Immediate effects of the toe spreader on the tonic toe flexion reflex.

Key Words: Electromyography, Hemiparesis, Temporal-distance gait characteristics, Toe spreader, Tonic toe flexion.

Tonic flexion of the toes is an abnormal reflex, which can occur as a result of a cerebrovascular accident or brain injury resulting in hemiparesis.[1-4] Because tonic toe flexion affects gait,[3,5,6] toe spreaders are sometimes used by physical therapists to inhibit tonic toe flexion during stance and gait (Fisher B, Utley J; unpublished notes; 1988).[7,8] Evidence for a positive effect of the toe spreader is needed to justify continued use of the device.

The tonic toe flexion reflex (TTFR) has been referred to by other names, such as "plantar grasp reflex."[9] For the purposes of this study, the term "tonic toe flexion reflex" will be used to describe a hollowing out of the sole and exaggerated curvature of the foot due to toe flexion and adduction with associated foot inversion, usually due to a cutaneous or proprioceptive stimulus.[1-6,10] Tonic toe flexion is considered abnormal after the age of 2 years and indicates damage to the contralateral frontal lobe or tracts descending from the frontal lobe, such as the corticospinal tracts.[1-5,10] The adequate stimulus used to elicit the TTFR varies among studies, but is usually cutaneous or proprioceptive stimuli (eg, muscle strength) applied to the plantar surface of the foot. In previous studies of adults, the stimulus intensity necessary to elicit the TTFR was not controlled.[1-6,10] The TTFR is frequently observed during gait.[3,5,6] During the stance phase of gait, both cutaneous and proprioceptive stimuli applied to the plantar surface of the foot could elicit the TTFR.[3,6] Studies of the TTFR using electromyography (EMG) indicate stimulation of the TTFR during gait may cause inappropriate timing and excessive activity of the intrinsic foot musculature and the toe flexors.[3,5,6] Inappropriate activation of other muscles in the lower extremity (eg, tibialis anterior, gastrocnemius, hamstring, and quadriceps femoris) has also been associated with the TTFR. The reported timing and amount of muscle activation during gait, however, differ among studies.[3,5,6]

Increased curvature of the foot due to the TTFR[5,6] may prevent normal plantar surface contact with the ground and weight transfer from the lateral border of the foot, across the metatarsal heads, to the ball of the great toe during stance.[11-14] The TTFR may also prevent normal dorsiflexion of the toes and foot at the beginning of swing phase.[7,9,11-13] Altered weight bearing and weight shift change the temporal-distance (T-D) characteristics of gait, commonly resulting in decreased stance time on the affected side, decreased contralateral step length, decreased stride length, and a decrease in cadence and velocity.[7,12,15-17]

The effect of the TTFR on gait is often treated by use of a toe spreader, a custom-fabricated orthotic device usually made from moleskin and closed-cell foam (Fisher B, Utley J; unpublished notes; 1988).[7,8] The toe spreader is used to inhibit excess the TTFR and pain (Fisher B, Utley J; unpublished notes; 1988).[7,8] Abduction of the toes by the toe spreader, made of foam rubber, appears to inhibit toe clawing and extensor spasticity (ie, increased resistance to flexion of the proximal interphalangeal joints) of the foot, and often inhibits extensor spasticity of the entire lower extremity in patients with hemiplegia.[7] Although clinical reports indicate positive effects with the use of the toe spreader, data supporting the use of the device have not been reported (Fisher B, Utley J; unpublished notes; 1988).[7,8] Use of the toe spreader, to inhibit the TTFR during standing and gait, may "normalize" muscle activity in the foot and in the lower extremity exhibiting the TTFR, and reduce the increased curvature of the foot caused by the TTFR. Normal plantar surface contact with the ground and normal weight transfer may be restored and may result in improved T-D gait characteristics.[7,12,15-17]

The purpose of our study was to answer the following question: Does use of a toe spreader alter gait characteristics or muscle activity of the lower extremity exhibiting tonic toe flexion during standing and gait in individuals with hemiparesis? The alternative hypotheses for this study were as follows: (1) Absence of the TTFR is related to wearing a toe spreader, and (2) there is a difference in gait characteristics and in integrated electromyographic (IEMG) activity from selected lower-extremity muscles in patients with hemiparesis exhibiting the TTFR between the conditions of toe spreader on and toe spreader off with the shoe on or with the shoe off.

Method

Design

To investigate the effectiveness of the toe spreader in a clinically relevant manner, the toe spreader was used with shoes on. However, we postulated that the toe spreader would reduce the effects of the TTFR, regardless of whether shoes were worn. Therefore, "shoes off" conditions in which TTFR could be visually observed, were included.

The design was a randomized, within-subject, between-condition comparison.

The study consisted of standing and gait phases. Each subject received four conditions for each phase of the experiment. The four conditions were (1) shoe off/toe spreader off, (2) shoe off/toe spreader on, (3) shoe on/toe spreader off, and (4) shoe on/toe spreader on. Each subject was randomly assigned to a sequence of conditions. The same random sequence of conditions was used for both standing and gait for a given subject. All conditions and testing for each subject were performed in one session.

Subjects

Eighteen adults who had hemiparesis secondary to supraspinal lesions and who met the selection criteria served as subjects for the standing portion of the study (9 men and 4 women with right hemiparesis, 3 men and 2 women with left hemiparesis). Two subjects did not complete the gait phase for reasons unrelated to the study. Sixteen of the 18 adults with hemiparesis served as subjects for the gait portion of the study (8 men and 4 women with right hemiparesis, 3 men and 1 woman with left hemiparesis). The characteristics of the subjects are presented in Table 1.

[TABULAR DATA 1 OMITTED]

Subject selection criteria were diagnosis of supraspinal lesion with secondary hemiparesis; exhibits TTFR; ambulatory (categorized as ambulation with or without an assistive device, an orthotic device, or manual assistance); does not exhibit flatfeet; age of greater than 2 years (the age at which the TTFR normally is integrated)[1-5,10]; no previous use of a toe spreader, able to follow commands; shoe covering at least 50% of the dorsum of the foot to hold the toe spreader in place, and the shoe does not interfere with the EMG electrodes on the foot; and medically stable (as indicated by physician's consent for subject participation in the study).

Sampling was nonprobability with voluntary participation by all subjects. All subjects signed an informed consent statement prior to participating in the study. Medical consent was also obtained for all subjects prior to participating in the study.

Measurement

A custom-made toe spreader was fabricated (by LRS), according to specific criteria, for each subject at the start of the experimental session (Fisher B, Utley J; unpublished notes; 1988). During "shoe-on" conditions, the toe spreader was worn inside of, and held in place by, the shoe. During "shoe-off" conditions, the toe spreader was held on the foot by tape on the dorsum of the foot. The toe spreader was donned immediately after the previous trial without the toe spreader and was doffed immediately after the last trial with the toe spreader, according to the random sequence of conditions. Testing was repeated for three trials during each condition. Results of the three trials were averaged for statistical analysis.

Subject characteristics and selection criteria were measured during an initial screening session and confirmed at the start of the experimental session. Age, gender, diagnosis, and etiology of supraspinal lesion with secondary hemiparesis, location of lesion, duration of lesion, side of involvement, previous use of a toe spreader, and physician's consent were determined by review of the medical record or by report from the physician, physical therapist, or subject. Ability to follow commands was determined by asking the patient to successfully complete a specific task when verbally directed.

Ambulatory status was determined by observing the subject walking with shoes on for approximately 9 m. Presence or absence of flatfeet was determined by observing the medial midfoot border while the subject stood barefoot with equal weight bearing for 1 minute.[18] Equal weight bearing was verified by observation. Presence or absence of flatfeet was determined because one exception to the typical pattern of intrinsic foot musculature EMG activity in nondisabled subjects is pronated, or "flat," feet. If the medial border was not in contact with the floor, the subject was classified as non-flat-footed. Presence or absence of the TTFR was determined by observation of the subject while standing barefoot for 1 minute and while walking barefoot approximately 9 m. If tonic plantar flexion and adduction of the toes, with hollowing of the sole and occasional foot inversion, were observed at any time during the observation period, the TTFR was considered present.

The ink footprint method is used to study gait characteristics.[19-21] This method of recording T-D measures of gait (eg, step length, stride length, step width, velocity, cadence)[19-21] has been shown to be simple and inexpensive and to yield valid and reliable measurements in nondisabled subjects.[19] Intrarater and interrater reliability of the ink footprint method of T-D gait analysis were also established in a study of subjects with neurological impairments.[20] Reliability does not appear to be influenced by patient diagnosis or ambulatory status.[20] Thus, we chose to use an ink footprint method for our study.

All T-D gait variables were measured during each of the four conditions of the gait phase of the study. Precut pieces of self-adhesive moleskin were placed on the heel of each bare foot or shoe. Footprints were obtained during each trial and used to analyze the T-D variables of step length bilaterally, stride length bilaterally, velocity, and cadence.[19-21] A computer recorded the walking time. A line was drawn through the most posterior edge of each heel-strike ink mark, with all of the lines being parallel to each other and to the heel-strike ink marks. The perpendicular distance was defined as the length of the perpendicular line between two heel-strike mark lines, as defined for step and stride length. Step length of the involved or uninvolved side was defined as the perpendicular distance from the residual ink marking of the heel-strike of the contralateral foot to the next residual ink marking of the heel-strike of the ipsilateral foot. Stride length of the involved or uninvolved side was defined as the perpendicular distance from the residual ink marking of the heel-strike of the ipsilateral foot to the residual ink marking of the next heel-strike of the ipsilateral foot. Velocity was calculated as the time to ambulate 6 m divided by six. Cadence was calculated as the number of steps per minute. The number of steps taken in 6 m was determined from the residual ink marking on the paper. Each heel-strike mark was counted as one step.[19-21]

Percentage of plantar surface contact was measured during each "shoe-on" condition for the gait phase of the study. The method used for recording plantar surface contact with "shoes on" was developed for this study. Percentage of plantar surface contact was not measured for the "shoe-of" conditions of the gait phase of the study, because application of moleskin applied directly to the ball of the foot might elicit the TTFR. The moleskin bar extended medially from the lateral border of the foot, across the metatarsal heads, to the ball of the great toe.[13,14] The length of the moleskin bar was measured with a metric ruler (possible plantar surface contact) prior to ambulation. The length of each ink marking caused by the moleskin bar of each foot was then measured for each step (actual plantar surface contact). The percentage of possible plantar surface contact was calculated by dividing the length of the residual ink marking from the moleskin bar contacting the paper walkway (the actual plantar surface contact) by the length of the moleskin bar itself (possible plantar surface contact) and then multiplying by 100.

To create a footprint record for measurement of T-D gait variables and plantar surface contact for each trial, ink was applied to each subject's bare feet or shoes prior to ambulation. Different colors of ink were used to distinguish the right and left feet. The subject was then asked to walk at a comfortable speed on a 9-m-long white paper walkway to create the footprint record. Analysis of the ink markings was performed only for the middle 6 m of the paper walkway to allow for acceleration and deceleration changes in gait.

Integrated (100-millisecond time constant) electromyographic (IEMG) activity was measured from the lower extremity exhibiting the TTFR for three test trials during each of the four conditions for both the standing and gait phases. The IEMG signal was processed (200 Hz) by an analog-to-digital converter(*) and stored by a personal computer. The IEMG data were not normalized against a maximal contraction because all comparisons were within-subject comparisons and because of documented problems using maximum contractions for subjects with hemiparesis.[22] Percentage of change in IEMG activity between toe spreader off and toe spreader on conditions was used for statistical testing to decrease variability of values. Pairs of silver-silver chloride surface electrodes([dagger]) were placed over five sites according to procedures described by Delagi et al[23] and Zipp.[24] Electrode locations for both active and ground electrodes were held constant across subjects, measured relative to predetermined and operationally defined bony landmarks on each individual. The sites chosen were the dorsum of the foot to record gross activity of the intrinsic foot musculature, the lateral gastrocnemius muscle, the tibialis anterior muscle, the long head of the biceps femoris muscle, and the vastus medialis muscle. Surface electrode sites were prepared by gently abrading the skin with a sterile needle followed by wiping the abraded area with alcohol. Interelectrode resistance was confirmed with an ohmmeter as being below 5,000 [omega]. The distance between the centers of the active electrodes was maintained at 1.0 to 1.5 cm. Electromyographic activity was measured for each of the four standing shoe on/off-toe spreader on/off conditions. To control the elicitation of the TTFR, the force necessary to elicit the TTFR was monitored by a Kistler force-plate platform,([double dagger]) which was interfaced to an oscilloscope. Gross IEMG recordings from the dorsum of the foot were used in an attempt to quantify the TTFR, and visual inspection was used to verify that the response had indeed occurred.

For the practice trials of standing, the barefoot subject was asked to assume a natural stance with his or her feet a comfortable distance apart, with only the involved foot on the Kistler force plate, which was located within a level walkway. Outlines of the position of the feet on the walkway and platform were traced. For all subsequent trials of standing, the subject was asked to place his or her feet within the outlines. For the practice trials of standing, the subject was instructed to shift his or her weight to the involved side until the TTFR was observed by the investigator. The investigator recorded the voltage output from the Kistler platform as an indication of the amount of force required to elicit the TTFR. The procedure to determine the amount of force required to elicit the TTFR was repeated four additional times. For all subsequent test trials of standing, the subject was instructed to assume the correct standing position, placing his or her feet the same width apart as during the practice trials by placing the feet within the reference foot outlines. The subject began with all of his or her weight shifted to the uninvolved side off of the Kistler force platform. The subject then shifted his or her weight to the involved side on the Kistler force platform until the largest force (voltage output from the platform) required to elicit the TTFR during the practice trials was recorded. This force (voltage output) triggered the computer to sample 5 seconds of EMG activity. The largest force was used to ensure that the stimulus was above the TTFR threshold.

Electromyographic activity during gait and T-D gait variables were measured for all conditions. For gait test trials, each subject was asked to ambulate 9 m. An event marker (a hand-held, DC offset switch) was used to activate/ deactivate computer analysis of EMG recordings for the middle 6 m of gait. A footswitch was taped on the subject's heel to mark the stance phase or swing phase of gait, signaling the computer to integrate swing and stance phase data separately. Raw EMG data were saved on FM tape. Integrated EMG data were stored on computer diskettes.

Interrater reliability of performance of protocols and intrarater/interrater reliability of all measurements were established prior to data collection and verified for every third subject during data collection. Compliance with protocols was increased by having a second investigator observe performance of the protocols by the primary investigator (LRS) during subject data collection. Intrarater and interrater reliability of measurements were maintained at 100% agreement of acceptable values obtained by repeated measurements by the same investigator or by a second investigator. Agreement was required within and between investigators as follows. Subject selection criteria, age, gender, etiology of diagnosis, area and duration of lesion, and presence or absence of TTFR measures agreed exactly. Step length, stride length, and possible or actual plantar surface contact values were [+ or -] 0.5 cm. Time to ambulate values were [+ or -] 2.0 seconds. Number of steps measures were [+ or -] 1 step. Electrode placement accuracy was [+ or -] 2 mm from the center of one electrode to the center of the other electrode. Afl equipment was calibrated before beginning the study and prior to data collection for every third subject.

Procedure

Subjects were randomly assigned to one of eight condition sequences for the standing and gait phases of the study. All subjects received the standing phase first, followed by the gait phase of the study. The study aim and method were explained to each subject at the data collection session. The subject then signed an informed consent statement. The subject was interviewed and evaluated/reevaluated for subject criteria and characteristics, as previously described. The subject then sat for 40 minutes while the toe spreader was fabricated and fitted.

The subject was given instructions on performing the standing tests and performed the practice and test trials of standing, as previously described. The Kistler platform was zeroed, and the EMG electrodes were checked for loose wires prior to and between standing trials. During standing test conditions, the subject rested while standing for 30 seconds between each of the three test trials. The subject sat and rested 3 minutes after the practice trials and after each condition of standing. While the subject rested, the shoe or the toe spreader were applied by the investigator to the involved foot, as appropriate, for the next randomly assigned standing test condition.

After the standing phase of the study, the subject sat and rested for 10 minutes. The subject then performed one practice trial of gait in his or her bare feet on the paper with all of the apparatus operational except the ink on the moleskin. Each subject walked 9 m with the use of an assistive device, if needed. The subject sat for a 5-minute rest period after the practice walk. While the subject rested before each gait condition, the shoe or the toe spreader were applied by the investigator to the involved foot, as appropriate, for the next randomly assigned gait test condition. Ink-saturated moleskin was applied to the bottom of the subject's shoe or bare heel for the footprint analysis. The paper was changed and saved, all of the EMG electrodes were checked for loose wires, and the ink on the moleskin was replenished prior to and between the gait trials. Between gait conditions, the subject sat for a 10-minute rest period. Within each condition, between each of the three test trials of gait, the subject sat and rested for 5 minutes. Throughout the session, all rest periods were timed with a stopwatch and held constant. During the shoe-off conditions, the TTFR was recorded as present or absent as determined by visual observation by the primary investigator. After all four conditions of gait, the toe spreader, moleskin, and EMG electrodes were removed from the subject, and the subject was free to leave. The investigator contacted the subject and/or the subject's physical therapist after the data collection session for their comments regarding the effects of the toe spreader on the TTFR and gait after continued use of the toe spreader.

Data Analysis

Data for plantar surface contact and T-D gait variables were collected for the 16 subjects who participated in the gait portion of the study. Electromyographic data were lost during the standing phase of the study for 1 subject due to computer malfunction. Electromyographic data for the intrinsic foot musculature and the tibialis anterior and gastrocnemius muscles were retrieved from the FM tape. Electromyographic data were lost during the gait phase of the study for 5 subjects due either to mechanical problems with the footswitch or computer malfunctions. For 4 of these subjects, however, EMG data for the intrinsic foot musculature and the tibialis anterior muscle were retrieved from the FM tape.

Data from the three trials in each condition were averaged for each of the T-D gait variables, plantar surface contact, and IEMG activity. Means and standard deviations were calculated per variable for each condition. Percentage of change values (means and standard deviations) for all variables were calculated for toe spreader on/ toe spreader off within the shoe-on and shoe-off conditions. Percentage of change was calculated as follows: Percentage of change =

(value for toe spreader on -

value for toe spreader off)

x (100/value for toe spreader off)

To reduce variability within and between subjects, only the percentage of change values were used in statistical analysis of the IEMG data.

The association between use of the toe spreader and visual observation of presence or absence of the TTFR within the shoe-off condition was determined using a McNemar's test. The McNemar's test is a test of equality of proportions in a 2 x 2 table (of two dichotomous variables) for a correlated sample.[25] In this study, the two dichotomous variables were the presence or absence of the TTFR and the toe spreader on/of condition. The effects of the shoes, the toe spreader, and the interaction between shoe and toe spreader on mean values of T-D gait variables and plantar surface contact during all four conditions of standing and swing and stance phases of gait were compared using a two-way, repeated-measures analyses of variance and Tukey's post hoc, multiple-comparison procedure for each variable. The significance of percentage of change in each gait variable and in IEMG activity of each muscle from toe spreader on to toe spreader off within each shoe condition was determined using a paired t test with the Bonferroni correction for multiple tests.[26] The normality of distributions was determined for each statistical test. If the assumption of normality was not met, a Wilcoxon rank sum for matched pairs was used. All tests were two-tailed, with a criterion alpha level of .05. Calculation of power was difficult because of the lack of a body of knowledge in the literature to use as a basis for determining significant change. Values of clinically significant change have not previously been reported for any of the variables studied. For this reason, the obtained means and standard deviations, the number of subjects, and an alpha level of .05 were used to calculate power.[26] The power values were [greater than or equal to] .30 for each test.[26] These power values are low and make inferences regarding findings prone to a Type II error. All computational techniques for conversion of raw data to usable form and statistical analyses of data were performed on a calculator or by computer.[27]

Results

The association of the toe spreader and presence or absence of the TTFR during gait was statistically significant ([[chi].sup.2] = 15, df = 1, P = .0001). Absence of the TTFR was defined as absence of toe flexion and adduction of the toes, with associated inversion of the foot and a hollowing out of the sole. Based on visual observation during standing in the shoe-off conditions, the TTFR was absent in all subjects when the toe spreader was on and present in all subjects when the toe spreader was of during gait. During gait in the shoe-off condition, the TTFR was absent in all subjects except one with the toe spreader on and was present in all subjects with the toe spreader off.

Gait Characteristics

Means and standard deviations for T-D variables and for percentage of change between toe spreader conditions for T-D measurements are presented in Table 2. Mean values for step length and stride length on both the involved and uninvolved sides were significantly (P[less than or equal to].05) greater with the shoes on than with the shoes off for both toe spreader conditions (Tabs. 3 and 4; Figs. 1A and 1B). Step length percentage of change between toe spreader conditions for the uninvolved leg was statistically significant in the shoe-on condition (Tab. 5).

[TABULAR DATA 2, 3, AND 4 OMITTED]

Table 5. Statistically Significant
Results of Paired t-Test and Wilcoxon
[Test.sup.a] (S) Comparisons of Percentage of
Change in Gait Variable Values During
Toe Spreader On Versus Toe Spreader Off
Conditions
Variable                 t        S       P
Step length
  Univolved: shoe on    2.15            .05
Velocity
  Shoe on               3.03           .008
Cadence
  Shoe off                       48     .01
  Shoe on               4.15           .001
(a) Wilcoxon rank sum test rsult reported only
in instances in which assumption of normal
distribution not met for the t test.

Mean values for velocity were significantly (P[less than or equal to] .05) greater with the shoes on than with the shoes off for both toe spreader conditions (Tabs. 3 and 4, Fig. 1C). The main effect of the toe spreader was statistically significant for velocity (Tab. 3, Fig. 1C). Percentage of change in velocity from toe spreader on to toe spreader off was statistically significant for the shoe-on condition (Tab. 5).

The main effect of the toe spreader was statistically significant for cadence (Tab. 3, Fig. 1D). Percentage of change in cadence from toe spreader on to toe spreader off was statistically significant for both the shoe-off and shoe-on conditions (Tab. 5). Comparison of mean and percentage of change values of plantar surface contact were not statistically significant.

Integrated Electromyographic Activity Characteristics

Examples of raw EMG activity during standing and gait phases of the study are shown in Figure 2. Means and standard deviations for percentage of change in IEMG activity from toe spreader off to toe spreader on are presented in Table 6. Percentage of change in IEMG activity was not statistically significant for either shoe condition.

[TABULAR DATA 6 OMITTED]

Discussion

Visual observation indicated absence of the TTFR with the toe spreader on (shoes off). In addition, use of the toe spreader during gait significantly increased gait velocity with the shoes on and cadence in both shoe conditions. Finally, use of shoes significantly increased bilateral step length and bilateral stride length, regardless of the toe spreader condition. Interpretation of these findings must be cautious because in each instance the magnitude of the difference in actual values or in percentage of change was less than the corresponding standard deviation.

Gait Characteristics

The results of the effects of the shoe and toe spreader on gait are similar to clinical observations (Fisher B, Utley J; unpublished notes; 1988).[7,8] No data for comparison, however, exist. Step length and stride length on the involved and uninvolved sides both increased significantly when the subjects' shoes were on versus when the shoes were off. The shoes may give added mechanical support and stability to the foot during the stance phase. This increased stability may, in turn, allow greater excursion of the non-weight-bearing limb. Another possibility is that the shoes may offer proprioceptive and cutaneous inputs, which inhibit the TTFR.

Both velocity and cadence were increased significantly with the toe spreader and shoe on. The toe spreader has been effective in inhibiting excess tone and pain in patients with a separate toe grasp reflex.[7,8] Altered foot mechanics can also affect cadence and velocity.[7,12,15-17] The toe spreader appears to mechanically position the metatarsophalangeal joints of the foot in neutral to slight extension and restricts the interphalangeal joints from going into extreme flexion. Thus, the toe spreader may have influenced velocity and cadence by altering the mechanics of the foot, rather than by altering IEMG activity, or by some combination of these factors. The clinical significance of the magnitude of the increases in velocity and cadence is difficult to determine and is perhaps dependent on the functional requirements of the individual patients. An increase in walking rate, however, is generally associated with improved function.

Use of the toe spreader did not significantly affect plantar surface contact during gait with shoes on. Presence or absence of the TTFR with shoes on was not determined in our study. If the TTFR did occur when shoes were worn, it did not appear to limit plantar surface contact (from the lateral border of the foot, across the metatarsal heads, to the ball of the great toe[13,14]), as shown by the high percentages of plantar surface contact (Tab, 2). The method of measuring plantar surface contact in this study, however, is questionable. The measurement of only contact with the ground by the plantar surface does not represent the actual force of weight transfer across the foot during stance. The subjects still may have altered weight bearing and weight shift due to the TTFR. These alterations may influence the T-D characteristics of gait.[7,12,15-17] Strength of the stimulus during the standing phase of the study was controlled (ie, weight-bearing force, as measured on the Kistler platform). The strength of the TTFR stimulus during the gait phase of the study (ie, amount of weight bearing or stance time for the lower extremity exhibiting the TTFR), however, was not controlled. In future studies, the amount or patterns of weight bearing through the foot during the stance phase may be correlated with the strength of the stimulus that elicits the TTFR or the strength of the TTFR. Changes in the actual translation or amount of force across the plantar surface during the stance phase of gait with use of the toe spreader may also be investigated in future studies.

Use of the toe spreader may be necessary for an extended period of time to realize the maximal effect on gait, All but two subjects continued to use the toe spreader periodically for at least 4 months. All subjects who continued to use the toe spreader believed that the toe spreader had allowed them to improve gait. Seven subjects reported a subjective decrease in the strength of the TTFR during gait without the toe spreader after having worn the toe spreader for varying periods of time ranging from 1 week to 5 months. For the subjects receiving gait training with the toe spreader, all of the physical therapists except one reported substantial positive changes in gait. Future study of repeated or extended use of the toe spreader, or use of the toe spreader in conjunction with gait training, may yield clinically relevant information.

Integrated Electromyographic Activity. Standing and Gait

None of the changes in IEMG were statistically significant. The finding of certain significant effects of the toe spreader on gait characteristics but not on IEMG activity may be due to several factors. A large sample size may be necessary to demonstrate statistically significant changes in IEMG activity secondary to use of the toe spreader. Electrode placement also may not have allowed recording from the appropriate part of the specific muscles, though documented placement sites[23,24] were used. Certainly, the placement of the electrode on the dorsum of the foot was not specific to the motor point(s) of the intrinsic foot musculature. At times, subjects did not have sufficient weight on the heel to trigger the heel switch for the entire stance phase. As a result, EMG activity during gait may not accurately represent the separation of the stance and stride phases. Finally, any effect of the toe spreader may be primarily or solely mechanical. These mechanical effects may not affect muscle activity. Regardless, the IEMG results are inconclusive.

Conclusions

The efficacy of the toe spreader in inhibiting the TTFR and in improving gait is usually determined in the clinic while the patient is wearing shoes. The results of our study indicate that wearing shoes without the toe spreader may positively affect gait characteristics, such as velocity and step or stride length. Use of the toe spreader with shoes on also may increase gait velocity and cadence and, thus, may be a useful treatment option to improve gait. Research on the effects of repeated or extended use of the toe spreader or other therapeutic interventions in combination with the toe spreader is recommended. (*) Keithley DAS, 349 Congress St, Boston, MA 02210, ([dagger]) Beckman Instruments Inc, 1630 S State College Blvd, Anaheim, CA 92806. ([double dagger]) Type 9218, Kistler Instrumente HG, CH-8408 Winterthur, Switzerland.

References

[1] Brain W. The grasp reflex of the foot. Med Clin (Barc). 1950;15:293-296. [2] Brain W, Curran R. The grasp reflex of the foot. Brain. 1932;55:347-356. [3] Cohen L, Iannone A. The tonic foot response. Arch Neurol. 1967;17:419-426. [4] Goldstein K. The tonic foot response to stimulation of the sole: its physiological significance and diagnostic value, Brain. 1938;61: 269-283. [5] Landau W, Clare M. Pathophysiology of the tonic innervation phenomenon in the foot. Arch Neurol. 1966;15:252-263. [6] Manfredi M, Sacco G, Sideri G. The tonic ambulatory foot response: a clinical and electromyographic study. Brain. 1975;98:167-180. [7] Bobath B. Adult Hemiplegia: Evaluation and Treatment 2nd ed. London, England: William Heinemann Medical Books Ltd; 1978. [8] Ford C, Grotz R, Shamp J. The neurophysiological ankle-foot orthosis. Clinical Prosthetics and Orthotics. 1986;10:15-23. [9] Bobath B. Abnormal Postural Reflex Activity Caused by Brain Lesions. 3rd ed. Rockville, Md: Aspen Publishers Inc; 1985. [10] Richter C, Hines M. Experimental production of the grasp reflex in adult monkeys by lesion of the frontal lobes. Am J Physiol. 1932; 101:87-88. [11] Kapandji I. The Physiology of the Joints, Volume 2. 2nd ed. London, England: Churchill Livingstone; 1970. [12] Perry J. The mechanics of walking in hemiplegia. Clin Orthop. 1969;63:23-31. [13] Woodburne R, Burkel W. Essentials of Human Anatomy. 8th ed. New York, NY: Oxford University Press Inc; 1988. [14] Williams P, Warwick R. Gray's Anatomy. 36th ed. Philadelphia, Pa: WB Saunders Co; 1980. [15] Brunnstrom S. Recording gait patterns of adult hemiplegic patients. Phys Ther. 1964;44: 11-18. [16] Brunnstrom S. Movement Therapy in Hemiplegia: A Neurophysiological Approach. Philadelphia, Pa: Harper & Row; 1970. [17] Lehmann J, Condon S, Price R, deLateur B. Gait abnormalities in hemiplegia: their correction by ankle-foot orthoses. Arch Phys Med Rehabil. 1987;68:763-771. [18] Gray E, Basmajian J. Electromyography and cinematography of leg and foot ("normal" and flat) during walking. Anal Rec. 1968;161:1-16. [19] Boenig D. Evaluation of a clinical method of gait analysis. Phys Ther. 1977;57:795-798. [20] Holden M, Gill K, Magliozzi M, et al, Clinical gait assessment in the neurologically impaired: reliability and meaningfulness. Phys Ther, 1984;64:35-40. [21] Holden M, Gill K, Magliozzi M. Gait assessment for neurologically impaired patients: standards for outcome assessment. Phys Ther. 1986;66:1530-1539. [22] Visser S, Aanen A. Evaluation of EMG parameters for analysis and quantification of hemiparesis. Electroencephalogr Clin Neurophysiol. 1981;21:591-610. [23] Delagi E, Iazzetti J, Perotto A, Morrison D. Anatomic Guide for the Electromyographer: The Limbs. Springfield, Ill: Charles C Thomas, Publisher; 1981. [24] Zipp P. Recommendations for the standardization of lead positions in surface electromyography. Eur J Appl Physiol. 1982;50:41-54. [25] Marascullo L, MacSweeny M. Nonparametric and Distribution-free Methods for the Social Sciences. Monterey, Calif: Brooks/Cole Publishing Co; 1977. [26] Zar J. Biostatistical Analysis. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall; 1984. [27] SAS/STAT User's Guide: Version 6, Volume 2. 4th ed. Cary, NC: SAS Institute Inc, 1989.

LR de Saca, PT, is a contract physical therapist. This work was completed in partial fulfillment o the requirements for her Master of Medical Science degree, Division of Physical Therapy, Department of Rehabilitation Medicine, Emory University School of Medicine, Atlanta, Ga 30322.

PA Catlin, EdD, PT, is Professor and Director, Division of Physical Therapy, Department of Rehabil Medicine, Emory University School of Medicine.

RL Segal, PhD, Pr, is Associate Professor, Division of Physical Therapy, Department of Rehabilitat Medicine, and Assistant Professor, Department of Anatomy and Cell Biology, Emory University School of Medicine, 1441 Clifton Rd NE, Atlanta, Ga 30322 (USA). Address correspondence to Dr Segal.

This study was approved by the Human Investigations Committee of Emory University School of Medicine.

This article was submitted July 20, 1992, and was accepted January 6, 1994.

Acknowledgments

We thank James Hudson for his advice and help with equipment; George Cotsonis for statistical consultation; Janet Delk for use of her computer; Nancy Heter for word processing; and Terrie Millard, Terri Reynolds, Sue Jacobs, and Gail Williams for their help with reliability testing.