Large scale modelling and large scale data Analysis

Level:Intermediate Prerequisites: Basic knowledge in linear algebra, calculus, probability theory, dynamical systems and programming. Description:The first starts by introducing the principles of large-scale neural modelling, and why and when simulations and analytical approaches are needed to model empirical observations of neural activity or behavior, or to implement specific computational goals. A focus will be on models of perceptual decision making, i.e. the question of how neural systems can make decisions based on incomplete or noisy sensory inputs. We will cover goal-driven optimization of neural computations in recurrent networks, as well as statistical and computational approaches for optimizing and constraining simulations of neural networks. The second part of the course covers principles of large-scale neural analysis, in particular Methods and tools for analysis of large-scale neural recordings and simulations, e.g. cross-correlations or clustering approaches. A particular focus is on methods and tools for dimensionality reduction and visualization of large-scale neural dynamics, including state-space models and time-series applications. Finally, it includes practical example applications and visualizations using large-scale data-sets, e.g. obtained using electrophysiological and optical multi-cell recording techniques. Learning Objectives: After successful participation, students are able to describe, understand and apply techniques for simulation and analysis of large- scale neural models. They understand differences between mechanistic and phenomenological models, and are able to select and use computational tools for different tasks in engineering and science. Additionally, students can describe, understand and apply statistical and mathematical tools for analyzing high-dimensional measurements or simulations of neural activity and/or connectivity (in particular multi-neuron neural activity measurements) and to evaluate which tools are appropriate for which measurement. Take home exam: students will have about a week.