This title appears in the Scientific Report : 2013 

New Perspectives on Workflows and Data Management for the Analysis of Electrophysiological Data
Grün, Sonja (Corresponding author)
Denker, Michael / Davison, Andrew / Wachtler, Thomas
Computational and Systems Neuroscience; INM-6
Theoretical Neuroscience; IAS-6
2013
INCF Workshop, Jülich (Germany), 2013-12-05 - 2013-12-06
Conference / Event
Supercomputing and Modelling for the Human Brain
Brain-inspired multiscale computation in neuromorphic hybrid systems
Signalling Pathways and Mechanisms in the Nervous System
Breakthroughs in recording technologies and analysis approaches have lead to unprecedented levels of complexity in electrophysiological experiments. Most notably, the ability to perform massively parallel recordings from hundreds of neurons in the brain, paired with a strong interest in complex, natural stimulation and behavioral paradigms, leads to a surge of intricately interwoven incoming data that should be analyzed with sophisticated methods to uncover the network dynamics by exploiting the parallel aspect of the data. The combination of these factors leads to highly complex projects that pose a challenge for researchers who, over the course of years, are confronted with planning of the analysis, organization of workflows in larger teams, programming of software, and bookkeeping of the results obtained by constantly evolving analysis methods. The complexity has reached a level where a professionalization of the data a nalysis workflow, both conceptually and in terms of supporting software infrastructures, has become a necessity. This workshop intends to provide a discussion platform for researchers and developers who work on components that are instrumental in such workflows. In particular, we focus on three major issues that need to be solved in the future: (i) the development of data structures and software libraries that enable interfacing of data from various sources and integration of methods for data manipulation and analysis; (ii) documentation and provenance tracking solutions to support reproducible analysis processes able to cope with the required levels of flexibility and data size; and (iii) workflow management systems that allow automatic and extensible processing of complex analysis workflows, in particular on high performance computers. While some of these aspects have already received attention during the last years, some have only been addressed in other fields where a less severe degree of heterogeneity of recording devices, simulators, data types, codes, and paradigms permitted the development of uniform frameworks. Therefore, we intend to bring together not only those researchers working on solving these issues for electrophysiology, but also from other areas of specialization in order to assess in how far their tools or concepts can be utilized. Following presentations that highlight the individual contributions of participants, ample opportunity for discussion is planned in order to identify possible links between projects, and potentially suggest collaborations between researchers. Scientists working with electrophysiological data contribute their experience from everyday data handling and analysis by providing direct requests, constraints and feedback on the practicability of suggested approaches. The outcomes of the discussions will lead to recommendations to the INCF on how to best support a rapid and coherent development of workflows for electrophysiology, and how to advertise them to the community.