KEY WORDS: Mental and physical practice, flexibility training.

An appropriately fluid range-of-motion in the body's various joint complexes is a desirable attribute for the comfortable completion of daily tasks. In more specific circumstances selected sports demand enhanced levels of flexibility thus it is usual for athletes to devote training time to attaining and maintaining desired levels of suppleness to ensure successful participation. Techniques that promote the most effective acquisition of the attributes in question are sought after and incorporated into the training program. Such a line of thought was the stimulus for the present research when one of the authors was assisting with the flexibility training of a group of women gymnasts. During stretching exercises targeted at improved hip mobility, several athletes commented to the effect that they were unsure as to the potentially achievable range of movement and that the uncertainty presented a barrier to improvement. To minimize ambiguity, a model skeleton was used to convey articular potential at the hip. In subsequent stretching sessions, the athletes were encouraged to visualize the range of motion demonstrated by means of the model skeleton and to attempt to incorporate the image formed into their own repertoire. The athletes reported that they were able to comply and, subjectively, claimed that their flexibility had improved. It was decided to test empirically the apparent enhanced effect of a PNF stretching program augmented with a visual imagery technique.

Various authorities differ in both the methods used to achieve targeted flexibility levels and the effectiveness of certain of those modalities, the technique referred to as proprioceptive neuromuscular facilitation (PNF) which originated in rehabilitation therapy (Voss et al., 1985) is widely practiced in athletic preparation and touted as successful in producing desired range-of-motion effects required in either sports performance or return to healthy status following injury (Alter, 1988, McAtee and Charland, 1999). Although the literature varies somewhat when explaining the exact cause of improved joint laxity, it is generally agreed that enhancement of range-of-movement is brought about through exploitation of a protective mechanism centred on the Golgi tendon organs (proprioceptors) whose function is to monitor tendon load. Stimulation of tendon load through stretch or forceful muscular contraction causes the muscles to relax via neurological inhibition in order to prevent tissue damage (McAtee and Charland, 1999). The resultant effect facilitates increased range-of-motion. Several studies and reviews of research support the view that PNF in its various forms is more successful than other techniques such as ballistic and static stretching for achieving desired range-of-motion (Sady et al., 1982, Shellock and Prentice, 1985, Surburg, 1981).

Less clear is the extent to which a momentarily attained range-of-motion can be retained without persistent training though preservation of a desired level of flexibility would be of interest to performers. In a comprehensive study targeted at retention of flexibility (range-of-motion). Zebas and Rivera (1985) compared the effects of three stretching methods, one of which was modified PNF, undertaken for six weeks on the ankle, neck, shoulder, hip and trunk laxity of coeducational students. Retention was assessed at two and four weeks after cessation of training. Although there were significant initial gains in flexibility, there were also significant losses of flexibility overall. Range-of-movement regression was least in the modified PNF group, and most notably between the end of the program and two weeks post-exercise. Clearly, the application of a training process that augments a selected stretching technique to facilitate both acquisition and, then, retention of a desired level of flexibility would be advantageous.

A plausible supplement to be used to improve both acquisition and retention of range of movement in a joint complex is to combine stretching with some variant of mental practice. The collated findings from a substantial body of research focusing mostly on motor skill acquisition inclines to the conclusion that some combination of physical and mental practice yields superior results to either physical or mental practice applied alone (Feltz and Landers, 1983). Could a similar effect be achieved by using the neuromuscular system to enhance range-of-motion? As indicated earlier, presenting gymnasts with a tangible perceptual frame-of-reference for the realization of potential range-of-motion at the hip using a skeleton model and requesting that they attempt to incorporate this percept into the physical process of PNF training apparently resulted in improvement.

Mental practice comes in a number of guises and its relationship with action is not fully understood. Nonetheless, sufficient support for the value of imagery as a representational process whereby physical tasks and related goals can be cognitively practiced with positive results renders the possibility of incorporating mental practice with PNF worthwhile (Jeannerod, 1994, Jeannerod and Frak, 2001). Mental imagery is defined here after Denis (Denis, 1985) as "a psychological activity which evokes the physical characteristics of an object either permanently or temporarily absent from the perceptual field." The process can apply to static or dynamic events, past, recent or yet to be accomplished.

Imagery is classified as either "external" which is predominantly visual and characterized as a third-person perspective (seeing oneself performing) or "internal" which is kinesthetic and explained as a first-person perspective (feeling oneself performing) (Epstein, 1980). Research has demonstrated that the two types of imagery were physiologically distinct by observing greater muscle activity during internal imagery (Epstein, 1980, Jacobson, 1932, Shaw, 1940). Furthermore it has been determined that the internal version is more effective when used in connection with motor performance (Weinberg, 1982). Jeanerrod and Frak (2001) found that neuron discharges in the parietal and premotor cortices 'map' a pattern of action even when the given action is not being physically performed. Further, they asserted that applied studies should be undertaken to explore the possibility that motor imagery could be a potent tool in improving the function of the motor system. Thus, an internal (motor) imagery technique was applied in this study. The working definition of motor imagery used in this study is akin to that proposed by Jeanerrod (2001) namely, that it is of the 'internal' type within which the performer represents the self in action. In this case, the participant imagines the feelings associated with the movement of the limb through a possible range-of-motion. The purpose was to determine whether a proprioceptive neuromuscular facilitation technique coupled with motor imagery, applied to elicit enhanced range-of-motion at the hip joint produced significantly better performance and retention than the proprioceptive neuromuscular facilitation training applied alone.

Subjects
A mixed gender sample of twenty-four undergraduate students mean age 21.01 (± 2.09) years participated with written informed consent in keeping with the University of Liverpool, UK regulations for research with human subjects. There were twelve men and twelve women who reported sound health and no neuromuscular ailments. Several other volunteer participants were eliminated because they had not fully recovered from recent sports injury.

ROM Measures
Both Range-of-Movement (ROM) and Reported Vividness-of-Imagery (RVI) were measured to control for its potential influence on performance (10) prior to the start of the training program. Both sets of data were used to form balanced treatment groups each comprising eight subjects, namely, Control, Proprioceptive Neuromuscular Facilitation (PNF), and PNF plus Mental Imagery (PNFI). The pre-testing group ROM means were 85.40 (± 13.30), 87.30 (± 13.85), and 86.33 (± 11.45) and not significantly different (p > 0.05). The RVI means were 8.00 (± 1.90), 8.32 (± 2.71), and 8.40 (± 1.62) respectively and not significantly different (p > 0.05).

The ROM measurements were taken from the dominant limb identified as the leg used for kicking. All measures were taken by one of the investigators using a photographic technique in which leg position was captured via a camera positioned at a lateral viewpoint and printed as a single frame, video-processed hard-copy (Mitsubishi P70B). ROM value was recorded by taking the table edge as the zero line, drawing lines to the joint markers then measuring the angle with a protractor. Each measure was cross-checked immediately with three types of goniometer, specifically, international standard (Bissell, #7512), spirit level (Medical Research Ltd., UK #8401841), and digital (Soar Nagano, DL 1000). For the purpose of reliability estimation, the ROM data were submitted to a one-way Analysis-of-Variance which resulted in F_3,92 0.05 (p > 0.01) and Kendall Coefficient-of-Concordance of W = 0.97 (p > 0.01). This result was taken to be indicative of both consistency and objectivity in the ROM measurement process. The photographic technique described was used to derive raw data in the experimental portion of the study because this method reduced necessity to maintain the limb at the end position for a lengthy period while the angle was measured.

Reported vividness of imagery was assessed using an "ability to image" rating scale developed specifically for limb positioning during related pilot work.

The protocol was as follows:
1. A four-minute warm-up of jogging and calisthenics without stretching.
2. Subject positioned supine on a firm table.
3. Reference markers placed at hip and lateral condyle of knee.
4. Subject instructed to keep both legs extended, arms crossed on the chest, and
head touching the table throughout.
5. Hip joint was actively flexed to static limit without swinging or bouncing. Non-dominant limb in contact with the table surface at all times.
6. Subject permitted one practice leg lift for familiarization purposes.
7. On the measurement trial, subject raised leg and confirmed when limit was reached.
8. ROM value was recorded by taking the table edge as the zero line, drawing lines to the joint markers on the photograph, then measuring the angle with a protractor.

For vividness of imagery assessment, participants viewed video clips of leg raising movements and, immediately following observation, were requested to close their eyes and attempt to imagine what they had seen, then rate the "clearness of the image seen in the mind's eye". The rating was made by promptly circling a number on a continuum ranging from 0 (no image at all) through 9 (very clear, vivid picture of the leg raising motion). The imagery ability of subjects assessed by this method was relatively high though statistical analysis showed that there were no significant differences between the three groups. The respective means and standard deviations were Control 8.0 (±1.9), PNF 8.3 (± 2.7) and PNFI 8.4 (±1.6).

Motor Imagery Training
First, subjects assigned to the motor imagery treatment viewed a short video presentation showing possible ranges of hip flexion demonstrated with both a skeleton and human model. An audio commentary drew the observer's attention to the manner in which the head of femur rotates freely within the hip socket to permit full range of movement and how hip flexion is constrained by the hamstring muscles when the knee is extended. Also, the commentary stated and showed that movement range as great as 150 degrees was attainable with appropriate training.

Following the video presentation, subjects were requested to assume a comfortable position, close their eyes, and mentally image, specifically, feel themselves producing the observed range-of-movement. It was confirmed verbally that they were able to engage in the imaging process (motor imagery) as requested. The subjects were instructed (and reminded) to use the practiced motor imagery technique during the upcoming flexibility training to imagine the motion felt as they moved their limb through the range demonstrated in the video. To counteract the possibility that the PNF-only group might also use motor imagery, these subjects were engaged in conversation during training. Such conversation consisted of typical 'patient-client' interactions that might occur during a treatment regimen timed so as not to interfere with full compliance in the process.

Flexibility Training
This was preceded with the standardized warm-up referred to above. An agonist-directed PNF technique was selected following consultation with the physical therapist member of the experimental team who trained another member of the team who, in turn, administered the flexibility program that was standardized for each subject and all sessions.

The procedure was as follows:
1. The subject's leg was positioned with hip extended, abducted, and internally rotated. The knee was extended and the ankle plantar flexed and everted.
2. The subject isotonically contracted the agonist through a flexion, adduction, external rotation pattern then returned the leg to its resting position.
3. The movement pattern was repeated five times. On the final attempt, the leg was kept at its maximum active range then repositioned to its extended, abducted, internally rotated state. Verbal confirmation of maximum range was provided on all trials by all subjects combined with a "certain" statement by the trainer.
4. The trainer's hands were readjusted and the subject repeated the same pattern of flexion, adduction and external rotation commencing from the limit of movement range, this time with isometric contraction brought about by the trainer's resistance to motion of the leg maintained for ten seconds.
5. The subject was instructed to relax for ten seconds while the trainer applied passive pressure to further stretch the extensor muscles.
6. The procedure was repeated four times, each time commencing at a new angle.

Post training flexibility measures were administered as described above, on the following day, then 3, 7, fourteen, and twenty-eight days after cessation of the program using the photographic technique described above.

The ROM data in degrees were acquired using the photographic technique explained above. Descriptive statistics for the three groups in each phase of the experiment are provided in Table 1.

To examine potential treatment effects, 'change' values that represented increment, zero change or decrement were derived from each subject's pre-test data. Inspection of these data showed relatively normal distribution (slight positive skew) with a reasonable balance of decrement (11 cases), zero change (6 cases) and increment (23 cases) in control subjects. By contrast, there were three decrements and four zero change values in the PNF treatment group and incremental changes only in the PNFI treatment group. The range of change was -6 through 24 degrees. The change values were submitted to analysis-of-variance with repeated measures on the 'time/ assessment' factor (Hinkle et al, 1979). The main effect of the training treatment was highly significant; F_2,238.78, p < 0.01. The main effect of the time factor (maintenance of ROM) was significant F_4,966.27, p < 0.05 with a significant linear trend as can be seen from the decreasing size of the values in Figure 1 and providing a clear picture of the trend (time periods on the x-axis are 1= 1 day, 2= 3days, 3= 7days, 4= 14 days, 5= 28 days post-training).

For the most part, ROM progressively diminished under both treatment conditions; a regression of some 50 per cent during the 28- day period of no training. Using the Tukey (HSD) procedure, a follow-up analysis was undertaken to examine differences between the mean scores for the two treatment effects (Hinkle et al, 1979). Statistically significant differences (p< 0.05) between the experimental conditions were found at the first and second assessments (1 and 3 days after cessation of training).

• A Proprioceptive Neuromuscular Facilitation (PNF) technique applied to enhance range-of-movement (ROM) at the hip joint was successful.
• The effect produced greater gains in participants who received and applied a motor imagery technique to supplement the regular PNF.
• Both effects receded by about 50% across a no-practice period of 21 days.
• Incorporation of motor imagery with physical therapy deemed worthwile.