| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
HIGHLIGHTED TOPICS
Neural Control of Movement
1Department of Biobehavioral Sciences, Teachers College, Columbia University, New York 10027; and 2Department of Rehabilitation Medicine, College of Physicians and Surgeons, Columbia University, New York, New York 10032
Submitted 7 October 2003 ; accepted in final form 17 November 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
proprioception; touch; muscle vibration; sensory integration; haptics
More recently, studies have highlighted tactile cues as an additional source of information used to localize the limb in space. Skin stretch orientation will excite mechanoreceptor activity associated with joint orientation (4, 6, 8, 9). Additionally, tactile cues from external surfaces can signal proximity to external spatial reference points. For example, tactile cues from glabrous skin may contribute to estimates of arm position during pointing movement. Reaching end point during Coriolis force perturbation is better adapted when end point contact is involved (21). Accuracy of reaching to a remembered position of the contralateral index fingertip improved when the "remembered" fingertip position was touching a surface (14). This effect is attributed to tactile cues, as opposed to deep pressure cues, because accuracy of reproducing end point locations degraded when the distal phalange of the index finger was anesthetized (23). Tactile information may also signal start position (22), which may be important for motor planning (1, 9, 24).
It is unknown what information is conveyed by the tactile cues during such motor tasks. In the Rao and Gordon study (23), the surface was uniform in texture, so tactile information may not convey information about locations on the touched surface. It is likely instead that the high resolution of cutaneous sensitivity in the fingertips affords information about arm orientation with respect to the surface. Such information could be integrated with other proprioceptive feedback, although the exact nature of such integration is unclear. An earlier study, in which light touch of the fingertip completely attenuated the postural destabilizing effects of ankle vibration, suggests that tactile feedback can, in certain circumstances at least, override muscle spindle feedback (20).
One way external tactile cues might contribute to the sense of body orientation is to allow the sense of body orientation on the basis of internal proprioceptive feedback (i.e., muscle spindles, joint receptors) to be matched against a spatial framework that is associated with the touched external cues. This depends on the interpretation of tactile inputs on the basis of knowledge of their spatial properties. Consider for example finding one's way in a familiar dark room. One's orientation in the room from tactile feedback depends on one's memory of the room from previous experience in the light. Features of the objects in the room constitute an environmental "context" that might carry various sensory and/or motor associations (26). If one is informed that all of the elements of the room were changed (i.e., unfamiliar context), these tactile cues might have very little meaning. Thus the meaningfulness of the cue may be dependent on an assumption that the tactile cue is stable within the environment.
The present experiment examines the integration of external tactile cues and muscle stretch cues in producing a final estimate of arm position. Primarily we tested the hypothesis that fingertip contact with an external reference surface can influence illusory arm movements that arise from muscle spindle vibration. Our strategy for determining the priority of tactile or muscle spindle feedback was to create tactile and muscle spindle feedback mismatches and observe from the subjects' reports (determined by the magnitude of the vibratory illusions) which cue was stronger. Subjects reported perceived orientation of the left arm under different combinations of deep proprioceptive cues (with or without vibration of biceps brachia, which causes illusions of elbow extension) and fingertip contact (with or without touch of the stationary cue surface). If this hypothesis, that touching a stationary surface can override spindle feedback from muscle vibration, is supported, then we may expect an effect of fingertip contact in which subjects are less susceptible to illusory arm movements produced during vibration.
We also manipulated cue context (stationary surface, movable surface) to test the hypothesis that a priori notions about whether the tactile cue surface is stationary would determine whether subjects would perceive arm displacement. If external tactile cues are interpreted with respect to a priori notions of the stability of the tactile cue surface, then touching the stationary surface that is known to be able to move (i.e., movable surface context) during biceps vibration may not help subjects reliably locate their arm in space. In these circumstances, subjects may report arm and surface motion even though the surface is stationary.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Experimental setup. Subjects sat with their elbows on a stationary table and their forearms oriented 
45° from horizontal (see Fig. 1). Their hands extended beyond the far edge of the tabletop. The right hand had nothing below it. The left hand was above a second horizontal surface ("cue surface"), which subjects could touch with the left index finger when instructed. The height of the cue surface was 20 cm higher than the table top. Its position was fixed during experimental trials, but it could pitch such that fingertip contact would be maintained during movements of the elbow when desired.
|
During trials with vibration (see below), a massage vibrator (Hitachi Magic Wand, Tokyo, Japan) was applied to the left biceps brachia transcutaneously by the experimenter throughout the trial (15 s). The vibrator ran at amplitudes of 1 mm at 100 Hz.
Procedure. Seated subjects with eyes closed reported the perceived orientation of their left ("cue") arm by "mirroring" its ongoing location across in the sagittal plane with their right ("report") arm. Before each trial, subjects oriented the left forearm 45° from horizontal, with the left index finger flexed to touch the top of the cue surface. For trials with no touch, they were instructed to hold the left fingertip just above cue surface. The experimenter began data collection at this point to acquire baseline discrepancies in fingertip positions. After data collection, the experimenter flexed the subjects' left arm at the elbow and the subjects rested the right forearm flat on the near table to prevent any visual awareness of arm mismatch.
The experimental design consisted of two blocks of experimental conditions. The experimental conditions varied biceps vibration (vibration, no vibration) and fingertip contact with the stationary cue surface (touch, no touch). Note that differences in control of the arm while maintaining fingertip contact vs. without fingertip contact might create differences in muscle activity (i.e., alpha-gamma coactivation) (e.g., Ref. 25), which could effect spindle feedback during touch and no touch conditions with vibration. Subjects need to overcome the tonic vibration reflex (e.g., Refs. 3, 5, 10, 11) while maintaining fingertip contact with an unrestrained arm undergoing muscle vibration, but not during a no-touch condition. Hence, we performed the vibratory and touch conditions both with and without manual restraint of the cue forearm above the wrist by the experimenter (restraint, no restraint). This condition was included to control for the potential confound of the need of the subject to overcome a tonic vibration reflex. Thus the eight conditions in each block were 1) free, 2) touch, 3) held by experimenter, 4) biceps vibration, 5) restraint + touch, 6) vibration + touch, 7) vibration + restraint, and 8) vibration + restraint + touch. Subjects were informed of these experimental manipulations before the experiment.
To change the context of tactile cue from "stationary" to "moving," subjects were initially naive of the potential of the cue surface to move but were made aware of its capacity to move after completion of the first block of trials. The cue surface's capacity of motion was disclosed in the following manner. The subjects were instructed to do an experimental "touch" trial without restraint or vibration and with eyes closed, and the cue surface was moved by the experimenter. When touching the cue surface with their index finger of the cue arm, subjects actually touched a small index card on the cue surface and were told to apply just enough pressure to hold the card against the cue surface. By keeping the card beneath the finger, but allowing the card to slide against the surface, subjects could minimize motion of the joints other than the elbow (i.e., wrist and finger) when the cue surface moved up and down (see below). The subject was reminded to keep the left finger on the surface and mimic the ongoing posture of the left hand and arm with the right hand and arm. The cue surface motion was slightly irregular in terms of speed and amplitude, so that in the block of trials after this demonstration the difference between this condition and the vibration conditions could not be distinguished by the regularity of movement feedback alone. The tracking of cue surface motion procedure was repeated four times.
After the demonstration of table motion, the cue surface was again fixed (unbeknownst to the subjects) and the eight conditions described above were repeated. Thus trials were run in two blocks: naive of potential cue surface motion (I) and aware of potential cue surface motion (II). The cue surface was always stationary throughout both blocks of trials. To minimize effects of fatigue from vibration, each condition was performed once per block. To control for possible fatigue effects, the trial order within blocks I and II was counterbalanced across subjects.
Data measurement and analysis. Electromagnetic three-dimensional sensors (Polhemus Fastrack; resolution = 0.75 mm) fastened to the dorsal surface of the wrists and elbows measured orientations of the forearms. Signals from the position sensors were sampled at 60 Hz by data-acquisition software (SC/ZOOM, University of Umeå, Umeå, Sweden) via the serial port. Data were sampled for 15 s.
The object of the study was not the accuracy of matching per se, but rather how the strength of the vibration illusion (excess elbow extension) is affected by touching. Therefore, we report the maximum displacement of the report forearm relative to the cue arm during the collection period. We also measured the relative displacement at the end of the 15-s trial. The results for this measure were qualitatively similar for all conditions except one. Therefore, we only report the maximum displacement except in the one instance in which these two measures differ.
Left and right arm orientations were quantified as the radial distance from the horizontal (elevation) along the arc prescribed by elbow articulation. A positive displacement corresponds to overextension of the report arm relative to the cue arm. Scores were unbiased by subtracting the error at the start of the trial. Thus the maximum relative displacement of arm orientation, Dmax, was calculated
 |
where C(t) is the cue forearm orientation at time t, R(t) is the report forearm orientation at time t, and C(0) and R(0) are the arm orientations at the start of data collection. Statistics were performed on subject scores for each condition.
A 2 x 2 x 2 x 2 ANOVA tested significance of effect of tactile cue context [stationary (block I) vs. moving (block II)], vibration (vibration vs. no vibration), fingertip contact (touch vs. no touch), and restraint (restraint vs. no restraint).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
Biceps vibration causes perception of elbow extension. Figure 3 shows that these results were representative of the subjects tested. On average there was no change in orientation of the report arm among any of the conditions without biceps vibration. Mean relative displacements for a representative nonvibration condition (no touch, no restraint, no vibration) are plotted at the far left of Fig. 3. The figure (all other bars) also shows all conditions with vibration, in which subjects reported displacements of various degrees in the direction of elbow extension as expected.
|
When the arm was unrestrained (left bars), biceps vibration caused elbow flexion in the cue arm. During trials with vibration without touch or restraint, elbow flexions of the cue arm during biceps vibration at the moment of greatest underestimation were in the range of 2.1-19.9° with a mean of 7.4° for all subjects. Subjects reported flexion with their right arms, but reports underestimated these movements by an average of 5°. Vibration-induced cue arm flexion continued past that in some cases, while the report arm caught up, reducing the discrepancy in forearm orientations to an average of 3.4°. This condition (vibration, no touch, no restraint) was the only condition in which the moment of greatest relative forearm discrepancy was not consistent with the end of the 15-s trial.
During restraint, the experimenter noted the resistance to restraint of the forearm during vibration but maintained the initial position as confirmed by the position sensor. Reports of displacement were greater by 10° during vibration when the cue arm was restrained (Fig. 3, right open bars), than when the arm was free (Fig. 3, left open bars) (restraint x vibration interaction, P < 0.001).
Before experience of table motion, touch suppressed perceived elbow extension. Touch of the stationary cue surface significantly reduced overestimation of elbow angle during biceps vibration with and without restraint (Fig. 3, solid shaded bars), but only during block I, when the context of the tactile cue surface was stationary (context x touch interaction, P < 0.001). Before any experience of the cue surface moving, touching the stable cue surface during biceps vibration reduced reports of displacement to 2° for both the restrained and unrestrained conditions (restraint, P > 0.05).
After experience of table motion, touch did not suppress perceived elbow extension. After disclosure of the cue surface's potential to move (moving cue context, block II) (Fig. 3, shaded hatched bars), fingertip contact with the same stationary surface no longer attenuated illusory elbow movement during biceps vibration with or without restraint (context x touch x vibration interaction, P < 0.016). When the context had changed to moving, reports of displacement while touching the stationary cue surface returned to levels observed during initial unrestrained vibration trials (8° without restraint; 10° with restraint), and the increased reports of displacement were unaffected by the restraint condition (P > 0.05).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The change in the response to fingertip contact during biceps vibration after table motion suggests that the context provided by tactile cue is updated in some fashion. One possible way for this update to occur stems from abstract knowledge of spatial properties of a tactile stimulus. That is the fingertip is localized on the basis of its contact with an object that is known to occupy a certain location, thus creating a "cognitive set." Once subjects are aware of the potential of surface motion, it is no longer trusted as a spatial cue. Alternatively, there might be a matching of the specific tactile and muscle stretch feedback experienced during tracking of the surface when it actually moved between blocks I and II. The pattern of stimuli in block II (fingertip contact, spindle feedback from vibration) recalls the fingertip contact and actual spindle stretch feedback from the recent experience when the elbow actually moved, thus elbow extension is reported. These possible mechanisms may be independent but may both underlie the present finding.
We note that the pattern of fingertip contact during the experience of actual table motion that altered table context may have differed from that of blocks I and II. The "high resolution" of tactile information about mechanical events at the fingertip through force changes at the fingertip should convey motion when the table occurs. It is also possible that slight movements of the wrist and finger may not have been identical during actual surface tracking and the moving cue context trials of block II when the cue was actually stationary. Despite these potential discrepancies in the patterns of feedback from the hand and fingertip during actual cue surface motion and block II when the table was stationary, subjects reported motion. Because the tactile feedback was similar for blocks I and II, we can attribute the differences in reports to the change in context brought about by the surface tracking procedure between trial blocks. In this light, the influence of context appears all the more significant because the subjects reported surface motion in the absence of actual dynamic tactile cues likely occurring during motion.
Spatial information conveyed by a stimulus is affected by assumptions about it. Our results are consistent with those of Jeka et al. (16), who demonstrated that assumptions about the stability of the tactile reference point are critical in the use of tactile feedback from light fingertip contact to control standing posture. Fingertip contact with a stable surface attenuates postural sway in the absence of other feedback such as vision (e.g., Refs. 15, 22). In the case that the touch surface oscillates below a critical amplitude, manual contact force resulting from motion of the tactile cue is attributed to postural sway of the subject and posture is adjusted to minimize motion relative to the tactile cue. The overall result is that postural sway will entrain to the frequencies of motion of the haptic reference. Once the possibility of motion of the surface is explicitly disclosed, subjects have a new problem and may mistake slip at the point of fingertip contact, which is actually due to postural drift, for motion of the haptic reference (16). This illustrates the usefulness of having redundant feedback from different modalities to resolve ambiguous stimuli in one modality. In the case of standing posture, the moving touch cue is only ambiguous if the amplitude of motion is below the vestibular or other proprioceptive thresholds. In the case of the moving cue surface in the present experiment, spatial ambiguity arises in the second half of the experiment because the tactile cue is associated with a greater range of spatial locations (i.e., is deemed an unreliable position cue), which overlaps with the range corresponding to the patterns of stimuli generated by the biceps vibration.
Conflicting stimuli were resolved according to what is physically possible and likely. Our results of paradoxical combinations of stimuli (stationary touch and moving muscle stretch) resolved with priority of touch over muscle spindle feedback cues is contrary to previous findings in which conflicting stimuli have been resolved in favor of the muscle spindle feedback, resulting in sensation of impossible distortions of body orientation. For example, Craske (5) reported that subjects felt their wrists extended far beyond what is physically possible, or multiple forearms including the actual static orientation and the moving "vibrated" orientation. These situations describe sensations arising from conflicts within muscle spindles and joint receptor feedback. Conflicts of muscle spindle feedback and tactile feedback include the "Pinocchio" illusion, in which subjects feel their nose grow when touching their nose with their finger while the biceps of the same arm is vibrated (17). In this scenario the unambiguously stationary tactile reference (the nose) does not prevent the illusory elbow movement or "nose growth." The present requirement of reporting the perceived orientation with the other single forearm effectively prohibited convincing reports of physically impossible scenarios such as hyperextension, elongation, or multiple spatially independent manifestations (5, 17).
Nevertheless, no verbal reports contradicted report-arm movements: perceptions of arm movement were unambiguous during vibration without fingertip contact, but, with the exception of one subject, fingertip contact with a stationary surface during vibration seemed to "cloud" elbow position sense at the elbow until the cue surface actually moved. When the arm was unrestrained, subjects reported that it was more difficult to maintain fingertip contact. Subjects described the experience of vibration during fingertip contact as their elbow feeling unusual ("weird"; "I can't tell what my elbow is doing"; or "numb"). They all reported that touching the cue surface was their primary cue of forearm orientation. One subject reported forearm movement verbally and with his report arm (12°) but only in the restraint condition. This subject felt that the table was moving in this condition but not in the condition without restraint. The one subject who experienced cue surface motion in the restraint-touch condition before experiencing table motion (block I) reported elbow extension that was also a 10° increase (22°) in the restraint-touch condition after experiencing the cue surface actually move.
This demonstrates not only an influence of cognitive assumptions but also a requirement for a singular coherent sense of arm position. Indeed, the aforementioned exceptional subject who experienced arm movement while touching with restraint before experiencing actual cue motion reported that the table was moving too, not that his hand was going through the table. To reconcile our findings with those of studies reporting impossible illusions, after our protocol and before explaining the effect of muscle vibration, we attempted to induce the Pinocchio illusion in six of our subjects with little success until they were verbally prompted (i.e., "do you feel your nose growing?"), at which point four of the subjects reported feeling the scenario described to them. Because our subjects reported the Pinocchio illusion only after verbal cues, we conclude that among naive subjects probable and likely sensations are preferred by the nervous system to unlikely or impossible ones. In this case the feasibility of the sensation of nose growth (as opposed to feasibility of the actual event) was established in this laboratory setting by the verbal cueing.
Effects of restraint. The greater errors with restraint than without may be related to joint angle sense resulting from comparisons of agonist vs. antagonist muscle stretch (e.g., Refs. 2, 10). In an unrestrained arm, the tonic vibration reflex causes elbow flexion, which unloads the biceps spindles and stretches the triceps (3, 5, 11, 12). The degree of resulting flexion is underestimated because of "extra" biceps feedback from the vibration. Joint immobilization during the tonic vibration reflex increases the magnitude of the extension illusion by preventing unloading of the vibrated muscle spindles and antagonist stretch. Although vibration-induced errors were greater with restraint than with the free arm, touch attenuated errors to near zero in both cases. Also, after experience of actual cue motion, subjects reported similar arm motion when touching when subjects maintain fingertip contact on their own, just like when the subjects' cue arm was restrained. This may be because requiring the subject to maintain fingertip contact is somewhat like having subjects administer their own restraint: the triceps will not lengthen, and the biceps will not be unloaded by shortening. However, the triceps will generate extra tension, and associated gamma activity (see Ref. 13), to counter the biceps tonic vibration reflex contraction.
The forearm restraint condition also provides another tactile cue to consider: that of the experimenter's hand on the wrist of the cue arm. It is noteworthy that the history of experience of the table (before or after actual table motion) determined how the cues provided by restraint contact, as well as spindle feedback, were interpreted. Like the cue surface, the restraint was a stationary contact cue. However, restraint without fingertip contact never attenuated reports of elbow movement. This suggests that the cue of manual contact by the experimenter is from the start a spatially ambiguous one, because it may be possible that the experimenter's hand can move the subject's arm.
We conclude that tactile cues associated with fingertip touch can contribute to position sense. Specifically, contact with an external reference provides a contextual influence on interpretation of the cues that inform limb position. A priori notions about the stability context of an external contact surface influence how this tactile cue is integrated with proprioceptive sensory modalities to generate an estimate of arm location in space. These findings support the notion that tactile cues are used to calibrate proprioception against external spatial frameworks.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Rabin, P. DiZio, J. Ventura, and J. R. Lackner Influences of Arm Proprioception and Degrees of Freedom on Postural Control With Light Touch Feedback J Neurophysiol, February 1, 2008; 99(2): 595 - 604. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. E. Hudson, L. T. Maloney, and M. S. Landy Movement Planning With Probabilistic Target Information J Neurophysiol, November 1, 2007; 98(5): 3034 - 3046. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Johannsen, A. M. Wing, and V. Hatzitaki Effects of Maintaining Touch Contact on Predictive and Reactive Balance J Neurophysiol, April 1, 2007; 97(4): 2686 - 2695. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
