Conflict control loops in multi-tasking

Shall I respond to the email that just pinged, write this blog, referee the kids, or take the dog out?

Every day, we are bombarded by multiple tasks that must be performed seemingly at the same time. Sometimes they are menial tasks that do not require much attention like a message coming into your phone or hearing your kid play video games. At other times, the tasks require more attention, like writing a Psychonomics post or handling fires at work. And behind each of these tasks is the running mental list of all the other tasks we need to still do, such as making an appointment, deciding on dinner, getting a car inspected, paying a bill, remembering to wish friends happy birthday, and so on.

My father-in-law referred to multitasking as semi-tasking, which reminds us that more often than not, trying to complete multiple activities at the same time causes performance of at least one, if not more, of the tasks to suffer.

Accordingly, much research has focused on the consequences of multi-tasking, and it is now commonly assumed that when two tasks compete for resources simultaneously, a so-called conflict-control loop is triggered.

The article focused on in this post, which was published by Stefanie Schuch, David Dignath, Marco Steinhauser, and Markus Janczyk in Psychonomic Bulletin & Review, reviewed existing evidence for conflict-control loops and how current theory might account for different multitasking contexts.

Cognitive control refers to the set of processes that direct our behavior to achieve a current objective. For example, I am writing this post while experiencing jet lag and multiple sensory inputs in the forms of my family’s voices, the cat meowing outside, and so on.

Being able to monitor ongoing activity that is first assessed for conflict and then adjusted to correct for the conflict is an example of cognitive control. To return to my example, I am struggling to ignore my family’s voices by listening to music as I write. The conflict occurring is that my attention keeps being drawn to their voices rather than the words I am trying to write. Thus, in my current conflict-control loop, I keep having to re-direct my attention to my thoughts about this post.

In the laboratory, a number of tasks have been developed to examine conflict-control loops experimentally. These tasks promote cognitive conflict by simultaneously activating two or more motor or cognitive representations that compete for control. The most famous example is the Stroop task, which we have covered numerous times on this blog, in which people must name the color ink of a word while ignoring the printed word. This is easy if the two streams are congruent but difficult if they are incongruent.

You can do a Stroop experiment by clicking here.

Two other tasks that are frequently used to assess response time to congruent and incongruent information are the Simon task and the Eriksen flanker task (see image below). Regardless of the specific task, incongruent conditions generally produce slower reaction times because of the conflict created between the stimulus and the response.

This well-known congruency effect comes with some nuanced wrinkles. For example, performance on subsequent trials can be affected by previous trials. Gratton and colleagues first noticed that performance is affected by the order of trial types. Specifically, the congruency effect in a trial is smaller after incongruent than after congruent trials (see figure above). The idea is that the irrelevant stimulus feature is no longer important for upcoming trials, and the conflict-control loop has kicked into gear.

The novel aspect of the review by Schuch and colleagues was how conflict-control loops operate in multitasking scenarios. Specifically, the authors argued that multitasking contexts likely involve multiple conflict-control loops.

As shown in the graphic below, conflict monitoring appears to operate within a task, like the Gratton effect during a flanker task. However, it seems that conflict monitoring also occurs across two different tasks under certain conditions (e.g., task set 1: attend to shape; if circle respond left, if square respond right. Task set 2: attend to color; if red, respond left, if blue, respond right). Those conditions arise when the tasks can be processed simultaneously in working memory without interfering with each other. This can occur if tasks sets are so similar to one another that they are perceived as the same task or, conversely, if they are so different that they do not interfere with one another.

In addition to task conflict monitoring, there are three other levels at which monitoring appears to operate: the responses needed, the stimuli being processed, and the errors being made.

Response conflict occurs when incongruent consecutive trials are experienced and a participant must determine if a left or a right response is needed. Typically, this incongruence slows down response times and produces different patterns of responses than task conflicts.

Conflict between stimuli is driven by determining which features are relevant and which ones are not (color or angle).

And finally, if multiple errors are made, a conflict is created, which is then assessed to produce some types of flexible adjustment or correction, whether that is with greater attention or slower responses times.

Switching back and forth between different tasks during multi-tasking contexts necessitates the processing of information at all four levels.  However, our current understanding of how all of these levels affect conflict monitoring and subsequently control adjustment is in need of substantial research.

To address the appearance of a conflict, an adjustment in control is needed. One line of research has investigated the presence of a processing bias. For a single task, there is evidence that after establishing a processing bias following a first experience, subsequent trials will produce stronger activation for the relevant stimulus dimension (shape) while inhibiting the irrelevant stimulus dimension (color).

Although less is known for multitasking, it is proposed that processing biases can occur at the stimulus level (both blue shapes) and perhaps at the task level (depending on the tasks themselves as described above).

A second proposed conflict-control loop in multitasking contexts may involve a speed-accuracy trade off.  Individuals will slow down to be more accurate, suggesting that they are evaluating their performance and adjusting it based on previous responses. Another proposed conflict-control loop involves feedback from various emotional states. It has been proposed that errors  are aversive and this emotional state may be monitored and triggered and adjusted to perform differently on a task so as to reduce the state of anxiety or conflict. This proposed loop needs more extensive study and may be related to the final hypothesized explanation, in which associative processes may be playing a role in conflict-control loops.

Specifically, Schuch and colleagues proposed the possibility that associative processes may be involved in evaluating conflict monitoring and conflict control. That is, each trial experienced may produce an, association between the current cognitive control state and the currently active trial features.  So if the trial produces an incongruent state, a conflict is registered and the feature with which the trial is currently activating will be associated. This association is then triggered by the next trial and will be evaluated using a conflict-control loop.

Whatever the answer, it is clear that when one’s resources are tapped, cognitive control is limited and the available resources for monitoring and adjustment are reduced even further. The authors suggest that the current body of evidence supports the possibility that multiple-conflict-control loops exist, but additional research is necessary to understanding the role of each proposed loop and the conditions in which each loop is triggered.

While we wait for the next set of experimental findings, my jet-lagged brain will continue to attempt to multitask and move towards the “slow down for increased accuracy” conflict-control loop while trying to ignore the competing and compelling voices of my family and the intrusive thoughts of sleep, so that this post may be completed!

Psychonomics article focused on in this post:

Schuch, S., Dignath, D., Steinhauser, M., & Janczyk, M. (2019). Monitoring and control in multitasking. Psychonomic Bulletin & Review, 26(1), 222-240. doi:10.3758/s13423-018-1512-z.

The Psychonomic Society (Society) is providing information in the Featured Content section of its website as a benefit and service in furtherance of the Society’s nonprofit and tax-exempt status. The Society does not exert editorial control over such materials, and any opinions expressed in the Featured Content articles are solely those of the individual authors and do not necessarily reflect the opinions or policies of the Society. The Society does not guarantee the accuracy of the content contained in the Featured Content portion of the website and specifically disclaims any and all liability for any claims or damages that result from reliance on such content by third parties.

You may also like