TITLE:

SPATIOTEMPORAL NETWORK DYNAMICS FROM HUMAN BRAINWAVE DATA

DATE:

Friday, October 30th, 2009

TIME:

3:30 PM

LOCATION:

GMCS 214

SPEAKER:

Richard E. Greenblatt, Ph.D,
Source Signal Imaging, Inc.

ABSTRACT:

Networks play an increasingly significant explanatory role in the social, physical, and biological sciences. In the brain sciences, hypothesized transient intentional networks, spanning a wide range of length and time scales, may underlie the perception/action cycle characteristic of human behavior. Pathological networks may underlie brain diseases such as epilepsy and schizophrenia.

We have been developing statistical signal processing tools, embodied in computer software, to identify and characterize brain networks from multichannel human electrophysiological time series data. The network identification methods are based on the estimation of bivariate instantaneous phase synchrony measures, combined with a deterministic clustering algorithm. The result of the analysis is a set of networks, each consisting of several nodes at known brain locations, along with a synchrony time series (the composite synchrony profile) associated with each of the networks. Within- and between-network interactions may then be characterized in the space-time and space-time-frequency domains.

These methods have been applied to intracranially recorded electrocorticographic (ECoG) data, obtained from patients undergoing presurgical evaluation for pharmacoresistant epilepsy. In this talk, I will describe some of these algorithms, and illustrate their application to the analysis of ECoG data, including seizure data, and data obtained during a word recognition task.

HOST:

Jose Castillo

DOWNLOAD:

PDF File