|Home | About | Journals | Submit | Contact Us | Français|
Brent L. Arnold, PhD, ATC, contributed to conception and design; analysis and interpretation of the data; and drafting, critical revision, and final approval of the article. Carrie L. Docherty, PhD, ATC, contributed to conception and design; acquisition and analysis and interpretation of the data; and drafting, critical revision, and final approval of the article.
Context: Functional ankle instability has been attributed to proprioceptive loss. However, in previous studies of proprioception, authors have not investigated the ability to sense force at the ankle. Additionally, previous investigators have viewed functional ankle instability as either a present or absent condition, rather than a continuum.
Objective: To determine the relationship of ankle giving-way frequency and perceived ankle instability to ankle eversion force sense.
Design: Cohort design.
Setting: Sports medicine research laboratory.
Patients or Other Participants: Twenty individuals (5 men, 15 women) with a history of unilateral ankle instability.
Intervention(s): We tested subjects with 2 loads: 10% and 30% of maximal voluntary isometric contraction.
Main Outcome Measure(s): We measured eversion force sense by calculating absolute, constant, and variable errors from a 3-trial force-matching procedure. Furthermore, subjects reported their frequency of giving way in units of times per day, week, or month, and these data were extrapolated to estimate annual giving-way frequency. Finally, subjects' answers to 6 questions about ankle stability during typical daily or sports activities were summed to create a perceived ankle instability index.
Results: Significant relationships were found for only the 10% maximal voluntary isometric contraction. For absolute error, a positive relationship existed between the number of self-reported episodes of giving way and eversion force sense for both ipsilateral ( r = .58) and contralateral ( r = .49) testing of the injured ankle. Constant error was correlated with giving way ( r = −.56) for ipsilateral testing of the injured ankle. The ankle instability index was also positively correlated with eversion force sense absolute error ( r = .51) for ipsilateral testing only.
Conclusions: Our results suggest that subjects with ankle instability had difficulty replicating eversion forces. Specifically, larger errors were related to both self-reported giving-way episodes and perceived ankle instability.
Although recent authors 1 have suggested that mechanical instability occurs concurrently with functional ankle instability, previous researchers 2–6 have shown that a significant number of individuals have functional ankle instability without mechanical instability. This finding suggests that some cause other than joint laxity produces functional ankle instability. Alternatively, ligament damage has been proposed to lead to deafferentation of the ligaments, impairing proprioception. 7 However, recent studies have cast doubt on this. Specifically, anesthetic injected into ankle ligaments failed to produce differences in postural stability 8, 9 or joint position sense. 9
Muscle damage is also possible during an ankle sprain. Evertor muscle strains occur in 15% of ankle sprains, 10 and 2% of patients with ankle sprains report maximal tenderness over the peroneal muscles or tendons rather than the ligaments. 11 Muscle receptors that are sensitive to tension (ie, the Golgi tendon organ) facilitate muscular activation in antagonist muscles but inhibit agonists. 12, 13 Muscle damage affects these receptors, creating a distorted sense of muscle tension. 14 This may lead to inadequate muscle tension generation, providing inappropriate joint stability and the ankle instability described by Freeman et al. 15
Joint position sense and joint movement sense are more common measures of proprioception and establish the accuracy of an individual's ability to detect joint position and movement. In contrast, force sense is the ability of an individual to detect muscular force. The sense of force is actually 2 related subsenses: tension and effort. Effort has been identified as being derived primarily from a central mechanism but influenced by peripheral receptors. 16–18 Specifically, when the motor command is generated, a copy of the command, or corollary discharge, is sent to the somatosensory centers of the brain. In essence, the size of this corollary discharge is perceived as effort. Conversely, the sense of tension is related to the tension developed by the muscle. 19 Although the receptor of origin for tension is unclear, studies of vibration imply that force sense is not derived from the muscle spindle's primary ending. 16 These findings suggest that tension is derived from the Golgi tendon organ, the secondary spindle ending, or both.
Originally identified in the psychophysical literature and one of the earliest senses studied by psychologists, 20 force sense has received little attention in the fields of orthopaedics and rehabilitation. Aside from our work at the ankle, to our knowledge no other authors have examined force sense after musculoskeletal injury. Establishing whether force sensory deficits exist after musculoskeletal injury is important because joint stability depends at least partially on muscle contraction and the appropriate force generation by the muscles.
For this study, we chose to investigate functional instability along a continuum of instability. Typically, researchers 21–25 of functional ankle instability have focused on injured and uninjured ankles or groups and have compared proprioception between the ankles or groups. We asked individuals to estimate their ankle instability by reporting their frequency of giving way and their perceived instability during various activities of daily life. Our purpose was to correlate these ankle instability measures with eversion force sense.
This study was part of a larger study examining relationships among ankle instability and joint position sense, force sense, and stiffness. 26
Twenty healthy college students (5 men, 15 women: age = 22.9 ± 5.8 years, height = 168.9 ± 9.7 cm, mass = 69.2 ± 13.6 kg) from a large university volunteered for this study. The University's Human Investigation Committee approved the study. Subjects provided written consent and completed the Ankle Instability Instrument (AII) 27 before testing. Subjects who reported at least 1 unilateral ankle sprain and subsequent episodes of giving way 15 were included in the study. Descriptive data about the injured subjects included 9 injuries to the right ankle and 11 to the left ankle; 11.5 ± 23.4 giving-way episodes per year; 1.9 ± 2.6 total number of sprains; sprain severity, grade 1 = 2, grade 2 = 7, grade 3 = 4, no grade reported = 7; and average time since last sprain = 1 to 2 years. Subjects were also free from any other lower extremity injury in the last 2 months. One subject was excluded because of an error in completing the AII.
Subjects came to the sports medicine research laboratory on 2 occasions. At the initial session, subjects were familiarized with the force reproduction task, performed eversion maximal voluntary isometric contractions (MVICs), and completed the AII. The second session included the ipsilateral and contralateral eversion force sense procedures, with ipsilateral testing performed before contralateral testing. For both procedures, both ankles were tested, with the ankle order randomly selected for each procedure. Ankle instability was determined by 2 methods. First, yes responses to 6 questions on the AII were summed to produce the AII6: (1) “Have you ever sprained your ankle?” (2) “Does your ankle ever feel unstable while walking on a flat surface?” (3) “Does your ankle ever feel unstable while walking on uneven ground?” (4) “Does your ankle ever feel unstable during recreational or sport activity?” (5) “Does your ankle ever feel unstable while going up stairs?” (6) “Does your ankle ever feel unstable while going down stairs?” These 6 questions were selected from the AII because they represented typical functional activities and were measured dichotomously and, thus, could be summed to produce a continuous measure. Second, subjects self-reported the frequency of their sense of giving way (eg, times/day, times/ week, times/month), which was subsequently extrapolated to times per year. The reliability (ie, intraclass correlation coefficients) for the AII6 and the frequency of giving-way items has been previously reported as .85 to .99. 27
Bilateral MVIC testing was conducted. We positioned each subject supine. The foot was secured to a Sensotec (Columbus, OH) 500-N load cell attached to a wall-mounted frame (Figure 1). The ankle was positioned in 0° of plantar flexion and subtalar neutral. The hips and knees were flexed to approximately 30° and 60°, respectively, with a foam bolster placed under the knees to maintain the position. Because of the positioning in hip flexion, any adduction of the hip also required hip internal rotation, which was prevented by the bolster. A belt with hook and loop fasteners was secured around both thighs just proximal to the knees. This prevented subjects from abducting the thighs and minimized quadriceps, hamstring, and gluteus involvement (see Figure 1). Subjects were instructed to maximally evert the foot 3 times for 5 seconds, and the average peak force of each ankle was identified as the MVIC.
Subjects' ankles were attached to the load cell as in the strength testing procedures. Subjects were tested at 2 randomly ordered loads: 10% 28 and 30% 14 of MVIC. We chose these loads because previous testing at higher loads (50% and 75% of MVIC) demonstrated no effect. 29 The target force was established using the load cell's digital display. Once the subject reached the specified load, he or she maintained the contraction for 5 seconds, using the digital display for feedback. The display was then covered, and the subject immediately reproduced the target force and held it for 5 seconds. Upon reaching the target, the subject depressed a switch that electronically marked the data file. A 1-minute rest was allowed between trials, and 3 trials were completed at each force. A 5-minute rest was provided before repeating the procedures for the second target force.
As in force sense studies at other joints, 14, 28 contralateral testing was completed with the uninjured ankle producing the target force. It was relaxed, and the injured ankle then reproduced the target force. Relaxing the uninjured ankle before reproducing the force with the injured ankle deviated from previous protocols 14, 28 but was done to maintain consistency between our contralateral and ipsilateral protocols.
For ipsilateral testing, the same basic procedures were followed except that the same ankle produced the target force, relaxed, and immediately reproduced the force. This procedure was repeated for the opposite ankle. For both testing procedures, the last second of the target force and the first second of the reproduction force were used for data analysis. The last and first seconds were selected to be consistent with previous procedures. 30
Three error scores were calculated for each subject: constant error (CE), variable error (VE), and absolute error (AE). These measures were selected to capture each subject's accuracy, consistency, and overall performance, respectively. The VE was calculated as the SD of the 3 reproduction forces. For CE and AE, the difference in newtons between a trial's target and reproduction force was defined as trial error. The CE was calculated as the average of the trial errors, and AE was calculated as the average of the absolute values of each trial error.
We calculated Pearson product moments to assess the correlations between eversion force sense errors (CE, VE, and AE) and the AII6 and the self-reported estimate of the frequency of the ankle giving way. The alpha level was set to .05 for all correlations.
Contralateral testing revealed no relationship between AII6 and any error measure (P > .05), but giving-way frequency and AE at 10% MVIC were positively associated (P < .05) (Table 1). For ipsilateral testing, giving-way frequency and CE and AE at 10% MVIC for the injured ankles were related (P < .05) (Table 2). Additionally, for ipsilateral testing, AII6 and AE at 10% MVIC for the injured ankles were significantly related (P < .05) (see Table 2). No significant relationships were noted for the uninjured ankles (see Table 2) or for any testing done at 30% of MVIC. Tables 3 and and44 provide descriptive data for the eversion force sense measures.
For this study, we chose to investigate ankle instability along a continuum of instability. Typically, authors 21–25 have focused on injured and uninjured groups or ankles and have compared proprioception between the ankles or groups. We sought to create a continuum of instability and correlate it with the subjects' ability to perceive eversion force. We found that subjects' ankle instability, as measured with the AII6 and giving way, was related to their ability to perceive force at 10% of MVIC and accounted for 24% to 34% of the variance. This value may seem relatively low, but we believe it is important for the following reasons. First, ours was not an experimental study; that is, subjects were not randomly assigned to groups, and we did not have control over the damage produced by the sprains. Second, ankle instability is likely the result of multiple factors (eg, joint laxity, muscle reaction time, strength, joint position sense). Third, the size of the correlations may be limited because of the limited variability in the predictor and criterion measures as shown in Figures 2 and and3.3. In both figures, it is apparent that the range of error scores is small. Furthermore, the range of AII6 is restricted because of the way it is calculated. Either or both conditions could restrict correlation size. Finally, converting our correlations to the Cohen d for the effect size of mean differences 31 produced effect sizes ranging from 0.8 to 1.0 SD. Effect sizes of this magnitude are detectable effects 32 and are equal to or greater than previously reported proprioception effect sizes. 33, 34
Our hypothesis was that measures of ankle instability were related to the subjects' ability to sense eversion force. The finding for AII6 was a significant relationship with AE at 10% of MVIC for the injured ankle. As seen in Figure 2, as errors increase, so do scores on AII6. This finding suggests that greater instability was related to the misperception of force.
Eversion force sense errors were also related to subjects' reported giving-way frequency (see Figure 3). For giving-way frequency, comparison of the correlations (see Table 2) reveals similar sizes but different directions for AE and CE. This is because of the way in which AE is calculated. Specifically, AE is calculated from the same errors used for CE, except that the negative values are converted to positive values for AE. Although not true for all data sets, for this data set, this manipulation has the effect of reversing the CE relationship. Thus, in this case, the CE relationship is the most informative.
Figure 3A illustrates the negative relationship between CE and giving-way frequency. Large negative errors (target undershoots) were associated with greater numbers of giving-way episodes. To undershoot the target during reproduction, 3 mechanisms may be possible. First, subjects may overperceive the target force. Second, subjects may perceive the force correctly but incorrectly replicate the force because of central processing factors. Third, subjects may perceive the force correctly but incorrectly replicate the force because of damage to the evertors.
Regardless of the mechanism, the spinal stretch reflex has been shown to be too slow to prevent an unexpected perturbation. 35–38 Thus, for the sensorimotor system to contribute to joint stability, it must do so before a perturbation. If force generated by the muscle before a perturbation is overperceived, loads generated by the muscle would be less than those potentially required to maintain joint stability. Alternatively, central factors or evertor damage may make replication of required forces difficult. Either way, the result would be inadequate stability of the joint. Furthermore, previous work 39 has demonstrated that subjects who underperceived force, regardless of injury status, had greater ankle stiffness. This suggests that force underperception, in contrast to overperception, may be beneficial.
Comparison of the plots of eversion force sense to AII6 scores (see Figure 2) and frequency of giving-way estimates (see Figure 3) illustrates a break in the range of giving-way estimates between 12 and 36 episodes per year that is not present in AII6 scores. This gap is created because of the reporting of the giving-way frequency. Subjects indicated how many giving-way episodes occurred daily, weekly, or monthly. When these data were extrapolated to an estimate of times per year, giving way reported once a month was converted to a 12, twice a month was converted to a 24, and 3 times a month was converted to a 36. This gap was not present on the AII6 because we specifically recruited subjects throughout the range: 5 subjects scored a 1, 5 subjects scored a 2, 3 subjects scored a 3, 3 subjects scored a 4, 2 subjects scored a 5, and 1 subject scored a 6.
One limitation of this study is the use of psychophysical measures of force. Although frequently used 9, 29, 39–42 and well accepted in proprioception research, these procedures require conscious perception of the target variable (eg, force or joint position). Subconscious aspects of proprioceptor function (eg, the muscle spindle's monosynaptic reflex) cannot be detected using these procedures. Thus, it is possible that force sense is used at multiple levels within the sensorimotor system, and our results represent only a limited perspective of the role of force sense in ankle instability.
It should also be noted that the significant relationships are partially because of 2 outliers: subjects 3 and 18. Subject 3 was an outlier for AE and giving way, and subject 18 was an outlier for AE and CE. The reasons for these outliers are unclear. However, ankle instability is typically studied by blocking subjects or ankles into injured and uninjured conditions. This blocking is subsequently analyzed by group mean comparisons. The use of blocking and mean comparisons can obscure extreme and potentially important data points. By not blocking and conducting a correlational analysis, we suspect that we included extreme individuals who normally occur as part of this injured population. We suggest that in the future, researchers should more closely explore and better report their data ranges, especially along the dimension of ankle instability.
For ipsilateral testing, the target force and the reproduction force were produced by the same ankle. Theoretically, this permits assessment of errors in a fashion similar to a test-retest reliability model. In contrast, the contralateral procedures were consistent with a validity-testing model, with the uninjured ankle serving as the objective standard. We had anticipated that with an internal standard (ipsilateral testing), subjects might be very accurate in reproducing the target force and no relationship would be found. Conversely, if damage to force receptors produced a uniform bias that increased in magnitude with the frequency of giving way, this bias would be detected only with contralateral testing and not with ipsilateral testing of the injured limb. In fact, the results for ipsilateral testing produced a greater number of significant relationships (3) than did contralateral testing (1). This finding suggests that the functionally unstable ankle behaves similarly to an unreliable force sensor. Because reliability must precede validity, the lack of stronger findings for contralateral testing is not surprising. Also, the range of errors was larger for contralateral testing, suggesting that the contralateral task may have been more difficult. Increased difficulty would be expected, because the task had to be transferred from one side to the other and probably represents error introduced from more complex processing by the central nervous system.
In addition to testing the participant's injured ankle, we also tested the uninjured ankle. We did this to determine whether deficits were controlled by a central mechanism. Balance is decreased when injured subjects are compared with uninjured subjects 43, 44 but not when the injured ankle is compared with the contralateral uninjured ankle. 43, 45, 46 This finding suggests a preexisting central mechanism. 46 Based on this, we expected that any existing relationship might occur bilaterally. This was not the case and casts doubt on a central control of force sense being responsible for the sense of instability.
All testing was done at 2 loads: 10% or 30% of MVIC. No statistically significant relationships were found for the 30% target load, but significant relationships were found for the 10% target load. Relatively larger errors appeared to be produced at the 10% target load. Specifically, when the injured ankle's AE was normalized to its respective target load, the errors at the 30% and 10% target loads were 11% and 19.8%, respectively. Similarly, normalized contralateral AEs at the 30% and 10% target loads were 23% and 40%, respectively. Why this is the case is unclear, but at least 2 mechanisms are possible. First, as muscle contraction increases, the number of active motor units also increases. The number of involved Golgi tendon organs and muscle spindles increases, increasing the amount of feedback from the periphery. Alternatively, larger loads would require larger motor commands with larger corollary discharges to the brain's sensory centers. Corollary discharges provide a sense of magnitude of the motor command. 19 Thus, loads lifted for prolonged periods of time (ie, fatiguing contractions) are perceived as heavier despite no actual change in load. Larger corollary discharges may provide greater feedback regarding muscular contraction to the sensory centers. In this case, the greater motor command necessary to produce the 30% target load would produce a larger corollary discharge and may decrease errors made by the sensorimotor system.
Our results suggest that eversion force sense errors at 10% of MVIC were related to 2 different measures of ankle instability: AII6 and giving-way frequency. The relationship between eversion force sense and giving-way frequency was absent when the uninjured ankle was tested, suggesting that the sense of giving way is probably because of a unique mechanism associated with the injured limb. Whether this is the result of damage to peripheral receptors or errors in the central processes is unclear. Future researchers should attempt to delineate which of the processes is primarily responsible.