Richard Aslin, Senior Scientist at Haskins Laboratories in New Haven, CT will be the keynote speaker for the second day of the CT Institute for the Brain and Cognitive Sciences (IBACS) Meet & Speak event. Dr. Aslin will be speaking on Wednesday, May 15th, at 5pm in Oak Hall 101. This event will be open to the public:
Title: "Mechanisms of prediction in the brain and prediction from the brain"
Abstract: Learning can be accomplished by different neural architectures that use associative mechanisms and/or feedback pathways. I will review one such feedback architecture that is particularly well suited for efficient prediction based on the statistics of the early sensory environment. This feedback architecture is present in human infants by 6 months of age, as assessed by a diffuse optical imaging technique called near-infrared spectroscopy (fNIRS). Premature infants, tested at 6 months corrected age, do not show evidence of this predictive pathway, despite having intact associative learning at the behavioral level. The utility of fNIRS to predict which of several stimuli are present on a trial-by-trial basis using Multivariate Pattern Analysis (MVPA) will also be summarized, for both 6-month-olds and adults. Properties of the brain's network architecture can also be used to predict behavioral outcomes. I will review the Connectome-based Predictive Modeling (CPM) approach that characterizes patterns of functional-connectivity (FC) in resting-state (and movie-viewing) fMRI from children and adults. These FC-patterns have been successful in predicting various phenotypes (e.g., high/low sustained attention, IQ, reading scores). An on-going study that uses CPM to predict monolingual from bilingual individuals will be summarized, with future extensions from fMRI to fNIRS.
For more information, contact: Crystal Mastrangelo at crystal.mastrangelo@uconn.edu