The Scientific Case for Brain Simulations

Abstract

In BriefA key element of several large-scale brain research projects such as the EU Human Brain Project is simulation of large networks of neurons. Here it is argued why such simulations are indispensable for bridging the neuron and system levels in the brain. AbstractA key element of the European Union's Human Brain Project (HBP) and other large-scale brain research projects is simulation of largescale model networks of neurons. Here we argue why such simulations will likely be indispensable for bridging the scales between the neuron and system levels in the brain, and a set of brain simulators based on neuron models at different levels of biological detail should thus be developed. To allow for systematic refinement of candidate network models by comparison with experiments, the simulations should be multimodal in the sense that they should not only predict action potentials, but also electric, magnetic, and optical signals measured at the population and system levels.

A comprehensive data-driven model of cat primary visual cortex

“…Sharing the model code that is defined in the highly modular Mozaik environment makes it particularly straightforward for other researchers interested in analyzing the given model or building upon the work presented here, thus supporting the long-term collaboration necessary for the integrative program. At the same time, although neuroscience should ultimately converge upon a single model, there are many paths through the space of partial models to arrive there, and for the health of the field it is good to have several "competing" models, each of which will make different approximations (Einevoll et al 2019); nevertheless, all models should pass the same validation tests. We therefore also make available our library of integration tests via a dedicated data store (http://v1model.arkheia.org) implemented in the recent Arkheia framework (Antolík and Davison 2018), developed to facilitate sharing model and virtual experiment specifications.…”

“…computations are multiplexed such that the same synapses participate in multiple simultaneous calculations. While further reductionist, hypothesis-led research is indisputably required, many of these questions require a systematic, integrative approach to progressively build a unified multi-scale theory of brain function (Sejnowski et al 1988;Einevoll et al 2019). A danger in detailed modeling is growth of free parameters, risking over-fitting and consequent lack of explanatory power.…”

Quantifying and modelling non-local information processing of associative brain regions

“…Large-scale models of the brain include the laminar cortex architecture 37 and detailed neuroanatomic models 38 . We focus on holistic function and properties of consciousness emerging from non-local information processing where holographic models are known to provide a strong foundation 1,2,5 .…”

Dense Computer Replica of Cortical Microcircuits Unravels Cellular Underpinnings of Auditory Surprise Response

“…With the increase in computer power we are able, for the first time, to connect via a dense computer replica of neuronal circuits, the subcellular (ion channels and synapses), the cellular (cell types) and the network (connectivity) levels, and obtain a new and lucid picture of the working of the brain at multitude levels of operation 44 . This will enable to resolve other long-standing open questions such as the emergence of "salt and pepper" organization in sensory systems of rodents [45][46][47][48] and the impact of different excitatory and/or inhibitory cell types on network dynamics 5,[49][50][51] .…”