Academic journal article Attention, Perception and Psychophysics

Parallel Response Selection in Dual-Task Situations Via Automatic Category-to-Response Translation

Academic journal article Attention, Perception and Psychophysics

Parallel Response Selection in Dual-Task Situations Via Automatic Category-to-Response Translation

Article excerpt

In contrast to the response selection bottleneck theory of dual-task performance, recent studies have demonstrated compatibility effects between secondary and primary responses on Task 1, suggesting that response information for two tasks may be generated in parallel. In two experiments, we examined the nature of Task 2 response activation in parallel with Task 1, using a psychological refractory period paradigm. Evidence of Task 2 to Task 1 response priming when each Task 2 stimulus was unique indicated that automatic parallel generation of response information occurred for Task 2 via abstract semantic category-to-response translation processes, independent of any direct stimulus-response influences. These findings are discussed in terms of their implications for the traditional response selection bottleneck theory of dual-task performance.

The psychological refractory period (PRP) paradigm (Telford, 1931) has long been used by psychologists as a tool for investigating people's ability to perform two tasks concurrently. In a typical PRP task, participants are presented with two stimuli in close temporal succession and are required to make a speeded response to each. The characteristic result in this paradigm is that the reaction time (RT) for Task 2 increases as the stimulus onset asynchrony (SOA) decreases, whereas the RT for the first task remains relatively unaffected by SOA. One common interpretation of this finding is that there exists a central-processing bottleneck in the stage of response selection, such that the selection of an appropriate response to Task 2 cannot begin until this operation is complete for Task 1 (Pashler, 1984). In this sense, PRP findings demonstrate that for at least the response selection stage of processing, people are strictly unable to do two things at once.

Pashler's (1994) response selection bottleneck (RSB) theory of dual-task performance relies on the locus-ofslack logic, which assumes that tasks can be divided into processing stages that are serial and discrete. Processing stages of the two tasks in a PRP paradigm can run in parallel until the response selection stage; here, there is a bottleneck that permits this processing stage to be completed for only one task at a time. The RSB theory therefore proposes that the response selection stages of the two tasks are serial and discrete with respect to each other, and this has generally accounted well for demonstrations of the PRP effect (for reviews, see Pashler, 1994; Pashler & Johnston, 1998).

Challenges to RSB Theory

More recently, several authors have taken an interest in testing this discreteness assumption. Meyer and Kieras (1997a, 1997b) proposed the executive-process interactive control (EPIC) architecture, which assumes that parallel processing is, in fact, possible in multiple-task performance and that delays in Task 2 performance at short SOAs in PRP studies are due to strategic response deferment. Tombu and Jolicoeur (2002, 2003) and Navon and Miller (2002) have argued for a central capacity-sharing or divisible resource model, in which PRP effects are due to capacity-limited processes that permit graded capacity sharing across tasks, rather than to a strict all-or-none bottleneck.

Cross-talk effects between tasks have also been examined. Logan and Schulkind (2000) presented four experiments demonstrating retrieval of semantic information about their Task 2 stimuli (S2) occurring in parallel with the processing of Task 1 stimuli (S1). They used a PRP paradigm with letter/digit (Experiment 1), magnitude or parity (Experiment 2), or word/nonword (Experiments 3 and 4) discrimination tasks and found that Task 1 RT was shorter when the category of S2 matched the category of S1 (e.g., both were letters) and Task 1 and Task 2 were the same. This finding suggests that participants were able to retrieve the semantic category of S2 early enough that this information was able to facilitate performance for S1. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.