Paivio (12) suggested that imagery plays a cognitive role (i.e., language and performance enhancement) in many life domains as well as a motivational role (i.e., controlling emotions or picturing specific goals and behaviors leading to these goals). As such, imagery has been a commonly used intervention in sport. Within each of these two roles, Paivio (12) also suggested that imagery can work at either general or specific levels, yielding four functions of imagery: cognitive specific (CS), cognitive general (CG), motivational specific (MS), and motivational general (MG). More recently, imagery research in sport has suggested that the motivational general imagery function can further be divided into two distinct functions: motivational general mastery (MG-M) and motivational general arousal (MG-A) (7). Subsequent work has continued to support the existence of these functions in sport and their relationship with variables such as confidence, anxiety, and performance (7,15) (Fig. 1).
Hall (6) was the first to suggest that imagery might be influential in changing exercise-related behaviors and cognitions as well. More specifically, he suggested that imagery might increase exercise motivation through its influence on self-confidence and outcome expectancy. Ultimately, he proposed that exercise imagery might be related to extreme exercise behaviors such as exercise dependence (Fig. 2). Whereas there has been some support for Hall's contentions (13), this model was proposed before any exercise imagery research was conducted, and the proposed model has not been revisited in light of the research subsequently completed.
Most of the exercise imagery research to date has relied on the Exercise Imagery Questionnaire (EIQ) (8). This measure was developed based on the idea that imagery may serve similar cognitive and motivational functions in exercise contexts as it does in sport, given that both are forms of physical activity (8). Development of this questionnaire, however, yielded only three functions of exercise imagery: energy, appearance, and technique. Energy imagery includes mental images related to becoming more energized or relieving stress. Appearance imagery includes images associated with a leaner, fitter, and healthier appearance. Technique imagery includes imagery related to the execution of proper body positioning and form while exercising. An examination of Paivio's model, however, suggests that the EIQ may not fully examine the range of functions exercise imagery may serve. Specifically, appearance imagery seems to assess MS imagery, technique assesses CS imagery, and energy assesses MG-A imagery. Consequently, CG and MG-M imagery are not represented. Further, even the functions that are assessed in the EIQ may be incomplete; for instance, MS imagery may assess goals beyond those related to weight and appearance and may also include fitness, health, or performance improvements. These weaknesses in the measurement of exercise imagery suggest that research findings to date must be interpreted with some caution. We may not fully understand the extent to which imagery can play a role in changing exercise-related behaviors and cognitions.
More recently, Giacobbi and colleagues (4,5) attempted to extend previous imagery research through grounded theory analytic methods. Their findings revealed eight themes, thus expanding on the three functions already established by other imagery researchers. These themes fit nicely into Paivio's framework of the functions of imagery and include: exercise technique (CS), aerobic routines (CG), exercise context (CG), appearance images (MS), competitive outcomes (MS), fitness/health outcomes (MS), emotions/feelings associated with exercise (MG-A and MS), and exercise self-efficacy (MG-M). Based on their findings, they addressed the need to revisit the EIQ. Ultimately, these authors developed a new measure of exercise imagery, yielding a four-factor model: appearance/health, exercise technique, exercise feelings, and exercise self-efficacy, which may tap into Paivio's model more fully. Together, these findings further support our contention that research to date has not fully captured the range of functions of exercise imagery. Given these concerns (i.e., lack of revision of the exercise imagery model to incorporate current research, potential problems with the measurement of exercise imagery, and evidence of more than three functions of exercise imagery), we propose a revised model of exercise imagery and how it may influence exercise behavior and exercise cognitions. Given the fact that sport and exercise settings both require people to perform physical activity in potentially evaluative settings, the proposed model was loosely based on the applied model of imagery of Martin, Moritz, and Hall (10) used in sport. However, we contend there is practical utility of this model in the exercise context—specifically, this model can potentially guide future research and imagery as an intervention to increase exercise behavior.
OVERVIEW OF THE MODEL
This proposed model of exercise imagery consists of five general components: the antecedents, the five functions of imagery, the outcomes (cognitive and behavioral) of imagery, the efficacy beliefs (as a mediator of the imagery-outcome relationships), and the potential moderating factors (Fig. 3). As this model illustrates, we suggest that the location of imagery as well as the experience level of the exercisers, the exercisers' goals, and their motivation to have others see them as exercisers may influence the type of imagery used. Further, each of the functions of imagery may lead to specific outcomes, a relationship mediated by changes in efficacy beliefs. These outcomes, in turn, can feed back into the efficacy beliefs. Finally, variables such as gender, activity type, frequency of exercise, imagery ability, age, physical health status, and personality variables may moderate these relationships.
The next sections review each of the five components of the proposed model and, when applicable, present research supporting proposed relationships.
Components of the Model
The first component of the model involves antecedents of exercise imagery. The first proposed antecedent is the setting, or context, in which the imagery occurs. Hausenblas et al. (8) and Giacobbi et al. (4) have found that exercise imagery can occur both in the exercise context and outside of it. It is likely that the nature of the imagery may differ with the setting, although this contention has yet to be investigated. For example, it may be that imagery in an exercise setting may tend to focus on cognitive functions (i.e., proper form, technique, routines), MG-A imagery to maintain energy and arousal during exercise, and MG-M imagery to stay motivated during exercise (e.g., when tired, or bored). By contrast, it may be that outside of exercise, imagery focuses on the outcomes of imagery (e.g., weight loss, improved appearance, enjoyment) to motivate one to actually exercise.
Moreover, experience (the length of time one has been exercising) may also impact the functions of imagery. For instance, an experienced exerciser may focus less on the cognitive functions of imagery given they have already mastered the task and more on the motivational aspects of imagery.
This model posits that one's goals may also influence the images people make. For example, individuals who wish to exercise to lose weight may be more likely to use appearance imagery (i.e., imaging their goal). However, someone who exercises to relieve stress may, by contrast, use energy imagery.
Finally, self-presentational concerns (i.e., impression motivation), or the desire for others to see you as an exerciser, may also serve as an antecedent to imagery use. Gammage and colleagues (3) found that exercise imagery was related to self-presentational concerns. Specifically, appearance and technique imagery accounted for 22.9% of the variance in self-presentational concerns. Thus, individuals who are more motivated to be seen as an exerciser by others (e.g., as fit, toned, attractive, or coordinated) may image themselves making these impressions.
This model proposes five functions of imagery in exercise consistent with the functions in sport settings (7). These functions are based on the original functions from the Sport Imagery Questionnaire (SIQ), and subsequent research. CS imagery involves imaging-specific skills and movements associated with exercise, such as form and body position, direction of movement, specific steps or skills, etc. This is consistent with the technique subscale of the EIQ, which encompasses CS imagery and the exercise technique theme (4,8). CG imagery involves imaging strategies or routines, such as aerobics routines, weight training routines, coordination with music, or running routes (4). MS imagery consists of images related to goals or the behaviors needed to achieve those goals such as improved appearance, toned muscles, increased strength, flexibility and endurance, improved health, increased energy, finishing a difficult workout, or sticking to an exercise program (4,8). MG-A imagery includes images of increasing motivation, excitement, relaxation, easing tension, or providing a distraction from exercise (4,8). Finally, MG-M imagery incorporates images of coping in difficult situations, staying focused throughout a workout, maintaining effort even when tired, or being confident and mentally tough (4).
The outcomes of exercise imagery can be divided into cognitive and behavioral outcomes. Behaviorally, it is hypothesized that imagery, and in particular CS and CG imagery, can help an exerciser learn and improve skills, routines, form, and technique during exercise. Further, imagery may also increase exercise behavior by encouraging the initiation of exercise in nonexercisers, or encouraging adherence in those who already exercise. Hausenblas et al. (8) and Gammage, Hall, and Rodgers (3) both found a positive relationship between imagery use and exercise frequency. This suggests that imagery may serve as a motivator for exercise behavior.
In addition to behavioral outcomes, it is possible that imagery may play a role in cognitive outcomes. Rodgers and colleagues (13) examined the relationship between imagery and exercise dependence and found that energy and technique imagery accounted for some variance in obligatory exercise. Hausenblas and Symons-Downs (9) found that energy and appearance imagery predicted exercise dependence in women, and only energy imagery predicted exercise dependence in men. Finally, Munroe-Chandler, Kim, and Gammage (11) found that all three functions of imagery predicted dependency to weight lift over and above that predicted by time spent weight lifting. These studies suggest energy imagery, not appearance, may be a key motivator in exercise dependence.
It is also likely that exercise imagery may be used to regulate anxiety levels. Imagery may be used to increase energy levels (e.g., to get oneself psyched up to exercise when tired, to increase energy at the end of a workout), or it may be used to decrease anxiety levels (e.g., moderate stress, decrease anxiety, calm down). Gammage et al. (2) found that appearance imagery accounted for significant variance in social physique anxiety; that is, imaging the desired outcomes of exercise may actually increase anxiety. However, it is possible that, if used effectively, MG-A imagery may serve to decrease anxiety, much as it does in sport (15). Unfortunately, this specific function of exercise imagery has yet to be investigated.
We believe that imagery may also impact other cognitive outcomes of exercise such as feeling states, intention/motivation, and body image. For example, if one images having fun during exercise, it may increase enjoyment and improve mood. Unfortunately, given that the EIQ does not currently assess these functions, it is impossible to examine this contention at this time. Rogers, Munroe, and Hall (14) also found that appearance imagery did predict intention to exercise in aerobics participants and a sample of general exercisers.
Finally, it is possible that body image may also be impacted by imagery use. It is likely that MS images related to appearance and weight are related to body image. People may image how they wish to look. Conversely, they may image how they actually look, which in turn may improve how accurately they perceive their bodies. Ultimately, this type of imagery may lead to decreased body dissatisfaction, as people learn to accept how their body appears or learn how to more accurately see their bodies. It is also possible that CS and CG imagery may also impact body image, related to how people behave toward their bodies. As people image themselves correctly performing specific skills and routines related to exercise, they may learn to appreciate what their bodies can do. Consequently, they may decrease anxiety related to their bodies and increase positive evaluations of what it can do.
It is important to note that both behavioral and cognitive outcomes of this model feed back into the efficacy beliefs, thereby demonstrating a reciprocal relationship—as efficacy increases, leading to improved exercise behaviors and cognitions, these changes serve as a source of mastery and emotional arousal, thereby serving to increase efficacy further.
In this model, efficacy expectancy, outcome expectancy, and outcome value are suggested to mediate the relationship between the functions of imagery and its behavioral and cognitive outcomes. This suggestion is consistent with Hall's (6) original model of exercise imagery. It is based on Bandura's (1) suggestion that imagery could be a potent source of self-efficacy information. Consequently, self-efficacy can influence behaviors (e.g., exercise) and cognitions (e.g., motivation, anxiety). It is possible that imagery may impact efficacy expectancies (e.g., imaging oneself performing a skill correctly may increase one's confidence in the ability to perform the behavior), outcome expectancies (e.g., imaging oneself exercising may help someone lose weight), or outcome value (e.g., imaging the health outcomes of exercise may make an individual realize the value of one's health).
Further, Rogers and colleagues (14) found scheduling and coping efficacy, but none of the imagery functions, directly predicted exercise behavior in two samples of exercisers. In addition, as mentioned, both types of efficacy and appearance imagery predicted behavioral intention. This finding suggests that self-efficacy is at least a partial mediator of exercise cognitions.
This model further suggests that another type of efficacy belief, specifically self-presentational efficacy, may also serve as a mediator. Gammage and colleagues (3) found that self-presentational efficacy expectancy, self-presentational outcome expectancy, and appearance imagery all accounted for significant variance in social physique anxiety. Again, this suggests that various manifestations of self-efficacy may at least partially mediate the relationship between imagery and its outcomes.
Finally, the model suggests several potential moderating factors—variables that may impact how effective imagery is as an intervention. Research has shown that the gender of the exerciser, the frequency of exercise, and the activity type all seem to influence imagery use (2,3,8), thus suggesting their moderating effects. Although imagery ability and age have yet to be examined in an exercise setting, sport research has generally shown that imagery ability and age moderate the effects of imagery use on outcomes. In addition to these proposed factors, physical health and personality variables may also moderate the effects of imagery use on behavioral and/or cognitive outcomes. For example, it is likely that an obese or previously sedentary individual may use different functions of imagery compared to a highly experienced and healthy exerciser. Perhaps for the obese exerciser, simply imaging finishing an exercise session will be sufficient, rather than imaging the long-term outcomes of exercise, such as weight loss or improved strength.
IMPLICATIONS OF THE REVISED MODEL
This model has implications for both basic and applied research. From a basic research perspective, this model yields specific testable relationships. It is possible to examine if specific functions of exercise imagery lead to specific outcomes, as they do in sport. For example, can MG-A imagery be used to reduce anxiety in exercisers, or does self-efficacy mediate the relationship between these functions and the outcomes? Second, this model also suggests that the EIQ needs to be revisited. It is likely that the current three-factor version of the EIQ is incomplete. Researchers (5) have recently developed the Exercise Imagery Inventory, a four-factor questionnaire assessing the frequency of imagery use. Although a wonderful start to revisiting the tool used to measure exercise imagery, further work must be performed so that the questionnaire fully captures the range of functions of exercise imagery and that each of the subscales is comprehensive.
From an applied research perspective, the model suggests imagery should be tailored to the desired outcomes. For example, to improve posture during weight training, CS imagery should be used; using MG-M imagery should have little effect on this specific outcome. Further, the potential moderators may also indicate that imagery programs may need to take these variables (e.g., age, gender) into account when developing any intervention.
The model has allowed us to provide some organization to an ever-increasing topic of research. With the development of an applied imagery use model, it is our hope that it will guide future research in the testing of specific hypotheses and application for exercise practitioners.
The authors acknowledge the contributions of their many collaborators during this program of research.