Comparing cognitive load and self-regulatory depletion: Effects on emotions and cognitions
Introduction
Self-regulation and working memory have been at the center of much psychological and educational research and have proven crucial to learning and academic success (Alloway et al., 2009, Tangney et al., 2004). Activities that deplete self-regulatory resources or burden working memory capacity can intensely influence both thinking and feeling processes, which may impact learning. Basic research that elucidates and differentiates the effects of self-regulatory depletion and cognitive load thus serves to inform effective teaching and instruction.
In the current research we focused specifically on effects of self-regulatory depletion and cognitive load on the processing of and reaction to emotion-laden information. Often, students must process information that contains emotional content, such as when reading a story for literature class, or that elicits automatic emotional reactions, such as when studying graphic illustrations of the human body in biology class. Learning is not only affected by the extent to which students attend to and process such emotion-laden information, but also by the emotional reactions students have to such information. Positive affect has long been theorized and shown to facilitate whereas negative affect has been thought to undermine educational growth (e.g., Boekaerts, 2007, Buff et al., 2011). Positive feelings have been linked with improvements in verbal fluency (Carvalho & Ready, 2010), attention to material (Plass & Brünken, 2015), and learning outcomes. Negative feelings have been linked with decrements in motivation, attention to material, overall achievement, and increased shallow processing of important information and task-irrelevant thinking (Pekrun, Goetz, Titz, & Perry, 2002). Limited self-regulatory and working memory resources may differentially affect responding to emotional information. Before developing our hypotheses, we define our terms and review relevant research from psychology on self-regulatory resources and working memory capacity—the two capacities implicated in ego depletion and cognitive load, respectively.
Self-regulation refers to the capacity to override a prepotent response and replace it with a response more in line with one's goals (Baumeister et al., 2007, Muraven and Slessareva, 2003). Self-regulation may thus be considered a general purpose capacity to be applied to many different challenges in life, from studying and learning challenging material to losing weight to managing one's emotions. Research has revealed that self-regulation is functionally limited: After using it on one task, people perform more poorly on subsequent tasks that also require self-control (Baumeister, Bratslavsky, Muraven, & Tice, 1998). This temporary deficit in the capacity for self-regulation is known as a state of ego depletion. To account for the ego depletion effect, theorists have proposed that people use self-regulation as though it relies on a limited inner resource or strength (Baumeister et al., 2007). This resource is temporarily depleted by effortful acts of self-regulation, and in the interim period before the resource is replenished, further efforts at self-regulation are prone to failure.
Working memory refers to the capacity to direct attention and consciously process and manipulate information. Working memory capacity is a reliable predictor of cognitive performance (Engle, 2002). Working memory is also functionally limited: People can manipulate or maintain only about seven pieces (or three or four chunks) of information at a time (Farrington, 2011, Miller, 1956). Procedures or tasks that occupy attention are said to create cognitive load. Under cognitive load, fewer processing resources are available for other information. For example, a student solving a new type of math problem must keep in mind the rules and steps by which to solve it, thereby creating a cognitive load that may reduce success at actually solving the problem (Sweller, Van Merrienboer, & Paas, 1998).
Both ego depletion and cognitive load are thought to reduce limited resources and tend to exert similar effects on behavior. For example, both ego depletion and cognitive load have been found to undermine performance on tasks that require deliberate, controlled, and complex cognitive processes (e.g., Ariely, 2000, Drolet and Luce, 2004, Schmeichel et al., 2003). Despite these commonalities, the present investigation undertook to show that there are important differences between depleting self-regulatory resources and overloading working memory.
Briefly, ego depletion and cognitive load differ insofar as they have different time courses: Ego depletion refers to a lagged or hang-over type effect (i.e., due to prior self-regulatory efforts), whereas cognitive load refers to a concurrent effect (i.e., due to concurrent cognitive processing). Recovery from ego depletion typically requires time for mental rest (Tyler & Burns, 2009), but a cognitive load can be lifted instantaneously (e.g., by processing requirements). Moreover, cognitive load may prevent even relatively simple cognitive processes such as short-term memory maintenance and attention to peripheral information (Lavie, Hirst, De Fockert, & Viding, 2004). Ego depletion, in contrast, does not interfere with attention or short-term memory (e.g., Schmeichel, 2007). Rather, ego depletion reduces the capacity to control attention effectively.
If attention can be conceived as a spotlight, then ego depletion undermines the capacity to control where the light shines, whereas cognitive load reduces the circumference of the spot. Put differently, the student under cognitive load (e.g., reading a new text message during a lecture) may forget what her instructor just said, whereas the student under ego depletion (e.g., having just resisted buying a tempting snack at the vending machine) may have increased difficulty managing her emotional response to a provocative question posed by a fellow student. With these considerations in mind, we conducted a series of experiments to compare cognitive load and self-regulatory depletion with regard to their respective effects on diverse responses to emotion-laden information, which may have important implications for learning and instruction.
After initial efforts at self-regulation, people may become less motivated or less able to exercise self-control on further tasks. Myriad experiments and field studies have supported the idea that self-regulatory capacities are limited and subject to short-term depletion or fatigue (for review, see Maranges & Baumeister, 2016, pp. 42–61). Although in modernity, and especially in the West, few people ever encounter the actual danger of exhausting their physical biological energy resources (e.g., glucose), the brain manages them as if it were vital to conserve. As with muscle tissue, the brain keeps track of its own energy expenditures. Via biological and physiological fatigue signals, the brain enforces conservation of resources by allotting fewer resources to metabolically expensive top-down cognitive processes, such as self-control (for a recent review, see Evans, Boggero, & Segerstrom, 2015). Other top-down influences such as motivation and rewards can override such signals to some extent because the resources are not actually limited (e.g., Baumeister et al., 2007). In this way, self-regulation is functionally limited. This state of limited self-regulatory capacity or energy is referred to as ego depletion, a term that pays homage to Freud, who was one of the first (and only) scientists to theorize an energy model for the self (Freud, 1961a, Freud, 1961b).
During ego depletion, automatic and intuitive thinking processes remain largely intact, but people tend to make cognitive errors because the capacity for conscious, deliberate, complex thinking is hampered (Masicampo and Baumeister, 2008, Pocheptsova et al., 2009). For example, depleted people perform more poorly relative to controls on logical reasoning, deduction, and inference tasks, but perform as well as control participants on simple, automatic cognitive tasks, such as rote memorization or retrieving general knowledge (Schmeichel et al., 2003). These findings fit with models of long-term memory insofar as information or procedures that have been deeply encoded in memory may arise and function automatically, even when the person is not consciously searching memory stores (Atkinson and Shiffrin, 1968, Shiffrin and Atkinson, 1969). The operation of long-term memory thus remains relatively unaffected under ego depletion, which appears to bias information processing toward heuristics, or mental shortcuts, to solve problems (Pohl, Erdfelder, Hilbig, Liebke, & Stahlberg, 2013) at the expense of more controlled or effortful processes.
Ego depletion also influences emotional processes, presumably by reducing success at emotion regulation and inhibition. For example, although negative feelings associated with thoughts of death are usually kept out of conscious awareness, ego depletion disinhibits thoughts and feelings associated with death (Gailliot, Schmeichel, & Baumeister, 2006). Similarly, individuals may suppress feelings of anxiety when taking consequential tests or exams, but this suppression becomes less successful under ego depletion. In one set of studies, depleted people with test anxiety were less successful at ignoring distracting worries and anxious feelings, which led them to perform more poorly on verbal learning and mental arithmetic tasks relative to non-depleted people (Bertrams, Englert, Dickhäuser, & Baumeister, 2013). Hence, ego depletion may have particular relevance for learning and performance in the context of negative emotional information.
A recent review of the neuroscience of self-regulation suggested that ego depletion disrupts top-down, frontal cortices-mediated control over automatic and implicit emotional processes resulting from lower brain regions, such as the amygdala (Heatherton & Wagner, 2011). In this view, top-down control keeps negative affect from interfering with other cognitive processes, but self-regulatory depletion undermines this process and hence may result in increased interference from negative affect. This shift toward automatic, emotional processes instead of more deliberate processes is not necessarily conscious. Indeed, Heatherton and Wagner (2011) proposed that when people are depleted, they become sensitized to cues in the environment that affect cognition and behavior through implicit and unconscious processes. Thus, ego depletion has been associated with decrements in higher level cognition, increases in automatic processing, and interference from negative feelings that individuals otherwise suppress.
Working memory refers to the use of attention to manage short-term memory or the capacity to manipulate and process transient bits of information, but this capacity is limited insofar as working memory can only handle about seven bits, or three to four chunks, of information at a time (for review, see Cowan, 2008). When working memory is burdened with too much information, conscious processing of additional information suffers and the mind relies increasingly on automatic retrieval from long-term memory. Similarly to ego depletion, cognitive loads impair controlled thinking and increases reliance on intuitive modes of thought (Ariely, 2000, Drolet and Luce, 2004). For example, under cognitive load people tend to rely more on simple principles rather than on complex reasoning when considering a moral decision (Greene, Morelli, Lowenberg, Nystrom, & Cohen, 2008). When working memory resources are burdened by load, learning and problem solving abilities also suffer (see Sweller, 1988, Sweller et al., 2011, Sweller et al., 1998, Van Merrienboer and Sweller, 2005). Accordingly, instructional designs that reduce cognitive loads on students have become the cornerstone to improved learning in technology-driven classrooms (Sweller et al., 2011, Sweller et al., 1998, Van Merrienboer and Sweller, 2005).
Whereas ego depletion may undermine the capacity to inhibit negative feelings, cognitive load seems to keep strong emotions—perhaps especially strong negative emotions—out of conscious awareness. For example, cognitive load has been found to protect people from feelings of anxiety associated with complex decision making (Drolet & Luce, 2004) and minor threats (i.e., of shocks, which never occurred; Vytal, Cornwell, Arkin, & Grillon, 2012). Further, fMRI studies have found that cognitive load not only reduces subjective experience of negative emotions but also down-regulates activity in brain regions associated with feelings, including the amygdala (Mitchell et al., 2007, Van Dillen et al., 2009). When under cognitive load—particularly a load that does not involve emotional content—emotional information is processed less thoroughly and hence exerts less influence on behavior. Hence one thrust of the present investigation was to establish these seemingly opposite effects of cognitive load and ego depletion on emotional responses to emotional information.
The present investigation also examined effects on pain. Recent work has suggested that the brain processes all negative feelings, including both pain and social distress, in the same areas and systems (for review, see Eisenberger, 2012), quite possibly because evolutionary processes coopted simple pain detection systems for use in managing social life. For example, both physical and emotional pain rely on mu-opioid-related signaling, the somatosensory cortices and posterior insula (which provide sensory information about the painful stimulus), and the dorsal anterior cingulate cortex (dACC) and anterior insula (which provide affective and distress information) (for review, see Eisenberger, 2012). The question of whether perceptions of physical pain are affected by cognitive load and ego depletion in the same way that perceptions of emotional pain are will be addressed by the current research.
Past work thus suggests that cognitive load and ego depletion have similar effects on controlled cognition but may have different effects on emotional processing. However, to our knowledge no prior studies have compared the effects of load and depletion directly. The current research compared the effects of cognitive load and self-regulatory depletion on cognitive processing of information with emotional content and subsequent emotional reactions. Based on the findings reviewed above, we predicted that cognitive load would distract attention away from negative emotional information and hence limit its impact, whereas ego depletion would not distract attention from negative information and may in fact weaken defenses against such information—thereby allowing it to remain influential and indeed potentially increase its impact.
Our first specific prediction was that people under cognitive load would be distracted from the experience of pain relative to people under ego depletion (Hypothesis 1). To induce the experience of pain in Study 1, we had student participants complete a cold pressor test, which involves immersing one's hand in ice water for as long as possible. Insofar as cognitive load distracts attention away from the experience of pain, this should enable individuals to persevere longer on the pain test, whereas people under ego depletion should be relatively less likely to persevere through the experience of pain.
In the subsequent two studies we focused on cognitive processes that engage the primary modalities by which instructional material is delivered: visual recognition memory and semantic processing, respectively. Most instructional tools (e.g., books, lectures, films, pictures) rely on visual stimuli, semantic stimuli, or both. Visual memory, including the process of encoding visual information, is modulated by the emotional content and context of what people see. Indeed, prior research has observed that people remember negatively-valenced visual information better than affectively neutral information (Kensinger et al., 2007a, Kensinger et al., 2007b, Kensinger et al., 2006). For example, negative objects (e.g., snake, grenade) are remembered with more visual detail than neutral objects (e.g., football, blender), due in part to increased activation of the amygdala (Kensinger, Garoff-Eaton, & Schacter, 2007b). Further, people are more likely to remember affective images, and perhaps especially negative affective images, better than neutral ones (Ochsner, 2000), even when they fixate equally long on all the images (Christianson, Loftus, Hoffman, & Loftus, 1991). A person's own emotional state can also affect visual memory. Research has observed that people more deeply encode and more often recollect emotionally arousing events, as indicated by increased amygdala activation and a strengthening of the interaction between the amygdala and temporal lobe regions implicated in memory (for review, see LaBar & Cabeza, 2006).
Semantic processing, including the encoding and accessing of words' meanings and connections, is also moderated by emotional content. For example, people naturally process emotional words, especially negatively-valenced words, more quickly than neutral words (Scott, O'Donnell, Leuthold, & Sereno, 2009). Further, the speed and conscious awareness of emotion word processing depend on top-down processes, motivations, and task demands (for review, see Kissler, Assadollahi, & Herbert, 2006). Like memory for visual information, semantic processing is shaped not only by the emotionality of the target stimulus but also by one's emotional state. For example, people in a negative mood are less likely than people in a positive mood to activate semantically related concepts (e.g., nurse/doctor) from their memories (Haänze & Hesse, 1993). Similarly, depressed individuals are less successful than others at ignoring negative information, even if it is task-irrelevant, when making lexical decisions (Sass et al., 2014).
Considering these findings and prior research suggesting that cognitive load distracts attention away from emotional information, whereas ego depletion may undermine top-down control over the cognitive and behavioral effects of emotional information, we predicted that cognitive load and ego depletion would differentially affect visual memory (Hypothesis 2) and semantic processing (Hypothesis 3) of negative emotional information. Specifically, we predicted that ego depletion would improve memory for pictures with negative emotional content (Hypothesis 2; Study 2) and increase the likelihood of matching words by negative emotional association (Hypothesis 3; Study 3)—processes implicitly influenced by negative affect—relative to control and cognitive load participants. Further, we predicted that people under cognitive load would report less negative affect than control or depletion participants after exposure to emotional images (Study 2), consistent with the general hypothesis that cognitive load reduces the processing of emotional information. We also tested the extent to which cognitive load reduces negative emotions even when emotions are not explicitly manipulated (Study 3).
Section snippets
Study 1: do cognitive load and ego depletion differentially affect the experience of negative physical feelings?
Study 1 tested how cognitive load versus ego depletion affects people's experiences of and capacity to tolerate pain, a negative visceral state that shares neural substrates with emotional pain. Pain elicits a prepotent response tendency to alleviate or escape the experience of pain. Because cognitive load reduces attentional resources, it should reduce attention to the experience of pain, thereby enabling persons to tolerate aversive stimulation for a longer period of time (Hypothesis 1). Ego
Study 2: do cognitive load and ego depletion differentially affect feelings and visual recognition memory?
Study 2 extends Study 1 using different procedures. We tested whether cognitive load and ego depletion differentially affect subjective responses to and recognition memory for emotional images. We manipulated ego depletion by having participants write a short essay with difficult directions. Specifically, they had to avoid words using A or N and then, in a second essay, avoid words using I or O. This constitutes self-regulation insofar as preparing to write would automatically generate many
Study 3: do cognitive load and ego depletion differentially affect semantic processing?
Study 3 tested again the effects of cognitive load and ego depletion on emotion and cognitive processing, this time in the context of a more complex cognitive process than visual recognition memory: semantic processing. A large portion of educational material is delivered via the written word. The extent to which a student processes text in connection to other words and existing cognitive schema affects how well he or she learns, evidenced by a long tradition of semantic network research (e.g.,
General discussion
Three studies investigated the differential effects of ego depletion and cognitive load on cognitive processing of emotional information and subsequent emotional reactions. We found that cognitive load distracts conscious awareness from aversive or negative emotional information, thereby minimizing negative emotional experience and reducing the effects of negative emotion on cognition, whereas ego depletion does not. Furthermore, ego depletion appears to disinhibit negative implicit, automatic,
Author Contributions
H.M. Maranges designed the overarching hypothesis. B.J. Schmeichel designed, oversaw data collection, and conducted analyses for Study 1. H.M. Maranges designed, oversaw data collection, and conducted analyses for Studies 2 and 3. R.F. Baumeister advised the study design and implementation processes. H.M. Maranges prepared the first draft of the manuscript, and all authors contributed to and approved the final version of the manuscript for submission.
References (74)
- et al.
Human memory: A proposed system and its control processes
Psychology of learning and motivation
(1968) - et al.
Effects of experimentally-induced emotional states on frontal lobe cognitive task performance
Neuropsychologia
(1999) Understanding students' affective processes in the classroom
- et al.
Activating positive affective experiences in the classroom: “Nice to have” or something more?
Learning and Instruction
(2011) What are the differences between long-term, short-term, and working memory?
Progress in Brain Research
(2008)- et al.
Cognitive load selectively interferes with utilitarian moral judgment
Cognition
(2008) - et al.
Cognitive neuroscience of self-regulation failure
Trends in Cognitive Sciences
(2011) - et al.
And deplete us not into temptation: Automatic attitudes, dietary restraint, and self-regulatory resources as determinants of eating behavior
Journal of Experimental Social Psychology
(2007) - et al.
Memory for specific visual details can be enhanced by negative arousing content
Journal of Memory and Language
(2006) - et al.
Effects of emotion on memory specificity: Memory trade-offs elicited by negative visually arousing stimuli
Journal of Memory and Language
(2007)