Information and Self-Organization of Behavior

The goal of Guided Self-Organization (GSO) is to leverage the strengths of selforganization while still being able to direct the outcome of the self-organizing process. GSO typically has the following features: (i) an increase in organization (structure and/or functionality) over some time; (ii) the local interactions are not explicitly guided by any external agent; and (iii) task-independent objectives are combined with task-dependent constraints. Over the last few years a mathematical framework has started to form around these features, promising to provide common organizational and guidance principles across multiple scales and contexts. This process is far from being complete, and every year an International GSO Workshop showcases new breakthroughs that diversify and reshape the field. Nevertheless, some themes and ideas withstand the test of time, maintaining the core of the GSO research. One of these themes is the role of information (understood as Shannon information, i.e. “reduction in uncertainty”) in guiding a self-organizing process. In particular, a lot of progress has been achieved in studying various aspects of information structure and information processing during self-organization of behavior (molecular, neural, cognitive, social, etc.). For example, several principles based on information flows through the perception-action loops of embodied cognitive systems were recently developed (Ay et al., 2012). These principles related GSO to the notion that adaptive behaviors emerge from interactions between brain, body, and environment while optimizing task-independent objective functions. Having a language that describes interactions is essential for a non-reductionist science (Gershenson and Heylighen, 2005). And so another common GSO trend is the use of graph theory in representing and analyzing interactions within a system, be it a cell, the brain, a social network, an ecological web, or a power grid. Several graph-theoretical measures have been devised and put to use in tracing various self-organization processes developing within networks, as well as in relating connectivity of the self-organizing systems to their function (Rubinov and Sporns, 2010; Gershenson, 2012). This topical issue presents a selection of papers following two GSO Workshops (Bloomington, Indiana, USA, in 2010 and Hertfordshire, UK, in 2011). These papers are grouped into three sections. The first section contains three studies characterizing neural dynamics with model-free techniques. It is followed