Lower-Limb Amputees Adjust Quiet Stance in Response to Manipulations of Plantar Sensation

Abstract

Interfering with or temporarily eliminating foot-sole tactile sensations causes postural adjustments. Furthermore, individuals with impaired or missing foot-sole sensation, such as lower-limb amputees, exhibit greater postural instability than those with intact sensation. Our group has developed a method of providing tactile feedback sensations projected to the missing foot of lower-limb amputees via electrical peripheral nerve stimulation (PNS) using implanted nerve cuff electrodes. As a step toward effective implementation of the system in rehabilitation and everyday use, we compared postural adjustments made in response to tactile sensations on the missing foot elicited by our system, vibration on the intact foot-sole, and a control condition in which no additional sensory input was applied. Three transtibial amputees with at least a year of experience with tactile sensations provided by our PNS system participated in the study. Participants stood quietly with their eyes closed on their everyday prosthesis while electrically elicited, vibratory, or no additional sensory input was administered for 20 s. Early and steady-state postural adjustments were quantified by center of pressure location, path length, and average angle over the course of each trial. Electrically elicited tactile sensations and vibration both caused shifts in center of pressure location compared to the control condition. Initial (first 3 s) shifts in center of pressure location with electrically elicited or vibratory sensory inputs often differed from shifts measured over the full 20 s trial. Over the full trial, participants generally shifted toward the foot receiving additional sensory input, regardless of stimulation type. Similarities between responses to electrically elicited tactile sensations projected to the missing foot and responses to vibration in analogous regions on the intact foot suggest that the motor control system treats electrically elicited tactile inputs similarly to native tactile inputs. The ability of electrically elicited tactile inputs to cause postural adjustments suggests that these inputs are incorporated into sensorimotor control, despite arising from artificial nerve stimulation. These results are encouraging for application of neural stimulation in restoring missing sensory feedback after limb loss and suggest PNS could provide an alternate method to perturb foot-sole tactile information for investigating integration of tactile feedback with other sensory modalities.

Keywords: balance perturbation; neuroprostheses; peripheral nerve stimulation; sensory feedback; somatosensation; standing balance; transtibial amputation; vibration.

FIGURE 1

Methods of delivering unilateral tactile stimuli. (A) Tactile percepts were elicited in the missing foot via peripheral nerve stimulation (PNS). Electrical stimulation was delivered through cuff electrodes implanted around the sciatic nerve and its branches. (B) Vibration was delivered to the intact foot via vibrating motors incorporated into an insole. Vibrating bar motors were used for participants LL01 and LL03, and vibrating coin motors were used for participant LL02. Illustrations in 1A are provided courtesy of the APT Center at the Louis Stokes Cleveland VA Medical Center.

FIGURE 2

Locations and intensities of tactile stimuli reported by participants. The participants were asked to draw the locations of the four tactile stimuli on an image of a healthy foot. PNS-evoked percepts in the missing foot are shaded in red and yellow and vibratory percepts are shaded in blue and teal. Participants verbally reported the intensity of each stimulus on a self-selected scale. Reported intensities were normalized by the lowest reported value x_i, where i = 1, 2, or 3 for participants LL01, LL02, and LL03, respectively. Images of the feet were also flipped for LL01, who was a left-side amputee, to facilitate comparison between participants.

FIGURE 3

Path length. The center of pressure path length over the 20 s stimulus was averaged across trials for the control condition (No Stim, gray, n = 6), both peripheral nerve stimulation conditions (PNS, orange, n = 12), and both vibration conditions (blue, n = 12) for each participant (LL01, LL02, LL03). Error bars indicate 95% confidence intervals and * indicates that the tactile stimulus was significantly different than the no stimulation condition (t-test, p < 0.05).

FIGURE 4

Center of pressure locations over the full 20 s trial. Tactile stimuli were delivered for 20 s with at least 3 s recorded before and after application. Positive mediolateral center of pressure (M/L CoP) values were in the direction of the amputated limb (Amp) and negative values were in the direction of the intact limb (Int). Positive anterior-posterior center of pressure (A/P CoP) values were in the forward direction (Fwd) and negative values were in the backward direction (Back). The mean CoP location across multiple trials is depicted with a bold line and the lighter traces depict CoP locations in individual trials. The center of pressure location for the no stimulation condition (gray) is shown on each plot for comparison against the locations for both peripheral nerve stimulation (PNS Condition 1, red; PNS Condition 2 yellow) and vibration conditions (Vibr Condition 1, blue; Vibr Condition 2, teal) for each participant (LL01, LL02, LL03). The perceived location of each stimulus is shown in the top left corner of each graph, while the angle of the center of pressure over the course of the whole trial is shown in the top right, where the bold line again indicates the mean CoP angle while the lighter lines indicate angles in individual trials.

FIGURE 5

Mediolateral CoP location initial responses. The initial responses (first 3 s) to tactile stimuli were averaged across six trials. The gray dashed line indicates the onset of the stimulus. Positive mediolateral center of pressure (M/L CoP) values were in the direction of the amputated limb (Amp) and negative values were in the direction of the intact limb (Int). The mean CoP across multiple trials is depicted with a solid line and the shaded regions around each line show the standard error. The center of pressure location for the control condition (No Stim, gray) is shown on each plot for comparison against the locations for both peripheral nerve stimulation conditions (PNS #1, red; PNS #2, yellow) and both vibration conditions (Vibr #1, blue; Vibr #2, teal) for each participant (LL01, LL02, LL03).

FIGURE 6

Mediolateral CoP location full responses. The responses to tactile stimuli in the full 20 s trial were averaged across six trials. The gray dashed line indicates the onset of the stimulus. Positive mediolateral center of pressure (M/L CoP) values were in the direction of the amputated limb (Amp) and negative values were in the direction of the intact limb (Int). The mean CoP across multiple trials is depicted with a solid line and the shaded regions around each line show the standard error. The center of pressure location for the control condition (No Stim, gray) is shown on each plot for comparison against the locations for both peripheral nerve stimulation conditions (PNS #1, red; PNS #2, yellow) and both vibration conditions (Vibr #1, blue; Vibr #2, teal) for each participant (LL01, LL02, LL03).

FIGURE 7

Anterior-posterior CoP location initial responses. The initial responses (first 3 s) to tactile stimuli were averaged across six trials. The gray dashed line indicates the onset of the stimulus. Positive anterior-posterior center of pressure (A/P CoP) values were in the forward direction (Fwd) and negative values were in the backward direction (Back). The mean CoP across multiple trials is depicted with a solid line and the shaded regions around each line show the standard error. The center of pressure location for the control condition (No Stim, gray) is shown on each plot for comparison against the locations for both peripheral nerve stimulation conditions (PNS #1, red; PNS #2, yellow) and both vibration conditions (Vibr #1, blue; Vibr #2, teal) for each participant (LL01, LL02, LL03).

FIGURE 8

Anterior-posterior CoP location full responses. The responses to tactile stimuli in the full 20 s trial were averaged across six trials. The gray dashed line indicates the onset of the stimulus. Positive anterior-posterior center of pressure (A/P CoP) values were in the forward direction (Fwd) and negative values were in the backward direction (Back). The mean CoP across multiple trials is depicted with a solid line and the shaded regions around each line show the standard error. The center of pressure location for the control condition (No Stim, gray) is shown on each plot for comparison against the locations for both peripheral nerve stimulation conditions (PNS #1, red; PNS #2, yellow) and both vibration conditions (Vibr #1, blue; Vibr #2, teal) for each participant (LL01, LL02, LL03).