Bottleneck theory’s enlightenment of practice-influence on

psychological refractory period

Stunning limiting mechanisms of the humans information processing, of perceiving abilities, working memory performance, sensomotoric reactions upon receiving point, can be classified as phenomenon, giving feedback about the humans assessment of reality. The psychological refractory period effect (PRP) is one of these phenomenon, representing the aspects of delayed reactions, two speed tasks are performed in rapid succession with resulting delay to the second task response (Craik, 1947, 1948; Hick, 1948; Telford, 1931). This increasing delay of the second task response is the high, the short the interval between the two tasks, task 1 and 2, is (Pashler, 1994). This PRP - effect can be observed even when modalities of the tasks (e.g. sensory, motor-driven) are distinct (Pashler, 1994).

Severely theories have tried to build up the PRP-effect’s underlying mechanisms. Following work will focus on one of those, the central bottleneck, trying to illuminate the influence of practice on PRPeffect.

What is the central bottleneck about? Welford (1952) discusses the human’s information process as a 3-stage process with succeeding precentral, central and postcentral stages. E.g. stimulus identification can be classified as an aspect of precentral stage, response selection as an aspect of central, response execution as an aspect of postcentral stage. According to the central bottleneck model pre- and postcentral stages , active during task 1 , are able to be carried out in parallel with any stage of task 2 ( with pre-, central and postcentral). Contrary, the central stage of task 2 shouldn’t be active as long as the central stage of task 1 has been finished. The resulting phenomenon is a waiting period, it has been called the bottleneck delay and should influence the cause of the PRP effect (Pashler & Johnston, 1989; Ruthruff, Johnston & Van Selst , 2001) . Jiang (2004) gave neurobiological support for the bottleneck account, observing significant correlations between lateral frontal, as well as medial frontal cortex activity and the magnitude of PRP.

A structural bottleneck theory (Pashler, 1994; Welford, 1952) and a strategic bottleneck theory (Meyer & Kieras, 1997) discuss the question of parallelism and seriality during central stage processes.

According the structural bottleneck theory, the PRP effect is due to an inability to carry out more than one central operation (e.g., response selection) at once. E.g., response selection needs access to one or more processors. These processors can only handle one input after another. If two response selection processes are dependent to one processor, only one of the two is able to get access. The activation of the second has to wait until the processor has finished with the first. (Pashler, 1994; Welford, 1952)

On the contrary, a strategic bottleneck theory, the adaptive executive control model (Meyer & Kieras, 1997a,1997b) doesn’t preclude parallel information processing in central stages. Under specific circumstances a simultaneous information transfer should be possible (Schumacher et al. 2001, p 102). The bottleneck delay should be caused by a central executive, controlling the reaction selection at task 1 and 2, inhibiting parallel information transfer, trying to avoid poor performance, errors in information processing(Meyer & Kieras, 1997,1997). FMRI studies, comparing one-task with dual-task performance, noticed a frontal and parietal cortical activation under dual-task condition. These activations could be the neural representation of the executive control, but could reflect executive processes too ( e.g. task-switching) Herath (2001).

Further important questions discuss the influence of practice on PRP-effect. Resulting assumptions could support enlightenment of bottleneck theories, their relevance for PRP-effect explanation. Several studies in literature found less influence (e.g., Davis 1956,1957; Dutta &Walker, 1995;Hick, 1948;Karlin & Kestenbaum, 1968; van Selst & Jolicoeur, 1997) . However, the general absence of a practice-effect is questioned for several reasons ( Van Selst et. al 1999). There should be an higher influence on the PRP-effect, if tasks are multimodal instead of modal ( DeJong, 1993; Fagot & Pashler, 1992; Pashler, 1990; Reynolds, 1964; Van Seslst & Johnston, 1997; Vidulich, 1988) Under special circumstances (e.g. in context of multimodal tasks), a decline of PRP effect from 350 ms to 40 ms has been found after practice. That’s a nearly 90% reduction in PRPeffect size ( Van Selst et. al 1999). Maquestiaux (2008) found a decline at 85% of the participants after practice. One reason for these findings could be the difference in difficulty of tasks. For example Van Selst (1999) used more different stimuli and more response alternatives, receiving a higher PRPeffect before, a lower after practice. The difference of difficulty made it possible to exclude, that participants associated the given stimuli (Stimulus Task 1 and 2) with specific responsepairs. It must be noticed, that Maquestiaux ( 2008) used simple tasks, worked with the association aspect after practice, got a significant practice effect too.

Another reason for different findings in practice influence on PRP-effect could be situated in task modality-differences. Practice may reduce the duration of central stages (e.g. response selection), could make it possible to reduce the interference of central stages. This interference-reduction should be higher in multimodal tasks (Welford, 1976) than in unimodal tasks. Further cross-talk between response selection processes in unimodal tasks could be (e.g. manualmanual tasks) increased (Wickens, 1991) and may inhibit practice-effects on PRP, even after extensive practice.

Another reason for interference amplication in PRP-paradigma, could be found in the manualmodality of task 1 with an absent influence of modality of task 2 ( Schvaneveldt, 1969). E.g. the control of manual responses implies the spatial attention of finger position, disrupting spatial processing of stimulus 2, whereas the vocal responses do not nead spatial attention. Further the manual task may needs mor central processing ressources than the vocal task need. On all accounts , the authors act on the assumption, that the response modality issue is important for a better understanding of dual task processing limitations.

If we assume practice- influences on PRP-effect, we encounter the question of its nature. Schumacher et al. (1996) assume a possible parallelism of operations in central stages after practice, whereas trials before practice should be dominated by serial operations.

Maquestiaux ( 2008) chase effect-interpretation in context of an automaticity hypothese. Accordingly participants learn to act more autonomously during central stages of only task 1 or only task 2 or even during both stages. Therefore bottleneck would still be existent, but it would be bypassed. Van Selst et. al (1999) prefering a stage-shortening stage. The automaticity hypothesis wouldn’t be able to explane their findings, e.g. give no good reason for the exact 1:1 relation between PRP-sizereduction and the reduction of task 1 duration after practice. Following the regarding automaticity hypothesesis , the difference of PRP-effects before and after practicing, should only be influenced by the task 2 central stage, independent the duration of task 1 stages. It would predict an higher decline