The aim of this paper is to study the relevance of simplicity and its formal representation as Kolmogorov or algorithmic complexity in the cognitive sciences. The discussion is based on two premises: 1) all human experience is generated in the brain, 2) the brain has only access to information. Taken together, these two premises lead us to conclude that all the elements of what we call `reality' are derived mental constructs based on information and compression, i.e., algorithmic models derived from the search for simplicity in data. Naturally, these premises apply to humans in real or virtual environments as well as robots or other cognitive systems. Based on this, it is further hypothesized that there is a hierarchy of processing levels where simplicity and compression play a major role. As applications, I illustrate first the relevance of compression and simplicity in fundamental neuroscience with an analysis of the Mismatch Negativity paradigm. Then I discuss the applicability to Presence research, which studies how to produce real-feeling experiences in mediated interaction, and use Bayesian modeling to define in a formal way different aspects of the illusion of Presence. The idea is put forth that given alternative models (interpretations) for a given mediated interaction, a brain will select the simplest one it can construct weighted by prior models. In the final section the universality of these ideas and applications in robotics, machine learning, biology and education is discussed. I emphasize that there is a common conceptual thread based on the idea of simplicity, which suggests a common study approach.
 
Attached: Slides of talk given at 

International Workshop on
Physiological Networks: Theory, Implementation and Application
 
The workshop/review will take place in The Castle of Rauischholzhausen Conference Center, 17-19 June, 2009