Skip to Content

Fri, Nov 18

Speaker: Pavan Ramkumar, Ph.D.

Title: How you see is why I get (HYSIWIG): Exercises in decoding visual information content from MEG signals

Multivariate pattern analyses of functional brain imaging data have proved successful in expanding our knowledge about the information resolution of said imaging methods. To what extent can this newfound knowledge about information resolution be translated into new knowledge about spatiotemporal organization of brain function? To discuss the above open-ended question, in this talk, I will present data from two ongoing (unpublished) MEG studies in low-level visual feature representation, and natural scene perception respectively. I will additionally offer theoretical speculations on how the framework of applying encoding and decoding models to brain imaging data in combination with psychophysical tags and computational representations of the environment can help us describe David Marr's famous three levels of a system: the computational, the algorithmic, and the mechanistic.

Host: Dr. Kording