Dynamic Field Theory: Conceptual Foundations and Applications in Cognitive Sciences

 

Gregor Schoener (gregor.schoener@rub.de)

Institut fur Neuroinformatik, Ruhr-Universitat Bochum, Germany

 

Objectives and scope

 

Dynamical Systems thinking has been influential in the way psychologists, cognitive scientists, and neuroscientists think about sensori-motor behavior and its development. The initial emphasis on motor behavior was expanded when the concept of dynamic activation fields provided access to embodied cognition. Dynamical Field Theory (DFT) offers a framework for thinking about representation-in-the-moment that is firmly grounded in both Dynamical Systems thinking and neurophysiology. Dynamic Fields are formalizations of how neural populations represent the continuous dimensions that characterize perceptual features, movements, and cognitive decisions. Neural fields evolve dynamically under the influence of inputs as well as strong neuronal interaction, generating elementary forms of cognition through dynamical instabilities. The concepts of DFT establish links between brain and behavior, helping to define experimental paradigms in which behavioral signatures of specific neural mechanisms can be observed. These paradigms can be modeled with Dynamic Fields, deriving testable predictions and providing quantitative accounts of behavior.

One obstacle for researchers wishing to use DFT has been that the mathematical and technical skills required to make these concepts operational are not part of the standard repertoire of cognitive scientists. The goal of this tutorial is, therefore, to provide the training and tools to overcome this obstacle.

I will provide a systematic introduction to the central concepts of DFT and their grounding in both Dynamical Systems concepts and neurophysiology. I will discuss the concrete mathematical implementation of these concepts in Dynamic Neural Field models, giving all needed background and providing participants with some hands-on experience using interactive simulators in MATLAB. I will review robotic implementations to make the ideas concrete. Finally, I will take participants through a number of selected, exemplary case studies in which the concepts and associated models have been used to ask questions about elementary forms of embodied cognition and their development.

The interactive simulators will be available at the tutorial.

 

 

Target audience

 

No specific prior knowledge of the mathematics of dynamical systems models or neural networks is required as the mathematical and conceptual foundations will be provided during the tutorial. An interest in formal approaches to cognition is an advantage.

 

Material covered in the course

 

  1. Conceptual foundations of Dynamical Systems Thinking and Dynamical Field Theory (DFT): Embodied and situated cognition; Stability as a necessary property of embodied cognitive processes; Distributions of population repre-sentation as the basis of spatially and temporally continuous neural representations;

 

  1. Dynamical Systems and Dynamic Field Theory Tutorial: Concept of dynamical system; Attractors and stability; Input tracking; Detection, selection, and memory instabilities in discrete neuronal dynamics; Dynamical Fields and the basic instabilities: detection, selection, memory, boost-driven detection; Learning dynamics; Categorial vs. graded mode of operation; Practical implementation of DFT in simulators; Interactive simulation; Illustration of the ideas through robotic implementations;

 

  1. Case study using DFT to understand embodied cognition and its development: visual and spatial working memory in children and adults; spatial precision hypothesis as a developmental mechanism in spatial recall, position discrimination, and change detection.

 

  1. Case study using DFT to understand how flexible action sequences can be generated: Dynamics of serial order and behavior organization; Coupling to real sensor and motor systems; Stability and flexible timing of actions in a sequence; Autonomy and executive control in neural and robotic systems.

 

  1. Case study using DFT toward an account for higher cognition: objet recognition, pose estimation, scene representation; grounding spatial language.

 

Schedule

 

  1. Conceptual foundations of Dynamical Systems Thinking and Dynamical Field Theory (DFT)

 

  1. Dynamical Systems Tutorial

 

  1. Case studies using DFT to understand embodied cognition and its development

 

  1. Case studies using DFT to understand flexible action sequences

 

  1. Case studies using DFT to understand higher cognition

 

 

Suggested Readings

 

(available online, see below)

 

  1. General review: Schoener, G.: Dynamical Systems Approaches to Cognition. In: The Cambridge Handbook of Computational Psychology, Ron Sun, (ed.), Cambridge University Press (2008), pages 101-126

 

  1. A shorter review: Schoener, G.: Dynamic Field Theory of Embodied Cognition. Encyclopedia of Computational Neuroscience, 2014

 

  1. For developmental aspects: Schoener, G.: Dynamical Systems Thinking — From metaphor to theory. In: Handbook of Developmental Systems Theory and Methodology. Peter C M Molenaar, Richard M Lerner, Karl M Newell (eds.). Guilford Publications, 2014, pp. 188-219

 

  1. For robotic use of DFT: Zibner, S. K. U., Faubel, C., Iossifidis, I., Schoener, G. : Dynamic Neural Fields as Building Blocks of a Cortex-Inspired Architecture for Robotic Scene Representation IEEE Transactions Autonomous Mental Development 3(1) 74-91 (2011)

 

  1. Toward higher cognition: Sandamirskaya, Y., Zibner, S., Schneegans, S., Schoener, G. (2013): Using Dynamic Field Theory to extend the embodiment stance toward higher cognition. New Ideas in Psychology 31 322339 (2013)

 

 

Lecturer

 

Gregor Schoener holds the Chair for Theory of Cognitive Systems and is the Director of the Institut fu¨r Neuroinformatik, Ruhr-Universita¨t Bochum, Germany. Following his PhD in 1985 in theoretical physics at the University of Stuttgart, he held positions at the Center for Complex Systems of Florida Atlantic University, the Institut fu¨r Neuroinformatik, and the Center for Cognitive Neuroscience of the CNRS in Marseilles, France before returning to Bochum, Germany in 2001 to assume his current position.

Gregor Schner brings bear his background in theoretical physics on interdisciplinary research in cognitive and neuroscience as well as in autonomous robotics and computer vision. Working closely with experimental groups, he develops theoretical concepts and models to understand movement, perception, and embodied cognition on the basis of dynamical systems ideas. Past contributions include predicting signatures of instability in movement coordination, the principle of the uncontrolled manifold in multi-degree of freedom movement, and Dynamic Field Theory accounts for sensory-motor decision making, perseverative reaching in infancy, habituation in infant looking, change detection and other elementary forms of cognition. He is also using such theoretical ideas to contribute to autonomous robotics, developing the attractor dynamics approach to robot motion planning, and neural dynamics approaches to sequence generation and behavioral organization.

 

 

Computer use

 

Participants who bring laptops with Matlab installed (student version is sufficient) will be able to follow demonstrations by actively working with the simulator during the lectures.

Online resources

Publications, lecture material, and interactive simulators can be found at the Neural Dynamics Summer School website http://www.robotics-school.org.

 

 

Files