- Email: [email protected]
Light touch and center of mass stability during treadmill locomotion
Gait and Posture 20 (2004) 41–47
Light touch and center of mass stability during treadmill locomotion Ruth Dickstein∗ , Yocheved Laufer Department of Physical Therapy, Faculty of Social Welfare and Health Sciences, University of Haifa, Mount Carmel, Haifa 31905, Israel Received in revised form 23 May 2003; accepted 16 June 2003
Abstract Purpose: To study the contributions of light fingertip touch on an earth-referenced object to body stability during treadmill locomotion. Method: Twenty young healthy adults were tested in two blocks of five testing conditions while walking on the treadmill at 3 km/h. In each condition, subjects were tested with eyes open (EO) and with eyes closed (EC). In each block, four separate conditions of heavy (H) or light (L) touch to either a left or to a right force sensor mounted on the respective side rail, as well as one condition of no touch (N), were randomly applied. The 3D positions of the center of mass (COM) and the midpoint of the posterior aspect of each leg were monitored via a kinematic ultrasonic system, while the anterior–posterior (AP) acceleration of the COM was measured with a uniaxial linear accelerometer. Results: Light touch had a similar stabilizing effect as vision and as heavy touch on COM sway. Thus, COM sway and AP acceleration were comparable in conditions of eyes open and eyes closed as long as touch was applied. Conversely, without vision and touch, subjects drifted backwards, with complete disruption of the coordinated stepping pattern. Conclusions: Somatosensory fingertip input from an external reference provides spatial orientation, which, similar to vision, enables the sustaining of body stability during treadmill walking. © 2003 Elsevier B.V. All rights reserved. Keywords: Light touch; Center of mass; Locomotion
1. Introduction Control of body balance is known to be largely dependent on sensory information from the visual, vestibular and lower limb somatosensory inputs [1]. In addition, during stance, somatosensory cues elicited through light fingertip touch of an external object can provide perceptual information that contributes to balance control. In the absence of visual input, light fingertip touch cues from an external reference can exert a stabilizing effect comparable to the effect of vision [2–7]. Similarly, light touch (L) can reduce the body sway of subjects with balance disorders emanating from vestibular deficits or from feet neuropathy [8,9]. Balance control is one of the main tasks required for purposeful and safe gait. Its completion is integrated with the execution of stepping and forward propulsion, as well as with the upholding of anti-gravitational posture [10]. The purpose of the present study was to examine the contribution of somatosensory touch inputs to body balance dur∗ Corresponding author. Present address: Department of Physical Therapy, Sackler Faculty of Medicine, Tel-Aviv University, Ramat-Aviv, Tel-Aviv 69978, Israel. Tel.: +972-4-824-9065; fax: +972-4-828-8140. E-mail address: [email protected] (R. Dickstein).
0966-6362/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0966-6362(03)00091-2
ing treadmill locomotion. The availability of a side or front rail on the treadmill makes the use of touch inputs easily accessible. Observation of subjects during treadmill walking, especially at low to moderate speeds, confirms occasional or constant touch of one or both rails. It could well be that these maneuvers are used for improving/correcting body balance. The reduction of physiological stress achieved by lightly touching the treadmill rail [11,12] may be related to the facilitation of balance control. However, biomechanical treadmill gait studies have generally failed to focus on whether or not subjects touched the rail. An exception is the study of Siler and associates, who demonstrated that sagittal plane kinematic parameters were not affected by grasp of the treadmill handrail [13]. Nevertheless, subjects’ balance during walking cannot be inferred from their results. Treadmill walking may well challenge the control of balance because subjects must adapt the execution of forward walking movements to zero optic flow and must use visual cues as an anchor for maintaining a stable walking base. Without vision, one is unable to distinguish between self-motion and treadmill motion or to determine how one is moving on the treadmill relative to the static environment [14]. Therefore, closing the eyes causes subjects to drift backward and to step off the treadmill [15,16].
42
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
2. Method
Measurements started after 5 min of familiarization, with subjects walking at a constant speed of 3 km/h (0.83 m/s). In view of the fact that the preferred walking speed of active adults is about 1.5 m/s [13], the speed chosen here can be considered as moderate to low. Immediately preceding each testing condition, the examiner informed subjects of the touch and side specifications involved. During the “L” testing conditions, subjects were asked to put their index finger on the force sensor without pressing on it. Pressure exceeding 200 g along any of the axes activated a warning signal alerting the subject to reduce pressure. During the “H” testing conditions, subjects were allowed to firmly press their index finger against the sensor. During the “N” conditions, they were asked to walk with their arms at the side of their body. Data collection was started when the subject performed the specific walking task fluently. Each testing condition was divided into three consecutive phases that differed in visual input, with each phase lasting 30–40 s: (1) walking with eyes open (EO); (2) walking with eyes closed (EC); and (3) walking with eyes open again (EOA). Subjects were notified to close their eyes (phase 2) and to reopen them (phase 3) by a high-pitched computer beep, which was simultaneously recorded in the data files. In the “no touch” conditions, phase 2 (EC) was always substantially shorter than 30 s and phase three (EOA) was largely absent because subjects walked backwards off the treadmill, requiring termination of testing. Subjects were allowed to rest between blocks, while conditions within each block were applied uninterruptedly with a negligible break time between them.
2.1. Sample and procedure
2.2. Data analysis
Participants were twenty university students, including 7 males and 13 females, with an average age of 23.7 (±2.6). They had no known health impairments, and all but three were right-handed (by self-report). Kinematic data were collected via an ultrasonic-based system (Vscope Ltd.), with mini-ultrasonic transducers attached to the center of the back at the level of the L5 spinuous process and to the midpoint of the posterior aspect of each calf. A uniaxial linear accelerometer (Zetra, Baxborough, MA, model 141), sensitive to accelerations along the anterior–posterior (AP) axis, was attached to the center of the back immediately below the ultrasonic transducer. Two triaxial force transducers were mounted on parallel spots on the left and right rails of the treadmill (height of 90 cm), positioned somewhat forward of the subject’s trunk. Subjects wearing shorts and sport shoes were tested during two blocks of five conditions applied in a random order (a computer random numbers generating program was used to randomize the order of testing conditions within each block). During each testing condition, either heavy (H), or light (L) touch was applied by the index finger to either the left or right rail; additionally one “no touch (N)” condition was tested.
Analysis of each testing condition was performed for two periods: (1) eyes open and (2) eyes closed. Data from the “eyes open again” period were not subjected to statistical analysis. In order to guarantee stable walking in each visual condition, all data collected during the first 5 s of each phase in each testing condition were excluded. Thereafter, 10 step cycles were delimited for analysis. The only exception was the condition of no touch with eyes closed, in which case only four to five steps were available for analysis, as further explained in the results. COM sway, which reflects the stability of the COM position was determined by the variance around the mean position of the COM transducer along the anterior–posterior and medio-lateral (ML) direction, and by the range (peak to peak) of movement of the COM along the anterior–posterior and medio-lateral axes for the same time period. Additionally, the mean rising slope of forward acceleration in each step was determined for the same time period. In order to characterize the effects of touch on the relative position of each leg with respect to the other leg, the linear correlation between the AP location of the two mid-calf transducers was calculated.
Nomenclature COM
center of mass
Touch modes H heavy touch L light touch N no touch Visual conditions EC eyes closed EO eyes open EOA eyes open again Sway direction AP anterior–posterior ML medio-lateral
Since the role of vision in treadmill walking is well understood, as is the contribution of light touch to body balance during stance, the present research was designed to complement current knowledge by focusing on the contribution of fingertip touch information to body balance during treadmill gait. In addition, separate tests were performed with touch cues provided by either the left or the right index fingertip in order to determine the effect of the side of the sensory input on balance control.
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
The distribution of the data was normalized by using the log of the raw values in the statistical analysis. Correlation values were Fisher z-transformed prior to statistical analysis. Unbalanced ANOVA [17] was used for analysis, applying two fixed effects (i.e. touch and vision) and one random effect (i.e. subject). This was followed by the Duncan post-hoc multiple comparisons procedure. ANOVA was also used for calculating the effect of side in conditions of heavy and of light touch. For descriptive purposes, raw values of the collected data were used.
3. Results 3.1. AP COM sway Mean values of AP COM variance are delineated in Fig. 1. The substantial increase observed in COM sway under the
43
condition of no touch with eyes closed was associated with backward movement of the whole body. It was quantified, however, for only four to five gait cycles because thereafter further movement of the treadmill had to be arrested for safety reasons, as subjects drifted backwards towards the treadmill edge. This phenomenon did not occur when touch was applied in the EC condition or in any of the EO conditions. The effects of touch and of vision on COM sway were significant (F = 33.5, P < 0.001 and F = 283.8, P < 0.001, respectively), as was the interaction between these effects (F = 75.5, P < 0.001). Using Duncan post-hoc multiple comparisons analysis, it was found that the effects of light and heavy touch on AP sway were not different from each other in decreasing AP sway and were similar to the effect of vision. However, the effect of no touch on AP sway was different, as sway was significantly increased by the combination of no touch and closure of the eyes (P < 0.05). The results also pointed to an interaction between the effects
180.0 160.0 140.0
EO EC
cm
120.0 100.0 80 .0 60 .0 40 .0 20 .0 0.0 (a)
HT
Heavy Touch
LT
NT
Light Touc h
No Touc h
13.5
Log (VarZ)
13.0 12.5 12.0 11.5 11.0 10.5 10.0 9.5 (b)
Eyes Open
Eyes Closed
Fig. 1. Mean (S.E.M.) AP sway (COM variance) under two visual (EO, EC) and three touch conditions (H, L, N) (a), and the effect of the interaction between touch and vision on AP sway (b).
44
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
of touch and subject (F = 2.8, P < 0.001), indicating that subjects differed in their usage of touch cues for stabilization. Values of sway range were more variable than were those of sway variance. Still, the statistical analysis yielded similar results: a significant effect of touch and of vision (F = 16.7, P < 0.001 and F = 289.9, P < 0.001, respectively), with a differential effect of light and heavy touch as compared to no touch (Duncan multiple comparisons, post-hoc analysis, P < 0.05). Once more, the interaction between the effects of touch and subject was significant (F = 4.1, P < 0.001), implying a non-uniform contribution of touch to different subjects. 3.2. ML COM sway Mean values of COM variance are outlined in Fig. 2. Although the effect of touch was less pronounced in the ML than in the AP direction, the overall effects of touch and of vision on COM variance were significant (F = 7.7, P < 0.002 and F = 28.7, P < 0.001, respectively). The interaction between these effects (F = 7.3, P < 0.001) pointed to greater sway under no touch conditions, particularly so when
Table 1 Slopes of COM forward acceleration during treadmill walking with heavy or light touch Touch mode
Slope (cm/s2 /s) mean
Slope (cm/s2 /s) standard deviation
Heavy Light
1.03 0.94
0.33 0.32
the eyes were closed. The effects of heavy and light touch were comparable to that of vision in decreasing ML sway, while the effect of no touch was again different (P < 0.05). Subjects’ effect was significant for both touch (F = 3.8, P < 0.001) and vision (F = 4.8, P = 0.000), again implying variance in use of these sensory inputs by different people. Similar results were found for the range of ML sway. 3.3. Acceleration of COM COM acceleration was measured in the AP plane only. Mean values of the rising slope of AP acceleration with either heavy or light touch are presented in Table 1. As long as either heavy or light touch was available, vision had no
14.00 12.00 10.00 EO
8.00
EC
6.00 4.00 2.00 0.00
(a)
HT
LT
NT
.8 11
Heavy Touch
Light Touch
No Touch
11.6
Log (VarY)
11.4 11.2 11.0 10.8 10.6 10.4 (b)
Eyes Open
Eyes Closed
Fig. 2. Mean (S.E.M.) ML sway (COM variance) under two visual (EO, EC) and three touch conditions (H, L, N) (a), and the effect of the interaction between touch and vision on ML sway (b).
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
significant effect on COM acceleration. However, the main effect of heavy touch was associated with a steeper increase in acceleration than was light touch (F = 9.6, P < 0.002). With no touch and no vision, acceleration slopes were the highest, though they could be measured for only a few seconds under these conditions due to the need to terminate testing as a result of subjects’ backward drift. Significant individual differences were noted between the slopes of the participants (F = 12.3, P < 0.001). 3.4. Side The side of the touched sensor had no significant effect either on COM sway (F = 0.024, P = 0.88 and F = 0.86, P = 0.36, for AP and ML sway, respectively), or on COM forward acceleration (F = 1.45, P < 0.23). Yet, as shown below, the force applied by the subjects’ right or left index finger differed. 3.5. Mean forces Mean forces were directed downward, forward, and to the right. The difference between heavy and light forces was significant (F = 224.2, P < 0.001; F = 171, P < 0.001; and F = 79, P < 0.001 for the vertical, AP, and ML axes, respectively), with vision having no effect on the magnitude of these forces in either direction. Significant individual differences were observed in the amount of force applied by subjects during downward light touch (F = 8.0, P < 0.001), as well as for AP forces (F = 5.2, P < 0.001). Still, complete differentiation between light and heavy touch was maintained by all participants. Regarding the impact of side, the vertical force applied to the right sensor was significantly greater than that applied to the left sensor during both heavy and light touch (F = 160.4, P < 0.001) (see Fig. 3). 3.6. Linear correlation Linear correlation between the AP position of the legs was calculated from the mid-calf markers. All correlations were 350 300
grams
250 200 150 100 50 0 EO
EC HLt
HRt
EOA LLt
LRt
Fig. 3. Index fingertip force applied during heavy and light touch on the right- and left-sided sensors.
45
negative, as consistent with the reciprocal AP movements of the legs during treadmill walking. The mean between-legs correlation with eyes open and with either heavy, light or no touch was −0.78, while the corresponding mean value with eyes closed and with either heavy or light touch was −0.69. Thus, closing the eyes was associated with a significant reduction in between-legs correlation (F = 170.9, P < 0.001). The decrease in the correlation between-leg positions was especially detrimental upon closing the eyes in the no touch condition (mean r = −0.06; F = 45.5, P < 0.001), pointing to derangement of the reciprocal gait pattern.
4. Discussion The main finding of this study is that touching an external object, even with a very light fingertip touch, provides treadmill walkers with a somatosensory anchor, which, similar to the visual anchor, permits spatial orientation and reduces body sway. Apparently, the similarities between neural control mechanisms of the two systems, with the operation of neuronal networks that involve vision and/or touch for spatial attention [18,19], account for their interchangeable role in balance control. The findings agree with those of previous research demonstrating that light touch has a comparable effect to that of vision in decreasing postural sway during stance [4,7,8,20,21]. The fact that either vision or touch (even a very light touch) of an external reference is needed for the optimal match of walking speed with velocity of the treadmill, indicates their inevitability for spatial orientation during treadmill locomotion. At the 3 km/h velocity that was applied here, the stabilizing effect of light touch was more enhanced in the AP than in the ML direction. This difference stems, most probably, from greater AP than ML destabilization in the tested velocity. Studies of stance stability have also indicated that the stabilizing effect of light touch cues are plane-specific, preferentially decreasing either ML or AP sway during tandem or normal bipedal stance, respectively [2,3]. However, the effect of touching a front rather than a side rail on COM sway during treadmill locomotion remains to be determined. Furthermore, as the relationship between ML and AP sway of the COM is speed-dependent, a desired complement to the current findings may be obtained by testing the effects of light touch on COM stability during a range of walking speeds. The results similarly show that the same orientation reference (i.e. vision or touch) is necessary for controlling forward COM acceleration, as well as for coordinating leg movements during treadmill gait. The normal AP coordination between the legs was somewhat reduced by the elimination of vision, but was completely disrupted due to simultaneous elimination of both vision and touch. This observation that kinematic patterns of gait are altered in the absence of vision or when there is inconsistent optic flow has also been made in previous studies [22,23].
46
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
Observation of treadmill walking among healthy subjects indicates that touch of the side rail is mainly used at slow and moderate speeds, while as walking shifts to running (about 7–8 km/h), subjects usually adduct their arms and flex their elbows, moving them in coordination with the movements of the lower extremities. Therefore, the current results are mainly applicable to slow treadmill walking, as frequently applied in rehabilitation. Here, touch cues can be used instead of, or in addition to vision for compromising balance difficulties. However, generalization of the findings to walking over ground with a cane could be misleading because neither can the treadmill belt be compared to a stationary floor nor can the touch of a stationary rail be judged against a cane that is moved by the walking subject him/herself. This notion is supported by previous findings that postural sway among standing subjects was decreased by touching an earth-fixed plate and increased by touching a sway-referenced plate [24]. In the current study, the side of the handrail that was touched had no statistical effect on the outcome variables; yet, the force applied by the participants to the right side rail was greater than that applied to the left side rail. Perhaps the superior haptic and motor abilities of the left hemisphere (dominant hand) could explain its “enhanced” use in comparison to the contralateral side [25]. These findings might have practical implications when considering the appropriate side for a rigid support provided to walking patients. However, further inquiry is needed on this issue. When considering the current findings, one should be aware of the fact that the data collected pertain only to translational movements of the body, whereas information on orientation and/or configuration of the body [26] is unidentified. Furthermore, the application of only one marker to represent COM location could undermine the accuracy of the measurements. Yet, the validity of that approach has previously been approved [27]. Additionally, the findings pertaining to the body COM should be supplemented by examining the contribution of touch cues to head stability. Head stability during treadmill locomotion has been characterized by fluctuations in the period and amplitude of motion in the sagittal plane across walking cycles. A dynamical modulation of head movements related to the phases of the gait cycle and aimed at minimizing force variations and increasing visual fixation has recently been proposed [28,29]. These considerations may be further enhanced by examining the effect of touch. In view of the above evidence, it appears necessary to consider the potential effects of vision and/or touch conditions on gait parameters in biomechanical treadmill studies. Accordingly, it is essential that visual and touch conditions be explicitly reported in the description of such research. The importance of the current analysis extends even further through its disclosure of significant individual differences, pointing to additional factors besides touch and vision that
may affect body stability during treadmill gait. Thus, further investigation is needed to shed more light on these factors as well.
References [1] Horak FB, Macpherson JM. Postural orientation and equilibrium. In: Shepard J, Rowell L, editors. Handbook of physiology. New York: Oxford University Press; 1996. p. 255–92. [2] Clapp S, Wing AM. Light touch contribution to balance in normal bipedal stance. Exp Brain Res 1999;125(4):521–4. [3] Jeka J, Lackner JR. Fingertip contact influences human postural control. Exp Brain Res 1994;100:495–502. [4] Jeka J, Oie K, Schoner G, Dijkstra T, Henson E. Position and velocity coupling of postural sway to somatosensory drive. J Neurophysiol 1998;79(4):1661–74. [5] Jeka J, Oie KS, Kiemel T. Multisensory information for human postural control: integrating touch and vision. Exp Brain Res 2000;134(1):107–25. [6] Lackner JR, Rabin E, DiZio P. Stabilization of posture by precision touch of the index finger with rigid and flexible filaments. Exp Brain Res 2001;139(4):454–64. [7] Rogers MW, Wardman DL, Lord SR, Fitzpatrick RC. Passive tactile sensory input improves stability during standing. Exp Brain Res 2001;136:514–22. [8] Lackner JR, DiZio P, Jeka J, Horak F, Krebs D, Rabin E. Precision contact of the fingertip reduces postural sway of individuals with bilateral vestibular loss. Exp Brain Res 1999;126(4):459–66. [9] Dickstein R, Shupert CL, Horak FB. Fingertip touch improves postural stability in patients with peripheral neuropathy. Gait Posture 2001;14(3):238–47. [10] Winter D. The biomechanica and motor control of human gait. Waterloo, Canada: University of waterloo press; 1991. [11] Manfre MJ, Yu GH, Varma AA, Mallis GI, Kearney K, Karageorgis MA. The effect of limited handrail support on total treadmill time and the prediction of VO2 max. Clin Cardiol 1994;17(8):445–50. [12] Christman SK, Fish AF, Bernhard L, Frid DJ, Smith BA, Mitchell L. Continuous handrail support, oxygen uptake, and heart rate in women during submaximal step treadmill exercise. Res Nurs Health 2000;23(1):35–42. [13] Siler WL, Jorgensen AL, Norris RA. Grasping the handrails during treadmill walking does not alter sagittal plane kinematics of walking. Arch Phys Med Rehabil 1997;78(4):393–8. [14] Lee DN. Getting around with light and sound. In: Warren K, Wertheim AH, editors. Perception and control of self motion. Hiilsdale, New Jersey: Lawrence Erlbaum Associates; 1990. p. 487– 505. [15] Durgin FH, Pelah A. Visuomotor adaptation without vision? Exp Brain Res 1999;127(1):12–8. [16] Paquet N, Watt DG, Lefebvre L. Rhythmical eye-head-torso rotation alters fore-aft head stabilization during treadmill locomotion in humans. J Vestib Res 2000;10(1):41–9. [17] Dean A, Voss D. Random effects and variance components. In: Design and analysis of experiments. Springer: Berlin; 1990. [18] Hsiao SS. Similarities between touch and vision. In: Morley JW, editor. Neural aspects of tactile sensation. Amsterdam: Elsevier; 1998. p. 131–63. [19] Macaluso E, Driver J. Spatial attention and crossmodal interactions between vision and touch. Neuropsychologia 2001;39(12):1304–16. [20] Jeka JJ. Light touch contact as a balance aid. Phys Ther 1997;77(5):476–87. [21] Jeka JJ, Schoner G, Dijkstra T, Ribeiro P, Lackner JR. Coupling of fingertip somatosensory information to head and body sway. Exp Brain Res 1997;113(3):475–83.
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47 [22] Konczak J. Effects of optic flow on the kinematics of human gait: a comparison of young and older adults. J Mot Behav 1994;26:225–36. [23] Prokop T, Schubert M, Berger W. Visual influence on human locomotion. Modulation to changes in optic flow. Exp Brain Res 1997;114:63–70. [24] Reginella LR, Redfern MS, Furman JM. Postural sway with earth-fixed and body-referenced finger contact in young and older adults. J Vestib Res 1999;9:103–9. [25] Nichols ER, Lindell AK. A left hemisphere, but not right hemispace advantage for tactual simultaneity judgments. Percep Psychophys 2000;62:717–25.
47
[26] Zatsiorsky VD. Kinematics of human motion. Champaign, IL: Human Kinetics; 1998. [27] Saini M, Kerrigan DC, Thirunarayan MA, Duff-Raffaele M. The vertical displacement of the center of mass during walking: a comparison of four measurement methods. J Biomech Eng 1998;120:133–9. [28] Holt KG. Head stability in walking in children with cerebral palsy and in children and adults without neurological impairment. Phys Ther 1999;79(12):1153–62. [29] Mulavara AP, Verstraete MC, Bloomberg JJ. Modulation of head movement control in humans during treadmillwalking 16 (2002) 271/282. Gait Posture 200;216:271–82.
Light touch and center of mass stability during treadmill locomotion Ruth Dickstein∗ , Yocheved Laufer Department of Physical Therapy, Faculty of Social Welfare and Health Sciences, University of Haifa, Mount Carmel, Haifa 31905, Israel Received in revised form 23 May 2003; accepted 16 June 2003
Abstract Purpose: To study the contributions of light fingertip touch on an earth-referenced object to body stability during treadmill locomotion. Method: Twenty young healthy adults were tested in two blocks of five testing conditions while walking on the treadmill at 3 km/h. In each condition, subjects were tested with eyes open (EO) and with eyes closed (EC). In each block, four separate conditions of heavy (H) or light (L) touch to either a left or to a right force sensor mounted on the respective side rail, as well as one condition of no touch (N), were randomly applied. The 3D positions of the center of mass (COM) and the midpoint of the posterior aspect of each leg were monitored via a kinematic ultrasonic system, while the anterior–posterior (AP) acceleration of the COM was measured with a uniaxial linear accelerometer. Results: Light touch had a similar stabilizing effect as vision and as heavy touch on COM sway. Thus, COM sway and AP acceleration were comparable in conditions of eyes open and eyes closed as long as touch was applied. Conversely, without vision and touch, subjects drifted backwards, with complete disruption of the coordinated stepping pattern. Conclusions: Somatosensory fingertip input from an external reference provides spatial orientation, which, similar to vision, enables the sustaining of body stability during treadmill walking. © 2003 Elsevier B.V. All rights reserved. Keywords: Light touch; Center of mass; Locomotion
1. Introduction Control of body balance is known to be largely dependent on sensory information from the visual, vestibular and lower limb somatosensory inputs [1]. In addition, during stance, somatosensory cues elicited through light fingertip touch of an external object can provide perceptual information that contributes to balance control. In the absence of visual input, light fingertip touch cues from an external reference can exert a stabilizing effect comparable to the effect of vision [2–7]. Similarly, light touch (L) can reduce the body sway of subjects with balance disorders emanating from vestibular deficits or from feet neuropathy [8,9]. Balance control is one of the main tasks required for purposeful and safe gait. Its completion is integrated with the execution of stepping and forward propulsion, as well as with the upholding of anti-gravitational posture [10]. The purpose of the present study was to examine the contribution of somatosensory touch inputs to body balance dur∗ Corresponding author. Present address: Department of Physical Therapy, Sackler Faculty of Medicine, Tel-Aviv University, Ramat-Aviv, Tel-Aviv 69978, Israel. Tel.: +972-4-824-9065; fax: +972-4-828-8140. E-mail address: [email protected] (R. Dickstein).
0966-6362/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0966-6362(03)00091-2
ing treadmill locomotion. The availability of a side or front rail on the treadmill makes the use of touch inputs easily accessible. Observation of subjects during treadmill walking, especially at low to moderate speeds, confirms occasional or constant touch of one or both rails. It could well be that these maneuvers are used for improving/correcting body balance. The reduction of physiological stress achieved by lightly touching the treadmill rail [11,12] may be related to the facilitation of balance control. However, biomechanical treadmill gait studies have generally failed to focus on whether or not subjects touched the rail. An exception is the study of Siler and associates, who demonstrated that sagittal plane kinematic parameters were not affected by grasp of the treadmill handrail [13]. Nevertheless, subjects’ balance during walking cannot be inferred from their results. Treadmill walking may well challenge the control of balance because subjects must adapt the execution of forward walking movements to zero optic flow and must use visual cues as an anchor for maintaining a stable walking base. Without vision, one is unable to distinguish between self-motion and treadmill motion or to determine how one is moving on the treadmill relative to the static environment [14]. Therefore, closing the eyes causes subjects to drift backward and to step off the treadmill [15,16].
42
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
2. Method
Measurements started after 5 min of familiarization, with subjects walking at a constant speed of 3 km/h (0.83 m/s). In view of the fact that the preferred walking speed of active adults is about 1.5 m/s [13], the speed chosen here can be considered as moderate to low. Immediately preceding each testing condition, the examiner informed subjects of the touch and side specifications involved. During the “L” testing conditions, subjects were asked to put their index finger on the force sensor without pressing on it. Pressure exceeding 200 g along any of the axes activated a warning signal alerting the subject to reduce pressure. During the “H” testing conditions, subjects were allowed to firmly press their index finger against the sensor. During the “N” conditions, they were asked to walk with their arms at the side of their body. Data collection was started when the subject performed the specific walking task fluently. Each testing condition was divided into three consecutive phases that differed in visual input, with each phase lasting 30–40 s: (1) walking with eyes open (EO); (2) walking with eyes closed (EC); and (3) walking with eyes open again (EOA). Subjects were notified to close their eyes (phase 2) and to reopen them (phase 3) by a high-pitched computer beep, which was simultaneously recorded in the data files. In the “no touch” conditions, phase 2 (EC) was always substantially shorter than 30 s and phase three (EOA) was largely absent because subjects walked backwards off the treadmill, requiring termination of testing. Subjects were allowed to rest between blocks, while conditions within each block were applied uninterruptedly with a negligible break time between them.
2.1. Sample and procedure
2.2. Data analysis
Participants were twenty university students, including 7 males and 13 females, with an average age of 23.7 (±2.6). They had no known health impairments, and all but three were right-handed (by self-report). Kinematic data were collected via an ultrasonic-based system (Vscope Ltd.), with mini-ultrasonic transducers attached to the center of the back at the level of the L5 spinuous process and to the midpoint of the posterior aspect of each calf. A uniaxial linear accelerometer (Zetra, Baxborough, MA, model 141), sensitive to accelerations along the anterior–posterior (AP) axis, was attached to the center of the back immediately below the ultrasonic transducer. Two triaxial force transducers were mounted on parallel spots on the left and right rails of the treadmill (height of 90 cm), positioned somewhat forward of the subject’s trunk. Subjects wearing shorts and sport shoes were tested during two blocks of five conditions applied in a random order (a computer random numbers generating program was used to randomize the order of testing conditions within each block). During each testing condition, either heavy (H), or light (L) touch was applied by the index finger to either the left or right rail; additionally one “no touch (N)” condition was tested.
Analysis of each testing condition was performed for two periods: (1) eyes open and (2) eyes closed. Data from the “eyes open again” period were not subjected to statistical analysis. In order to guarantee stable walking in each visual condition, all data collected during the first 5 s of each phase in each testing condition were excluded. Thereafter, 10 step cycles were delimited for analysis. The only exception was the condition of no touch with eyes closed, in which case only four to five steps were available for analysis, as further explained in the results. COM sway, which reflects the stability of the COM position was determined by the variance around the mean position of the COM transducer along the anterior–posterior and medio-lateral (ML) direction, and by the range (peak to peak) of movement of the COM along the anterior–posterior and medio-lateral axes for the same time period. Additionally, the mean rising slope of forward acceleration in each step was determined for the same time period. In order to characterize the effects of touch on the relative position of each leg with respect to the other leg, the linear correlation between the AP location of the two mid-calf transducers was calculated.
Nomenclature COM
center of mass
Touch modes H heavy touch L light touch N no touch Visual conditions EC eyes closed EO eyes open EOA eyes open again Sway direction AP anterior–posterior ML medio-lateral
Since the role of vision in treadmill walking is well understood, as is the contribution of light touch to body balance during stance, the present research was designed to complement current knowledge by focusing on the contribution of fingertip touch information to body balance during treadmill gait. In addition, separate tests were performed with touch cues provided by either the left or the right index fingertip in order to determine the effect of the side of the sensory input on balance control.
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
The distribution of the data was normalized by using the log of the raw values in the statistical analysis. Correlation values were Fisher z-transformed prior to statistical analysis. Unbalanced ANOVA [17] was used for analysis, applying two fixed effects (i.e. touch and vision) and one random effect (i.e. subject). This was followed by the Duncan post-hoc multiple comparisons procedure. ANOVA was also used for calculating the effect of side in conditions of heavy and of light touch. For descriptive purposes, raw values of the collected data were used.
3. Results 3.1. AP COM sway Mean values of AP COM variance are delineated in Fig. 1. The substantial increase observed in COM sway under the
43
condition of no touch with eyes closed was associated with backward movement of the whole body. It was quantified, however, for only four to five gait cycles because thereafter further movement of the treadmill had to be arrested for safety reasons, as subjects drifted backwards towards the treadmill edge. This phenomenon did not occur when touch was applied in the EC condition or in any of the EO conditions. The effects of touch and of vision on COM sway were significant (F = 33.5, P < 0.001 and F = 283.8, P < 0.001, respectively), as was the interaction between these effects (F = 75.5, P < 0.001). Using Duncan post-hoc multiple comparisons analysis, it was found that the effects of light and heavy touch on AP sway were not different from each other in decreasing AP sway and were similar to the effect of vision. However, the effect of no touch on AP sway was different, as sway was significantly increased by the combination of no touch and closure of the eyes (P < 0.05). The results also pointed to an interaction between the effects
180.0 160.0 140.0
EO EC
cm
120.0 100.0 80 .0 60 .0 40 .0 20 .0 0.0 (a)
HT
Heavy Touch
LT
NT
Light Touc h
No Touc h
13.5
Log (VarZ)
13.0 12.5 12.0 11.5 11.0 10.5 10.0 9.5 (b)
Eyes Open
Eyes Closed
Fig. 1. Mean (S.E.M.) AP sway (COM variance) under two visual (EO, EC) and three touch conditions (H, L, N) (a), and the effect of the interaction between touch and vision on AP sway (b).
44
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
of touch and subject (F = 2.8, P < 0.001), indicating that subjects differed in their usage of touch cues for stabilization. Values of sway range were more variable than were those of sway variance. Still, the statistical analysis yielded similar results: a significant effect of touch and of vision (F = 16.7, P < 0.001 and F = 289.9, P < 0.001, respectively), with a differential effect of light and heavy touch as compared to no touch (Duncan multiple comparisons, post-hoc analysis, P < 0.05). Once more, the interaction between the effects of touch and subject was significant (F = 4.1, P < 0.001), implying a non-uniform contribution of touch to different subjects. 3.2. ML COM sway Mean values of COM variance are outlined in Fig. 2. Although the effect of touch was less pronounced in the ML than in the AP direction, the overall effects of touch and of vision on COM variance were significant (F = 7.7, P < 0.002 and F = 28.7, P < 0.001, respectively). The interaction between these effects (F = 7.3, P < 0.001) pointed to greater sway under no touch conditions, particularly so when
Table 1 Slopes of COM forward acceleration during treadmill walking with heavy or light touch Touch mode
Slope (cm/s2 /s) mean
Slope (cm/s2 /s) standard deviation
Heavy Light
1.03 0.94
0.33 0.32
the eyes were closed. The effects of heavy and light touch were comparable to that of vision in decreasing ML sway, while the effect of no touch was again different (P < 0.05). Subjects’ effect was significant for both touch (F = 3.8, P < 0.001) and vision (F = 4.8, P = 0.000), again implying variance in use of these sensory inputs by different people. Similar results were found for the range of ML sway. 3.3. Acceleration of COM COM acceleration was measured in the AP plane only. Mean values of the rising slope of AP acceleration with either heavy or light touch are presented in Table 1. As long as either heavy or light touch was available, vision had no
14.00 12.00 10.00 EO
8.00
EC
6.00 4.00 2.00 0.00
(a)
HT
LT
NT
.8 11
Heavy Touch
Light Touch
No Touch
11.6
Log (VarY)
11.4 11.2 11.0 10.8 10.6 10.4 (b)
Eyes Open
Eyes Closed
Fig. 2. Mean (S.E.M.) ML sway (COM variance) under two visual (EO, EC) and three touch conditions (H, L, N) (a), and the effect of the interaction between touch and vision on ML sway (b).
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
significant effect on COM acceleration. However, the main effect of heavy touch was associated with a steeper increase in acceleration than was light touch (F = 9.6, P < 0.002). With no touch and no vision, acceleration slopes were the highest, though they could be measured for only a few seconds under these conditions due to the need to terminate testing as a result of subjects’ backward drift. Significant individual differences were noted between the slopes of the participants (F = 12.3, P < 0.001). 3.4. Side The side of the touched sensor had no significant effect either on COM sway (F = 0.024, P = 0.88 and F = 0.86, P = 0.36, for AP and ML sway, respectively), or on COM forward acceleration (F = 1.45, P < 0.23). Yet, as shown below, the force applied by the subjects’ right or left index finger differed. 3.5. Mean forces Mean forces were directed downward, forward, and to the right. The difference between heavy and light forces was significant (F = 224.2, P < 0.001; F = 171, P < 0.001; and F = 79, P < 0.001 for the vertical, AP, and ML axes, respectively), with vision having no effect on the magnitude of these forces in either direction. Significant individual differences were observed in the amount of force applied by subjects during downward light touch (F = 8.0, P < 0.001), as well as for AP forces (F = 5.2, P < 0.001). Still, complete differentiation between light and heavy touch was maintained by all participants. Regarding the impact of side, the vertical force applied to the right sensor was significantly greater than that applied to the left sensor during both heavy and light touch (F = 160.4, P < 0.001) (see Fig. 3). 3.6. Linear correlation Linear correlation between the AP position of the legs was calculated from the mid-calf markers. All correlations were 350 300
grams
250 200 150 100 50 0 EO
EC HLt
HRt
EOA LLt
LRt
Fig. 3. Index fingertip force applied during heavy and light touch on the right- and left-sided sensors.
45
negative, as consistent with the reciprocal AP movements of the legs during treadmill walking. The mean between-legs correlation with eyes open and with either heavy, light or no touch was −0.78, while the corresponding mean value with eyes closed and with either heavy or light touch was −0.69. Thus, closing the eyes was associated with a significant reduction in between-legs correlation (F = 170.9, P < 0.001). The decrease in the correlation between-leg positions was especially detrimental upon closing the eyes in the no touch condition (mean r = −0.06; F = 45.5, P < 0.001), pointing to derangement of the reciprocal gait pattern.
4. Discussion The main finding of this study is that touching an external object, even with a very light fingertip touch, provides treadmill walkers with a somatosensory anchor, which, similar to the visual anchor, permits spatial orientation and reduces body sway. Apparently, the similarities between neural control mechanisms of the two systems, with the operation of neuronal networks that involve vision and/or touch for spatial attention [18,19], account for their interchangeable role in balance control. The findings agree with those of previous research demonstrating that light touch has a comparable effect to that of vision in decreasing postural sway during stance [4,7,8,20,21]. The fact that either vision or touch (even a very light touch) of an external reference is needed for the optimal match of walking speed with velocity of the treadmill, indicates their inevitability for spatial orientation during treadmill locomotion. At the 3 km/h velocity that was applied here, the stabilizing effect of light touch was more enhanced in the AP than in the ML direction. This difference stems, most probably, from greater AP than ML destabilization in the tested velocity. Studies of stance stability have also indicated that the stabilizing effect of light touch cues are plane-specific, preferentially decreasing either ML or AP sway during tandem or normal bipedal stance, respectively [2,3]. However, the effect of touching a front rather than a side rail on COM sway during treadmill locomotion remains to be determined. Furthermore, as the relationship between ML and AP sway of the COM is speed-dependent, a desired complement to the current findings may be obtained by testing the effects of light touch on COM stability during a range of walking speeds. The results similarly show that the same orientation reference (i.e. vision or touch) is necessary for controlling forward COM acceleration, as well as for coordinating leg movements during treadmill gait. The normal AP coordination between the legs was somewhat reduced by the elimination of vision, but was completely disrupted due to simultaneous elimination of both vision and touch. This observation that kinematic patterns of gait are altered in the absence of vision or when there is inconsistent optic flow has also been made in previous studies [22,23].
46
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47
Observation of treadmill walking among healthy subjects indicates that touch of the side rail is mainly used at slow and moderate speeds, while as walking shifts to running (about 7–8 km/h), subjects usually adduct their arms and flex their elbows, moving them in coordination with the movements of the lower extremities. Therefore, the current results are mainly applicable to slow treadmill walking, as frequently applied in rehabilitation. Here, touch cues can be used instead of, or in addition to vision for compromising balance difficulties. However, generalization of the findings to walking over ground with a cane could be misleading because neither can the treadmill belt be compared to a stationary floor nor can the touch of a stationary rail be judged against a cane that is moved by the walking subject him/herself. This notion is supported by previous findings that postural sway among standing subjects was decreased by touching an earth-fixed plate and increased by touching a sway-referenced plate [24]. In the current study, the side of the handrail that was touched had no statistical effect on the outcome variables; yet, the force applied by the participants to the right side rail was greater than that applied to the left side rail. Perhaps the superior haptic and motor abilities of the left hemisphere (dominant hand) could explain its “enhanced” use in comparison to the contralateral side [25]. These findings might have practical implications when considering the appropriate side for a rigid support provided to walking patients. However, further inquiry is needed on this issue. When considering the current findings, one should be aware of the fact that the data collected pertain only to translational movements of the body, whereas information on orientation and/or configuration of the body [26] is unidentified. Furthermore, the application of only one marker to represent COM location could undermine the accuracy of the measurements. Yet, the validity of that approach has previously been approved [27]. Additionally, the findings pertaining to the body COM should be supplemented by examining the contribution of touch cues to head stability. Head stability during treadmill locomotion has been characterized by fluctuations in the period and amplitude of motion in the sagittal plane across walking cycles. A dynamical modulation of head movements related to the phases of the gait cycle and aimed at minimizing force variations and increasing visual fixation has recently been proposed [28,29]. These considerations may be further enhanced by examining the effect of touch. In view of the above evidence, it appears necessary to consider the potential effects of vision and/or touch conditions on gait parameters in biomechanical treadmill studies. Accordingly, it is essential that visual and touch conditions be explicitly reported in the description of such research. The importance of the current analysis extends even further through its disclosure of significant individual differences, pointing to additional factors besides touch and vision that
may affect body stability during treadmill gait. Thus, further investigation is needed to shed more light on these factors as well.
References [1] Horak FB, Macpherson JM. Postural orientation and equilibrium. In: Shepard J, Rowell L, editors. Handbook of physiology. New York: Oxford University Press; 1996. p. 255–92. [2] Clapp S, Wing AM. Light touch contribution to balance in normal bipedal stance. Exp Brain Res 1999;125(4):521–4. [3] Jeka J, Lackner JR. Fingertip contact influences human postural control. Exp Brain Res 1994;100:495–502. [4] Jeka J, Oie K, Schoner G, Dijkstra T, Henson E. Position and velocity coupling of postural sway to somatosensory drive. J Neurophysiol 1998;79(4):1661–74. [5] Jeka J, Oie KS, Kiemel T. Multisensory information for human postural control: integrating touch and vision. Exp Brain Res 2000;134(1):107–25. [6] Lackner JR, Rabin E, DiZio P. Stabilization of posture by precision touch of the index finger with rigid and flexible filaments. Exp Brain Res 2001;139(4):454–64. [7] Rogers MW, Wardman DL, Lord SR, Fitzpatrick RC. Passive tactile sensory input improves stability during standing. Exp Brain Res 2001;136:514–22. [8] Lackner JR, DiZio P, Jeka J, Horak F, Krebs D, Rabin E. Precision contact of the fingertip reduces postural sway of individuals with bilateral vestibular loss. Exp Brain Res 1999;126(4):459–66. [9] Dickstein R, Shupert CL, Horak FB. Fingertip touch improves postural stability in patients with peripheral neuropathy. Gait Posture 2001;14(3):238–47. [10] Winter D. The biomechanica and motor control of human gait. Waterloo, Canada: University of waterloo press; 1991. [11] Manfre MJ, Yu GH, Varma AA, Mallis GI, Kearney K, Karageorgis MA. The effect of limited handrail support on total treadmill time and the prediction of VO2 max. Clin Cardiol 1994;17(8):445–50. [12] Christman SK, Fish AF, Bernhard L, Frid DJ, Smith BA, Mitchell L. Continuous handrail support, oxygen uptake, and heart rate in women during submaximal step treadmill exercise. Res Nurs Health 2000;23(1):35–42. [13] Siler WL, Jorgensen AL, Norris RA. Grasping the handrails during treadmill walking does not alter sagittal plane kinematics of walking. Arch Phys Med Rehabil 1997;78(4):393–8. [14] Lee DN. Getting around with light and sound. In: Warren K, Wertheim AH, editors. Perception and control of self motion. Hiilsdale, New Jersey: Lawrence Erlbaum Associates; 1990. p. 487– 505. [15] Durgin FH, Pelah A. Visuomotor adaptation without vision? Exp Brain Res 1999;127(1):12–8. [16] Paquet N, Watt DG, Lefebvre L. Rhythmical eye-head-torso rotation alters fore-aft head stabilization during treadmill locomotion in humans. J Vestib Res 2000;10(1):41–9. [17] Dean A, Voss D. Random effects and variance components. In: Design and analysis of experiments. Springer: Berlin; 1990. [18] Hsiao SS. Similarities between touch and vision. In: Morley JW, editor. Neural aspects of tactile sensation. Amsterdam: Elsevier; 1998. p. 131–63. [19] Macaluso E, Driver J. Spatial attention and crossmodal interactions between vision and touch. Neuropsychologia 2001;39(12):1304–16. [20] Jeka JJ. Light touch contact as a balance aid. Phys Ther 1997;77(5):476–87. [21] Jeka JJ, Schoner G, Dijkstra T, Ribeiro P, Lackner JR. Coupling of fingertip somatosensory information to head and body sway. Exp Brain Res 1997;113(3):475–83.
R. Dickstein, Y. Laufer / Gait and Posture 20 (2004) 41–47 [22] Konczak J. Effects of optic flow on the kinematics of human gait: a comparison of young and older adults. J Mot Behav 1994;26:225–36. [23] Prokop T, Schubert M, Berger W. Visual influence on human locomotion. Modulation to changes in optic flow. Exp Brain Res 1997;114:63–70. [24] Reginella LR, Redfern MS, Furman JM. Postural sway with earth-fixed and body-referenced finger contact in young and older adults. J Vestib Res 1999;9:103–9. [25] Nichols ER, Lindell AK. A left hemisphere, but not right hemispace advantage for tactual simultaneity judgments. Percep Psychophys 2000;62:717–25.
47
[26] Zatsiorsky VD. Kinematics of human motion. Champaign, IL: Human Kinetics; 1998. [27] Saini M, Kerrigan DC, Thirunarayan MA, Duff-Raffaele M. The vertical displacement of the center of mass during walking: a comparison of four measurement methods. J Biomech Eng 1998;120:133–9. [28] Holt KG. Head stability in walking in children with cerebral palsy and in children and adults without neurological impairment. Phys Ther 1999;79(12):1153–62. [29] Mulavara AP, Verstraete MC, Bloomberg JJ. Modulation of head movement control in humans during treadmillwalking 16 (2002) 271/282. Gait Posture 200;216:271–82.