A key assumption underpinning kanban is that context switching in work performed by people—the inevitable by-product of multitasking—leads to considerable waste. When individuals perform knowledge work, the impact of context switching resolves, in part, into the question of whether the human brain can multitask efficiently. Therefore, it is useful to understand the neurological basis for why attempts to multitask lead to lower efficiency and effectiveness.
What is cognitive multitasking?
What do we really mean by cognitive multitasking in the brain? In fact, there are two related concepts. In the first concept, multitasking is the true, simultaneous execution of multiple cognitive activities in the brain. Imagine that two different people simultaneously pose a mathematical problem to a third person and that third person solves both problems at the same time. To use the analogy of a computer, it is as if there were two separate computing devices running in parallel, or a quantum computer whose qubits can compute multiple problems simultaneously.1
The second concept refers to the case where the brain handles only one problem at a time, but switches among multiple problems. If, in this case, the brain completes both problems at about the same time, it will appear as if it had been working simultaneously. However, only the time between the start and the end of each task overlap.
Both of these concepts may be contrasted with monotasking, where the brain completely solves one problem before going on to another problem.
I am not a neuroscientist and do not pretend to be able to challenge and to assess independently the results of neuroscience research. My purpose is simply to present to the lean and kanban community findings, such as they appear today in the literature.
No true multitasking
The first finding is that there is no true multi-tasking in the human brain, when it comes to cognitive activities. Neuroscientists refer to the work of their colleague Earl Miller, who said that our brains are “not wired to multitask well… When people think they’re multitasking, they’re actually just switching from one task to another very rapidly. And every time they do, there’s a cognitive cost in doing so.”
How is this conclusion drawn? Neuroscientists assume that brain activity is detectable by the imaging of the brain, using such techniques as functional magnetic resonance imaging (see Fig. 1). They can measure and locate brain activity and relate it to what a subject says he was thinking, doing or they can perform other types of controls.2
Now, I do not understand how this approach can be used to demonstrate that our example of solving two math problems simultaneously is not really happening, since the same part of the brain ought to light up for both problems. I can understand, for example, that the solution of a mathematical problem might light up one portion of the brain and the analysis of the aesthetics of a painting might light up a different area.
But what about the case of two separate mathematical problems, which presumably use the same area of the brain? I will take on faith that there is some truth to what many neuroscientists believe.
Be that as it may, the implication is that all of our subsequent references to multitasking are really referring to the second case, where the brain is switching among multiple problems.
Multitasking is usually less efficient than monotasking
One of the common games used by kanban coaches illustrates the penalty of context switching, that is, trying to multitask. One example would be the name game. Is it faster to spell five different names, one after the other (monotasking), or rather to say the first letter of each name, then the second letter of each name, and so on until all names are spelled (multitasking). Try it (see Fig. 2). Monotasking is five to ten times faster, in most cases, given that there is not only a lot more context switching—there are also defects that need to be corrected.
Neuroscientists have found this to be true in the brain. There is usually a penalty to trying to multitask. I say “usually” because, curiously, about 2.5% of the cases tested did not show a penalty in a series of experiments. The reason for this difference is not known.
However, before everyone starts to try to change how their brains’ function and become like that 2.5%, note that while the rare cases show no penalty to multi-tasking, there is no premium, either.
Practice can improve multitasking performance
Nonetheless, some improvement in multitasking performance is possible. It should be no surprise that we are less penalized when we perform a series of actions with which we are very familiar and which are not too complex. I have not seen a neurological explanation for this phenomenon, but the evidence has been summarized here.
In fact, I suspect that the frequency of task repetition is more important than simplicity of the task for improving multitasking performance. For example, when I sight read a complicated piece of music, I stumble, am not fluid and I make mistakes. But practice leads to fluidity and the ability to express nuance without having to “think” about it. And the more I master the breadth and scope of a musical style, the easier even complicated sight reading of that style becomes.
Think, for example, of that section of Igor Stravinsky’s ballet, Le sacré du printemps, where the rhythm of the dance is extravagantly syncopated, to the point of making it impossible to predict when the next syncopation will fall (unless you have memorized the music in your muscles).
Intuitively, as we practice tasks we reduce the number of conscious decisions we must make. Unfortunately, in a business context most workers never really practice their tasks before performing them in production. When a new type of task is assigned, there might be a brief “training” session (i.e., an introduction via a slide show with commentary—with an exercise if you are lucky). And then it’s off to work.
It should be mentioned that certain types of multitasking appears to be pretty efficient, even without practice. Apparently, people are good at repeating back what they hear, while they continue to listen to the first person speaking. Is this an exception or an indication that we can multitask better without specific training? I have not found an answer.
Impact of multitasking on working memory
It appears that the brain has a working memory, much like the live memory of a computer. It seems to be located in the prefrontal cortex, but perhaps in other places, too. The memory’s capacity is about two to four items (neuroscientists used to say it was about seven items, but this assessment has been downgraded).
Like computer memory chips, which have prescribed environmental operating ranges, the neurons in the prefrontal cortex become less effective when the person is stressed. This stress can be the result of multitasking. Thus, multitasking can lead to stress, which can lead to less effective and efficient working memory, which leads to less efficient work and defects. Thus, the memory capacity of the brain is not a simple bottleneck; the multitasking brain behaves as if that memory were not limited. When cognitive context is switched, items in the brain’s working memory may be replaced by new items. When the context switches back to the previous cognitive process, the old items need to be restored to the working memory.
Task switching depletes oxygenated glucose
Brains require energy to function effectively and this energy is delivered to neurons via oxygenated glucose. The rate at which this glucose becomes available is limited. Thus, a lot of thinking might deplete it faster than it is supplied. The fact of switching tasks itself uses up the glucose, energy that might otherwise be available for more productive thinking. This results in loss in the ability to make neuron connections.
In short, we can make better use of our energy and perform for longer periods if we do not try to multi-task. In addition to being less efficient, multi-tasking also requires more frequent or longer rest periods to restore the oxygenated glucose levels.
So we need to pause from time to time to restore energy levels to our brain. But what happens if, during those “pauses”, we simply end up performing some other form of cognitive activity, such as reviewing our social network accounts? The result is that our thinking becomes ever more fragmented and becomes ever more inefficient.
However, this conclusion may be in conflict with another theory I have seen, coming from those researching emotional intelligence. Namely, positive emotional attractors, such as the support and encouragement that friends might provide, stimulate the parasympathetic nervous system, provide more energy for cognitive activities.3
I suppose that the brain and nervous system, like most of the rest of biological systems, are subject to various cycles and homeostatic mechanisms.
Note that the effects of napping on work performance are regularly discussed in the non-scientific literature. While a nap is a pause that allows for glucose replenishment, it apparently has various other benefits to a working person, benefits not directly related to the question of cognitive multitasking.
Impact of multitasking on memory
It appears that information handled during multitasking is less well remembered. I refer here to the longer term memory, not the brain’s working memory. Good memory requires that the information be transmitted to a certain part of the brain, which transmission apparently either does not occur, or occurs less frequently, during multitasking.
The idea is that information stored in the hippocampus benefits from organizational processing and classification that makes the memory more usable, more retrievable. This occurs during monotasking. However, during multitasking, that information may be sent, instead, to the striatum, thereby failing to gain the benefits provided by the hippocampus.
Dopamine addiction feedback loop
Daniel J. Levitin, the seeming apostle of anti-multitasking, has described in his book, The Organized Mind: Thinking Straight in the Age of Information Overload, the dopamine-addiction feedback loop engendered by multitasking. As the brain is rewarded with dopamine for context switching and losing focus, it seeks to reproduce those same conditions to obtain the same reward.
In another study, so-called heavy (frequent) media multitaskers were found to be less able to ignore distractions than light (infrequent) media multitaskers.
One is reminded of the classic experiment wherein rats would starve to death, spending all their time pressing a button that rewards their brains with pleasurable stimuli.
Prefrontal cortex novelty bias
The prefrontal cortex of the brain has been understood to play a co-ordinating role in cognitive processes (the executive functions). Imagine what might happen when an air traffic controller is distracted by something new, something not directly concerned with the safety of the multiple aircraft within the controlled space. Levitin (see his book, cited above) wrote that the prefrontal cortex’s attention can be easily hijacked by something new.
This is not an issue that distinguishes multitasking from monotasking. However, it does explain, in part, our attraction to attempting to multitask. An open question is whether resisting that temptation uses up energy that might be available for useful work.
The temptation to multitask
As Oscar Wilde famously wrote, “I can resist anything except temptation.” Does the mere temptation to multitask produce the same effects as multitasking? Was Oscar right when he said,
The only way to get rid of a temptation is to yield to it. Resist it, and your soul grows sick with longing for the things it has forbidden to itself, with desire for what its monstrous laws have made monstrous and unlawful.
Should we read “brain” for “soul”?
A 2005 study commonly cited in the press and in blogs reported that multi-tasking reduces IQ by 10 points, noticeably greater than the 4 point reduction due to smoking marijuana. Its author, Glenn Wilson, referred not only to reading email while doing other work, but being impacted by the mere knowledge that an unopened message was in your in-box!
When interviewed on this subject, Wilson responded, “Oh, that damned thing” and went on to say, “It didn’t prove much of anything, of course.”
This is a useful reminder that:
- Factoids cited in the blogosphere and in the press often have one or more errors
- Most of the research on which multitasking analysis is based has not been repeated and is applicable to a limited scope.
Continued multitasking triggers a stress response
One-off multitasking, such as answering the telephone while driving a car, does not have the same impact as multitasking that continues over a period of hours. A study of multitasking women, compared to their husbands working in an office, has shown a significant increase in the stress response for those multitasking women.4
This response, which includes raised levels of cortisol and adrenaline, can overstimulate the brain and cause mental fog or scrambled thinking.
Apparently, certain commentators have extrapolated from this research to make general statements about multitasking. I have not found additional evidence to support this assertion, although it might be perfectly reasonable.
There is a wide body of neuroscience research supporting the idea that multitasking, and perhaps even the thought of multitasking, has a significant negative impact on cognition.
At the individual’s level, this phenomenon would help to explain why team-based multitasking, with no limits to work in progress, has been found to result in much longer lead times. In short, when multitasking we:
- think slower
- make more mistakes
- are distracted by irrelevant topics.
These results are not the only causes of longer leads times. For sure, interpersonal and team dynamics account for much of the slowness. It would be interesting to analyze and even quantify the relative contributions of the neurological/psychological dimensions to the contributions of the social dimensions.
The article Multitasking, kanban and neuroscience by Robert S. Falkowitz, including all its contents, is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
1 I cannot help but wonder if the model of how the brain works, with its neurons that trigger (or not), is too influenced by the binary model of computing based on electronic transistors. Is it possible that something much more complicated is happening in the brain?
2 See also Russell A. Poldrack, “Can cognitive processes be inferred from neuroimaging data?”, Trends in Cognitive Sciences, Volume 10, Issue 2, pp. 59–63, February 2006; Russel A. Poldrack, J.A. Mumford, T.E. Nichols. Handbook of fMRI data analysis. Cambridge University Press 2011; Russell A. Poldrack, “Inferring mental States from neuroimaging data: from reverse inference to large-scale decoding”. Neuron, 72, pp. 692-7.
3 See the course offered by Richard Boyatzis on Coursera. Boyatzis has an extensive bibliography on the subject.
4 Shira Offer, Barbara Schneider. “Revisiting the Gender Gap in Time-Use Patterns Multitasking and Well-Being among Mothers and Fathers in Dual-Earner Families”. American Sociological Review Volume 76, Issue 6, December 2011, pp. 809–833.
The video clip was downloaded from YouTube: https://www.youtube.com/watch?v=NOTjyCM3Ou4. It features Pina Bausch and company.
Fig. 1: By Schematic_of_cortical_areas_involved_with_pain_processing_and_fMRI.jpg: Borsook D, Moulton EA, Schmidt KF, Becerra LR.derivative work: Anthonyhcole (talk) – Schematic_of_cortical_areas_involved_with_pain_processing_and_fMRI.jpg, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=10555524
Fig. 4: By Basal_Ganglia_and_Related_Structures.svg: John Henkelderivative work: Leevanjackson (talk) – Basal_Ganglia_and_Related_Structures.svg, Public Domain, https://commons.wikimedia.org/w/index.php?curid=6338580