Mass Research (Volume 2 - Issue 5)
Mental imagery training is an intriguing field of research because it truly highlights the power of the mind and just how important the contribution of the nervous system to performance can be. This meta-analysis (1) is the most up-to-date and most comprehensive review of the research examining the effect of mental imagery on maximal strength. In it, the authors followed accepted best practice guidelines for meta-analytic procedures, managed to include more studies than in previous reviews, and also performed a meta-regression (an analysis that allows you to provide dose-response relationships with the outcome in question and the moderating variables in question). This analysis showed quite clearly that mental imagery training was superior to doing nothing for enhancing strength, but that actual physical training was better than only mental training. Additionally, the authors were unable to discern a clear effect when comparing combined physical and mental imagery training to purely physical training, largely because there are too few studies that make this comparison. In addition to these broad strokes, the authors assessed the relationship of different volumes, intensities, frequencies, and imagined contraction types of mental imagery training, and discussed the potential limitations of the data set to provide a much clearer picture of this topic than had previously existed.
The purpose of this review was to provide an evidence-based synthesis of the currently published research on mental imagery’s effect on maximal strength and to address the following questions:
Based on functional equivalence theory (2) – which more or less is the (largely proven) theory that imagined movement and actual movement use similar neural pathways – the authors hypothesized that both imagined and physical training would share common modifying variables in a dose-response manner, such that an “optimal” frequency, intensity, and volume of mental imagery could be prescribed based on their analysis.
Search Strategy, Inclusion Criteria, and Exclusion Criteria
The authors followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines to conduct a search of peer-reviewed journals for any randomized controlled trials assessing the effect of mental imagery on strength performance in healthy adults. Included studies had to be on healthy adults (male or female), must have lasted at least one week (more than three training sessions), must have had at least one control group either doing physical training or no training, and must have assessed maximal voluntary strength. Studies not in English or German, on unhealthy populations, uncontrolled or non-randomized trials, and studies in which training variables or effect sizes couldn’t be extracted were excluded.
Quality Control, Data Extraction, and Analysis
Two authors independently conducted the searches and included or excluded studies based on consensus. When consensus couldn’t be achieved, a third researcher was consulted. The Cochrane review group’s guidelines were followed to ensure accurate data extraction from the included studies prior to analysis and were checked and rechecked for accuracy by multiple authors. A 10-point modified version of the Physiotherapy Evidence Database (PEDro) scale was used to assess the methodological quality of all included studies. Based on my functional but non-expert knowledge of research statistics, the authors did a commendable job with their analysis. They used a random-effects model in all comparisons, and to test the robustness of these analyses, they also used and reported the outcomes of a fixed-effects model for major comparisons (in all cases, the findings were very similar). They performed a sensitivity analysis to ensure no single study greatly influenced the findings, and they performed a random-effects meta-regression to assess the dose-response relationship between changes in maximal strength and training variables (volume, frequency, intensity, muscles trained, contraction mode). Also, on top of standard statistical analyses (p values), they provided effect sizes and the probabilities a beneficial, harmful, or trivial effect was observed (based on how much of the variability in the effect was positive, negative, or trivially above or below zero), in order to give more applicability to their findings (3). Finally, they performed a funnel plot analysis to detect publication bias in the data set.
First let’s start with the boring (but still important) stuff that you can skip but that you should probably read if you want a complete idea of the context of the findings. The funnel plot analysis did actually find evidence of publication bias. Meaning, the distribution of the data in the studies suggests that there may be null findings that haven’t been published, skewing the results of the meta-analysis and potentially making the effect of mental imagery appear stronger than it would be in actuality. However, they used a threshold of p < 0.10 versus the standard and more stringent 0.05 to detect bias, and the test they specifically used (Egger’s) has been criticized for high rates of false positives (4). Thus, I still think the overall findings of this meta-analysis are broadly accurate. However, it is possible the magnitudes of the effect sizes are slightly higher than they are in actuality; by how much though, I can’t say.
Regarding the people actually analyzed, the pooled sample size of the final analysis included 13 studies with a total of 370 participants; the typical sample size of the individual studies ranged from 8 to 15 participants per group. Four of the studies were on males and females, four did not define their subject pool (meaning they most likely recruited males and females because the authors felt sex would not be a confounding factor), one study was on females, and the remaining four were on males. This is a pretty encouraging blend of male and female participants, likely indicating that the findings can be applied to lifters of either sex. The average age of participants was 28.5 years (age range 18-83 years), and two studies examined a population of older adults (mean age of 72.9 years), so a broad range of age groups was assessed. Except for one study on athletes from both individual and team sports, all of the included studies assessed untrained individuals. Additionally, all of the studies assessed individuals who had never previously performed any structured mental training. As a final note, the studies included in this analysis overall had high scores for methodological soundness, which further reduces the likelihood that the meta-analyzed results are skewed.
For those of you who skipped to this, I can’t blame you. For those who read the previous section, your rewards are the main outcomes in simple bullet points:
This is a pretty cool meta-analysis with a pretty impressive overall message: Mental imagery training is an effective tool to enhance strength even while you are doing no actual physical training, and in fact, is only slightly less effective than physical training. While that is awesome, I have to rain on the parade a little bit by pointing out some important elements of this analysis. First of all, beginners make rapid gains in strength largely due to the initial neuromuscular and motor control adaptations they experience, while the structural changes to their bodies contribute far less to initial strength gains in a relative sense. This meta-analysis largely assessed untrained participants and, given that mental imagery can only improve the neurological and motor control pathways to strength, the finding that physical training was only slightly more effective than mental imagery is unsurprising when you think about it. Unfortunately, this means well-trained lifters would probably benefit less from mental imagery training. In trained lifters, mental imagery training compared to doing nothing might just serve to reduce losses in strength if large periods of physical detraining occurred, versus actually increasing strength (as was shown in this analysis). I don’t know that for sure, but if you squat twice your bodyweight and read this study and think you can take three months off training completely and get stronger by just visualizing squats, I think you’ll be disappointed.
To continue raining on the parade, the effect of mental imagery on dynamic contractions (what we do during actual weight lifting) was small and only roughly one-third of the effect shown for isometric contractions. This also makes sense. If you close your eyes and think of flexing your bicep as hard as you can without actually performing a movement, it’s easy to imagine the feeling and experience. If you imagine yourself doing a free weight bench press or squat, going through the setup and completing the down and up phase of each repetition while imagining the feeling of the bar in your hands or on your back, the change in your visual field, and the contractions in all the associated muscle groups, it’s a lot harder.
You may have also found it a little disappointing to see that combined visualization and physical training was no better than physical training alone. However, in this case, there is a glimmer of hope for trained lifters. The reason visualization is probably so effective in beginners is the same reason the combination of visual and physical training is probably not additionally effective in beginners. Sure, physical training beat mental imagery in beginners by a small margin, but I would speculate this marginal superiority of physical training is not primarily because of added structural changes to the body like one might assume, but rather, because actual training provides better neuromuscular adaptations than imagined training. In physical training, you don’t have to try to visualize your surroundings, the feeling of the bar on your back or in your hands, or the feelings in your muscles when you actually lift…because it’s actually happening. Also, I would guess that trained lifters who have been lifting for years are probably better at re-creating visual images of their training compared to lifters who have just been newly introduced to lifting, such as is the case with most of the subjects in this meta-analysis. To put a little data behind these speculations, you might recall Greg’s article on mental training in which kickboxers with at least a year of resistance training experience made better gains in bench press and half-squat strength when performing visualization after training and positive self-talk during their rest periods, compared to doing physical training only. This study wasn’t included in the present meta-analysis (presumably because it was published after their search or possibly because of the confounding factor of incorporating positive self-talk), but nonetheless, it shows that trained lifters may be able to get additional strength gains from including mental training.
As a final note, I wouldn’t interpret the training-variable ranges associated with the highest effect sizes for strength shown in Table 2 as directly prescriptive. For example, 25 repetitions per set is likely an artefact of the finding that the more total mental imagery practice that was performed in aggregate over the course of the study, the greater the strength gains. Remember, in the meta regression, the variables that had a significant effect on increasing strength were high relative intensities, the number of repetitions imagined per session and per study, and the duration of each training session. So basically, overall volume and intensity. The reason some of those variables look funky (like reps per set) is because this meta-analysis was on a mixture of studies from different disciplines and included stuff like imagined isometric pinky training, ankle dorsiflexion, and finger abduction. So unsurprisingly, the corresponding training variables used by the researchers in these studies don’t necessarily reflect strength and conditioning protocols. It’s also worth pointing out that you can do too much mental imagery in a single session. Despite the finding that the study in which the most total time (in minutes over the course of the entire study) spent visualizing produced the greatest strength gains, the authors were careful to point out that some research has shown that doing a large volume of visualized training in a single session can cause mental fatigue and alter the motor patterns of both actual and imagined movement (5). Thus, when considering actionable prescriptive data from Table 2, I think it probably makes the most sense to focus on the number of sessions per week, training intensity, minutes per session, and sets per session (or perhaps view that as sets per movement so long as you keep the total duration under 15 minutes). One final thing I think is pretty cool is that imagined maximal intensity was found to result in greater strength gains than submaximal intensities. This makes sense. Since you don’t have to experience the stress (both physical and mental) of doing an actual maximal lift when just visualizing it, you can visualize – and benefit from – the impossible, such as doing triples at 100% of your 1RM.
The authors did mention that previous research has found differences favoring first- versus third-person visualization (imagining seeing yourself lifting through your own eyes is better than imagining yourself as seen by someone else) and better outcomes in individuals who score higher on tests discerning their innate ability to visualize. Given these two factors, I think future meta-analyses should strive to quantify these differences. Finally, I would love to see more work on trained lifters to see if visualization is something that should be included in addition to training, or if it would be more beneficial during periods where you can’t train physically in an ideal way (like during travel, when you’re injured, or perhaps imagining heavy singles when you’re on a deload or in the middle of a block of high-volume, low-load training).