ABSTRACT
In recent years, neuroscience has become an increasingly computational discipline. This is a natural consequence of the complexity of the systems under study combined with the constant improvement and development of novel experimental techniques, which allow for increasingly detailed observations. Neurophysiologists can now record simultaneous neural activity, at temporal resolutions of tens of kilohertz, from tens to hundreds of intracranial electrodes (Csicsvari et al., 2003). From each electrode, both action potentials of individual neurons (reecting the output of a cortical site, see Chapters 2, 4, and 5) and local eld potentials (LFPs, reecting both population synaptic potentials and other types of slow activity, see Chapter 3) can be extracted. Moreover, electrophysiological recordings can now be accompanied by simultaneous measurements of other brain signals, such as those recorded with optical imaging, electroencephalography (EEG, see Chapter 22), or functional magnetic resonance imaging (fMRI, see Chapter 24). Such developments make managing the collection, storage, preprocessing, and analysis of such data a signicant computational challenge. Further, increasingly detailed large-scale modeling produces sizable quantities of synthetic data that must also be carefully analyzed to give insight into the behavior of the models and to provide meaningful comparisons to experiments.
