Complex Systems Science potentially offers Cognitive Science a unifying framework that would bring together the study of brain, mind, and behavior. The Theoretical framework of Complex Systems defines a system as a collection of interconnect components that operate together in a way that generates the behavior of the system as a whole. Each system is also a unit, embedded within a larger system, and interacts with other such units in a way that defines its function. There are multiple quantitative techniques that have been developed to study systems as such, including Dynamical Systems Theory, Information Theory, and Network (Graph) Theory.

I am applying this approach to the study of cognitive systems by evolving minimal models of brain-body-environment systems through computer simulation, and then analyzing the resulting models with the previously listed quantitative methods.

Some of my research interests within this framework are:

• How multiple functions can arise from a dynamical neuronal circuit coupled with a body and environment.
• Goal Setting and Switching: how such a dynamical brain-body-environment system can set goals, pursue them, and switch to different goals
• How multiple informational inputs get integrated in a dynamical circuit.
• How agents in an environment can improve their performance through a process of learning.

For my Masters Thesis, I created a computational model of a brain region in mormyrid electric fish called the Electrosensory Lateral Line Lobe (ELL). This brain regions functions as a primary processing area for the sensory modality of electroreception in these fish.  It acts as an adaptive filter that learns to predict reoccurring stimuli and removes them from its sensory stream, passing only novel inputs to other brain regions for further processing. 

Individual activation tendencies of the different known cell types within the ELL were modeled with dynamical systems equations, and were interconnected according to the connectivity pattern of real ELLs.  Several of the input patterns recorded in mormyrid ELLs were modeled and incorporated in the simulation. 

The modeling approach claims that if all of the relevant components of a system are captured and interconnected accurately in a computer program, then when provided with accurate representations of the inputs a simulation should produce functional patterns similar to those of the real system.   Such simulated patterns were generated by the ELL model, and were compared to recordings from real mormyrid ELLs. This comparison validated the model’s integrity due to its success in reproducing some of the adaptive filter function seen in real ELLs. 

Download (PDF)