Frank W. Grasso
Position: Associate Professor
Campus Affiliation: Brooklyn College
Degrees/Diplomas: Ph. D., University of Massachusetts
Training Area: Animal Behavior and Comparative Psychology|Behavioral and Cognitive Neuroscience
Research Interests: Neural Systems for control and coordination of behavior: BioMimetic Robotics, Quantitative Animal Behavior, Biological and Artificial Neural networks architectures.
The BioMimetic and Cognitive Robotics Laboratories study the neural control and coordination of behavior using combined empirical and theoretical approaches. The core concept in our studies is to use constrained models of behavior implemented in robots or in simulations to evaluate hypotheses of nervous system function. We focus on the behavior of whole organisms. We make quantitative measures of the behavior of live animals and use those to infer models of the underlying cognitive information processing mechanisms. We then implement these models into biomimetically scaled robotst that we test under the same conditions that we studied the animals. The animal behavior becomes the yardstick for evaluating the validity of any given theory. Thus, under one roof we combine computational neuroscience, animal behavior and robotics.
We focus these methods on cognitive information processing in the "simple" nervous systems of invertebrates. Spatial memory, navigation and orientation using olfactory, visual and tactile sensory guidance information are areas of focus. The relative roles of instinctive ( or "hard-wired" ) versus plastic ( or "learned" ) processes in the self organization of developing behavioral systems is central to our approach. Species studied and modeled include cephalopods (octopuses and cuttlefishes) and crustaceans ( lobsters, crabs and crayfishes).
Use of invertebrates and our BioMimetic approach contribute to understanding the evolution of behavior by making clear the neuro-computational innovations in species on evolutionary paths not followed by vertebrates. The cross-species differences and similarities of the solutions provide insight into the design of nervous systems as computational devices.