Angus Silver

Angus Silver is a Wellcome Trust Principal Research Fellow, an ERC advanced investigator and Professor of Neuroscience at UCL. His research has contributed to our understanding of the fundamental properties of both chemical and electrical synaptic signalling. He has also discovered how neurons perform distinct mathematical operations and has extended our understanding of the relationship between network structure and function. Several of these discoveries were made possible through his development of new methods. These include statistical methods for quantifying the properties of central synapses, tools for modelling neurons and networks and the development of a new type of high speed 3D two-photon microscope. His Lab is highly multidisciplinary with research that spans the fields of neurophysiology, microscope development and neuroinformatics.

University College London, United Kingdom

Talk title
Open source tools and resources for collaborative modeling of brain function

Talk abstract
Computational models are key for understanding how the biophysical mechanisms and synaptic connectivity within neural circuits generate sensory representations and control behaviour. Given their potential power for understanding brain function it is therefore surprising that models are not more widely used across the neuroscience community. A key factor that has impeded adoption of computational modelling is skepticism about the scientific value of model-based results. For a model to be useful as a scientific tool it is essential that the results it produces are robust and reproducible and that it can be critically evaluated by a wide range of scientists. However, this is difficult as models are implemented using a diverse set of modelling approaches, simulation tools, and computer languages. I will discuss how standardized model descriptions (e.g. NeuroML2) can be used to build modular well-structured self-consistent models. Furthermore, I will show how the internal structure and functional properties of NeuroML2-based models can be transformed using into web-browser-based representations that are familiar to a wide range of neuroscientists. Moreover, OpenSourceBrain-based simulations and analysis also enable the functional properties of the models to be explored through the browser. I will argue that standardization and open source web-based solutions that enable a collaborative approach and open, critical evaluation of models by the wider neuroscience community are essential for improving the scientific utility of models of brain function.