Document Actions

You are here: Home / The Network / Network Partners / D - USA Collaborations / Magdeburg - Hanover / Princeton

Magdeburg - Hanover / Princeton

German - US-American Collaboration in Computational Neuroscience

"Development of  general high-dimensional models of neuronal representation spaces"

Multivariate analysis methods nowadays allow to decode information from brain activation patterns. Yet, little is known about how neural codes differ between individuals. Previous methods have the disadvantage that decoding models must be created separately for each individual brain, since the functional fine structure of two brains can not be aligned with sufficient accuracy. In this project, methods are developed with which it is possible to discover and to describe similarities in neural information representation by projecting individual brain activity patterns into a common high-dimensional space in order to create models for the representation spaces of various brain areas that are valid for a wide variety of environmental stimuli and individuals. This also includes complex cortical networks that do not respond consistently to external stimulation (e.g. for social cognition).

The algorithms "hyperalignment" and "hyperalignment of functional connectivity" are developed, which are optimized for the application to functional brain activation profiles, patterns or connectivity between brain areas, respectively. For validation, functional magnetic resonance data are collected during complex visual and auditory stimulation to analyze the underlying representation space.


The following scientists take part in this collaboration:

  • Michael Hanke, Otto von Guericke University Magdeburg
  • James V. Haxby, Dept. Psychological and Brain Sciences, Dartmouth College, Hanover, USA
  • Peter J. Ramadge, Dept. Electrical Engineering, Princeton University, Princeton, USA


German Coordinator:

Dr. Michael Hanke
Otto von Guericke University Magdeburg      
Institute for Psychology II
PF 4120
39106 Magdeburg

Phone: +49 (0) 391 / 67 18481
Fax: +49 (0) 391 / 67 11401