requirements must be eventually resolved to reach a compromise that fulfils in the best possible way the stated objectives of the project, within its time duration and resources.
The online analysis of data will be effective  only  if the data that are analysed are optimised for such analysis.  The optimization must first recognize the items of key  importance to each particular case: key clinical requirements and critical technical issues.
The report begins with  a discussion of possible ways to resolve the  conflicting requirements emerging from some theoretical considerations imposed on the data and practical considerations related to cost, comfort and adequacy  of  solutions. It then continues with the identification of time stamps as a key critical parameter and provides  an overview of the methods that could control its quality.
The report proceeds to describe how the  online data  analysis methods will be evaluated based on existing and to-be-collected sleep MEG/EEG data. In this document we also consider the data fusion platform and describe it as is delivered at this early stage in the project. Finally, the report ends with a summary section, where we briefly summarize the requirements for the real-time data in a tabular easy-toreference form.

 

D2.3: ARMOR Middleware Requirements

 

D2.3: ARMOR Middleware RequirementsThis document is part of the WP2, whose main objective is to evaluate the most efficient body sensors available and adapt them in order to be able to collect the required physiological data and store them in a proper way.
This document defines the requirements related to the middleware  platform within ARMOR. From device sensors to electronic health record interface and notification applications, the middleware address the needs of the interoperability that is required among the different technologies in place within ARMOR and its functional goals.

 

 

D2.4: ARMOR Sensor Requirements

 

D2.4: ARMOR Sensor RequirementsThis document is part of the WP2, whose main objective is to evaluate the most efficient body sensors available and adapt them in order to be able to collect the required physiological data and store them in a proper way.
The main goal of this document is to define all the requirements related to the sensor platform. From sensor types to data format and interfaces, this deliverable covers all  the sensor related requirements.

 

 

 

D2.5: ARMOR Middleware Service

 

D2.5: ARMOR Middleware ServiceThis document is part of the Work Package 2, whose main objective is to evaluate the most efficient body sensors available and adapt them in order to be able to collect the required physiological data and store them in a proper way.
This document defines ARMOR middleware service. Taking into account requirements collected in [Ref.1], this documents provides insights about the service functionalities of the middleware and the interfaces connecting the different technologies in place within the ARMOR System.
The document is divided in three main blocks, the first and second block describe the functional and non-functional aspects of the ARMOR Middleware, while the third block is a set of annexes where all necessary information to setup and run ARMOR middleware is provided.
ARMOR middleware runs within the Home Gateway, it is the ARMOR's ICT part responsible to connect sensor data to upper software layers (like patient health record service and the tele-alarm and messaging manager), as well as provide the necessary infrastructure (thanks to the inclusion of a data stream management system called ARMOR Insight) for the development of the future on-line multi-parametric data processing and analysis that will take place in WP5. Initial ARMOR upper software layers connected are the ARMOR's Tele-alarm and Messaging Manager and the ARMOR's Personal Health Record. All ARMOR Middleware components follow a SOA (service oriented architecture) paradigm that allows easy scalability of the system.

 

D2.6: ARMOR Sensors

 

D2.6: ARMOR SensorsThis document is part of the WP 2, whose main objective is to evaluate the most efficient body sensors available and adapt them in order to be able to collect the required physiological data and store them in a proper way.
This document is separated in two parts. In the first part the intergration of the sensor and the FPGA module for the secure data transmission is discussed. There each component is presented and the approach for their intergration is shown. Finally a possible solution for the complete intergration of those modules under one single housing is shown.
In the second part the definition of the sensors that will be used in WP6, where the final system will be tested is presented. This is based both on the scenarios and the sensor requirements as defined in previous deliverables. The sensor selection is further separated between the sensors to be used for the online and the ones for the offline scenarios.

 

D2.7: Data Acquisition

 

D2.7: Data AcquisitionThe deliverable D2.7"Data Acquisition" has been produced in Work PackageWP2: Multi-parametric Data Collection (WPL: KIT) which runs in months 4-20 of the project. Particularly it reflects work described under Task 2.4: Data Acquisition (Task leader KCL) which runs in months 8-20 of the project. We hereby report on the experience gained and progress made on data acquisition and the procedures adopted for this acquisition, which ARMOR will use in the case of real time monitoring (scenario #4). We describe the entire spectrum of procedures of data acquisition. Task 2.4 aimed to present and make available already acquired data which would be studied as for their type, the modalities they included, quantity, medical targets, etc. and select the most appropriate acquisition procedures for guiding the development of ARMOR towards the stated objectives. It further called for new recordings so that the totality of data will face in a complimentary way the main medical aspects of epilepsy. The acquisition procedures for each modality are presented separately and then in combination (polygraphy). This analytical presentation enables consideration and qualified decisions on the compromise needed between the richness of acquired data and limitations imposed by the real time monitoring. It is concluded that the great variety of seizure types and the possible physiological markers by which each of them can most reliably be detected imposes analogous versatility in ARMOR's data acquisition procedures. ARMOR has to be developed as a system equipped with the full potential of all acquisition procedures, from which the clinical epileptologist will choose to activate for real time monitoring the set which is most appropriate for the particular patient and his/her type of seizures.

 

D3.1: Data Privacy and Security Requirements 

 

This document is part of the WP 3, whose main objective is to tackle all aspects of ARMOR coordination and communication system emphasizing on security and privacy issues.
The main goal of this document is to identify all the related patient and medical data security and privacy aspects based on input from technical and medical experts and according to all relevant legislations and EU ethical recommendations.
In this document all the possible security and privacy issues have been identified. The main focus was to underline all the possible attack points in form of risk analysis and then introduce the technical measures that will be taken against them. Therefore issues such as authentication, authorization, secure communication, cryptography and anomynisation have been discussed extensively. The security checklist at the end summarises the most important aspects that will taken into account in form of a checklist.

 

D3.2: Data Management Processes 

 

This document is part of the WP3, whose main objective is to tackle all aspects of ARMOR coordination and communication system emphasizing on security and privacy issues.
This deliverable focuses on data management processes concerning both technical related controls to protect against respective privacy/security issues and attacks as well as data management procedures aiming to defend ARMOR platform against soft security issues such as information misuse, unauthorized access, accidental error etc.

 

D3.3: End to End Security Infrastructure 

 

This document comprises the deliverable for Task 3.2 entitled "End-to-End Security Infrastructure Configuration". In that context the main objective is to study security provision of ARMOR platform components based on respective requirements' deliverable, evaluate respective capabilities of already existing platforms/systems within the ARMOR consortium and propose appropriate adaptations.
Different approach will be evaluated based on their suitability and efficiency concerning the ARMOR specific objective. The aim is to identify the best options and indicate how already existing platforms can support them. Aiming towards end-to-end infrastructure such evaluation will cover issues from the acquisition of raw data from diferent ARMOR sensors to back-office system provided to the medical users and caregivers. Adequate techniques, algorithms, mechanism and technologies are evaluated based various aspect such as cost, complexity, security efficiency, effect on system performance and feasibility. Furthermore, respective evaluations and enhancements are made taking into consideration requirements made in relative deliverable D3.1 as well as the end-toend system architecture as that is defined in the context of the ARMOR consortium.

 

D3.4: Real Time Data Communication Infrastructure

 

D3.4: Real Time Data Communication InfrastructureThe deliverable D3.4 reports on the outcomes of the Task 3.3 of WP3. It focuses of the communication aspect of the ARMOR system that is expected to enhance the performance of already existing platforms/systems (provided by involved partners) in the direction of efficient and real-time communication capabilities.
In order to achieve these objectives the appropriate adaptation concerning sensors'communication, data aggregation and data fusions in aggregation points (access points) is considered. Furthermore, the effect of backhaul data transfer to the end-to-end communication is evaluated based on existing platforms, considering different technologies and configurations leading to the proposal of appropriate adaptations.

 

 

D3.5: Set UP of the ARMOR Service Delivery Platform

 

D3.5: Set UP of the ARMOR Service Delivery PlatformThis document produced as part of the Work Package 3 activities, whose main objective is to tackle all aspects of the ARMOR coordination and communication system emphasizing on both, security issues related to patient data, and real-time capabilities based on available systems and existent platforms.
In line with the Task 3.4 objective, the main goal of this document is to define the communication capabilities and services needed so that the ARMOR system meets all requirements identified so far and described in previously prepared deliverables.
Technical specifications of the ARMOR Service Delivery Platform are presented in this document. The main focus is to describe how communication messages will be exchanged among different platform components, and provide the data flow among components.

 

 

D4.1: Medical Data Monitoring And Analysis Requirements and Specifications 

 

In the process of delineating the pertinent biological functions and the optimal number of sensors for the needs of ARMOR, and according with the objective of Task 4.1, clinical and neurophysiological (in the clinical context) provisional analysis of already acquired neurophysiological data from 20 patients with focal epilepsies, 20 patients with idiopathic generalised epilepsies and 20 patients with nonepileptic paroxysmal disorders from KCL, and neurophysiological analysis of all night multichannel sleep recordings of 10 healthy subjects from UoP has been performed. In addition, and although prospective recordings had not been planned for this early phase of the project, we performed 24-hour video polygraphic recordings with 32 electrodes in 5 patients with epilepsies and another 5 with sleep disorders without epilepsy; these recordings include all night polysomnography with full (21 channels) EEG montage and comprehensive measurements of all parameters that record autonomic functions.
Analysis of these data have given clear indications about the provisional requirements for a number of different clinical scenarios, and the relevant specifications. These scenarios are not to be used to test ARMOR, but to offer guidelines for further clinical research and data collection and analysis in order to establish the minimal and most effective set of sensors that can pragmatically detect the key changes of the body functions that are pertinent for diagnosis. It is understood that not all parameters will be required for the final product. Ongoing clinical and neurophysiological analysis of the prospectively aquired polygraphic data from the heart and other autonomic functions, is expected to reduce the number of body functions that need to be monitored (and therefore the number of the relevant sensors), while analysis using MEG is expected to reduce the number of EEG electrodes. Narrowing down the number of sensors and closely cooperating with our technology partners will enable the development of technologically feasible ways to record and transfer the necessary for the diagnosis data, and the building of ARMOR.
The work is progressing without problems and is well within the original plan.

 

D4.2: Report on Data Acquired and Pre-processed 

 

Upon the completion of the first year of the project we hereby report on the experience gained (by mainly the teams in AAISCS, KIT and UOP) from off line pre-processing of multimodal data acquired from control subjects and from epileptic patients during wakefulness and sleep.
The characteristics of acquired data, several procedures of pre-processing methodology and the results from implementing these preprocessing techniques on data from the different recording modalities are presented. The modalities employed covered the electric activity of the brain, the body homeostasis and its autonomic control as well as the overt activity of the subject. The presentation is not exhaustive but covers the most essential of the preprocessing techniques, offering examples of results. Techniques were evaluated and adopted or originally developed, which allowed data cleaning and baseline correction, data transformations (i.e., filtering), dimensionality reduction (i.e., downsampling), summarization and integration (of different derivations), etc. Others allowed automatic or manual detection of specific data features (i.e., heart's QRS), and so lead to higher order data representations (like Heart Rate Variability, Hypnograms, etc). Finally methods were developed and applied to magnetoencephalographic data relying on tomographic source analysis to clean the data from artifacts, focus on activity in specific brain regions of interest and extract their activation time courses in every single trial. This approach aims to facilitate acquisition and preprocessing of data in specific groups of patients.
It is therefore believed that sufficient experience has been acquired with offline pre-processing of all planned but independently recorded modalities. This experience along with that of the following steps in data fusion, analysis, etc, will adequately serve the purpose of designing and developing the relevant parts of ARMOR.

 

D4.3: Techniques and Results of Data Fusion 

 

This deliverable describes mainly the work of task 4.3 dealing with techniques used for multi-parametric data fusion and integration, their implementation details for the purpose of ARMOR, and describes the result of data fusion so far. The underlying idea is that the synthesis of information gathered from various knowledge sources and multimodal sensors can provide a better understanding of the underlying processes, e.g. significant epilepsy related events. The technical content of the work is neatly separated at three levels: (1) fusion of the sensor signals of the same and closely related modalities (e.g. EEG and EOG, EMG, etc.); (2) fusion of electrographic (e.g. EEG or MEG) and anatomical (e.g. MRI or approximate head model) data; (3) fusion at the level of ARMOR middleware (related also to the deliverable D2.5). These relatively independent levels of the work in Task 4.3 proceeded in parallel in UoP, AAISCS and STMA initially, but they were brought together in this output.
The data fusion of the sensor signals is a key component of ARMOR, which relies on multimodal recordings and existing knowledge on epilepsy. This part of the work recruits ICT technology to mimic what is done intuitively by a diagnosing physician. The application of the work completed and the ongoing work that will follow will not only be a key ingredient of the ARMOR tools but could also help us understand the mental processes that a clinician is trained to do.
Techniques considered for the fusion of the sensor signals from the EEG and peripheral sensors (restricted so far to EOG, ECG and EMG) included measures of local (electrode) signal power changes at different temporal scales and changes in synchronicity and connectivity of sensor-level networks. The fusion of electrographic data with brain anatomy is a self-contained task, it is completed and the tool produced is fully adapted for the future work of ARMOR. Some of the data fusion work relies on information of clinical nature (e.g. data annotations) and formats that are not yet finalized, but they are part of ongoing work in WP5 and WP6. It is therefore inevitable that some iterations of the data fusion work will continue in the WPs 5 and 6.

 

D4.4: Data Management

 

D4.4: Data ManagementThis document presents the work on Task 4.4: Development of Tools for Offline Data Management. The most important outcomes of our work presented in this deliverable include a) the development of a database schema able to store all the available information about the multimodal recordings (including relationships and correlations between clinical and neurophysiological data of interest) and the most common analysis tasks, b) the development of the necessary infrastructure that supports the necessary data transfer requests among the various databases involved in ARMOR (nASSIST middleware database, PHR, offline analysis database), c) the initial steps of embedding content-based data retrieval into the database infrastructure, and d) the data compression and summarization issues investigated.

 

D4.5: Offline analysis of data

D4.5: Offline analysis of dataThis deliverable (D4.5) describes the methods that have been developed for offline data analysis and the results obtained from such analysis. On the basis of this and previous clinical experience it also provides recommendations for polygraphic online monitoring of patients. This deliverable is mainly based on the work of tasks 4.5 and 4.6, dealing with the development and application of offline data analysis methods, respectively.

 

 

D4.6: Development of models and personalized Patient Health Profile

D4.5: Offline analysis of dataThis Document is part of the WP4 and its main goal is to describe the Patient Health Profile (PHP) functionnalities and modules. PHP includes the systematic integration of the sum of information related to each patient. It ensures the unified, complete, organized regulation of data flow that can be personalized depending on the clinical question, the medical history and the results of use of detection tools.The structure of the PHP is such that systematically describes all the related patient information. This information, which consists of static and dynamic data, is structured in such a way that can be automatically or semi-automatically accessible from other components of the ARMOR system. The Health Record Summarizer is also presented and the form and information included are reported. The personalized patient health record is periodically updated by information extracted from the analysis. This analysis is based on the tools developed within the corresponding WP4 and WP5 tasks for offline and online analysis respectively. Additionally valuable information deriving from sleep evalutaion tools are included. Moreover the models for the investigation of the events of interest are reported. Specifically the associations and rules for different types of events and the differentiation between epielptic and non epileptic events are stated in terms of sensors' findings and furthermore the combination of certain sensors' caracteristics is defined to make the differential diagnosis between focal and generalized epileptic events. Finally the implementation of metadata and the link of PHP to data fusion framework is presented.

D4.7: Decision Support for Diagnosis

D4.5: Offline analysis of dataThroughout this deliverable, we cover the actions taken within the "Computer Assisted Diagnosis" task. Here, we present the tools developed aiming to support medical diagnosis, such as the discrimination of epileptic and non-epileptic events, identification of generalized and focal seizures, along with the detection of status epilepticus. In addition, we focus on the compression hardware implementations for the diagnosis support, correlation analysis and offline processing of physiological data. At the end of the present deliverable, guidelines for medical decision making are provided.

 

D5.1: Medical Data Requirements and Specifications

 

D5.1: Medical Data Requirements and SpecificationsThe objective of Task 5.1 is to establish the requirements for the online medical data monitoring  and analysis. The work in this task relies on and extends the work conducted in Task 4.1, focusing the description on the online requirements. The deliverable D4.1 of the corresponding task was restricted to the consortium partners only, while the current deliverable (D5.1) is public.
Section 2 provides background information on: (2.1) the current state-of-the-art methods for online monitoring and analysis of multimodal data from epilepsy patients; (2.2) specificities of physiological data acquisition for online analysis; and (2.3) a brief description of theoretical and practical aspects of relating the electrophysiological  signals measured non-invasively outside the head to detailed tomographic neural source estimates.
The  main  work of the involved partners is presented in section 3: (3.1) the modern polysomnography measurements (UoP) and (3.2) application of real-time tomographic source analysis of MEG/EEG data to epilepsy monitoring, something that is not routine clinical practice, so it is approached here from the basics, i.e. demonstrating the relationships that must exists between signal based measures and tomographic estimates (AAISCS). Section 4 summarises the results of this work and provides the preliminary requirements and specifications for online  medical data monitoring and analysis. Section 5 presents this deliverable (D5.1) in the overall context of ARMOR. Although this deliverable describes the work of a relatively small task (5 PM) it serves as a basis for the key work of the project that follows in WP5 and links back to foundation work done and to be done in WP4.

 

D5.2: Report on Real-Time data Pre-processing

 

D5.2: Report on Real-Time data Pre-processingDuring the first half of the project, WP5 concentrated on making the initial steps for the development of the tools that will support online analysis, with real-time pre-processing being one of the first to consider. In this document, we present the progress achieved on the real-time data pre-processing.
The work focused on investigating existing techniques and developing new ones for real-time multimodal data pre-processing, data reduction and data fusion. Corresponding counterparts from the offline data pre-processing and fusion were tested and optimized so that they can run in real time. Relevant conclusions drawn from offline data preprocessing and fusion were used in preprocessing and fusion in real time. Online algorithms for real time seizure detection/classification as well as detection of other waveforms of interest, such as sleep spindles, K-Complexes, were studied and developed. Several methods for pre-processing physiological signals such as step detection and the necessary tasks for HRV analysis were also developed. The online data pre-processing algorithms were developed using the available functionalities of XAffect, Microsoft™ StreamInsight™ (a Data Stream Management System) and the Hidden Markov Model Toolkit (HTK). Methods of utilizing the results of tomographic analysis for supporting ARMOR's real-time tasks were also introduced. Data reduction and summarization methods that can be adjusted to work in real time for streaming data were investigated and used. Several implementation details of the developed online data pre-processing algorithms are discussed in this deliverable.

 

D5.3: Report on Data Manager of Online Support

 

D5.3: Report on Data Manager of Online SupportThis deliverable describes the work on managing ARMOR's sensor data streams in the context of facilitating online analysis. Specifically, it presents the basic concepts and tools that will be utilized in order to transform the data into the appropriate format and continuously stream them to the online analysis algorithms. A data stream management system, Microsoft StreamInsight, has been employed as the platform for implementing these algorithms and interfacing with xAffect, the online streaming tool that aggregates the data from the sensors. These efforts aim to establish the streaming infrastructure that will provide both the data sources and the development platform for realizing ARMOR's critical real-time tasks.

 

D5.4: Real time (online) analysis of data

D4.5: Offline analysis of dataThe work inWP5 relates mainly tothe development of the tools that support the online analysis of the acquired multimodal data. In this document, we present the progress achieved on the development of novel real time analysis methods and S/W tools for multi-parametric stream data. Trade-offs for automated analysis taking place at the local site of each patient (instead of at the Health Center) aiming to reduce processing time and storage requirements were taken into account. The implementation of hardware computational intensive algorithms allowed real time analysis of patient's data. Results of the online analysis of multiparametric data and the evaluation of the methodology are also reported. During this task, we mainly targetedthe development of online (real-time) seizure detection, alpha rhythm detection and analysis of skin conductance. The online seizure analysis included mainly time and frequency domain features from EEG and ECG. Inaddition to this, algorithms for the analysis of the electrodermal activity were developed. The developed online tools were based on patient-specific models of ictal patterns, however the software framework allows the extension to inter-patient models. An online alpha rhythm detection tool was developed in order to support the development and testing of the end-to-end ARMOR online architecture and assist the debugging of it. The algorithmic part of the online software tools was mainly developed in the Mathworks' Matlab environment. In order to meet the specifications of the ARMOR platform and the need for online operation the S/W toolswere later introduced in the Microsoft's StreamInsight environment.This enabled processing of data streams in a time and memory efficient way. The EEG device was integrated in xAffect and xAffect was extended to be able to stream data from a file in order to be able to simulate the online analysis while streaming data throughout the whole ARMOR platform. As part of the work of T5.4, hardware implementation of feature vector extraction and seizure detection algorithms was performed. The hardware implementation will result in significant reductionof the amount of wirelessly transmitted data, thus minimizing power consumption as well as enhance the robustness and efficiency of the WSN communication capabilities.

 

D5.5: Risk Assessment and Decision Support for Emergency Situations

D4.5: Offline analysis of dataIn D5.5 we present the risk assessment and decision support mechanisms for emergency situations. Specifically, we describe the design and implementation details of the online tools for risk assessment. Risk assessment involves low and high risk situations of interest. Moreover, it involves PHR notification mechanisms for alarms and/or emails, depending on the level of risk. The implementation and validation of the decision support mechanism, workflows for each event and rule encoding necessary for implementing the decision support are described in detail. In addition, we describe the conclusions of the correlational analysis of the clinical neurophysiological data obtained from the retrospectively collected and prospectively acquired multimodal recordings. The conclusions of the correlation analysis data obtained during the whole night sleep recordings, which would indicate possible premonitory signs and risk for seizure development are also discussed here. Additional information about the correlation analysis and its significance through reference to how these might relate to regional changes in activity, as this is manifested in the tomographic solutions of MEG signals, is provided. Moreover, the results of hardware implementation of algorithms' developed in Tasks 4.5 and 5.4, integrated into the ARMOR decision support system are presented.

 

D6.1: Definition of ARMOR Metrics 

 

The deliverable D6.1 "Definition of ARMOR metrics" has been developed under the scope of Work Package 6, "Test Cases" which runs under the leadership of KCL in MMs 18-36 of the project. Particularly it reflects work described under Task 6.1: "Definition of metrics" (Leader: UoP) [M19-M22]. This deliverable follows two axons: (a) combined ICT and medical approach and (b) bottom-up approach, i.e. from the metrics of simple EEG events to those of complex seizures expression.
Experience gained in previous tasks related to normal and epileptic data multimodal acquisition, handling and analysis allowed the definition of the "metrics of seizures", i.e. which are the cardinal features of ictal (i.e. seizures) and interictal epileptic expressions, which can be detected and quantified, so as to serve clinical diagnosis and the other goals of ARMOR. These include directly assessed elementary events, secondarily derived parameters as well as multimodal integrated ones from normal and epileptic subjects. In the 8 appendices of this report we give details of the normal ranges and the description of abnormal patterns. We focus on (but are not limited to) the most common and important events taking examples mainly from the electrical recordings of the brain.
The biomedical metrics can be separated to metrics of elementary EEG events and metrics of epileptic expression. We summarize the metrics of elementary EEG events (waveforms, rhythms, common artifacts), of derived variables of EEG (i.e. frequency, voltage, synchrony, periodicity, paroxysmity) and of derived multimodal data characteristics (i.e. time frequency analysis, sleep staging, quantifying microarousals and Cyclic Alternating Patterns). From all above one can derive metrics of sleep quality. Further on we summarize metrics of epileptic expression starting with elementary EEG events and derived variables for interictal and ictal epileptic EEG waveforms and continuing to metrics of derived multimodal data characteristics which are expected to help (a) Diagnosis of epileptic seizures and differential diagnosis from non-epileptic paroxysmal events, (b) Differentiation of nocturnal epileptic seizures from other paroxysmal non-epileptic events (sleep disorders) and (c) Detection / prevention of nonconvulsive status epilepticus. Based on the clinical relevance and characteristics of some biomarkers and guided by MEG and EEG tomographic studies, efforts will be made to further define the characteristics of these biomarkers and their reliability for different sensor combinations. Included here are normal; sleep biomarkers as well as seizure related EEG biomarkers.
Next in this report we summarize the metrics of ARMOR performance, i.e. the metrics against which the ARMOR system will be evaluated. . We start by describing the metrics of Wireless Sensor Network Performance: (a) life time capabilities of sensors, (a) communication capabilities, (b) Sensor Lifetime Capabilities, (c) Bluetooth Radio Security features overhead, (d) ARMOR Encryption Module overhead and (e) compression algorithm performance and overhead. Under metrics of data analysis we describe the metrics for (a) event detection, (b) event classification and (c) knowledge discovery/data mining and correlation/ ssociation rules. Finally we take into account the metrics for data management: (a) DSMS: detection latency, memory requirements, produced events/sec and (b) DBMS and content-based retrieval: Query processing time, memory requirements and offline storage requirements.
Metrics always depend on what is measured as well as on how. Therefore Task 6.1 demanded and gave an excellent opportunity for collaboration and common understanding by the medical and ICT teams (of which there is never enough). Since some of the data analysis tools are still in development (i.e. Tasks 4.5 - 4.8 are due to finish in M30 and M32), some of the ICT metrics are described from state of the art rather than practical experience with ARMOR data. The final list of tools with their metrics should be ready well before the end of these mentioned tasks and finalized at the Barcelona meeting in 11/2013. For the purpose of convenience to the reader, all biomarkers and performance metrics reported in this deliverable are indicated in bold.

 

D6.2: Test Cases Requirements and Specifications 

 

This deliverable "Advertising materials" is under work package 7 "Exploitation and Dissemination". It lists the advertising materials produced during the first two years of the ARMOR project. These materials help to raise awareness of the project and communicate its achievements so far (at the second of the 3 years of the project). A brief summary of planned marketing activities for the last year of the project is also given.

D6.3: Test Cases Implementation

D4.5: Offline analysis of dataThe deliverable D6.3 "Test Cases Implementation" has been produced in Work Package 6, "Test Cases" which runs under the leadership of KCL in MMs 18-39 of the project. Particularly it reflects work described under Task 6.4: "TEST Cases Implementation" (Leader: KCL) [M23-M39].

 

 

 

D6.4: Test Cases Evaluation

D4.5: Offline analysis of dataThe current deliverable (D6.4) is based on Task 6.5 and details the results of the evaluation of the ARMOR platform and its methods. Recordings from patients have included two events of sleep microstructure: K-Complexes and spindles. ARMOR's seizure classification tool was used to evaluate ARMOR methods for their ability to detect epileptic spikes and seizures (offline and online), and to differentiate epileptic from non-epileptic brain activity. The outputs of ARMOR methods were compared with the ones obtained through the visual assessment by KCL clinicians. Whenever applicable, the overall performance of the ARMOR methods was compared with other state-of-the-art laboratory tests. In addition to real data recorded from patients and healthy individuals, ARMOR was evaluated also based on simulated EEG data with and without seizures. As part of the ARMOR evaluation, two datasets were subjected to detailed tomographic source analysis. These analyses revealed estimates of brain activity related to ARMOR biomarkers, established their usefulness and provided objective measures for assessment of the ARMOR methods' performance. Moreover, the tomography-based EEG data analysis was used in a ‘smart' way providing clinically useful information and guidance for the future measurements. The commentary and conclusions written by KCL clinicians regarding the different aspects of ARMOR performance is the central component of this deliverable.

D7.1: Dissemination Plan

D4.5: Offline analysis of dataThis document is part of the WP 7, whose main objective is to implement a set of dissemination and exploitation activities and to facilitate the successful exploitation of the results.
This first release of the dissemination plan describes the overall activities that the ARMOR Consortium has carried out or plans to carry out in order to disseminate the project results among as large a number of people as possible. At the current stage, partners have already made some progress in order to disseminate the project. The development of the project website and logo were among the first main dissemation activities. Partners have already presented the project in national and international conferences such as the conference on Systems Neuroscience and Rehabilitation (SNR2012) held in Japan and plan to participate in coming conferences in the ARMOR research field such as the 10th European Congress of Epileptology in London. Apart from this flyers have already been created and distributed among researchers, whose expertise is related to the ARMOR research field. The document is mainly divided into two parts. In the first part all the means that are and will be used in order to disseminate the results obtained throughout ARMOR project are presented. In the second part all involved partners have developed and presented their own dissemination plan in order to ensure that the project continues to achieve full potential impact on target groups. All the planned or already carried out actions are included in this document.

D7.2: Project Website

D4.5: Offline analysis of dataThis document is a short report on the technology used for the ARMOR's website (D7.2). It is a short description on operating system of server, database, server type and the platform that is used to build the portal. Also it describes the basic specifications and structure of portal. The reader of this report must be familiar with operating systems and web techologies.

 

 

D7.3: Advertising Materials

D4.5: Offline analysis of dataThis deliverable "Advertising materials" is under work package 7 "Exploitation and Dissemination". It lists the advertising materials produced during the first two years of the ARMOR project. These materials help to raise awareness of the project and communicate its achievements so far (at the second of the 3 years of the project). A brief summary of planned marketing activities for the last year of the project is also given.

 

 

D7.4: Exploitation Plan

D4.5: Offline analysis of dataThe aim of this document is to prepare the Exploitation Plan including the preliminary Exploitation Claims and Strategies for each ARMOR partner and for the whole Consortium. The document "ARMOR_Initial market survey", produced by Mark Richardson (KCL) on 12th June 2013 is summarized in Chapter 2; while the "ARMOR_innovations_FINAL", describing the most relevant achievements of the project (contributions by the corresponding ARMOR partners), is reported in Chapter 3. Both documents underline the added value of the ARMOR results with respect to current State of the Art solutions and they are the basis to build the preliminary list of exploitable results. In Chapter 4 an overview of the scenario described in DoW regarding the possible exploitation strategies for each ARMOR partner is reported, as a starting point and reference for the definition of the ARMOR exploitation strategies. Moreover the draft version of individual exploitation plans for the industrial partners AAISCS, S&C and ICOM is also included in Chapter 5, summarizing the contents of the document "Exploitation of ARMOR technology". The preliminary list of possible Exploitable Results is reported in Chapter 6. The list of the Exploitable Results and the related information will be regularly updated and optimized till the end of the project, discussed during the WP7 and Exploitation meetings and reported in the final progress report. An overview of the whole ARMOR Exploitation opportunities is finally reported in Chapter 8, including a list of potential investors, the ARMOR Business model and possible funding schemes for "ARMOR 2" proposal submission.