Currently, there are no drugs which can be used for preventing the development of epilepsy in patients exposed to risk factors such as brain injuries or infections of the central nervous system.
In order to foster the development of new treatments, it is important to identify indicators (biomarkers) of the time window during which uncovered mechanisms lead the brain to develop epilepsy – a period called epileptogenesis. These biomarkers refer to one or more measurable events that: 1) are early factors predictive of the emergence of epilepsy and 2) vary depending on the severity of the illness. These biomarkers could be exploited for the development of disease-modifying therapies in individuals exposed to risk factors. The success of such therapies would reduce the incidence of epilepsy or relieve its burden.
The main goal of our investigation is to identify a marker of epileptogenesis in the brain electrical activity, as measured by the Electroncephalogram (EEG) – an assessment tool used for epilepsy investigations. We have recently shown that the brain electrical activity following the exposure to a risk factor is characterised by the emergence of specific patterns of behavior. This aspect may be exploited to identify one or more markers of epileptogenesis and we are currently investigating such possibilities.
To this aim, EEG tracings are analysed by the Recurrence Quantification Analysis (RQA) tool – a mathematical tool suitable for detecting and quantifying specific behaviors of brain electrical activity embedded in noisy time-series, such as the EEG. Several steps are involved in the implementation of the RQA, some of which are computational intensive.
An important contribution to support this activity is the recent integration of the Elastic Cloud Computing Cluster (EC3), operated by the Polytechnic University of Valencia (UPV), with the HNSciCloud vouchers to access commercial cloud providers provided by Exoscale. The voucher scheme was proposed by Exoscale, the IaaS computing resources for the RHEA Cloud platform, as a result of the collaboration with the HNSciCloud Pre-Commercial Procurement (PCP) project. Thanks to this tighter integration, users of the Mario Negri Institute in Milan were able to allocate the resources needed to configure the cluster to analyse the EEG with the RQA mathematical tool. Fully-dedicated high-performing CPUs strongly contribute to reduce the calculation times in implementing the RQA of EEG tracings, so that long-running jobs that normally take months to complete can be accomplished in a matter of days. This allows researchers to get results much quicker and to increase the amount of information gained by the analyses.
Thanks to EGI, the voucher scheme offered by the HNSciCloud PCP project and the EC3 framework, we already planned to enrich our investigation by using further analytical tools. In the future, we hope to significantly improve our understanding of mechanisms leading to the development of epilepsy and to discover new approaches of prevention.
HNSciCloud has made a number of Exoscale vouchers available to EGI to support the work of researchers.
Massimo Rizzi is a researcher of the Department of Neuroscience at the Mario Negri Institute for Pharmacological Research, Milan, Italy.
As a reference example, the deployment of a small EC3 cluster made of 2 front-end CPUs plus 2 working-nodes CPUs, 4 GB RAM, allowed to reduce the time of analysis of 1000 EEG epochs (equivalent to 24h of EEG tracings) from 2 weeks (time required with an ordinary desktop PC) to 26 hours.