Evaluation of changes in kinematic measures of three dimensional reach to grasp movements in the early subacute period of recovery from stroke

Abstract

This study examines longitudinal data of subjects initially examined in the early subacute period of recovery following a stroke with a test of reach to grasp (RTG) kinematics in an attempt to identify changes in movement patterns during the period of heightened neural recovery following a stroke. Subjects (n=8) were a convenience sample of persons with stroke that participated in an intervention trial. Baseline Upper Extremity Fugl Meyer Assessment (UEFMA) scores ranged between 31 and 52 and ages were between 49 and 83. The UEFMA and RTG test were collected prior to intervention, immediately after the intervention (approximately 18 days later post baseline) and one month after the intervention. RTG data for the uninvolved UE was collected at the one-month session. Subjects reached for objects placed on a table 10 cm from their sternums, picking them up and placing them on a target 30 cm from their acromioclavicular joints. Data was collected using an optical motion capture system. Active makers were placed on each fingertip, metacarpophalangeal, and proximal interphalangeal joint. Four additional passive markers were placed on the dorsum of the hand, the elbow, the shoulder, and the sternum. Subjects demonstrated statistically significant improvements in reaching duration, reaching trajectory smoothness, time after peak velocity and peak grip aperture. All of these measures correlated significantly with improvements in UEFMA.

Clinical relevance—

Kinematic measures of reaching and grasping collected early in the subacute period of recovery from stroke may offer insight into specific aspects of the recovery of upper extremity motor function that differ from the information gleaned from clinical scales.

I. Introduction

Impaired hand function caused by stroke has a profound impact on disability levels in persons with stroke [1] and is particularly recalcitrant to intervention [2]. One of the factors contributing to making the rehabilitation of hand function challenging is the complexity of normal upper extremity (UE) function. The upper extremity is an interdependent system that requires the shoulder, elbow and hand to act in coordination with each other [3]. The role of the upper extremity is constantly changing from primary mover to stabilizer to manipulator as one interacts with an object [4]. Each of these roles varies based on the physical, spatial and temporal characteristics of a task. The process of rehabilitating UE motor behavior is further complicated by the heterogeneous constellations of impairments caused by strokes[5].

This complexity presents challenges in the development of rehabilitation interventions, but also in the development of measures to monitor the effectiveness of these therapies. Clinical measurements of motor function are widely used to evaluate rehabilitation interventions. While ecologically valid, clinical assessments provide uncertain measures of a patient’s progress and improvement due to a lack of sensitivity and floor or ceiling effects[6]. In addition, measures that focus purely on task outcomes fail to differentiate between neural recovery processes and the development of efficient, but abnormal compensatory movement patterns [7]. By definition motor recovery is characterized by the use of premorbid motor patterns to accomplish functional tasks [8]. Multiple authors cite kinematic analysis of three dimensional reach to grasp movements as a viable means to identify the normalization of motor function in persons with stroke [9, 10].

A majority of the studies of reach to grasp kinematics examine subjects’ movement patterns at a single point in time or they are used as outcome measures in the late subacute or chronic stages of recovery as outcome measures for rehabilitation interventions[3]. These studies provide important insight into abnormal motor control and responses to training that are a combination of later stage neural recovery and the amelioration of mechanical impairments such as diminished range of motion or disuse atrophy. This study will examine longitudinal data of subjects initially examined in the early subacute period of recovery (range 6 to nineteen days post CVA, mean 12 days) with a test of reach to grasp (RTG) kinematics in an attempt to identify changes in movement patterns during the period of heightened neural recovery following a stroke [11].

II. Method

A. Participants

Subjects that participated in an intervention trial were in the early subacute phase of recovery from stroke. Subjects were selected for the intervention study based on the following inclusion criteria: a) 18–95 years old, b) unilateral right or left sided stroke, c) score of 22 or greater on the Montreal Cognitive Assessment [22], d) no hemispatial neglect or severe proprioceptive loss, e) Upper Extremity Fugl-Meyer Assessment (UEFMA) of 10–49/66, f) no receptive aphasia, g) intact cutaneous sensation. Exclusion criteria were a) orthopedic pathology limiting the ability to perform upper extremity movements without pain b) other central nervous system pathology. Subjects were included in the analyses described in this abstract if they were able to lift at least three of the four test objects from a table during their preintervention examination. All subjects received an extra 10 hours of upper extremity rehabilitation therapy sessions in addition to an inpatient rehabilitation program, that was transferred into home care and/or out-patient rehabilitation for all 8 subjects.

B. Reach-to-Grasp setup

RTG test with the affected side was conducted at the same test points as clinical measurements. RTG test with the unaffected arm was conducted once at one of the three test points. RTG test were administered using an optical motion capture system (Prime 13 cameras, OptiTrack,USA). Fifteen active makers were placed on each fingertip, metacarpophalangeal, and proximal interphalangeal joint (Figure 1). Additionally, four passive markers were placed on the dorsum of the hand, the elbow, the shoulder, and the sternum. Subjects were seated at a table with their hips and knees in a 90-degree position. Their semi pronated forearm and palm rested on a table, with the hand in front of and 10 cm away from the acromioclavicular (AC) joint. Test objects were positioned at the subjects’ midline approximately 15 cm from the subject’s chest. Objects were transported to a target approximately 30 cm from their AC joints. Each to reach to grasp movement was repeated up to 10 times based on the subject’s tolerance. The four objects included 1” and 3” sized cubic objects and, 2.5” and 4.5” diameter circular objects. Figure 1 shows the subjects’ seating position and active maker placement during the RTG test.

Fig. 1.

Reach-to-Grasp setup. A. Participant seating setup during the RTG. B. Active markers were placed on each finger joint and finger tip.

C. Metrics

Clinical and RTG related kinematics measures were conducted at baseline, immediately post 10 hours of extra therapy training, and one month after training. In addition, kinematic data was collected from the unaffected side as well.

1). Clinical measurements:

Charlson Comorbidity Index (CCI) [12] was evaluated at baseline to quantify the presence and severity of comorbidities. The National Institute of Health Stroke Scale (NIHSS) [13], the Upper Extremity Fugyl Meyer Assessment (UEFMA)[14] were conducted at all three test sessions.

2). Kinematic measurements:

The reaching movement onset was designated as the time at which the 3-dimensional wrist velocity exceeded 5 percent of the peak. The end of the reaching movement was defined as the point at which the wrist changed the movement direction to initiate transport of the grasped object. This movement was processed to calculate wrist trajectory length ratio and smoothness, time to and after peak velocity, peak velocity, peak grip aperture, and time to peak grip aperture. Trajectory Length Ratio (TLR) was calculated as the ratio between the actual reaching trajectory length and the linear distance from the starting position to the object location. Reaching Trajectory Smoothness (RTS) was evaluated by Log transformed dimensionless jerk [15]. Peak Aperture (PAp) was the largest index finger – thumb distance measured during the reaching movement. [16]. Time to peak aperture (TPAp) was calculated from the initiation of reaching to peak aperture. Peak velocity (PV) was the highest tangential velocity of the wrist measured after the reaching movement was initiated. Time to Peak Velocity (TTPV) was calculated from the initiation of reaching to peak wrist velocity. Time After Peak Velocity (TAPV) was calculated from wrist peak velocity to the initiation of the transport phase of the task. Reach Grasp Coupling (RGC) of the reaching and grasping movements was evaluated by subtracting time to peak index – thumb aperture from time to peak wrist velocity [17].

D. Data Analysis

A 1 × 3 repeated measures ANOVA was calculated to evaluate change between pre-intervention, post-intervention and one month retention test scores. Correlation between all above kinematic measurements and UEFMA were calculated using Pearson correlation coefficients. Models of Pre, Post, and one-month retention test values for six kinematic measures (three measurements for each subject) and the UEFMA (three measurements for each subject) were utilized to examine the relationship between combinations of the kinematic measures we studied and their clinical testing score. We chose an exploratory analysis using best subsets regression to evaluate all possible combinations of the six kinematic measures for the strongest correlation with the UEFMA.

III. Results

A. Participants

Eight subjects (3 females), who had a stroke within 15.87 (SD = 13.3) days, completed all three testing sessions. The average age is 60.87 (SD = 14.19) years old. See Table 1 for baseline data.

TABLE I.

Subjects baseline data

Subj ID	Days since CVA	Age	NIHSS	Charlson	UEFMA
1	11	49	4	0	45
2	11	45	2	1	37
3	15	62	6	3	40
4	15	76	4	3	46
5	6	83	6	6	31
6	13	61	2	2	41
7	19	48	1	0	52
8	8	78	4	10	35

B. Clinical Results

Mean NIHSS improved from 3.8 (SD=1.5) at baseline (Pre) to 2.3 (SD=2) immediately after training (Post) and 1.3 (SD=1.4) at one-month retention. Mean UEFMA increased from 41(SD=6) at Pre to 57 (SD=6) at Post and 60 (SD=4) at one-month retention. All eight subjects demonstrated Pre to Post improvements in UEFMA that exceeded the 10-point minimum clinically important difference (MCID) for this measure [19].

C. Kinematic Results

Mean RGD decreased from 1.6s (SD = 0.7) to 1.14s (SD = 0.6) to 0.97s (SD = 0.2). TTPV and TAPV improved from 0.43s (SD = 0.29) to 0.36s (SD =0. 27) to 0.26s (SD = 0.07) and 1.17s (SD = 0.45) to 0.79s (SD = 0.45) to 0.7 (SD = 0.23) respectively. RTS during the reaching movement was improved from 3.8 (SD =0.5) to 3.4 (SD =0.4) to 3.3 (SD =0.4). TLR decreased from 2.0 (SD =0.7) to 1.8 (SD =0.5) to 1.6 (SD =0.1). Mean and SD at one month were equal to that of the unimpaired arm for this measure. PAp increased from 0.08m (SD = 0.01) to 0.10m (SD = 0.01) and remained 0.10m (SD = 0.01) at one month. RGC decreased from 0.44s (SD = 0.15) to 0.37s (SD = 0.24) to 0.24s (SD = 0.14). Figure 2 illustrates the kinematics change over time compared with unaffected hand.

Fig. 2.

RTG kinematics were measured pre and post the training, and one month after the training. Average of all kinematics were improved and approached the kinematics measured from the unaffected side. Smaller smoothness scores indicate better performance. Error bars indicate standard deviation.

Pre to Post changes for RGD, PAp, RTS, and TAPV were statistically significant. All other measures demonstrated consistent decreases in the mean score and/or standard deviation that approached values demonstrated subjects’ unaffected hand. RGD, TPV, TAPV, TLR, RTS and PAp all demonstrated moderate to strong, statistically significant correlations with the UEFMA score. RGC showed a weak but statistically significant correlation with the UEFMA score.

Best subsets regression identified two linear combinations of variables with the strongest, statistically significant relationships with UEFMA score. The first combination, TPAp and Pap, demonstrated an R-Sq (Adj) of 59.8, an R-Sq (Pred) of 51.9 and a Mallow’s Cp of 1.2 (See Figure 4. left panel). The second combination, TAPV and TPAp, demonstrated an R-Sq (Adj) of 61.9, an R-Sq (Pred) of 52 and a Mallow’s Cp of 0.2 (See Figure 4, right panel).

Fig. 4.

Scatter plots describing the relationship between predicted UEFMA scores from best subsets regression models of RTG kinematics and actual UEFMA scores collected on the same day

IV. Discussion

Across the board improvement in NIHSS and UEFMA scores were demonstrated by all eight subjects, lending support to our hypothesis that this is an appropriate stage of recovery to study improvements in hand function associated with neural recovery. Despite the tiny sample size, the group demonstrated statistically significant improvements in trajectory smoothness as well as a trend toward significance in time to peak velocity. These metrics are both well studied indices of neural recovery in persons with stroke [3]. Changes in these measures also demonstrated statistically significant correlations with changes in UEFMA score, strengthening the case that these measures are consistent with neural recovery. Peak aperture during the reaching movement, an indicator of movement planning effectiveness [18], demonstrated statistically significant improvements and correlations with UEFMA scores as well. It was interesting to see improvements in distal motor function this early after stroke. Time to peak velocity and time after peak velocity both improved while peak velocity kept relatively consistent suggesting that improvements in total duration were related to improvements in motor planning and initiation of the task as well as improvements in the correction of the initial plan and the grasping phase of the task.

The measure we chose to evaluate the coupling between reaching and grasping movements demonstrated a trend toward unaffected hand and decreased variability over one-month retention. A weak correlation between this measure and UEFMA score suggests that this metric requires further study. This measure may not be related to recovery or could measure an aspect of neural recovery that clinical measures do not capture. Along these lines, all subjects made their largest improvements in UEFMA score during the Pre to Post training interval, but the group demonstrated larger improvements in time after peak velocity and coupling during the Post to one-month retention interval, suggesting that changes in aspects of motor control not captured by the UEFMA might have occurred.

Regression models of our kinematic data demonstrated a strong relationship between the recovery of distal function and overall recovery. Of the three variables included in the two best models produced, only TAPV is affected by a combination of proximal and distal function with the other two measures, TPAp and PAp, reflecting only on hand function. The exploratory nature of this approach to analysis precludes strong conclusions but identifies an interesting venue for future inquiry.

Limitations of this study include the small sample and the use of the unimpaired arm of persons with stroke as an indicator of “normal” function. Future studies comparing reaching performance to persons without stroke, the changes in hand and arm kinematics demonstrated by subjects in later stages of stroke recovery, and examining the correlation of kinematic measures with changes in neurophysiological measures during early recovery, will all allow for more robust evaluations of these measures as biomarkers of neural recovery.

Fig. 3.

Time to and after Peak Velocity, Peak Aperture, Time to Peak Aperture, Reach and Grasp Coupling and Trajectory Smoothness are significantly correlated with UEFMA.