Neural sources of performance decline during continuous multitasking

Abstract

Multitasking performance costs have largely been characterized by experiments that involve two overlapping and punctuated perceptual stimuli, as well as punctuated responses to each task. Here, participants engaged in a continuous performance paradigm during fMRI recording to identify neural signatures associated with multitasking costs under more natural conditions. Our results demonstrated that only a single brain region, the superior parietal lobule (SPL), exhibited a significant relationship with multitasking performance, such that increased activation in the multitasking condition versus the singletasking condition was associated with higher task performance (i.e., least multitasking cost). Together, these results support previous research indicating that parietal regions underlie multitasking abilities and that performance costs are related to a bottleneck in control processes involving the SPL that serves to divide attention between two tasks.

Keywords: Attention; Cognitive control; Multitasking.

Fig. 1

NeuroRacer tasks and experimental design. Screenshots of experimental conditions. A.) The Drive Only (DO) condition was used to establish a titrated driving level setting of no distraction for each participant in the scanner. B.) The Sign Only (SO) was used to establish an individualized response time discrimination window for each participant in the scanner. C.) The Sign with Road (SWR) condition contained the discrimination task with the driving task presented passively in the background. The car followed an idealized, autopiloted path while the participant engaged in the discrimination task. Note: the displayed text “auto-pilot on” is for display purposes only and was not present during the experiment. D.) The Sign & Drive (SD) contained both tasks simultaneously and was and was interrogated along with SWR during the testing phase of the experiment as our multitasking condition.

Fig. 2

Whole-brain, cluster-corrected univariate analysis between the two NeuroRacer conditions, SD and SWR. (z = −29, −10, 18, 36, 40, 50).

Fig. 3

Whole-brain cluster-corrected neurobehavioral correlation analysis conjoined with the contrast analysis. (z = x = 21, y = −34, z = 45).

Fig. 4

Correlation plots of RT differences (SD – SWR) versus β-value differences (SD – SWR).