Cognitive load theory (CLT) should be considered a key instructional theory for medical educators because of its strong evidence base and its relevance to our understanding of human cognitive architecture. This theory is founded on the fact that working memory has a limited capacity when dealing with novel information and, when requirements surpass this capacity, then learning is impaired.1 Initially, much of the enthusiasm for healthcare simulation was related to its grounding in adult learning theory and, in particular, Kolb’s cycle of experiential learning.2 It is now well established that simulation-based training techniques are effective3 and the research agenda has shifted to one of how to best apply this educational technique. With this shift, CLT is being increasingly cited as another important theoretical framework3–5 and we feel that a discussion of the origins of CLT, emerging evidence, and its limitations is timely for both researchers and educators. In this review, we use examples from immersive simulations (see Table 1 for definition) wherever possible because these complex learning environments place a high demand on the cognitive resources of the learners and CLT is most relevant in this context. However, the principles of CLT are also important to those using standardized patients (SPs), task trainers, and other simulators because the cognitive architecture of all learners is limited by a finite working memory capacity. Most of the data available to support the theory come from experiments in high school and college learners in traditional learning environments,8–10 as well as trade apprentices in the work place,11 so we strongly advocate for the repetition of empirical CLT studies in the simulation context. In the interim, our recommendations for healthcare simulations are generalizations and predictions derived from the available evidence supporting this influential learning theory.
COGNITIVE LOAD THEORY
Cognitive load theory1 is used to devise instructional procedures based on our knowledge of human cognitive architecture. Cognitive load theory differs from many other educational theories in that it emphasizes explicit rather than minimal guidance during instruction12 and assumes that the knowledge acquired in academic areas such as medicine is domain specific rather than generic.13 Foundational to this theory is the fact that working memory has a limited capacity and duration when dealing with novel information. When the capacity or duration of working memory is surpassed, learning is impaired.14 In contrast, working memory has no known limits when dealing with previously organized information that is retrieved from long-term memory. Immensely complex bodies of information can be dealt with by working memory providing it first has been organized and stored in long-term memory. Accordingly, expertise and skill derive from the storage of immeasurable amounts of domain-specific knowledge in long-term memory.15 It follows that the function of instruction is to build the large knowledge bases in long-term memory required for skilled performance.
Evidence for the size and influence of long-term memory in activities such as problem solving comes from pioneering research using the game of chess. In his classic work, De Groot16 found that the only difference between chess masters and weekend players was in memory of chess board configurations taken from real games. When shown a board configuration taken from a real game for 5 seconds, masters could reproduce the board with approximately 80% accuracy whereas weekend players had an accuracy rate of approximately 30%. Both weekend players and masters had an accuracy of approximately 30% for random board configurations.17 The implication is that chess grand masters win their games because they have learned to recognize tens of thousands of board configurations along with the best moves for each configuration. Expert decisions concerning chess moves depend on previous knowledge just as expertise in medicine relies on huge amounts of domain-specific information stored in long-term memory.15 An experienced emergency department practitioner can analyze an arterial blood gas and rapidly recognize acute respiratory acidosis because they have previously spent many hours training on the individual components of the task including the basic science of acid-base balance, gas exchange principles, and common presentations of respiratory diseases.
Based on this cognitive architecture, the function of instruction is to change the content of long-term memory so that working memory can ultimately use stored information for activities that otherwise would be impossible. However, instruction generally occurs before information being transferred to long-term memory, and so, to be effective, instruction must consider the limitations of working memory when dealing with novel information. The working memory load imposed by instruction consists of the intrinsic complexity of the material, called intrinsic cognitive load, plus the nature of the instruction, that can lead to extraneous cognitive load. Extraneous load is defined as working memory resources required for task completion that do not enhance learning and may be due to poor instructional design. Instructional design that engages learners in processing directed to the construction of schemas is termed germane load, a subtype of intrinsic load. The key concept is that if total cognitive load exceeds working memory capacity, then learning is impaired.
The effect of cognitive load on learning is not linear, and the challenge for all educators is to provide instruction that is adequately stimulating and motivating for learners without pushing them off the far end of the curve as depicted in Figure 1. This figure is taken from an experiment in which the perceived cognitive load of first year medical students after simulation training and debriefing was plotted against performance outcomes.4 Those students with “very high” loads performed less well on subsequent testing compared with their colleagues who did not rate such high loads. Of course, the exact point at which working memory becomes overloaded is multifactorial, individual, and known to neither the student nor the instructor. Nonetheless, it is generally agreed that intrinsic cognitive load should be appropriate to the level of the learner, extraneous cognitive load should be reduced, and germane load increased so long as the limits of working memory are not surpassed.
Readers of this journal will recognize similarities between CLT and constructs discussed in related disciplines such as human factors, cognitive engineering, and decision-making sciences. Each of these fields also acknowledges the limitations of human working memory during the development of complex technological systems or work environments. Thus, the concepts discussed in this paper are not just important to those who educate using simulation techniques but also for those who use simulation to optimize the complex healthcare systems in which humans must perform.
Measuring Cognitive Load
Finding a reliable measure of cognitive load has been of great interest to CLT researchers and more recently to simulation educators.18 To date, the most commonly used measure has been the 9-point Likert scale developed by Paas19 with which he originally asked high school students to rate their mental effort during learning. He found that for instructional materials designed to reduce cognitive load, participants rated their perceived mental effort lower than on materials with no strategy to reduce cognitive load. The self-rating instrument was able to detect differences in effort required for different strategies. In this case and many studies that followed,20 lower cognitive load corresponded to higher test performance. However, it should be noted that too little effort can also lower learning outcomes21 indicating that this relationship is not linear.
When compared with other methodologies, Paas and van Merriënboer22 reported that the self-rating scale was more sensitive than a physiologic measure (spectral analysis of heart rate) and far less intrusive. Since then, the 1-item self-rating scale has proven to be highly reliable and able to differentiate the cognitive load imposed by different instructional procedures and task complexities.20,23 Some CLT researchers24,25 have developed multi-item scales to try to measure the different categories of cognitive load (intrinsic, extraneous, and germane), and the very recent work of Leppink et al26 suggest that learners can indeed reflect on the cognitive load associated with task complexity, instructional design, and their own learning efforts, if not individual loads per se. That novel multi-item subjective instrument holds potential for future research in the area of instructional design.
Researchers in psychology have long used dual-task methodologies in which the learner’s ability to react to an secondary (usually unrelated) stimulus or task, while simultaneously responding to a primary task (learning the content), is used to estimate working memory load. This technique has been found to be useful by CLT researchers particularly when instructing with multimedia materials.27,28 Rojas et al29 recently assessed the cognitive load of medical learners while performing a psychomotor primary task (knot tying) by asking them to respond to a predetermined change in the patient’s heart rate by pressing a foot pedal. They were able to differentiate between novices and experts using response times and error rates, in addition to the subjective ratings (Paas scale). Of particular note was the investigator’s novel use of a clinically relevant stimulus as the secondary task because this begins to address one of the major concerns about using dual task methodology that the secondary task will itself interfere in the learning outcome of interest.
Researchers have also advocated the use of continuous and instantaneous measures of cognitive load during a task because subjective measures can only indicate load retrospectively, after a task has been completed. Antonenko and Niederhauser30 recently found that electroencephalogram (EEG) measures were more sensitive than a subjective measure, whereas Van Gerven et al31 found that pupil dilation increased with higher levels of memory load. Commentators have also advocated using techniques such as functional magnetic resonance imaging32 and eye tracking,33 but these research areas are in their infancy and applications to the physically active learning environments in healthcare simulation will be challenging. Simulation researchers have also measured salivary cortisol34 and heart rate variability as markers of stress, but the complex relationship between stress and cognitive load is not completely understood (see later); further research into the best physiologic markers for each is needed.
In summary, several methods have been used to measure cognitive load, although Paas’ subjective scale of mental effort has had the most use and is currently the easiest to measure in the setting of simulation training. The increasing use of physiologic methods to obtain instantaneous and continuous readings, particularly during task training, is promising for future study of the complex simulation environment where we would like to understand the temporal changes in load occurring across key stages of training (briefing, scenario performance, and debriefing). Ideally, we need to measure cognitive load in addition to learning outcomes in response to manipulations of instructional design if we hope to understand the true value of CLT for designing simulation training.
SOME COGNITIVE LOAD PRINCIPLES RELEVANT TO SIMULATION DESIGN
Cognitive load theory has been used to generate a range of cognitive load principles (CLT refers to these as “effects”) and several applications to medical education have recently been reviewed.35 We will discuss some of the effects in the context of healthcare simulation training in the following sections. Table 2 provides several specific examples of how these design principles might be applied to teaching with simulation. Because CLT has evolved from key aspects of human cognition, there is significant overlap between the principles that we will discuss and other theories or conceptual frameworks. We highlight these similarities in Table 2 with associated references. However, CLT is unique from many of the referenced frameworks in that its effects were generated from the results of randomized and controlled trials in which cognitive load, hypothesized to be on the causal pathway between the intervention and the outcome, was measured. We believe that CLT will prove to be a useful tool to the simulation community for organizing, explaining, and justifying some of the multitude of theoretical perspectives that have already been applied to the practice of SBE.
Medical simulations are highly complex learning environments. Before participating in a simulation, the healthcare learner has spent considerable time studying the background information required for performance of clinical duties. Once starting the simulation, that domain-specific knowledge and skill are required, as well as knowledge and skills about the simulated clinical environment. Consequently, intrinsic load can be high; as well, potentially high levels of emotion and anxiety will also impact cognitive load. As previously discussed, if cognitive load is too high, then learning is likely to be reduced significantly. Cognitive load theory has identified many strategies to reduce intrinsic and extraneous load, which are briefly discussed next with reference to their applications to healthcare simulations.
Managing Intrinsic Cognitive Load
When to-be-learned materials are relatively simple with few interacting elements, intrinsic cognitive load is low and cognitive load issues are not likely to arise. However, interacting elements cause high cognitive load because they must be considered simultaneously in working memory.50 Nonetheless, complex problems can be addressed by bringing previously processed, organized, and stored information from long-term memory into working memory in the form of schemata or chunks. These elements do not take up precious working memory resources because they have already been learned and can be automatically applied. Through understanding this cognitive architecture, specific instructional strategies have been demonstrated to be effective for optimizing intrinsic load.
One common approach to dealing with high intrinsic load is through segmenting of information. By reducing tasks into manageable chunks and practicing each chunk until it is effectively stored in long-term memory, a large amount of information can eventually be manipulated in working memory in the form of these chunks.51 This concept is applied routinely in healthcare training when students are instructed in the basic sciences, clinical skills, and paper-based cases before they move onto immersive simulation experiences. The intrinsic load of the simulation experience comes from knowledge or skills that have not yet been consolidated in long-term memory, the simulated clinical environment in which the learner needs to function and interactions between the two. If data required for the exercise have been previously learned, then reactivating it from an existing schema will not have a significant cognitive load effect; however, reinforcing the concept through a novel application will impose intrinsic load and learning will occur as long as cognitive overload is avoided.
In a recent simulation application, Brydges et al52 demonstrated the effectiveness of segmenting for teaching clinical intravenous catheterization. He found that a group of students assigned to “progressive learning” in which simulator fidelity and task difficulty were gradually increased and learned more efficiently than those students who trained with either a low-fidelity or high-fidelity simulator alone. The low-fidelity simulator presumably avoided cognitive overload in these novice learners compared with the high-fidelity group who learned poorly (although load was not measured). Once the task was “learned” in the low-fidelity situation and therefore stored in long-term memory, then working memory resources were available for incorporation of novel elements imposed by the next level of realism. Why then is progressive learning actually better than low fidelity alone? The answer is most likely because of sequencing effects. By gradually building knowledge in a low-high complexity sequence, learners were able to increase their knowledge stored in long-term memory and ultimately were able to tackle the highest complexity situations and gain the necessary exposure to the highest complexity. Other important mechanisms for learning that are unrelated to cognitive load, including motivation, reflection, and encoding specificity, may also have contributed to improved outcomes. This experiment demonstrates that the complex cognitive processes involved in consolidating, retrieving, and transferring knowledge might not be possible if the initial working memory processing is hindered by cognitive overload.
In considering a simulation environment, there are multiple elements that are likely to be novel to the learner and on their own and could potentially require large amounts of working memory for task completion. For a medical student who has never been in any clinical environment, the interaction with another healthcare professional and with a bedside patient monitor (with several interacting elements of information displayed) could be overwhelming. Even for experienced clinicians, there could be a several new items of information related to the simulation experience: Where are the pulses and heart sounds on this particular mannequin? Do the pupils react? Do I ask the nurse or the patient for a list of medications? Although the best way to introduce the learners to the simulation environment has not been studied, there are some data from cognitive load experiments that can guide our practice.
Mayer et al53 used narrated animations to teach college students how brakes work. Learners had to understand what individual components did (how the brake piston moves), as well as an overall causal model (relations between the piston movement and the brake fluid), placing heavy demands on working memory. Best results were obtained by pretraining on components (names and behaviors) initially before learning about the causal effects. Mayer et al54 also applied a cognitive apprenticeship model37 for teaching geology to college students, and they demonstrated that those who received pretraining on illustrations of key geologic features (eg, a ridge) showed superior problem-solving performance. In a study of mathematics instruction using spreadsheet analysis,38 high school students learned more when they were instructed in the use of Excel spreadsheets before instruction in mathematics rather than learning both concepts concurrently. Training on some elements initially allowed for schema construction to occur before teaching interactions.
Similarly, the information related to the simulation environment (and not the case content) is a potential source of high intrinsic load for healthcare learners. Adhering to a practice of routine and detailed briefing of the simulation environment as part of each session will enable the creation of “simulation schemata,” thereby freeing working memory resources for optimal learning during the scenario. It follows that the time spent briefing a simulation should not be trimmed when time is, all too frequently, constrained. Maintaining the same “ground rules” between sessions for returning learners is strongly recommended. If students need to reconsider instructional details each day, such as whether they are expected to deliver medications in real-time versus pretend to push drugs, then that new information will need to be processed in working memory on each occasion, engendering higher intrinsic load.
Connections to Scaffolding
These strategies and others described by CLT theory1 advocate a simple-to-complex sequence, and sequencing has long been considered an important teaching strategy both for individual learning episodes (CLT) and in full training programs.36 Effective sequencing has close associations with the concept of scaffolding. Generally, scaffolding has a broad meaning and can refer to any type of guidance given to a learner. For example, Clark et al55 refer to scaffolding as an instructional strategy that provides enough support to reach the intended learning outcomes and can take a variety of forms such as visual aids and on-screen pedagogic agents. Such aids are specifically designed to help learners deal with high intrinsic cognitive load due to task complexity. There are also parallels with the work of Vygotsky44 who believed that there is a zone of proximal development, which represents a gap between the existing developmental level of a learner and a potential level. Learners cannot reach this desired level alone without help from teachers or peers. Consequently, they need to be scaffolded through this zone with well-designed instruction. Whereas most of the CLT strategies can be considered scaffolds, the ones that reduce intrinsic load are specifically aimed at reducing problem complexity.
Managing Extraneous Cognitive Load
Most work within a CLT framework has investigated the negative consequences of extraneous cognitive load caused by inappropriate instructional procedures. As indicated later, many of the established effects are relevant to teaching with medical simulators and we describe additional practical applications in Table 2. The recently described cognitive load consequences of emotional activation4 are of particular relevance to simulation instructors.
Worked Example Effect
When presented with a problem or goal, novice learners will typically engage in problem-solving strategies, such as means-end analysis,56 that require a high degree of mental effort. The result can be that inadequate cognitive resources remain for actual learning. Several lines of evidence clearly support the superiority of an instructional method in which the procedure for solving the problem is provided for study.19,57–60 For novices, studying worked examples is usually more effective than solving the equivalent problem and this has been designated as the worked example effect. Worked examples are another form of scaffolding, because learners would not be able to learn the materials through their own problem solving and need the assistance of expert help presented through the worked example. While learners acquire more practice and develop schemata that allow them to recognize and categorize a problem, then guidance is typically faded.61 Fading typically involves leaving out steps in the given solutions. While expertise develops, more steps are left out for the learners to complete until the students are able to solve the problems unaided.
One way in which a worked example can be provided in the simulation laboratory is with the “call for help,” often incorporated as an opportunity for students to practice effective resource management. When the instructor enters the scenario and demonstrates the appropriate problem-solving strategy (scaffolding), then the high cognitive load imposed on learners trying to solve the problem is reduced and working memory resources can be reallocated to learning. The literature on worked examples also suggests that pairing a worked example with a subsequent similar problem to be solved in its entirety is an excellent strategy for learning.62 This type of approach has been advocated by Zigmont et al49 who recommend that simulation learners be given a second opportunity to experience the same case after debriefing. This approach is also consistent with the active experimentation phase of Kolb’s learning cycle, frequently referenced in simulation education.2 Thus, simulation instructors can be reassured that these common instructional techniques have a sound theoretical basis—students will often learn better when they are provided with the full answer to a problem or a partial answer depending on the levels of expertise.
The split-attention effect occurs when learners must divide their attention between 2 or more sources of information that are separated either spatially or temporally. If the 2 sources of information presented are essential to understanding the content of the exercise, then working memory resources are required to mentally integrate the disparate elements. This mental load is extraneous to learning the targeted task, and studies have repeatedly shown improved learning when sources are presented in an integrated format.8 Cierniak et al24 provided a relevant example of this effect when he asked university students to learn about the structure and physiology of human kidneys. Very dense explanatory text positioned alongside a diagram (split-attention format) led to poorer learning outcomes than using the same text fully integrated inside the diagram.26
In clinical medicine, the reality is that attention is often divided. Educators need to think very carefully about the overall objectives of a given scenario when they design the flow of information. If a very realistic scenario is fraught with split sources of information, then integrating those sources might impair realism, but it will reduce the extraneous load caused by split attention and potentially improve learning. Distractions are common in medical practice, and simulation studies have shown that they cause errors, reduce performance, and irritate practitioners.63 Managing divided attention is a real concern for practicing clinicians,40 but we cannot find any evidence that intentionally providing split-source information during simulation training benefits learning outcomes or the ability to manage such distractions in practice. This is an area for future research but for novice learners, CLT research has clearly shown that unnecessary interruptions and distractions should be minimized to optimize learning outcomes.8
Rather than integrating a diagram with written text to overcome the split-attention effect, the modality strategy advocates the concomitant use of spoken text.64 By providing explanatory text verbally through some recording or directly by the teacher, learners are able to listen and focus on the visual presentation simultaneously. Search processes are reduced and split attention is avoided. It is assumed for this effect to occur that both sources of information are relevant for learning in combination. The dual modality strategy engages both the auditory and visual channels of working memory described by Baddeley6 and also taps into the dual-coding theory of Paivio41 who argues that stronger representations of the concepts are formed by using both channels. One excellent example of the modality effect is from a study in the domain of electric engineering in which investigators found that learning was superior using a spoken text and diagrams compared with identical written text with the same diagrams.65 A second group replicated this effect in an interactive agent-based environment for learning electronics.66
Evidence for this cognitive load effect has been summarized in a meta-analysis of 43 published studies.9 The modality effect certainly supports immersive simulation as an instructional technique given its multimodality nature; however, it reminds those instructing with SPs or task trainers to also engage both sensory channels when feasible. For example, SP cases are often scripted as primarily auditory experiences, focusing on communication skills. If the SP has a history of rash or joint swelling, then providing a picture or moulage instead of a verbal description can reduce the load on the auditory channel. When creating instructional videos for technical skills, narration will be more effective than concomitant text instruction.42
The need for learners to integrate and cross-reference redundant instructional materials can also impose an extraneous cognitive load. Cognitive load theory defines redundant materials as those that are independently intelligible but not essential to the learning activity. This might be duplicate information in a different format (eg, providing both text words and speech during a PowerPoint presentation) or distracting nonessential information (eg, PowerPoint sounds and visual effects). The type of information that contributes to redundancy is that which can be understood on its own and, in this way, differs from the split attention materials in which the 2 (or more) sources of complementary information must be integrated to comprehend the materials. Mayer et al67 described the redundancy effect specifically for multimedia learning as the situation in which the use of text and speech was worse than presenting words solely as speech. Chandler and Sweller11 further demonstrated the redundancy effect in other instructional formats including written materials for learning biology and for electric circuitry. In healthcare simulations with multiple learners, providing laboratory results on paper (or a small computer) could potentially create redundancy because the written information often needs to be shared aloud for all to hear. Providing results on a large screen available for everyone to view at their own pace might be preferable because some learners will not have to attend to the same information twice.
Expertise Reversal Effect
Simulation instructors frequently need to adapt their scenarios and teaching techniques for varying learner groups or levels of learners. This challenge can be framed and potentially eased by an understanding of the well-established expertise reversal effect, a CLT effect caused by differing knowledge levels.10 According to the expertise reversal effect, an instructional design that is beneficial to a novice learner may be detrimental, rather than just neutral, to a more experienced learner and vice versa. For example, in the case of split attention discussed previously, more experienced learners who have gained experience and knowledge might actually find an integrated source of information to be less helpful and, at some point, deleterious to learning. The best instructional format changes with expertise, and this change is explained by the organization of acquired knowledge into highly complex and variably automated schemata in long-term memory.67,68
Consider the following example of the potential effects of split attention, expertise reversal, and redundancy during simulation training. Medical educators understand that healthcare practitioners must obtain patient histories from a variety of sources. Accordingly, in simulation, the nurse “actor” often provides the case introduction with demographics and chief complaint and then learners are expected to extract further information from the patient, family members, and/or the chart. This requires the learner to maintain some information in working memory, while searching for novel elements and processing the interactions between these sources of information. This results in a large working memory load due to the split attention effect. For novice learners, it might be cognitively more efficient to provide the learners with a single source of information, such a comprehensive patient history. Cognitive load theory tells us that for novices, this integrated information will engender a lower cognitive load and result in superior learning compared with the conventional split-source format.8
However, more senior learners such as residents or nurse practitioners could find such an integrated format to be onerous. As evolving experts, they are already adept at processing information from multiple sources because of preformed schemata regarding where to search and find specific data. For example, a list of medications from a reliable database can be quickly scanned for pertinence to the overall clinical presentation and would likely be preferable to having to verbally extract the medication name, route, and dosing schedule from the patient. The advanced learner already has some knowledge; for example, the typical dose of a baby aspirin for primary prevention, and the redundancy effect states that collecting such redundant information can impose an additional cognitive load.69 For the more experienced learners, it is preferable to use conventional data-gathering procedures rather than an integrated procedure that was designed to reduce split-attention effects in novices. Thus, the best instructional procedure reverses depending on the expertise of the learners, consistent with the expertise reversal effect. Simulation instructors need to recognize that how the information is presented will affect the mental resources available for learning. In addition, the effect of the presentation method on learner’s cognitive load will depend on learner previous experience. Again, these cognitive load implications for scenario design must be considered alongside techniques to motivate learners to effortfully engage in the learning opportunity. Whereas novice learners may be unmotivated by activities beyond their level of development, more experienced learners will be motivated by realistic challenges.
Cognitive Load Due to Emotions
The previous sections outlined the 2 most common forms of cognitive load (intrinsic and extraneous) as investigated by CLT researchers. Very little CLT research has been conducted into the affective domain and how it impacts on working memory and learning. In the past few years, medical simulation researchers have contributed to an understanding of the role of emotions on cognitive load. The full relationship between emotions and learning is complex, but the established effects include the following70,71: (1) positive emotions generally increase learning mediated through motivation and attention, (2) positive emotional states improve learning by enhancing cognitive processes such as problem solving and creativity, (3) recall of learned information can be dependent on whether or not the mood of the learner trying to recall information is similar to the mood at the time of encoding, and (4) negative emotions can have differential effects on learning, depending on the situation. The following sections describe how emotions generate both extraneous and intrinsic load.
Emotion as Extraneous Load
Interesting details have traditionally been added to simulations for a variety of excellent pedagogic reasons,72,73 but some of those additions come with emotional content, and it is becoming increasingly clear that emotions can exert an extraneous cognitive load. Recently, Um et al74 induced positive emotions during learning in a multimedia environment (by using specific shapes and colors) and they showed that emotion-induced changes in comprehension and knowledge transfer were related to cognitive load.71
The simulation laboratory provides a unique opportunity for studying the relationship between affect and learning because participants’ emotions are inevitably activated. The broad range of emotions experienced in simulation training can include those related to the patient’s health and psychological status, peer pressure, frustration with the environment, and self-monitoring. Fraser et al4 asked students to rate their emotions after a simulation session on an 8-item Likert type scale anchored by the bipolar descriptors of emotions described by Feldman Barrett and Russell.76 Results from first year medical students revealed a positive association between activating emotions and cognitive load. Very high levels of cognitive load were associated with poorer learning outcomes.4
Obviously, patient outcome can impact learner emotions in simulation, particularly if the simulated patient dies. A recent review77 of mannequin death described the potential benefits of the practice as a way to teach about death and dying, to increase realism, and to generate buy-in. Cons included the potential for psychological harm to students, particularly novice learners, although cognitive load effects on learning were not discussed. In a recent randomized controlled trial of final year medical students participating in simulation, the unexpected death of the simulated patient increased the report of negative-activating emotions compared with the control intervention, mannequin survival.75 Learners who experienced the mannequin death were also found to have higher self-rated cognitive load78 and poorer learning outcomes measured objectively 6 weeks later, compared with the control learners. Among the potential untoward effects of mannequin death during simulations, it now seems that some learners are less likely to reach desired learning goals because of the extraneous cognitive load imposed by associated emotions. Of course, the expertise reversal effect10 reminds us that the outcome could be completely different when teaching a group of more or less experienced learners.
Emotions as Intrinsic Load
The potential for emotions to overwhelm mental resources must be considered when instructors augment the emotional experience to meet goals of reflection and realism.79–81 However, within a cognitive load framework, emotional load might not always be extraneous. Extraneous load is defined as load that is not essential to the learning task. If the specific task to be learned is an emotional one, such as a scenario on delivering bad news, then the related content, including the emotional experience, is actually intrinsic to that task and cannot be omitted. In the aforementioned randomized controlled trial on patient death during a simulation, the clear objectives of the exercise were related to the diagnosis and management of a patient with an altered level of consciousness. In that case, simulated patient death adversely affected learning of the objectives because the evoked emotions were not required to achieve the instructional goals. The emotional manipulations in simulation scenarios are often more subtle than mannequin death, such as an upset family member or an unhelpful colleague. Nonetheless, the decision to add such details should be made by weighing up the importance of the emotional context for achieving learning objectives versus the risk that learning will not be successful if working memory limitations are surpassed.
Acute stress is a common learner experience that can contribute to emotional extraneous load. In psychological terms, stress refers to the biological and emotional responses when encountering a threat that one feels that he/she might not have the resources with which to cope.82 Stress is a highly individualistic experience that depends on specific psychological determinants that trigger the physiological response. The complex relationships between stress, learning, and performance have been studied primarily by measuring cortisol levels and the response of the sympathetic nervous system. One established mechanism for impaired learning and performance under stress is a “reduction” in the size of working memory because valuable working memory capacity is allocated to the anxiety and cognitive processes associated with the perceived threat,83 consistent with an extraneous cognitive load effect. However, this reduction in working memory has not been previously framed as a cognitive load effect, nor has CLT generally concerned itself with effects of emotions on learning. Fortunately, CLT researchers are increasingly interested in applying their theory to more authentic, complex learning environments84 just as medical educators are appreciating the importance of applying the cognitive sciences to research in SBE.85
Importantly, stress is also known to exert effects on many of the other processes involved in memory and learning as depicted in Figure 2. Specifically, memory consolidation (ie, long-term memory) seems to be enhanced by stress (and the accompanying cortisol response) that is directly related to the task to be learned86,87; however, memory retrieval is generally impaired.88,89 These differential effects of stress on memory processes are incompletely understood and likely account for the inconsistent reports of stress either enhancing or impairing learning. In a recent comprehensive review of the stress effects in simulation training, LeBlanc90 recommends that to enhance memory consolidation, one must ensure that “the source of the stress is intrinsically linked to the information that is to be learned. If the source of the stress is peripheral to the task, then learning will not be enhanced.” These recommendations are in keeping with a cognitive load perspective of minimizing stress when it imposes an extraneous load. We would add that even when stress is deemed to be relevant to the training goal, the overall capacity of working memory should not be exceeded.
Another source of emotional activation during simulation training is that of “achievement (academic) emotions,” defined as those emotions that are directly linked to learning, instruction, and performance.91 Pekrun’s control-value theory7 suggests that all academic emotions arise from a cognitive appraisal of the controllability of success or failure and the value assigned to the task by the learner. For example, during learning in simulation, a student might experience enjoyment as an achievement emotion because he or she is engaging in an authentic experience that will improve future clinical performance (high value). Alternatively, students could become frustrated if they feel that the simulation mannequin is not a realistic representation of a human patient and they have difficulty, for example, with hearing the breath sounds (low control of success). The perceived attitudes of peers and mentors during the simulation exercise could also affect the value assigned to the task by the trainee, and, accordingly, the academic emotions experienced.
Accordingly, from a cognitive load perspective, the importance of the “safe learning environment” cannot be overstated. Instructors generally establish group trust and respect by setting explicit confidentiality rules, clear expectations, and a commitment to collective learning from errors, rather than blame or humiliation of individuals.92 Educators expect that students will engage in deeper reflection when they are feeling psychologically comfortable and pleasant student experiences should contribute to the ongoing integrity and success of simulation programs.93 Furthermore, human factors research suggests that healthcare teams perform better when members report feeling “psychologically safe.”94 We now propose that this well-established simulation principle of ensuring learner safety probably also improves learning outcomes by optimizing academic emotions and reducing the associated extraneous cognitive load. We need to better understand how to build trust within healthcare teams during simulation training and in practice,95–97 and we believe that measuring emotions and cognitive load will provide insight into the effect of psychological safety on learning and performance.
In summary, emotions and stress are well known to affect learning outcomes and they do so, in part, through effects on working memory. The additional demands placed on working memory to attend to the emotions of the learner could potentially overload mental resources. Because emotions are intentionally manipulated in simulation sessions, it behooves us to consider the balance of detrimental with beneficial effects.
LIMITATIONS AND FUTURE DIRECTIONS
Although CLT has been able to generate new findings and explain old findings, it does have clear limits. The primary concern in applying this theory to healthcare simulations is that the framework has been largely derived from experiments in traditional classrooms with relatively novice learners. Nonetheless, several research articles on teaching with simulation in the past 2 years cite CLT as an important conceptual framework for their findings. Because this trend continues, we will learn more about how the limits of human cognition affect simulation trainees; but to do so, we must first understand the underlying basis for this theory including the details of its established effects.
The major difficulty in applying CLT principles to simulation is in the complexity of simulation environments as compared with the relatively simple teaching platforms in which CLT has traditionally been tested. Defining “extraneous load” as that due to poor instructional design makes sense when the problem to be solved is completely under the control of the instructor, for example, in a mathematics classroom. However in simulation settings, there are multiple “uncontrollable” variables that can be significant sources of extraneous load, including mannequin malfunction, over-zealous actors, or students doing something that was not anticipated by the simulation designer. However, an instructor with an understanding of CLT principles can try to identify and mitigate such potential sources of extraneous load in a proactive fashion. For example, dry runs of scenarios can identify and remediate technical difficulties, actors can be carefully scripted, and “what-if” contingency plans can be provided to all individuals involved in the scenario delivery.
One of the most controversial aspects of CLT has been the term “germane load,” which was added to the CLT framework in the mid 1990s to explain how some additional cognitive loads could be beneficial to learning, when earlier work in the field had always advocated reductions in cognitive load. It was introduced when empirical findings demonstrated that not all cognitive load is bad; indeed, learning will not occur without effortful cognitive processing and its associated mental workload.98 Classic examples of “germane load” include variability in practice and self-explanation, through which learners are encouraged to process data more deeply, creating more flexible and elaborate schemata. In that way, the additional load of the task might be beneficial to learning and especially to transfer, so long as working memory limitations are not surpassed. Since then, germane load has been reconceptualized as a component of intrinsic load. For example, Paas and Van Merriënboer22 improved learning outcomes in college students studying geometry by adding variability to the training materials. The additional “germane” mental workload was intrinsic to the task because the task itself was expanded to include recognition and classification of the problem types. Brydges et al99 applied this principle to orthopedic task training for medical students and residents by assigning them to practice tasks in random versus blocked order. In that case, the expected benefit of contextual interference was not realized and the authors hypothesized that the high complexity of the task with the added variability challenge might have overloaded learner cognitive resources. When germane load is conceptualized as intrinsic material that encourages deeper processing of content, then realism and/or emotion can potentially improve learning through a germane-load effect, particularly in the more advanced learner for whom the intrinsic nature of the original task might not be high.
Until recently, CLT has been largely silent on affective and personality issues such as emotions and motivation. To the simulation instructor, these issues are integral to training and cannot be ignored. New conceptions associated with emotion during simulation have been introduced here, and there are many unanswered questions including how best to measure emotions and whether emotions themselves affect measurements of load. Clearly, motivation is essential to learner engagement, and without it, cognitive load issues are actually irrelevant. However, manipulations in simulation that aim to motivate learners, largely based on adult learning theory, can carry cognitive load consequences particularly for novices. At present, we would recommend that extraneous details introduced for purposes of motivation be carefully considered for their potential to overload learners based on learner experience and previous knowledge. The effect of simulation on motivation and, in turn, on learning is an area that is in need of much study.
One new promising direction for CLT concerns collaboration, because most of the research to date has been conducted with individuals. Collaborative learning environments have a long history of being highly advocated, usually based on a motivational perspective—students are motivated by being grouped together, or from a social constructivist point of view, knowledge is best constructed by discourse between students.44 In healthcare circles, interprofessional training is emerging as an essential component of educational programs aiming to prepare trainees for a collaborative work environment. In recent studies, Kirschner et al100,101 conceptualize collaborative learning from a CLT perspective. A key aspect of their hypothesis is that for complex problem-solving situations, group members can effectively reduce their collective working memory load by working together. Instead of learners trying to solve problems that might overload individual working memories, by sharing the load through collaboration, individual loads are reduced and learning improves. Recently, CLT researchers showed that worked examples are an effective strategy for middle school children in collaborative settings.102 Similarly, medical students training in pairs (as compared with alone) on a medical simulator reported subjectively that they “benefitted from a united memory of the prior instruction.”47 Salas et al103 has clearly demonstrated that the “sharing of mental models” improves team performance through better communication, understanding of common goals, and anticipation of the actions of others. Could an additional advantage of training healthcare learners in groups be improved learning outcomes?104
To sum up, CLT has dramatically improved instruction in many educational domains and it is time for the simulation community to examine its practices under this lens. The simulated learning environment is complex with a potentially high cognitive load that must be recognized and carefully managed in relation to learner previous experience. Recent data have shown that a failure to learn in the simulation setting can be related to cognitive load, but much research is required into how this finding should inform the design of simulations. The incorporation of cognitive load measures into ongoing research in simulation could help quantify the relevance of this theory for our field and provide an additional framework upon which to interpret emerging results.
1. Sweller J, Ayres P, Kalyuga S. Cognitive Load Theory
. New York: Springer; 2011.
2. Kolb DA. Experiential Learning: Experience as the Source of Learning and Development
. Englewood Cliffs, NJ: Prentice Hall Inc; 1984.
3. Cook DA, Brydges R, Hamstra SJ, et al. Comparative effectiveness of technology-enhanced simulation versus other instructional methods: a systematic review and meta-analysis. Simul Healthc
2012; 7: 308–320.
4. Fraser K, Ma I, Teteris E, Baxter H, Wright B, McLaughlin K. Emotion, cognitive load and learning outcomes during simulation training. Med Educ
2012; 46: 1055–1062.
5. Hatala R, Cook DA, Zendejas B, Hamstra SJ, Brydges R. Feedback for simulation-based procedural skills training: a meta-analysis and critical narrative synthesis. Adv Health Sci Educ Theory Pract
2014; 19: 251–272.
6. Baddeley A. Working memory. Science
1992; 225: 556–559.
7. Pekrun R. The control-value theory of achievement emotions: assumptions, corollaries, and implications for educational research and practice. Educ Psychol Rev
2006; 18: 315–341.
8. Ginns P. Integrating information: a meta-analysis of the spatial contiguity and temporal contiguity effects. Learn Instr
2006; 16: 511–525.
9. Ginns P. Meta-analysis of the modality effect. Learn Instr
2005; 15: 313–331.
10. Kalyuga S, Ayres P, Chandler P, Sweller J. The expertise reversal effect
. Educ Psych
2003; 38: 23–31.
11. Chandler P, Sweller J. Cognitive load theory and the format of instruction. Cogn Instruct
1991; 8: 293–332.
12. Kirschner P, Sweller J, Clark R. Why minimal guidance during instruction does not work: an analysis of the failure of constructivist, discovery, problem-based, experiential and inquiry-based teaching. Educ Psych
2006; 41: 75–86.
13. Tricot A, Sweller J. Domain-specific knowledge and why teaching generic skills does not work. Educ Psych Rev
2014; 26: 265–283.
14. Miller GA. The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychol Rev
1956; 63: 81–97.
15. Ericsson KA, Charness N. Expert performance: its structure and acquisition. Am Psychol
1994; 49: 725–747.
16. De Groot A. Thought and Choice in Chess
. The Hague, Netherlands: Mouton. (Original work published 1946); 1965: 463.
17. Chase WG, Simon HA. Perception in chess. Cogn Psychol
1973; 4: 55–81.
18. Issenberg SB, Ringsted C, Ostergaard D, Dieckmann P. Setting a research agenda for simulation-based healthcare education: a synthesis of the outcome from an Utstein style meeting. Simul Healthc
2011; 6: 155–167.
19. Paas FG. Training strategies for attaining transfer of problem-solving skill in statistics: a cognitive-load approach. J Educ Psych
1992; 84: 429.
20. Paas F, Tuovinen JE, Tabbers H, Van Gerven PW. Cognitive load measurement as a means to advance cognitive load theory. Educ Psych
2003; 38: 63–71.
21. Ayres P. Impact of reducing intrinsic cognitive load on learning in a mathematical domain. Appl Cogn Psych
2006; 20: 287–298.
22. Paas FG, Van Merriënboer JJ. Variability of worked examples and transfer of geometrical problem-solving skills: a cognitive-load approach. J Educ Psych
1994; 86: 122.
23. Van Gog T, Paas F. Instructional efficiency: revisiting the original construct in educational research. Educ Psych
2008; 43: 16–26.
24. Cierniak G, Scheiter K, Gerjets P. Explaining the split-attention effect: is the reduction of extraneous cognitive load accompanied by an increase in germane cognitive load? Comput Hum Behav
2009; 25: 315–324.
25. DeLeeuw KE, Mayer RE. A comparison of three measures of cognitive load: evidence for separable measures of intrinsic, extraneous, and germane load. J Educ Psych
2008; 100: 223.
26. Leppink J, Paas F, van Gog T, van der Vleuten CP, van Merriënboer JJ. Effects of pairs of problems and examples on task performance and different types of cognitive load. Learn Instr
2014; 30: 32–42.
27. Brünken R, Steinbacher S, Plass JL, Leutner D. Assessment of cognitive load in multimedia learning using dual-task methodology. Exp Psychol
2002; 49: 109–119.
28. van Gerven PW, Paas F, van Merriënboer JJ, Schmidt HG. Modality and variability as factors in training the elderly. Appl Cogn Psychol
2006; 20: 311–320.
29. Rojas D, Haji F, Shewaga R, Kapralos B, Dubrowski A. The impact of secondary-task type on the sensitivity of reaction-time based measurement of cognitive load for novices learning surgical skills using simulation. Stud Health Technol Inform
2014; 196: 353–359.
30. Antonenko PD, Niederhauser DS. The influence of leads on cognitive load and learning in a hypertext environment. Comput Hum Behav
2010; 26: 140–150.
31. Van Gerven PW, Paas F, Van Merriënboer JJ, Schmidt HG. Memory load and the cognitive pupillary response in aging. Psychophysiology
2004; 41: 167–174.
32. Paas F, Ayres P, Pachman M. Assessment of Cognitive Load in Multimedia Learning
. Charlotte, NC: Information Age Publishing Inc; 2008: 11–35.
33. van Gog T, Scheiter K. Eye tracking as a tool to study and enhance multimedia learning. Learn Instr
2010; 20: 95–99.
34. Bong CL, Lightdale JR, Fredette ME, Weinstock P. Effects of simulation versus traditional tutorial-based training on physiologic stress levels among clinicians: a pilot study. Simul Healthc
2010; 5: 272–278.
35. van Merrienboer JJ, Sweller J. Cognitive load theory in health professional education: design principles and strategies. Med Educ
2010; 44: 85–93.
36. Reigeluth CM. The elaboration theory: Guidance for scope and sequence decisions. In: Reigeluth CM, ed. Instructional Design Theories and Models: A New Paradigm of Instructional Theory
. New Jersey: Lawrence Erlbaum; 1999: 425–453.
37. Collins A. Cognitive apprenticeship and instructional technology, educational values and cognitive instruction: implications for reform. Idol L, Jones BF, eds. Hillsdale, NJ: Lawrence Erlbaum; 1991: 121–138.
38. Clarke T, Ayres P, Sweller J. The impact of sequencing and prior knowledge on learning mathematics through spreadsheet applications. Educ Technol Res Dev
2005; 53: 15–24.
39. Knowles M. The Modern Practice of Adult Education: From Pedagogy to Andragogy
. Englewood Cliffs: Cambridge Univ Press; 1980.
40. Rivera-Rodriguez AJ, Karsh BT. Interruptions and distractions in healthcare: review and reappraisal. Qual Saf Health Care
2010; 19: 304–312.
41. Paivio A. Mental Representations: A Dual Coding Approach
. Oxford: Oxford University Press; 1986.
42. Mayer RE, Moreno R. Nine ways to reduce cognitive load in multimedia learning. Educ Psych
2003; 38: 43–52.
43. Leahy W, Sweller J. Cognitive load theory, modality of presentation and the transient information effect. Appl Cogn Psych
2011; 25: 943–951.
44. Vygotsky LS. Mind in Society: The Development of Higher Mental Processes
. In: Cole M, John Steiner V, Scribner S, Souberman E, eds. Cambridge: Harvard University Press; 1978.
45. Ellis HC, Ashbrook PW. Resource allocation model of the effects of depressed mood states on memory. In: Fiedler K, Forgas J, eds. Affect, Cognition and Social Behavior
. Toronto: Hogrefe; 1988: 25–43.
46. Eysenck MW, Calvo MG. Anxiety and performance: the processing efficiency theory. Cogn Emotion
1992; 6: 409–434.
47. Räder SB, Henriksen AH, Butrymovich V, et al. A study of the effect of dyad practice versus that of individual practice on simulation-based complex skills learning and of students’ perceptions of how and why dyad practice contributes to learning. Acad Med
2014; 89: 1287–1294.
48. Bandura A. Social Foundations of Thought and Action
. Englewood Cliffs: Prentice-Hall; 1986: 50–107.
49. Zigmont JJ, Kappus LJ, Sudikoff SN. The 3D model of debriefing: defusing, discovering, and deepening. Semin Perinatol
2011; 35: 52–58.
50. Sweller J, Chandler P. Why some material is difficult to learn. Cogn Instruct
1994; 12: 185–233.
51. Van Merrienboer JJG, Kirschner PA, Kester L. Taking the load off a learner’s mind: instructional design for complex learning. Educ Psychol
2003; 38: 5–13.
52. Brydges R, Carnahan H, Rose D, Rose L, Dubrowski A. Coordinating progressive levels of simulation fidelity to maximize educational benefit. Acad Med
2010; 85: 806–812.
53. Mayer RE, Mathias A, Wetzell K. Fostering understanding of multimedia messages through pre-training: evidence for a two-stage theory of mental model construction. J Exp Psychol Appl
2002; 8: 147–154.
54. Mayer RE, Mautone P, Prothero W. Pictorial aids for learning by doing in a multimedia geology simulation game. J Exp Psych
2002; 94: 171.
55. Clark RC, Nguyen F, Sweller J. Efficiency in Learning: Evidence-Based Guidelines To Manage Cognitive Load
. San Francisco: John Wiley & Sons; 2006.
56. Sweller J. Cognitive technology: Some procedures for facilitating learning and problem solving in mathematics and science. J Educ Psych
1989; 81: 457.
57. Atkinson RK, Derry SJ, Renkl A, Wortham D. Learning from examples: instructional principles from the worked examples research. Rev Educ Res
2000; 70: 181–214.
58. Van Merriënboer JJG. Strategies for programming instruction in high school: program completion vs. program generation. J Educ Comput Res
1990; 6: 265–285.
59. Oksa A, Kalyuga S, Chandler P. Expertise reversal effect
in using explanatory notes for readers of Shakespearean text. Instr Sci
2010; 38: 217–236.
60. Kyun S, Kalyuga S, Sweller J. The effect of worked examples when learning to write essays in English literature. J Exper Educ
2013; 81: 385–408.
61. Renkl A, Atkinson RK. Structuring the transition from example study to problem solving in cognitive skill acquisition: a cognitive load perspective. Educ Psychol
2003; 38: 15–22.
62. Sweller J, Cooper GA. The use of worked examples as a substitute for problem solving in learning algebra. Cogn Instruct
1985; 1: 59–89.
63. Pluyter JR, Buzink SN, Rutkowski AF, Jakimowicz JJ. Do absorption and realistic distraction influence performance of component task surgical procedure? Surg Endosc
2010; 24: 902–907.
64. Low R, Sweller J. The modality principle in multimedia learning. Mayer RE, ed. The Cambridge Handbook of Multimedia Learning
. Cambridge: Cambridge University Press; 2005: 158.
65. Tindall-Ford S, Chandler P, Sweller J. When two sensory modes are better than one. J Exp Psychol Appl
1997; 3: 257.
66. Mayer RE, Dow GT, Mayer S. Multimedia learning in an interactive self-explaining environment: what works in the design of agent-based microworlds? J Educ Psych
2003; 95: 806.
67. Mayer RE, Heiser J, Lonn S. Cognitive constraints on multimedia learning: when presenting more material results in less understanding. J Educ Psych
2001; 93: 187–198.
68. Kalyuga S, Chandler P, Sweller J. Levels of expertise and instructional design. Hum Factors
1998; 40: 1–17.
69. Sweller J, Chandler P. Evidence for cognitive load theory. Cogn Instruct
1991; 8: 351–362.
70. Bower GH, Forgas JP. Mood and social memory. In: Forgas JP, ed. Handbook of Affect and Social Cognition
. New York: Oxford University Press; 2000.
71. Pekrun R. The impact of emotions on learning and achievement: towards a theory of cognitive/motivational mediators. Appl Psychol
1992; 42: 359–376.
72. Gordon JA, Hayden EM, Ahmed RA, Pawlowski JB, Khoury KN, Oriol NE. Early bedside care during preclinical medical education: can technology-enhanced patient simulation advance the Flexnerian ideal? Acad Med
2010; 85: 370–377.
73. DeMaria S Jr, Bryson EO, Mooney TJ, et al. Adding emotional stressors to training in simulated cardiopulmonary arrest enhances participant performance. Med Educ
2010; 44: 1006–1015.
74. Um ER, Plass JL, Hayward EO, Homer BD. Emotional design in multimedia learning. J Educ Psyc
2012; 104: 485–498.
75. Fraser K, Huffman J, Ma I, et al. The emotional and cognitive impact of unexpected simulated patient death: a randomized controlled trial. Chest
2014; 145: 958–963.
76. Feldman Barrett L, Russell J. Independent and bipolarity in the structure of current affect. J Pers Soc Psychol
1998; 74: 967–984.
77. Corvetto MA, Taekman JM. To die or not to die? A review of simulated death. Simul Healthc
2013; 8: 8–12.
78. Paas F, Van Merrienboer J. The efficiency of instructional conditions: an approach to combine mental effort and performance measures. Hum Factors
1993; 35: 737–743.
79. Schön D. The Reflective Practitioner
. San Francisco: Jossey-Bass; 1983.
80. Boud D, Keogh R, Walker D. Reflection: Turning Experience Into Learning
. London: Kogan Page; 1985.
81. Rudolph JW, Simon R, Raemer DB. Which reality matters? Questions on the path to high engagement in healthcare simulation. Simul Healthc
2007; 2: 161–163.
82. How to measure stress in humans? Fernand-Seguin Research Centre of Louis Lafontaine Hospital, 2007. Available at: http://www.stresshumain.ca/documents/pdf/Mesures/physiologiques/CESH_howMesureStress-MB.pdf
Accessed January 15, 2015.
83. Beilock SL, Kulp CA, Holt LE, Carr TH. More on the fragility of performance: choking under pressure in mathematical problem solving. J Exp Psychol Gen
2004; 133: 584–600.
84. Choi HH, van Merriënboer JG, Paas F. Effects of the physical environment on cognitive load and learning: towards a new model of cognitive load. Educ Psych Rev
2014; 26: 225–244.
85. Mayer RE. Applying the science of learning to medical education. Med Educ
2010; 44: 543–549.
86. Cahill L, Gorski L, Le K. Enhanced human memory consolidation with post-learning stress: interaction with the degree of arousal at encoding. Learn Mem
2003; 10: 270–274.
87. Wolf OT. The influence of stress hormones on emotional memory: relevance for psychopathology. Acta Psychol (Amst)
2008; 127: 513–531.
88. Kuhlmann S, Piel M, Wolf OT. Impaired memory retrieval after psychosocial stress in healthy young men. J Neurosci
2005; 25: 2977–2982.
89. Buchanan TW, Tranel D, Adolphs R. Impaired memory retrieval correlates with individual differences in cortisol response but not autonomic response. Learn Mem
2006; 13: 382–387.
90. LeBlanc VR. The effects of acute stress on performance: implications for health professions education. Acad Med
2009; 84: S25–S33.
91. Artino AR Jr, Holmboe ES, Durning SJ. Can achievement emotions be used to better understand motivation, learning, and performance in medical education? Med Teach
2012; 34: 240–244.
92. Rudolph JW, Raemer DB, Simon R. Establishing a safe container for learning in simulation: the role of the presimulation briefing. Simul Healthc
2014; 9: 339–349.
93. Truog RD, Meyer EC. Deception and death in medical simulation. Simul Healthc
2013; 8: 1–3.
94. Edmondson A. Psychological safety and learning behavior in work teams. Admin Sci Quart
1999; 44: 350–383.
95. Zhao N. Learning from errors: the role of context, emotion and personality. J Organ Behav
2011; 32: 435–463.
96. Tomaka J, Blascovich J, Kelsey RM, Leitten CL. Subjective, physiological, and behavioral effects of threat and challenge appraisal. J Pers Soc Psychol
1993; 65: 248–260.
97. Williams M. Building genuine trust through interpersonal emotion management: a threat regulation model of trust and collaboration across boundaries. Acad Manag Rev
2007; 32: 595–621.
98. Kalyuga S. Cognitive load theory: how many types of load does it really need? Educ Psychol Rev
2011; 23: 1–19.
99. Brydges R, Carnahan H, Backstein D, Dubrowski A. Application of motor learning principles to complex surgical tasks: searching for the optimal practice schedule. J Mot Behav
2007; 39: 40–48.
100. Kirschner F, Paas F, Kirschner PA. Individual and group-based learning from complex cognitive tasks: effects on retention and transfer efficiency. Comput Hum Behav
2009; 25: 306–314.
101. Kirschner F, Paas F, Kirschner PA. A cognitive load approach to collaborative learning: united brains for complex tasks. Educ Psychol Rev
2009; 21: 31–42.
102. Retnowati E, Ayres P, Sweller J. Worked example effects in individual and group work settings. Educ Psychol
2010; 30: 349–367.
103. Salas E, DiazGranados D, Weaver SJ, King H. Does team training work? Principles for health care. Acad Emerg Med
2008; 15: 1002–1009.
104. Kirschner F, Paas F, Kirschner PA. Task complexity as a driver for collaborative learning efficiency: the collective working-memory effect. Appl Cogn Psych
2011; 25: 615–624.