Graduation / internship project: “Spike sorting” for ultra-high density silicon probe array data
Research interest of your company / institute:
The group “Neuronal Networks of Memory” studies the dynamics of brain areas that are involved in memory processes, and their interactions during active and inactive behaviours (sleep) leading to memory encoding, consolidation and retrieval.
We develop and use the most advanced experimental tools to monitor brain activity in rodents, and apply a vast array of data analytical and theoretical techniques to make sense of these complex data.
Advances in our understanding of brain processes depend critically on gathering large samples of neural activity. For this, multi-site electrophysiological recording is one of the most important tools, that has led to important discoveries, such as place and grid cells. Microfabrication techniques allow us to massively increase the density of our recordings, with the promise of expanding the yield of our recordings manifold.
We are involved in the EU NeuroSeeker project that is building some of the high density (ranging from hundreds to thousands of recording sites). How to sort signals from these very large arrays into contributions from single cells is however still an open problem. We will make use of existing algorithms (clustering, template matching based) for spike sorting, refine them, and adapt them to very large data sets (data rates up to ~400 MB/second), also making use of some “ground truth” available data sets.
Key skills student:
Proficiency in programming: python or Matlab (required), C/C++, GPU programming, knowledge of high-performance computing environments a plus. Knowledge of advanced statistical/machine learning tools) Difficulty: Medium/Hard (for a 3-months internship) Hard (for a master thesis project).
Dr. Francesco P. Battaglia
Donders Centre for Neuroscience