Delays without Mistakes: Response Time and Error Distributions in Dual-Task

Kamienkowski, Juan E.; Sigman, Mariano · 2008 · OpenAlex-citations

DOI: 10.1371/journal.pone.0003196

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the cognitive architecture of dual-task performance, specifically testing the sequential bottleneck model of the Psychological Refractory Period (PRP). The central research question addresses whether concurrent task interference affects only the timing of cognitive processes or also degrades the quality of decision-making. The sequential model posits that while sensory and motor operations can occur in parallel, central processing stages are serialized, creating a bottleneck that delays the second task without altering the characteristics (precision, variability) of the processing stages themselves. To test this, the authors examined whether error rates and error distributions remain unaffected by dual-task interference, a prediction that had not been rigorously tested in previous studies due to low error rates. The experimental design involved 16 participants performing two non-symbolic comparison tasks in rapid succession: a visual number comparison task (judging if a dot array was larger or smaller than a reference of 20) and an auditory tone discrimination task (judging if a tone was higher or lower than a 700 Hz reference). The Stimulus Onset Asynchrony (SOA) between the two stimuli varied from 100 to 1250 ms, and the order of presentation was unpredictable. Participants were instructed to respond as quickly and accurately as possible. The study manipulated the difficulty of both tasks parametrically to ensure error rates ranged from chance to near-perfect, allowing for a detailed analysis of error distributions as a function of evidence strength (logarithmic distance from the reference). The results confirmed the classic PRP effect on response times (RTs): RTs for the first task were largely unaffected by SOA, while RTs for the second task decreased significantly as SOA increased, showing a slope close to -1 for short intervals. Crucially, however, dual-task interference had no significant effect on error rates. Neither the total number of errors nor the distribution of errors as a function of stimulus distance changed with SOA or task order. The authors modeled the decision process using a Log-Gaussian representation of numerosity to estimate the internal Weber fraction (a measure of decision resolution). They found that the Weber fraction remained constant regardless of whether the number task was performed first or second, or across different SOAs. While the internal reference point shifted slightly when the number task was performed second, the precision of the decision process itself was preserved. These findings provide strong evidence in favor of a sequential model of task execution. The data indicate that dual-task interference acts as a scheduling delay rather than a source of noise or degradation in the decision-making process. The decision mechanism is delayed but its internal workings—specifically the accumulation of evidence and the resolution of the decision—remain unaffected by concurrent tasks. This distinction clarifies the nature of cognitive bottlenecks, suggesting that the brain queues central processing stages without compromising the quality of the decisions made within those stages.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• dual task multitasking
• dual task performance
• task switching