Structural Limits of Information Processing: ¿What is possible and not possible to learn?

JUAN E KAMIENKOWSKI; STANISLAS DEHAENE; HAROLD PASHLER; MARIANO SIGMAN

Buzios, Brazil

Congreso; Ier CONGRESO IBRO/LARC DE NEUROCIENCIAS DA AMERICA LATINA, CARIBE E PENÍNSULA IBÉRICA; 2008

Objectives: When two tasks are presented within a short interval, a delay in the execution of the second task has been systematically observed. Psychological theorizing has argued that while sensory and motor operations can proceed in parallel, the coordination between these modules establishes a processing bottleneck. An intriguing open question is which aspects of this interference are subject to changes with learning and experience and to subject to subject variability. Here we set up an experiment to explore these issues. Methods: In this experiment we asked subjects to perform multiple sessions. In each session participants performed two simple decision tasks: a Tone Discrimination and a Number Comparison Task. Participants performed these tasks either in isolation or superposed in time. This allowed us to address specific aspects of learning related to single task performance or to the coordination of both tasks. Results First, we found that most independent variables were subject to changes with learning: RTs to both tasks, as well as other aspects which involve the coordination of both tasks improved with training. In striking contrast, the main interference effect remained completely unchanged. Subject by subject analysis reveals remarkable differences which were very stable across different sessions. We observed a clustered pattern of dependencies with very tight correlation in the different observed variables. A group of subjects was extremely fast in single task performance, showed consistent learning in the dual-task performance, moderate effects of interference and very reliable across session performance. Another group showed a consistently distinct behavior in all these variables. Conclusions Our work has shed light on which robust observations in psychology may be variable across individuals or, within each individual, to changes related to practice. While most variables are subject to change with practice, the dual-task interference, which implies a sequential organization of certain cognitive operations seems to be unchanged by learning. This argues in favor of a purely structural limit probably related to architectonic properties of the human brain. In addition the observed pattern with very strong correlations in different presumably independent measurements argues in favor of common processes which underlie different aspects of performance in a broad range of cognitive tasks.