Anatomical correlates of proprioceptive impairments ... · used more modern quantitative robotic assessments to measure proprioception....

Page 1

Journal of the Neurological Sciences 342 (2014) 52–61

Contents lists available at ScienceDirect

Journal of the Neurological Sciences

j ourna l homepage: www.e lsev ie r .com/ locate / jns

Anatomical correlates of proprioceptive impairments following acutestroke: A case series

Jeffrey M. Kenzie a, Jennifer A. Semrau a, Sonja E. Findlater a, Troy M. Herter b, Michael D. Hill c,Stephen H. Scott d,e, Sean P. Dukelow a,c,⁎a Division of PhysicalMedicine and Rehabilitation, Department of Clinical Neurosciences, Hotchkiss Brain Institute, Faculties ofMedicine and Kinesiology, University of Calgary, 2500 University Dr.NW, Calgary, AB T2N 1N4, Canadab Department of Exercise Science, University of South Carolina, 1300 Wheat St, Columbia, SC 29208, USAc Calgary Stroke Program, Department of Clinical Neurosciences, Hotchkiss Brain Institute, University of Calgary, 2500 University Dr. NW, Calgary, AB T2N 1N4, Canadad Department of Anatomy and Cell Biology, Queen's University, Botterell Hall, Room 219, Kingston, ON K7L 3N6, Canadae Providence Care, St. Mary's of the Lake Hospital, 340 Union St, Kingston, ON K7L 5A2, Canada

⁎ Corresponding author at: 1403 29th St NW, South ToT2N 2T9, Canada. Tel.: +1 403 944 5930; fax: +1 403 94

E-mail address: [email protected] (S.P. Dukelow).

http://dx.doi.org/10.1016/j.jns.2014.04.0250022-510X/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 15 January 2014Received in revised form 31 March 2014Accepted 21 April 2014Available online 27 April 2014

Keywords:StrokeProprioceptionKinesthesiaRoboticsMagnetic Resonance ImagingVisually Guided Reaching

Background: Proprioception is the sensation of position andmovement of our limbs and body in space. This senseis important for performing smooth coordinated movements and is impaired in approximately 50% of strokesurvivors. In the present case series we wanted to determine how discrete stroke lesions to areas of the brainthought to be critical for somatosensation (thalamus, posterior limb of internal capsule, primary somatosensorycortex and posterior parietal cortex) would affect position sense and kinesthesia in the acute stages post-stroke.Given the known issueswith standard clinicalmeasures of proprioception (i.e. poor sensitivity and reliability)weused more modern quantitative robotic assessments to measure proprioception.Methods:Neuroimaging (MRI, n= 10 or CT, n= 2)was performed on 12 subjects 2–10 days post-stroke. Propri-oception was assessed using a KINARM robot within the same time frame. Visually guided reaching was alsoassessed to allow us to compare and contrast proprioception with visuomotor performance.Results and Conclusions: Proprioceptive impairments were observed in 7 of 12 subjects. Thalamic lesions (n= 4)were associated with position sense (n = 1) or position sense and kinesthesia (n = 1) impairments. Posterior

limb of the internal capsule lesions (n= 4) were associatedwith primarily position sense (n= 1) or kinesthesia(n=2) impairments. Lesions affecting primary somatosensory cortex and posterior parietal cortex (n=2)wereassociated with significant position sense and kinesthesia impairments. All subjects with damage to hypothe-sized structures displayed impairments with performance on the visually guided reaching task. Across theproprioceptive tasks, we saw that position sense and kinesthesia were impaired to differing degrees, suggestinga potential dissociation between these two components of proprioception.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Proprioception is a term used to describe the knowledge of thelocation and movement of our limbs in space [1]. Classically, it hasbeen considered to have two subcomponents: position sense and kines-thesia. Position sense is the perception of static limb location, whereaskinesthesia is the sensation of limb or joint motion [2].

Proprioceptive impairment following stroke has been reported tooccur in approximately 50% of patients [3–5]. While many post-strokestudies focus on motor function, impairments in proprioception havebeen linked to postural instability [6], impairedmotor recovery [7], safe-ty concerns, as well as longer hospital stays and decreased functional

wer Room 905, Calgary, Alberta4 0977.

independence at discharge [8,9]. Proprioception can also predict long-termmotor recovery after stroke [10,11] and has been strongly correlat-edwithmotor recovery of the hemiplegic arm after stroke [12]. Howev-er, the relationship between lesion location and specific proprioceptiveimpairment remains poorly understood.

Prior studies attempting to link neuroimaging and proprioception instroke have relied on finger position sense or proprioception measuredby standard clinical assessment [13–17]. These studies have reportedproprioceptive impairments following thalamic lacunar stroke [14,15,17], posterior limb of the internal capsule stroke [16] and cortical stroke[13,16].

Most clinical tests of proprioception involve a patient’s ability todiscriminate between the upward or downward position of a digitwhen passively moved [18]. This test and other clinical assessments ofproprioception such as the thumb localizing test [19], are based on ordi-nal scales, show relatively poor inter-rater reliability and lack sensitivity

Page 2

53J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

[20,21]. Further, these tools were not designed to differentiate betweenimpairments in position sense and kinesthesia following stroke.

Robotic assessments have recently been developed to quantitatesensorimotor impairments following stroke [5,22,23]. These robotic as-sessments are relatively quick to administer and can quantitate positionsense reliability [5], while also providing insight into other aspects ofproprioception, such as kinesthesia [24]. Further, subjects can easilycomplete a brief battery of robotic tasks to assess various aspects of be-havior, including motor function. This allows the observer to compareand contrast differences in sensory versus motor performance after astroke. Pairing these assessment methods with lesion analysis afterstroke may allow for an improved ability to interpret the behavioralconsequences of a particular stroke lesion location.

The present study examined twelve stroke survivors with acutelesions to structures believed to be involved with proprioception andevaluated their performance on three different robotic tasks. Compari-sons were made of the subjects’ performance on tasks measuringposition sense, kinesthesia and a standard motor task (visually guidedreaching). We hypothesized that damage to the following structures:ventral posterior lateral (VPL) nucleus of the thalamus, the posteriorlimb of the internal capsule (PLIC), the post-central gyrus (S1) and pos-terior parietal cortex (PPC)would producemeasureable impairments insensorimotor function. Further, we made comparisons to two acutestroke subjects without damage to these brain areas to demonstrate abehavioral dissociation based on lesion location.

2. Methods

2.1. Subjects

A total of 12 subjects with first diagnosis of clinical stroke were re-cruited from the Foothills Medical Centre (FMC) or the Dr. Vernon Fan-ning Centre (VFC) in Calgary, Alberta, Canada. Ten cases were chosenbased on lesions identified on neuroimaging to VPL thalamus, PLIC, S1or PPC. Two caseswere chosen because they had no damage to these re-gions. Subjects had no other neurological diagnoses (including previousstroke) and cognition and language were sufficient to follow the in-structions required to complete the assessments. Neuroimaging wasconducted amean of 1.5 days (SD 2.2) post-stroke and clinicalmagneticresonance imaging (MRI) or computed tomography (CT) scanswere ob-tained according to the standard acute stroke protocol at the FoothillsMedical Centre for use in the present study. Axial T2-weighted FLAIR(fluid attenuated inversion recovery) images were used to depict lesionlocation. Subjects participated in both a clinical and robotic assessmentamean of 7.3 days (SD 3.7) post-stroke. This studywas approved by theUniversity of Calgary Research Ethics Board.

2.2. Lesion Delineation

Although we based inclusion in the study on the appearance of thestroke lesions on the clinical neuroimaging scans, in order to accuratelyquantify the burden of lesion in our hypothesized areaswe performed aregion of interest (ROI) analysis. Regions of interest for each area (thal-amus, posterior limb of the internal capsule, post-central gyrus and su-perior parietal lobule)were first drawn on the T1MontrealNeurologicalInstitute (MNI) template brain with MRIcron [25] (www.mricro.com)using both a white matter atlas [26] and Myeloarchitectonic Atlas [27].Lesion location of each subject was then demarcated directly on corre-sponding slices of the T1-weighted MNI template brain in MRIcron byclosely examining the FLAIR and diffusion weighted imaging (DWI)for those with MRI and CT for those without. This procedure is consis-tent with previously reported methods [28–30]. ROI analysis wasperformed using Non-Parametric Mapping (NPM) software (availablewith the MRIcron software package). This provided the percentage ofeach ROI that was damaged in each subject and the percentage of an

individual’s lesion that was located within the borders of each ROI. Le-sion volume was also obtained through NPM.

2.3. Robotic Assessment

Robotic assessmentwas performed using a KINARMrobotic exoskel-eton (BKIN Technologies Ltd., Kingston, Ontario, Canada) (see Fig. 1A).Subjects were seated in the wheelchair base with both arms supportedagainst gravity by the robotic exoskeleton in the horizontal plane (~80°shoulder flexion) and the exoskeleton was adjusted to fit each subject’sbody dimensions (height, limb segment length) by the study therapist.The robot allowed subjects to move freely in the horizontal plane withflexion and extension movements of the elbow joints and shoulderjoints. The KINARM monitored and recorded arm movement, and ap-plied mechanical loads to the shoulder and elbow joints during passivemovements used in the position matching and kinesthesia tasks.

2.4. Arm position matching task

The arm position matching task (Fig. 1B) was used to quantify posi-tion sense of the upper extremities. This taskwas performedwithout vi-sion, as previously described in detail by Dukelow and colleagues [5].The position matching task required the subject to move his/her unaf-fected/less affected arm (active arm) to mirror match the end positionof the stroke affected arm that was passively moved by the KINARM(passive arm). The robotmoved the subjects’ passive arm to nine differ-ent spatial locations pseudorandomly, with the subject matching eachlocation with the active arm before moving onto the next trial. Thepassive arm was moved to each target 6 times for a total of 54 trials.Three parameterswere derived from the endpoint position of the activehand for all trials. Variability in the x and y direction (varxy) measuredtrial to trial consistency of the end position of the active arm. Spatialcontraction/expansion (cont/expxy) measured the ratio of the totalarea of workspace matched by the active arm relative to the passivearm. Systematic shifts (shiftxy) measured consistent errors between ac-tive and passive arms. Consistent errors were measured as the meanerror between passive and active hands for each target location acrossall trials in the x direction, y direction and combined xy. The averageof these mean errors in the combined xy coordinate then denoted themagnitude of systematic shift. Normal ranges for each parameter werederived from the 95% confidence interval from 170 control subjectswith consideration for age, sex and handedness [31]. Overall, wefound that 95% of controls failed 1 or fewer parameters on the arm po-sition matching task. Thus, we created a task failure threshold forsubjects with stroke of 2 or more parameters [31].

2.5. Arm kinesthesia task

This kinesthesia task (Fig. 1C) was used to examine kinesthesia ofthe upper extremities. This task (also previously described) [24]was ad-ministered without vision of the upper extremities. This task requiredthe subjects to use their active (unaffected/less affected) arm to mirrormatch the movement of their passive (stroke affected) arm that wasbeingmoved by the KINARM. Prior to the start of each trial, the subject’shands were moved to one of three pre-set mirrored locations in theworkspace. The passive armwas thenmoved to one of two other targetlocations by the robot and subjects were instructed to use their activearm tomirrormatch the speed, direction and distance of themovementas soon as they felt the robot move, thereby attempting to mimic thepassive movement in real time. A total of 6 movement directions wereperformed 6 times each for a total of 36 trials.

Four parameters were used tomeasure the temporal and spatial (x,y)aspects of movements of each subject [24]: Response Latency (RL)—the time betweenmovement initiation (point where subject reached10% of hand speed maximum) of the passive arm and active arm.Peak Speed Ratio (PSR) — the ratio of maximum passive arm speed

Page 3

A

D

B

C

KINARM exoskeleton (BKIN technologies, Kingston, ON)

Active arm Passive arm

Active arm Passive arm

Left arm Right arm

time (s)spee

d (c

m/s

)

time (s)

spee

d (c

m/s

)

time (s)

spee

d (c

m/s

)

Fig. 1.Control exemplarsA.Picture of KINARM robotic exoskeleton (BKINTechnologies, Kingston,ON). Subjects are seated in thewheelchair basewith both arms supported against gravityby the robot in arm troughs.B.Overhead viewof exemplar subject for PositionMatching task. Passive arm (roboticallymoved) positions are represented byfilled symbols (connectedwithblack lines). Active arm (subject moved) mirror matched positions are represented by open symbols (connected with grey lines). Ellipses represent one standard deviation. C. Overheadviewof an exemplar subject performing the kinesthesia task in one of the six possible directions tested. Temporalmeasurements of active arm shown on the left. Passive armmovement isrepresented by the black lines. Active armmovements are represented by the gray lines.D.Exemplar control data for the Visually GuidedReaching task. Black circles represent targets, greylines represent subjectmovement. Subjects were required to reach from centre target to peripheral targets as quickly and accurately as possible. Temporal measurements for the forwardreaching movement only (centre target to top target) are displayed for both arms.

54 J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

and maximum active arm speed. Initial Direction Error (IDEKIN) — theabsolute angular deviation between passive and active hands at thepoint of hand speed maximum. Path Length Ratio (PLR)— the ratio be-tween total movement length of the subject’s active arm and passivearm. The end of active movement was the point where the subject’sarm slowed to 10% of maximum hand speed after reaching maximumspeed. Variability for all four of these parameters was also calculatedas the standard deviation across all movements (RL variability (RLv),PSR variability (PSRv), IDEKIN variability (IDEvKIN), PLR variability(PLRv)) producing a total of eight parameters. Normal ranges foreach measure were determined from the 95% confidence interval of74 healthy control subjects [24]. Overall, we found that 95% of controlsfailed 2 or fewer parameters on the arm kinesthesia task. Thus, wecreated a task failure threshold for subjects with stroke of 3 or moreparameters [24].

2.6. Visually Guided Reaching

Motor behavior was assessed with a visually guided reaching task(Fig. 1D) that has been previously described [23]. With full vision ofarms and hands, subjects made movements with their stroke-affectedarm from a central target located at approximately the centre of thearm’s workspace to one of eight peripheral visually presented targetsas quickly and accurately as possible (Fig. 1D). All targets were red cir-cles with a 1 cm radius. For the duration of the task, a white circle wasused to represent the subjects’ index finger. To start each trial the sub-ject placed the tip of their index finger in the central target for 1250 to1750 ms. One of the eight peripheral targets was then illuminated and

subjects were given 3000 ms to complete the reach (central targetextinguished when hand exits target). Targets were presented in arandom-block design along with two catch trials where no peripheraltarget was illuminated. This was repeated 8 times for a total of 80 trialsfor each arm.

Five movement parameters were recorded. These parameters havepreviously been described in detail [23]. Postural control was deter-mined based on postural hand speed (PS) for 500 ms prior to targetillumination. Reaction time (RT) was the time between peripheral tar-get illumination and movement onset. Initial movement directionerror (IDEVGR) was the angular deviation between subject’s initialmovement path (point of first hand speed minimum) and the straightpath to the target. The number of corrective movements was countedas the number of speed peaks per reach (NSP). Finally, total movementtime (MT) was used as the total movement parameter. Normal controlranges for each parameter were determined as the 95% confidence in-terval derived from 231 healthy controls [32]. Overall, we found that95% of controls failed 1 or fewer of these parameters on the visuallyguided reaching task. Thus, we created a task failure threshold forsubjects with stroke of 2 or more parameters [32].

2.7. Clinical Assessment

Clinical assessments were performed by a trained study therapist.The Thumb Localizing Test (TLT) [19] was used to test upper extremityposition sense of the stroke affected arm. It has been used previously ina number of studies on individualswith stroke [5,33,34]. During this testthe subject’s eyes were closed and the therapist manipulated the

Page 4

Table 1Subject demographics, imaging results, clinical measures and robotic measures.

Subject Age Sex Vascular Vol. Clinical Measures # of Robotic ParametersFailed

Territory (cm3) Hand. TLT FIM CMSA (1–7) PPB (#) BIT PM KIN VGR

1 65 M L PCA (I) 0.83 R 0 105 A: 5, H: 6 R: 9, L: 11 143 0/3 0/8 3/5⁎

2 72 F L PCA (I) 2.03 R 1 102 A: 6, H: 5 R: 7, L: 10 143 1/3 1/8 4/5⁎

3 58 F L PCA (H) 4.14 R 1 115 A: 5, H: 6 R: 8, L: 10 145 2/3⁎ 2/8 3/5⁎

4 59 M R MCA (H) 15.29 R 2 77 A: 2, H: 4 R: 11, L: 0 143 2/3⁎ 8/8⁎ 5/5⁎

5 49 M R MCA (I) 4.18 R 1 105 A: 6, H: 5 R: 11, L: 2 143 2/3⁎ 1/8 3/5⁎

6 41 M R MCA (I) 2.53 R 0 118 A: 7, H: 6 R: 12, L: 4 141 1/3 4/8⁎ 4/5⁎

7 48 M L MCA (I) 3.11 R 0 78 A: 1, H: 1 R: 0, L: 14 145 1/3 4/8⁎ 5/5⁎

8 58 M L MCA (I) 2.11 R 0 82 A: 4, H: 4 R: 0, L: 10 142 0/3 2/8 3/5⁎

9 56 M R MCA (I) 57.3 R 2 62 A: 5, H: 5 R: 8, L: 0 86 2/3⁎ 6/8⁎ 5/5⁎

10 65 M R MCA (I) 118.4 R 2 67 A: 5, H: 5 R: 9, L: 2 115 2/3⁎ 7/8⁎ 3/5⁎

11 31 F L MCA (I) 10.3 R 0 120 A: 7, H: 7 R: 15, L: 14 145 0/3 0/8 0/512 33 F R PCA (I) 77.3 L 1 125 A: 7, H: 7 R: 15, L: 14 139 0/3 0/8 1/5

MCA:Middle Cerebral Artery, PCA: Posterior Cerebral Artery, I: Ischaemic stroke, H: Hemorrhagic stroke. Hand: handedness prior to stroke. CMSA: ChedokeMcMaster Stroke AssessmentImpairment Inventory for the Upper Extremity (affected side), greater score = increase in motor function, A = arm, H = hand. FIM: Functional Independence Measure. TLT: ThumbLocalizing Task, 0 = no difficulty finding thumb, 3 = completely unable to locate thumb (best of two trials for affected side). PPB: Purdue Peg Board, best score of two trials reported.BIT: Behavioural Inattention Test. PM: Position Matching, KIN: Kinesthesia, VGR: Visually Guided Reaching.⁎ Indicates failure of the robotic task.

55J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

shoulder and elbow joint of their stroke affected arm to place it some-where in space above eye level. The subjects were then asked to usetheir unaffected side to pinch the thumb of the stroke affected armbeing held in a static position by the therapist. Subjects were scoredon a scale of 0 (accurately performs the task) to 3 (unable to locatethumb).

The Modified Edinburgh Handedness Inventory determined hand-edness of each subject prior to stroke [35]. The Chedoke–McMasterStroke Assessment (CMSA) Impairment Inventory [36] was used tomeasure movements of the arm and hand. Subjects were scored on ascale from 1 (flaccid paralysis) to 7 (normal motor function) based ontheir ability to perform a set of pre-determined movements. The Func-tional Independence Measure (FIM) [37] was used to assess functionalabilities. A maximum score of 126 indicates functional independencewith daily activities. The Behavioural Inattention Test [38] was used asan indicator of visuospatial neglect. Finally, to assess manipulative dex-terity of the hands, the Purdue Pegboard (PPB) test was administered(LaFayette Instrument Co, LaFayette, IN) [39]. Using one hand at atime subjects were required to place as many metal pegs into apegboard as possible in 30 seconds.

3. Results

3.1. Clinical and Control data

Clinical assessment scores and demographics for each subjectwith stroke are described in Table 1. Ten subjects had MRI while twounderwent CT. The average age of all subjects with stroke was 53 years(31–72) and all subjects were right handed prior to the stroke exceptfor subject 12. The smallest lesion volume was in subject 1 (0.83 cm3)and the largest in subject 10 (118.4 cm3). Seven out of twelve subjectswere observed to have proprioceptive deficits of the stroke affected armbased on the Thumb Localizing Test. The average FIM score of the samplewas 96 (62–125) and the average CMSA score was 5 (1–7) for both thearmand hand (see Table 1). Themaximumnumber of pegs placed duringthe Purdue Peg Board test by subject’s stroke affected arm ranged from 0(Subjects 4, 7, 8 and 9) to 15 (Subject 11). Only subjects 9 and 10 withparietal lobe damage were identified as having visuospatial neglectaccording to the BIT.

Four subjects presented with lesions involving the thalamus(Fig. 2A), four with lesions involving the posterior limb of the internalcapsule (Fig. 3A), and two with lesions involving the primary somato-sensory cortex and posterior parietal lobe (Fig. 4A). It is important tonote that while we have classified these subjects based on lesion loca-tion, variability in lesion location within groups is observed as well as

some lesion extension into surrounding structures in some cases. Forexample, the lesion in subject 4 primarily involved the thalamus yetthere is some extension into the internal capsule as seen in Fig. 2. Wealso present two subjects with lesions outside of the areas hypothesizedto be involved in proprioception (Fig. 5) for comparison.

Fig. 1 presents exemplar healthy, non-stroke control data for all 3robotic tasks. Note the relatively small variability in errors made bythe control subject in the positionmatching task as denoted by the stan-dard deviation ellipses (Fig. 1B). The control subject easily mirrors thedirection and path of movement in the kinesthesia task (Fig. 1C). Thecontrol exemplar also demonstrates relatively straight line reachingmovements to all eight targets (Fig. 1D).

3.2. Thalamus

Fig. 2B displays performance for those subjects with lesions primar-ily involving the thalamus on the position matching task. The ROIanalysis revealed both the percentage of the thalamus that was dam-aged (subject 1 = 1.6%, subject 2 = 12.1%, subject 3 = 21.7%, subject4 = 54.5%) and the percentage of lesion that was located within thethalamus (subject 1 = 94.3%, subject 2 = 82.8%, subject 3 = 42.0%,subject 4 = 43.4%). Subject 1 was within normal ranges for all pa-rameters tested. Subject 2 was found to contract the workspace(cont/expxy = 0.41) yet fell within normal ranges for the othertwo parameters. Subject 3 fell outside the normal range for variabil-ity between trials (varxy = 10.6 cm) and contraction/expansion (cont/expxy = 0.12) whereas shift fell within the normal range (Fig. 2B).Finally, Subject 4 had increased variability (varxy = 8.8 cm) andcontracted the workspace (cont/expxy = 0.03) with little evidenceof a significant systematic shift. In summary, although abnormalitieswere observed in three subjects, only two of the four subjects withprimarily thalamic lesions were considered to have failed the positionmatching task.

Fig. 2C displays performance on the kinesthesia task. Subject 1 fellwithin normal ranges for all measurements. Subject 2 had an increasedresponse latency (RL = 970 ms) and fell within normal ranges on allother parameters of the kinesthesia task. Subject 3 had increased initialdirection error (IDEkin = 22.5°) yet fell within normal ranges for allother parameters. Subject 4 reported he was unable to feel the robotmoving his arm during this task, therefore no subject movement isdisplayed. In summary, although 3 subjects demonstrated some abnor-malities, subject 4was the only subject with a thalamic lesion that failedthe kinesthesia task.

Fig. 2D displays performance on the visually guided reaching task.Subject 1 was outside the normal range on PS (0.79 cm/s), RT (0.5 s)

Page 5

4

1

3

C DBA

2

56 J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

Page 6

57J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

and MT (1.26 s). Subject 2 was outside the normal range on IDE (9.4°),NSP (3.25) and MT (1.4 s). Subject 3 was outside the normal range onPS (1.26 cm/s), RT (0.5 s) and IDE (4.5°). Finally, subject 4 failed all pa-rameters. Since all 4 of these subjects failed 2 or more parameters, theywere all considered to have failed the visually guided reaching task [32].

3.3. Posterior Limb of the Internal Capsule

Subjects 5 to 8 presented with lesions primarily involving theposterior limb of the internal capsule. ROI analysis of these subjectsrevealed percentage of PLIC damage (subject 5 = 19.3%, subject 6 =29.0%, subject 7 = 8.3%, subject 8 = 19.6%) and the percentage oflesion locatedwithin the PLIC (subject 5= 29.5%, subject 6= 45.0%, sub-ject 7 = 22.8%, subject 8 = 42.1%). For position matching (Fig. 3B), sub-ject 5 demonstrated increased variability (varxy = 7.8 cm) with a shiftaway from midline (shiftxy = 10.8 cm). Subject 6 also had increasedvariability (varxy = 6.5 cm) between trials while falling into normalranges for the other measurements. Subject 7 was only found to havean increased contraction/expansion ratio (cont/expxy = 1.38). Subject 8fell within normal ranges for all parameters of this task. Thus, while ab-normalities were noted in three of the four subjects in this group, subject5 was the only subject with a lesion to the internal capsule that failed theposition matching task.

On the kinesthesia task (Fig. 3C), subject 5 had increased responselatency (RL = 1007 ms). Subject 6 had increased IDE (35.2°) withlarge IDEv (34.1°), PLRv (0.3) and PSRv (0.4). Subject 7 had increasedPLR (1.6) and PSR (1.7) with increased variability on both of theseparameters (PLRv = 0.3, PSRv = 0.4). Subject 8 had increased IDE(28.9°) and was variable on this parameter (IDEv = 40.3°) with allother parameters falling within normal ranges. In summary, althoughsome abnormalities were seen on all subjects, only two subjects [6,7]with lesions of the internal capsule failed the kinesthesia task.

During visually guided reaching (Fig. 3D), subject 5 demonstratedabnormalities in PS (1.0 cm/s), RT (0.5 s) and IDE (9.8°). Subject 6 felloutside the range of normal in RT (0.5 s), IDE (11.2°), NSP (3.6) andMT (2.4 s). Subject 7 was impaired all 5 parameters. Subject 8 wasimpaired on IDE (10.5°), MT (1.9 s) and NSP (2.98). Thus, all 4 subjectswere considered to have failed this task.

3.4. Somatosensory Cortex and Posterior Parietal Cortex

Subjects 9 and 10 experienced the largest impairments inmotor andproprioceptive function of all the subjects with stroke (Fig. 4). ROIanalysis informed us of the percentage of both post-central gyrus andsuperior parietal lobule damage (subject 9 = 24.3%, subject 10 =42.4%) and percentage of lesion located within these two regions (sub-ject 9 = 15.9%, subject 10 = 8.6%). During the position matching task,subject 9 had increased variability between trials (varxy = 7.3 cm)and severely contracted the workspace (cont/expxy = 0.02) (Fig. 4B).Subject 10 had the highest variability of all subjects (varxy = 26.9 cm)and was also found to contract the workspace (cont/expxy = 0.23).

On the kinesthesia task subject 9 was outside the normal range onIDE (59°), PLR (0.8) and PSR (0.35) as well as had increased IDEv(27.5°), PLRv (0.4) and RLv (0.7 s). Subject 10 had abnormal IDE(90.1°) and IDEv (42.2°), PLRv (0.7), RL (0.9 s) and RLv (0.8 s) and PSR(0.6) and PSRv (0.5).

Both subjects 9 and 10 showed impaired visually guided reachingperformance (Fig. 4A). Subject 9 fell outside the normal range on all 5

Fig. 2 Thalamus lesionsA.Axial T2weighted FLAIR images of subjects with thalamic damage disindicate lesion location. Images are oriented anatomically (left is left and right is right). B. Positioplots of the passive (affected) armposition for display purposes. Symbols and colors are the samby the dashed black line to one of six possible directions. Gray lines indicate subjects’movemenmovements of the passive arm for display purposes. Temporal measurements are shown direcmovements are displayed. Symbols and line demarcations as in Fig. 1D. Temporalmeasurementtop target).

parameters (PS = 0.3 cm/s, RT = 0.6 s, IDE = 9.2°, NSP = 4.23, MT =1.9 s) and subject 10 performed abnormally on 3 parameters (PS =0.8 cm/s, IDE = 9.6°, NSP = 3.1). In summary, both subject 9 and 10failed all three tasks.

3.5. Superior Temporal Gyrus and Lingual Gyrus

We present two subjects as acute stroke controls, one with a lesioncentered on the left superior temporal gyrus and another with a lesioncentered on the right lingual gyrus. Both of these subjects, with lesionsin brain areas presumed not to be involved in proprioception, per-formed within the normal range across all three robotic tasks (Fig. 5).Only subject 12 was impaired on one parameter of the visually guidedreaching task (IDE = 3.79°).

4. Discussion

We examined 10 subjects in the acute stage after strokewith lesionsto areas traditionally thought to be involved in somatosensory process-ing. The majority of these 10 subjects demonstrated some impairmentof position and/or kinesthetic sense as measured by the KINARM butnot always by standard clinical assessment. The two subjects with le-sions to somatosensory and posterior parietal cortex demonstratedthe most significant impairments in proprioception. Interestingly, posi-tion sense and movement sense were impaired to differing degrees infour subjects (subjects 3, 5, 6 and 7). These subjects only failed eitherthe position sense or kinesthesia tasks. All 10 subjects with lesions tosomatosensory structures demonstrated some difficulty performing vi-sually guided reaching. The 2 subjectswe examinedwith lesions outsideof our hypothesized somatosensory areas performed normally on allthree robotic tasks.

In our sample, lesions to the thalamus produced impairments inposition sense, kinesthetic matching and visually guided reachingto varying degrees. Pure sensory strokes resulting from thalamiclesions have been previously reported [14,15,40]. These resultsare not surprising as the VPL nucleus of the thalamus is known toproject somatosensory information from the extremities to thecortex [41,42]. Therefore, lesions to the VPL would directly disrupttransmission of proprioceptive information. The impairments ob-served in visually guided reaching are consistent with the knownimportance of sensory feedback from the limb to cortex for thecontrol of voluntary movement [43]. However, it may also reflectthat the area of infarct affected the descending cortocospinal tractswhich run adjacent to the thalamus. We suspect the variability ofthe performance on the three robotic tasks likely results fromboth the inter-subject differences in lesion location and size, butalso in variability in the individual neuroanatomic representationof given neurological function [44,45].

Damage to the PLIC resulted in impairments in visually guidedreaching with less significant impairment of position sense and kines-thetic sense in our sample. The internal capsule is known to transmitboth ascending sensory signals [46,47] from the thalamus aswell as de-scending motor commands from the cortex [42,48,49]. Interestingly,subject 7 presented with almost exclusively motor impairments whilepreserving position sense in the upper extremities. Similarly, subject 8showed impaired visually guided reaching of his affected arm with rel-atively intact position and kinesthetic sensation of the same arm. This issomewhat consistent with “pure” motor strokes having been observed

played in left column. Subject number is indicated on left side of each image.White arrowsnmatching task: the subject’s active (unaffected) arm position ismirror transformed ontoe as Fig. 1B. C. Kinesthesia task: movement of the passive (stroke affected) arm is depictedt of the active (unaffected arm). Movements of the active arm have beenmirrored onto thetly below movement paths. D. Visually guided reaching task: subject’s stroke affected arms are shown belowmovement paths for forward reachingmovement only (centre target to

Page 7

B C

5

6

7

A

8

D

Fig. 3. Posterior limb of the internal capsule lesions A.Axial T2 weighted FLAIR images of subjects with internal capsule damage displayed in the left column. Subject number indicated onthe left side of each image. White arrows indicate lesion location. Images are oriented anatomically (left is left, right is right). B. Position Matching Task. C. Kinesthesia Task. D. VisuallyGuided Reaching.

58 J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

Page 8

DCA

10

9

B

Fig. 4. Primary somatosensory cortex and posterior parietal cortex lesions A. Axial computed tomography images of subjects with parietal cortex damage. Subject number is indicated onleft side of image, white arrows indicate area of ischemia. B. Position Matching Task. C. Kinesthesia Task. D. Visually Guided Reaching.

59J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

with lesions involving the internal capsule previously [50,51]. Damageto the posterior limb of the internal capsule in our sample was more in-dicative of motor impairments rather than proprioceptive impairments,

BA11

12

Fig. 5. Stroke control subjectswith lesions outside of hypothesized somatosensory structuresA.insula damage and subject 12 with right lingual gyrus and occipital lobe damage B. Position M

perhaps supporting the well-known division between ascending pro-prioceptive pathways and descending motor pathways at the level ofthe internal capsule and into the corona radiata.

C D

Axial T2weighted FLAIR images of subject 11with left superior temporal lobe andposterioratching Task. C. Kinesthesia Task. D. Visually Guided Reaching.

Page 9

60 J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

It was observed that subjects 7 and8 had lesions that likely appear tohave preferentially damaged motor fibres. This was contrasted by sub-ject 6 who was less impaired in visually guided reaching and positionmatching, yet demonstrated significant impairments in kinestheticmatching (Fig. 3). While the dissociation between sensory and motorimpairments might be predicted based on the anatomy of the PLIC,the dissociation between position sense and kinesthesia is unusual,given both are thought to originate in the same end-organ receptors(mainly muscle spindles [52,53]) and ascend in the dorsal columnsthrough the thalamus and up to somatosensory cortex [41,54]. Anatom-ic distinction of the pathways for position sense and kinesthesia has notbeen reported in humans. However, we can see from subject 6 thatposition sense and kinesthetic sense may be affected separately andtherefore, consideration should be given to assessing them separately.Further anatomic mapping in humans is necessary to clarify thisseparation.

Lesions involving S1 and PPC in this sample appeared to have thegreatest impact on proprioceptive and motor function. These two sub-jects also had the largest volume of lesion (subject 9: 57.3 cm3, subject10: 118.4 cm3). S1 receives somatosensory information directly viathe VPL nucleus of the thalamus [54,55] and Brodmann’s areas 5 and 7in the posterior parietal cortex are important for integration of multiplesensory signals (ie. Visual, tactile, proprioceptive) in order to performsmooth coordinated movements [43]. These areas of parietal cortexare involved in numerous functions related to sensory processing. In aprevious lesion analysis study [56], subjects with damage to primarysomatosensory cortex and superior parietal cortex generally demon-strated impairments in proprioception. In a separate study of healthyparticipants assessed with functional neuroimaging, kinesthetic sensa-tion evoked bymuscle tendon vibrationwas shown to involve the supe-rior parietal cortex [57]. Thus it was, perhaps, not surprising to see themagnitude of the impairments seen in subjects 9 and 10. Another po-tential contributor to the impairment seen in these two subjects couldbe the presence of spatial neglect. This is especially apparent in the visu-ally guided reaching performance of subject 10 (Fig. 4) who misses leftsided targets on multiple trials. Right parietal cortex lesions have beenassociated with left sided spatial neglect in the past [58,59], which isconsistentwith our results.We are unable to drawassociations betweenlesion location, spatial neglect and proprioceptive impairment here dueto the small sample size, however this remains an important topic forfuture research.

Another factor that may have influenced our overall results could belesion volume. Subjects 1 to 4 had increasing lesion size in the thalamus,which was associated with increasing proprioceptive impairment. Thistrend was not observed in subjects 5 through 8 (PLIC), as their lesionvolumes were quite similar (Table 1). Subjects 9 and 10 (parietal lobe)had substantially larger lesion volumes (57.3 cm3 and 118.4 cm3, re-spectively) compared with the rest of the sample and also had signifi-cant proprioceptive impairment. Interestingly, subject 12 had thesecond largest lesion volume (77.3 cm3) in our sample yet demonstrat-ed normal proprioception and visually guided reaching. Current litera-ture has disputed whether or not lesion volume is related to upperextremity motor deficits post-stroke [60–62]. However, the exact rela-tionship between lesion volume and proprioceptive deficit post-strokehas yet to be determined. When we look at the percentage of each re-gion that has been damaged in our subjects we see large variability inthe amount of damage aswell as large variability in their proprioceptiveimpairment. This suggests that even a small amount of damage to thesesensory areas may have large effects on proprioceptive function afterstroke.

The two stroke controls (subjects 11 and 12) both had lesions locat-ed outside of our predicted somatosensory areas. These subjects had rel-atively large lesions (subject 11 = 10.3 cm3, subject 12 = 77.3 cm3)compared to the rest of our sample yet performed normally on theposition matching, kinesthesia and visually guided reaching tasks(Fig. 5). Subject 11 had damage primarily in the left superior temporal

gyrus, which is an area typically thought of as processing auditoryinformation and for language comprehension. Subject 12 had damageprimarily to the right lingual gyrus, an area believed to be involved inprocessing visual information. This strengthens our findings by demon-strating lesion locations that spare our hypothesized areas that are notassociated with proprioceptive impairment.

Severity of proprioceptive deficits after stroke may also be hemi-sphere dependent. Previous studies have observed that right hemi-sphere damage may result in more severe proprioceptive deficits inthe subacute phase post-stroke [5,24]. Additionally, hand dominancemay affect processing of proprioceptive feedback, as it has been pro-posed that the non-dominant hemisphere is involved in processingproprioceptive information [63–66]. This is consistent with our resultsas we found that only two (subjects 3 and 7) out of seven subjects withdominant hemisphere lesions demonstrated proprioceptive impairmentwhile all five (subjects 4–6, 9, 10) individuals with non-dominanthemisphere lesions had proprioceptive deficits (Table 1).

5. Conclusions

We have used three robotic assessments to systematically evaluateour subject’s position sense, kinesthesia and visually guided reachingperformance acutely after stroke [5,23,24] in order to draw the follow-ing conclusions. Firstly, as shown in the individual case examples inthe present study, these abilities following stroke can be impaired todiffering degrees and as such consideration should be given to assessingthem accordingly. Secondly, lesions that appeared very similar in locationon neuroimaging presented with different behavioural impairments,indicating the need to utilize a variety of assessment techniques to fullyappreciate the extent of injury after stroke. Lastly, while our results onlypertain to the acute stages of recovery following stroke future researchthat monitors changes in performance over time with quantifiablemeasures, such the ones used in this study, will inform on how specificlesions and sensory impairments impact long-term recovery.

Disclosure

Dr. Scott is cofounder and chief scientific officer of BKIN Technolo-gies, the company that commercializes the KINARM robotic device.

Acknowledgements

The present work was supported by a Canadian Institutes of HealthResearch (MOP 106662), a Heart and Stroke Foundation of Alberta,Northwest Territories and Nunavut Grant-in-Aid, an Alberta Innovates- Health Solutions Team Grant, and an Ontario Research Fund Grant(ORF-RE 04-47). J.K. was supported by a Canadian Institutes of HealthResearch Master’s Award. Special thanks to Megan Metzler and JaniceYajure for subject recruitment and assessment.

References

[1] Sherrington C. On the Proprio-ceptive System. Esecially in its Reflex Aspect. Brain1907;29(4):467–82.

[2] McCloskey D. Differences Between the Senses of Movement and Position Shown bythe Effects of Loading and Vibration of Muscles in Man. Brain Res 1973;63:119–31.

[3] Connell L, Lincoln N, Radford K. Somatosensory impairment after stroke: frequencyof different deficits and their recovery. Clin Rehabil 2008;22:758–67.

[4] Carey LM. Somatosensory Loss after Stroke. Crit Rev Phys Rehabil Med1995;7(1):51–91.

[5] Dukelow SP, Herter TM, Moore KD, Demers MJ, Glasgow JI, Bagg SD, et al. Quantitativeassessment of limb position sense following stroke. Neurorehabil Neural Repair2010;24(2):178–87.

[6] Smith DL, Akhtar AJ, GarrawayWM. Proprioception and Spatial Neglect After Stroke.Age Ageing 1983;12:63–9.

[7] Abela E, Missimer J, Wiest R, Federspiel A, Hess C, Sturzenegger M, et al. Lesions toprimary sensory and posterior parietal cortices impair recovery from hand paresisafter stroke. PLoS One 2012;7(2):e31275.

[8] Prescott RJ, GarrawayWM, Akhtar AJ. Predicting functional outcome following acutestroke using a standard clinical examination. Stroke 1982;13(5):641–7.

Page 10

61J.M. Kenzie et al. / Journal of the Neurological Sciences 342 (2014) 52–61

[9] Tyson SF, Hanley M, Chillala J, Selley AB, Tallis RC. Sensory loss in hospital-admittedpeople with stroke: characteristics, associated factors, and relationship withfunction. Neurorehabil Neural Repair 2008;22(2):166–72.

[10] Wadell I, Kusoffsky A, Nilsson BY. A follow-up study of stroke patients 5–6 yearsafter their brain infarct. Int J Rehabil Res 1987;10(4):103–10.

[11] Feys H, De Weerdt W, Nuyens G, Van De Winckel A, Selz B, Kiekens C. Predictingmotor recovery of the upper limb after stroke rehabilitation: value of a clinicalexamination. Physiother Res Int 2000;5(1):1–18.

[12] deWeerdt W, Lincoln NB, Harrison MA. Prediction of armand hand function recoveryin stroke patients. Int J Rehabil Res 1987;10(4):110–2.

[13] Derouesne C, Mas JL, Bolgert F, Castaigne P. Pure sensory stroke caused by a smallcortical infarct in the middle cerebral artery territory. Stroke 1984;15(4):660–2.

[14] Sacco RL, Bello JA, Traub R, Brust JC. Selective proprioceptive loss from a thalamiclacunar stroke. Stroke 1987;18(6):1160–3.

[15] Ja Gutrecht, Zamani AA, Deepak NP. Lacunar thalamic stroke with pure cerebellarand proprioceptive deficits. J Neurol Neurosurg Psychiatry 1992;55:854–6.

[16] Shintani S, Tsuruoka S, Shiigai T. Pure sensory stroke caused by a cerebral hemor-rhage: clinical–radiologic correlations in seven patients. AJNR Am J Neuroradiol2000;21(3):515–20.

[17] Lee MY, Kim SH, Choi BY, Chang CH, Ahn SH, Jang SH. Functional MRI finding byproprioceptive input in patients with thalamic hemorrhage. NeuroRehabilitation2012;30(2):131–6.

[18] Bickley LS. Bates' guide to physical examination and history taking; 2013.[19] Hirayama K, Fukutake T, Kawamura M. 'Thumb localizing test' for detecting a lesion

in the posterior column-medial lemniscal system. J Neurol Sci 1999;167:45–9.[20] Carey LM, Oke L, Matyas T. Impaired limb position sense after stroke: a quantitative

test for clinical use. Arch Phys Med Rehabil 1996;77:1271–8.[21] Lincoln N, Crow J, Jackson J, Waters G, Adams S, Hodgson P. The unreliability of

sensory assessments. Clin Rehabil 1991;5(4):273–82.[22] Gilliaux M, Lejeune T, Detrembleur C, Sapin J, Dehez B, Stoquart G. A robotic device

as a sensitive quantitative tool to assess upper limb impairments in stroke patients:a preliminary prospective cohort study. J Rehabil Med 2012;44(3):210–7.

[23] Coderre AM, Zeid AA, Dukelow SP, Demmer MJ, Moore KD, Demers MJ, et al. Assess-ment of upper-limb sensorimotor function of subacute stroke patients using visuallyguided reaching. Neurorehabil Neural Repair 2010;24(6):528–41.

[24] Semrau JA, Herter TM, Scott SH, Dukelow SP. Robotic identification of kinestheticdeficits after stroke. Stroke 2013;44(12):3414–21.

[25] Rorden C, Karnath HO, Bonilha L. Improving lesion-symptom mapping. J CognNeurosci 2007;19(7):1081–8.

[26] Oishi K, Faria A, van Zijl PC, Mori S. MRI atlas of human white matter. 2nd ed.Elsevier; 2011.

[27] Mai JK, Paxinos G, Voxx T. Atlas of the Human Brain. Elsevier; 2008.[28] Baier B, Mueller N, Fechir M, Dieterich M. Line bisection error and its anatomic

correlate. Stroke 2010;41(7):1561–3.[29] Rousseaux M, Honore J, Vuilleumier P, Saj A. Neuroanatomy of space, body, and

posture perception in patients with right hemisphere stroke. Neurology2013;81:1291–7.

[30] Karnath HO, Fruhmann BergerM, KukerW, Rorden C. The anatomy of spatial neglectbased on voxelwise statistical analysis: a study of 140 patients. Cereb Cortex2004;14(10):1164–72.

[31] Debert CT, Herter TM, Scott SH, Dukelow S. Robotic assessment of sensorimotordeficits after traumatic brain injury. J Neurol Phys Ther 2012;36(2):58–67.

[32] Dukelow SP, Herter TM, Bagg SD, Scott SH. The independence of deficits in positionsense and visually guided reaching following stroke. J Neuroeng Rehabil 2012;9(72).

[33] Kishi M, Sakakibara R, Nagao T, Terada H, Ogawa E. Thalamic infarction disruptsspinothalamocortical projection to the mid-cingulate cortex and supplementarymotor area. J Neurol Sci 2009;281(1–2):104–7.

[34] Bohannon RW. Evaluation and treatment of sensory and perceptual impairmentsfollowing stroke. Top Geriatr Rehabil 2003;19(2):87–97.

[35] Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory.Neuropsychologia 1971;9:97–113.

[36] Gowland C, Stratford P, Ward M, Moreland J, Torresin W, Hullenaar SV, et al.Measuring physical impairment and disability with the Chedoke–McMaster strokeassessment. Stroke 1993;24:58–63.

[37] Keith RA, Granger CV, Hamilton BB, Sherwin FS. The functional independencemeasure: a new tool for rehabilitation. Adv Clin Rehabil 1987;1(6):6–18.

[38] Wilson BA, Cockburn J, Halligan PW. Behavioural inattention test: manual. ThamesValley Test Company (TVTC); 1987.

[39] Tiffin J, Asher EJ. The Purdue Pegboard: norms and studies of reliability and validity.Division of Education and Applied Psychology, Purdue University; 1948.

[40] Kim JS. Pure sensory stroke. Clinical–radiological correlates of 21 cases. Stroke1992;23(7):983–7.

[41] Bowsher D, Albe-Fessard D. The anatomophysiological basis of somatosensorydiscrimination. Int Rev Neurobiol 1965;8:35–75.

[42] Yamada K, Nagakane Y, Yoshikawa K, Kizu O, Ito H, Kubota T, et al. Somatotopicorganization of thalamocortical projection fibers as assessed with MR tractography.Radiology 2007;242(3):840–5.

[43] Bernier PM, Grafton ST. Human posterior parietal cortex flexibly determinesreference frames for reaching based on sensory context. Neuron 2010;68(4):776–88.

[44] Schmahmann JD. Vascular syndromes of the thalamus. Stroke 2003;34(9):2264–78.[45] Shintani S. Clinical–radiologic correlations in pure sensory stroke. Neurology

1998;51:297–302.[46] Omae T, Tsuchiya T, Yamaguchi T. Cheiro-oral syndrome due to lesions in the corona

radiata. Stroke 1992;23(4):599–601.[47] Isono O, Kawamura M, Shiota J, Araki S, Hirayama K. Cheiro-oral topography of

sensory disturbances due to lesions of the thalamocortical projections. Neurology1993;43:51–5.

[48] Buhmann C, Kretschmann H-J. Computer-assisted three-dimensional reconstructionof the corticospinal system as a reference for CT and MRI. Neuroradiology1998;40:549–57.

[49] Kretschmann H-J. Localisation of the corticospinal fibres in the internal capsule inman. J Anat 1988;160:219–25.

[50] Arboix A, Padilla I, Massons J, Garcia-Eroles L, Comes E, Targa C. Clinical study of 222patients with pure motor stroke. J Neurol Neurosurg Psychiatry 2001;71:239–42.

[51] Staaf G, Lindgren A, Norrving B. Pure motor stroke from presumed lacunarinfarct: long-term prognosis for survival and risk of recurrent stroke. Stroke2001;32(11):2592–6.

[52] Goodwin GM, McCloskey DI, Matthews PB. Proprioceptive illusions induced by musclevibration: contribution by muscle spindles to perception? Science 1971;175:1383–4.

[53] Proske U, Gandevia SC. The proprioceptive senses: their roles in signalingbody shape, body position and movement, and muscle force. Physiol Rev2012;92(4):1651–97.

[54] Weber DJ, Friesen R, Miller LE. Interfacing the somatosensory system to restoretouch and proprioception: essential considerations. J Mot Behav 2012;44(6):403–18.

[55] Jones EG, Friedman DP, Hendry SHC. Thalamic basis of place- and modality-specificcolumns in monkey somatosensory cortex: a correlative anatomical and physiologi-cal study. J Neurophysiol 1982;48(2):545–68.

[56] Kim JS. Patterns of sensory abnormality in cortical stroke. Neurology 2007;68:174–80.[57] Romaiguere P, Anton J-L, Roth M, Casini L, Roll J-P. Motor and parietal cortical areas

both underlie kinaesthesia. Cogn Brain Res 2003;16:74–82.[58] Karnath HO, Rorden C. The anatomy of spatial neglect. Neuropsychologia

2012;50(6):1010–7.[59] Molenberghs P, Sale MV, Mattingley JB. Is there a critical lesion site for unilateral

spatial neglect? A meta-analysis using activation likelihood estimation. Front HumNeurosci 2012;6:78.

[60] Schiemanck SK, Post MWM, Kwakkel G, Witkamp TD, Kappelle LJ, Prevo AJH. Ische-mic lesion volume correlates with long-term functional outcome and quality of lifeof middle cerebral artery stroke survivors. Restor Neurol Neurosci 2005:257–63.

[61] Schiemanck SK, Post MW, Witkamp TD, Kappelle LJ, Prevo AJ. Relationship betweenischemic lesion volume and functional status in the 2nd week after middle cerebralartery stroke. Neurorehabil Neural Repair 2005;19(2):133–8.

[62] Page SJ, Gauthier LV, White S. Size doesn't matter: cortical stroke lesion volume isnot associated with upper extremity motor impairment and function inmild, chron-ic hemiparesis. Arch Phys Med Rehabil 2013;94(5):817–21.

[63] Roy EA, MacKenzie C. Handedness effects in kinesthetic spatial location judgements.Cortex 1978;14:250–8.

[64] Goble DJ, Brown SH. Upper limb asymmetries in the perception of proprioceptivelydetermined dynamic position sense. J Exp Psychol Hum Percept Perform2010;36(3):768–75.

[65] Goble DJ, Lewis CA, Brown SH. Upper limb asymmetries in the utilization ofproprioceptive feedback. Exp Brain Res 2006;168(1–2):307–11.

[66] Goble DJ, Noble BC, Brown SH. Proprioceptive target matching asymmetries in left-handed individuals. Exp Brain Res 2009;197(4):403–8.