抄録
When we try to solve a task, brain cortical areas that have different but necessary functions to the task are selectively activated depending on the task. Their activation automatically produces a system function that is tuned to the task realization. The selection of those areas is dependent on the task and changes when the task changes. We should think that the brain searches for a new combination of necessary areas every time it encounters a new task. So, in our model, we assume a set of functional parts that correspond to the cortical areas with different functions, and a control system that searches for a combination of those areas. The function of control system is alike to that of attention working on the internal behavior. We call the idea of system function formation as the task dependent functional parts combination model (TDFPC). The specific feature of TDFPC model is an incremental nature of resolvable task types by an acquisition of new functional circuit for the available functional parts set. The possible range of realizable task function increases depending on the type and the number of available functional parts. In this study, we present a computer simulation result of TDFPC application on a set of navigation tasks, and evaluate a possibility of the model as the basic principle of brain intelligence. [Jpn J Physiol 54 Suppl:S14 (2004)]
