The Neural Basis of Simultaneous Multitasking

Abstract

As humans, we show striking performance costs when attempting to complete more than a single task at a time. This is perhaps surprising given the inherent complexity and processing power of the human brain. In this chapter, we synthesise the in-roads that have been made into understanding the neural basis of multitasking costs and their practice-induced remediation. We propose a taxonomy of the theoretical traditions and empirical insights that have informed our understanding of why the brain struggles to multitask. Cognitive architecture approaches pose cognitive operations that may give rise to multitasking costs and seek neural correlates of those operations. System architecture approaches leverage insights regarding the function of brain circuits to propose why such systems may struggle to multitask. Last, neural architecture approaches use principles of neural computation to build networks that are constrained in multitasking performance as a consequence of their computational features. These approaches converge to demonstrate the importance of prefrontal cortical and subcortical interactions in the production of multitasking costs and, surprisingly, demonstrate the computational advantages afforded by multitasking costs. In the final sections, we outline a new framework for characterising multitasking costs, demonstrating that the brain may leverage shared information between tasks to gain representational efficiency, which occurs at the expense of multitasking performance. Practice serves to attenuate shared environmental information, thereby separating representations and facilitating multitasking performance.

Visualisations by K.G. Garner & David Lloyd.