Evolving Structure-Function Mappings in Cognitive Neuroscience Using Genetic Programming

Fernand Gobet, Amanda Parker
2005 Swiss Journal of Psychology  
A challenging goal of psychology and neuroscience is to map cognitive functions onto neuroanatomical structures. This paper shows how computational methods based upon evolutionary algorithms can facilitate the search for satisfactory mappings by efficiently combining constraints from neuroanatomy and physiology (the structures) with constraints from behavioural experiments (the functions). This methodology involves creation of a database coding for known neuroanatomical and physiological
more » ... ints, for mental programs made of primitive cognitive functions, and for typical experiments with their behavioural results. The evolutionary algorithms evolve theories mapping structures to functions in order to optimize the fit with the actual data. These theories lead to new, empirically testable predictions. The role of the prefrontal cortex in humans is discussed as an example. This methodology can be applied to the study of structures or functions alone, and can also be used to study other complex systems. Keywords Complex systems, evolutionary computation, prefrontal cortex, scientific discovery, structure-function mapping, theory formation 18/5/07 3 Evolving structure-function mappings in cognitive neuroscience using genetic programming A primary aim in science is to develop theories that summarize and unify a large body of experimental data. However, there is no overarching theory in psychology (or even in subfields of psychology, such as the study of memory, emotions or perception) which imposes order on the mass of data and makes it possible to derive quantitative predictions, in the way, for example, quantum mechanics can be used to organize empirical data in chemistry. To compound the difficulty, there are currently around 1,500 journals devoted to scientific psychology. A substantial proportion of these journals publish mainly experimental results. Psychology is not exceptional: in sciences from astrophysics to meteorology to biology, technological progress has enabled the rapid collection of huge amounts of data. How can scientists keep track of this exponentially increasing amount of information, in spite of their bounded rationality? While progress in database management of scientific results is notable, there remains the question of how this new information can foster scientific understanding, as opposed to simple accumulation of knowledge. A particularly interesting approach is to develop theories implemented as computer programs which account for, and therefore summarize, empirical data. How scientific theories are developed has been the focus of a number of studies in psychology, philosophy, history, and, more recently, artificial
doi:10.1024/1421-0185.64.4.231 fatcat:2ovdavn3tfbe5k7okxhcfgi2vy