Deep Learning in Neuroscience

Course Description

Recent advances in experimental techniques allow more detailed measurements of biological systems than ever before particularly in neuroscience. These methods result in vast amounts of complex, multimodal, and dynamic data that creates a strong demand for nonlinear, large-scale analysis tools such as machine learning, particularly through the application of deep learning methods. In the last decade, we have seen a surge of deep learning applications in analyzing large-scale neuroscience data, and this trend will only continue to grow.

In this course, we will explore some of the latest applications of deep learning methods in modeling the complex but exciting data from the brain, and how such a model could play a significant role in driving the future of neuroscience research. Students will gain hands-on experience working with real brain recordings and will learn to develop a deep neural network (DNN) model of the recorded responses. All students with prior experiences in DNNs who are interested in but have no prior exposure to neuroscience data analysis are welcome to join the course.

Course tools

  • Python
  • NumPy
  • PyTorch
  • Pandas
  • Jupyter Notebook

Prerequisites

  • Linear Algebra
  • Calculus
  • Basic familiarity with neural networks (MLP, CNN)
  • Experience in Python programming and development/training of NN in PyTorch

Level of complexity of course

Intermediate

Lecturer

Dr. Edgar Walker
Computational/theoretical neuroscientist and a machine learning researcher at the University of Tübingen.

Edgar is fascinated by how our brain can seemingly effortlessly interact with the complex sensory world and give rise to sophisticated decision-making and behavior. His research interest lies in understanding such sensory-driven behavior from the perspective of probabilistic computation, with a specific goal of capturing the brain’s generative model of the world. Edgar develops and applies machine learning, particularly deep learning-based techniques to analyze large-scale multi-modal neuronal data in combination with rich behavior and to teach apart competing models of probabilistic computations in the brain.

Fields of interests: Computational neuroscience, Bayesian inference, Deep Learning

Contacts:[email protected]