Recent studies have emphasized on the idea that attention is a multi-faceted phenomenon that emerges from interactionbetween a number of different selection-based processes, and is influenced both by the expectations from the environmentas well as the constrains of the underlying cognitive system. Dynamical system approach enables us to look at temporalstructure of behavior and talk about the underlying system. With help of three experiments, the study looks at how thetemporal structure of reaction time is influenced by predictability of the environment as well as the task , manipulating bothspatial scope of attention as well as temporal scope of attention. Reaction time of participant is treated as a time-series andHurst component is estimated to measure nature of long-range temporal correlations. Results show an interaction betweentask-demands and predictability of the environment on LRTC, suggesting that task-related constraints and environmentalconstraints are handled by interdependent processes.

Hierarchical control theories of perception-action conceptualize action as control of input, occurring simultaneously atmultiple levels. These levels differ in terms spatio-temporal proximity of the perception controlled. However, it is not clearhow this interaction between different levels in a control hierarchy can be measured from the behavior of the organism.We propose that Long Range Temporal Correlations (LRTC) in RT data can be used as a measure of coupling betweendifferent control levels within such complex system. Participants perform the task of controlling a hierarchical stimuluseither at global level or at local level in a noisy presentation, while the level of control and noise are manipulated. Theresults suggest that LRTC in control task is higher for global level of control compared to local level of control in the nonoise condition. We discuss implications of the results for understanding of perception-action interactions as a complexdynamic system.