Il Memming Park

Research Interests/Expertise

Many have proposed theories of how the brain works, but few of them have been rigorously tested due to difficulties in obtaining sufficiently rich neural datasets. However, recent advances in recording technology allow simultaneous observation of a large population of neurons, providing the opportunity to use statistical and machine learning tools to bridge the gap between theory and data in neuroscience. These neural signals that are noisy and high-dimensional pose a significant challenge for analysis, because (1) neural signals are non-linear & non-gaussian, and (2) the number of parameters in models usually grows with the number of neurons, causing difficulties in estimation.

I am interested in how information is encoded as spatiotemporal patterns of neural activity, and how information is processed to perform specific computations within and across brain areas. For instance, I analyze how visual motion is represented by visual cortices, and how it is subsequently integrated over time to form decisions that are represented in higher order cortices. I collaborate with experimental neurophysiologists who record neural activities, and I model their data to find structures hidden in noisy observations; structures within the signals themselves, and to relate neural activity with external stimulus and behavior.

 

You may visit my blog for more info @ http://memming.wordpress.com