Discovering Informational Biomarkers for Cognitive Tasks: Recent Examples and Lessons from Neurocomputation

Lecture / Panel
For NYU Community

Speaker: Larry M. Manevitz, University of Haifa

Machine Learning tools and their multivariate tools have made dramatic successes in recent years. In this talk, we will discuss, as time allows, some recent works involving various biological signals (such as fMRI) and their relationship to various cognitive functions. Amongst the areas are:  

  1. Helping to establish the existence of a secondary declarative memory system in human adults
  2. Early potential discover of Parkinson's disease directly from the voice signal
  3. The ability to tell from a brain scan if a free recall memory is one with positive or negative valence
  4. Ability to tell from a scan if one is reading text with "deep" grammatical elements or not.

While the talk will be at the general colloquium level (i.e. no previous technical exposure to machine learning is needed) we will nonetheless try to explain some of the difficulties involved.

Bio: For the last fifteen years, Larry Manevitz has been the Director of the Neurocomputation Laboratory at the University of Haifa, located in Israel.   This year he has been on a sabbatical visit to the U. of Otago in Dunedin, New Zealand.  Currently, his laboratory (  focuses mostly on machine learning of cognitive mental functions and some brain modeling. His Ph.D. was from Yale University, Department of Mathematics with a specialty in mathematical logic (under Prof. Abraham Robinson of "non-standard analysis" fame.)

He has worked in applied (to other areas of mathematics) model theory, foundations of combining uncertainty (a sub-field of Artificial Intelligence) and theoretical and applied areas of neurocomputation and their relationship to cognition.