Multidisciplinary neuroinformatics problems for execution in distributed computing infrastructures

Neuroinformatics lies at the intersection of computer science and neuroscience, making it possible to use methods and tools from one domain for accumulating, processing, analyzing, and managing data and modeling techniques from another. Nowadays, neuroinformatics is evolving very fast, this leads to a rapid expansion of the range of scientific problems that need to be solved. This article deals with a number of urgent problems in the area of cognitive functions modeling of the neurophysiology domain. Problems are analyzed from the point of view of neuroinformatics. Common pitfalls, methods, processing tools, and implementation issues are examined. In total, four problem statements are discussed with data sets residing in the resting state and task functional magnetic resonance imaging as well as in electroencephalograms. The methods vary from simple linear models to highly sophisticated deep neural networks. Justifications for using distributed computing infrastructures are discussed for each problem, including high dimensionality in data that requires, on the one hand, distributed implementation and, on the other hand, using computationally extensive methods that require low-level GPU-based parallelization.