Research Seminar
A thermodynamical perspective on the two-choice decisionmaking processStijn VerdonckKU Leuven | |
| Abstract: | In psychological literature, linear diffusion models constitute a popular framework for choice RT modeling. Initially conceived as abstract noisy information accumulators, there is an increasing trend towards a neurological interpretation/explanation of these models. In recent years, the field of biological neural networks has advanced to a point where the key aspects of decisionmaking can be accounted for ab initio. The biologically inspired integrate and fire networks typically used for this kind of research, result in a rather intricate model for two-choice decisionmaking (Wong and Wang, 2006). Apart from being quite cumbersome, this model does not seem to sprout a simple realistic speed versus accuracy trade-off parameter, an important feature in the field of choice RT. In this presentation a new Hopfield (Ising) based neural network model for a two-choice decisionmaking process will be proposed. Its properties will be discussed, as well as its ability to address the problems described above. Its equivalence to a specific nonlinear extension of a two-dimensional Ornstein–Uhlenbeck process will be shown. |
| Date: | Tue Dec 7, 12:15 pm - 1:15 pm |
| Place: | room 03.60 (Department of Psychology, Tiensestraat 102, 3000 Leuven) |