1

Bilateral hand representations in human primary

proprioceptive areas

Svenja Borchersa, Till-Karsten Hauserb, Marc Himmelbacha*

a Division of Neuropsychology, Hertie-Institute for Clinical Brain Research, Eberhard Karls

University, Hoppe-Seyler-Str. 3, 72076 Tübingen, Germany

b Department of Neuroradiology, Eberhard Karls University, Hoppe-Seyler-Str. 3, 72076

Tübingen, Germany

*Corresponding author at: Division of Neuropsychology, Hertie-Institute for Clinical Brain

Research, Eberhard Karls University, Hoppe-Seyler-Str. 3, 72076 Tübingen, Germany, Tel.:

+49 7071 87600; fax: +49 7071 4489, Email: marc.himmelbach@uni-tuebingen.de

2

Abstract

Sensory representations in the postcentral gyrus are supposed to be strictly lateralised and to

provide spatially unbiased representations of limb positions. However, electrophysiological

and behavioural measurements in humans and non-human primates tentatively suggested

some degree of bilateral processing even in early somatosensory areas. We report a patient

who suffered a small and confined lesion of the hand area in the postcentral gyrus that

resulted in a proprioceptive deficit without any concomitant primary motor impairment. We

performed a finger position-matching task with target locations being defined

proprioceptively. Without visual feedback of either hand, the patient demonstrated a

significant leftward shift of perceived locations when reaching with the ipsilesional right hand

to her contralesional left hand and an opposite rightward shift when reaching with the left

hand to the position of the right hand. Although these directional errors improved when vision

of the active hand was allowed, errors were still significantly larger than those of age-matched

healthy controls with unconstrained view of the active contralesional hand. Reaching to visual

targets without visual online feedback the patient revealed comparable errors with both hands.

Reaching to visual targets with full visual feedback, she was as accurate as controls with

either hand. In summary, our data demonstrate an effect of the right postcentral lesion on

proprioceptive information processing for both hands. The results suggest an integration of

contralateral and ipsilateral proprioceptive information already at this early processing stage

possibly mediated by callosal connections.

Keywords: proprioception, reaching, lesion, postcentral gyrus, lateralisation

3

1. Introduction

The postcentral gyrus is presumably one of the best understood cortical structures in primates.

It was investigated in numerous human and animal studies ranging from electrical stimulation

in patients suffering from epileptic seizures (e.g., Cushing, 1909; Penfield and Rasmussen,

1950) to single cell recordings in monkeys (e.g. Inase et al.,1989) and functional

neuroimaging in healthy humans (e.g., Simões-Franklin et al., 2010). Going back to the

earliest descriptions of a sensory homunculus the sensory representations of the upper limbs

in primates are usually presumed to be strictly lateralised with virtually no direct interaction

between hemispheres (Gardner & Kandel, 2000; Penfield and Boldrey, 1937). However,

bilateral responses and functions are rarely explicitly addressed in research on early

proprioceptive and somatosensory respresentations because of this established and accepted

knowledge. In contrast to this view some electrophysiological experiments reported a

considerable proportion of neurons at the postcentral gyrus that responded to ipsilateral

stimulation in the non-human primate (Iwamura et al., 1994). Histological examinations

showed that this neuronal population was located in BA 2 and extended into BA 5 (Iwamura

et al., 1994).

This evidence for ipsilateral somatosensory representations from studies with monkeys is

supported by very few patient studies. It has been shown that motor impairments can occur

both contralateral as well as ipsilateral to a lesion after unilateral stroke (Schaefer et al., 2009;

Jones et al., 1989; for an overview see Gonzalez et al., 2004). But also ipsilateral

somatosensory deficits have been shown in patients with lesions of the postcentral gyrus

(Corkin et al., 1970; Boll, 1973). These studies, however, have tested only for simple sensory

recognition, or discrimination deficits in these patients but not for proprioceptive

4

localisation. Furthermore, the patients in these studies typically suffered an extensive damage

of multiple sensory and motor areas affecting multiple functional aspects of efferent

sensorimotor control and afferent feedback. Therefore, it is currently unclear whether

functionally relevant ipsilateral representations of the upper limb could also be found in

humans at a very early stage of processing, especially for the proprioceptive system.

We used a proprioceptive matching and a visual reaching paradigm to investigate the

performance of a patient (RW) who suffered from an intracerebral bleeding that almost

exclusively affected her sensory hand representation (Fig. 1). At the time of testing (3

measurements between 15 to 23 months post-stroke) she did not show any signs of

hemiparesis or other sensorimotor difficulties and she correctly recognised passive

movements of her left and right index finger and arm, as being routinely tested in neurological

examinations, without a single error. However, she reported to have persistent difficulties to

localise her left hand in space whenever it was concealed from vision.

2. Materials and Methods

2.1 Participants

Patient RW is a 67-year old, right-handed woman who suffered from a bleeding caused by

thrombosis of a cortical blood vessel 15 months prior to our first measurements. An MR scan

revealed an atypical, horseshoe-shaped, intracerebral hemorrhage with a length diameter of

3.6 cm in her right parietal cortex centered on the postcentral gyrus (Fig. 1). The chronic

damage was confined to the postcentral gyrus with a small extension across the postcentral

sulcus into the putative location of BA 7 (Fig. 1). When tested neuropsychologically, patient

RW showed neither symptoms of apraxia, neglect, nor extinction, and she had no difficulties

with visual perception/recognition as assessed by the Visual Object and Space Perception

5

Battery (VOSP, Warrington & James, 1991). Upon clinical screening she demonstrated signs

of optic ataxia with a small but significant number of reaching errors in the contralesional

field with the contralesional hand. However, this impairment resolved soon after the incidence

and was not observed in the chronic stage. During the acute phase RW also showed a slight

hemiparesis of the left arm (level of strength 4-5/5) and difficulties to detect and distinguish

up- or down-movements of the passively displaced left finger and wrist. Sensitivity and motor

functions recovered fully; at the time of testing (15-23 months later) she correctly identified

up- and down- movements of the left and right finger, wrist, elbow and shoulder. The power

she was able to exert with her hands was in the normal range (right: 18.2 kg, left 15.6 kg).

When blindfolded, RW correctly identified the site of cutaneous stimulation to each digit of

her right and left hand. In a two-point discrimination task with a distance of 3 mm, she

showed an equally accurate performance for her left and right index finger and she showed no

signs of agraphesthesia or astereognosis with either hand. She also performed equally well

with either hand in tasks of steadiness, line pursuit, aiming, and tapping that are part of the

MLS (Die Motorische Leistungsserie, Hamster, 1980). When asked for any noticeable extant

deficits, she only reported problems localising her left arm during dancing when her arm was

concealed from vision.

{please insert figure 1 about here}

Twelve healthy age-matched right-handed control subjects (9 females, 3 males) were tested

on both tasks. The age of the control group (62.67 years, Range: 57-76) was not significantly

different from RW's age (t = 0.644, p = 0.266) as tested with a significance test for single-case

statistics (Crawford et al., 2010). The control subjects had no history of head injury or

neurological disorder and had normal or corrected-to-normal vision. All participants gave

their informed consent prior to testing according to the 1964 Declaration of Helsinki and the

6

study was approved by the local ethical committee of the medical faculty.

2.2 Procedure

In experiment 1, participants were seated at a table upon which rested a miniature table (70

cm x 39.5 cm). Six targets were arranged with a distance of 10 mm to each other in a circular

array (three targets on the right, three on the left side, Fig. 2) centred around the start position

that was located at the midsagittal near end of the platform. One hand of the participant was

placed below the miniature table with the palm of the hand oriented upwards. The tip of this

hand's index finger was then passively positioned at one of the target positions by the

experimenter before each trial. The other (active) hand was placed on the start position above

the table prior to each trial. When a verbal start signal was given, subjects reached with the

upper index finger to the felt position of the opposite index finger below the miniature table.

They were encouraged to execute an uninterrupted smooth movement and were not permitted

to correct inaccurate reaches after they made contact with the table surface. However, they

were not instructed to execute especially fast movements but move at a comfortable speed.

This procedure was performed in three different viewing conditions that for each subject

occurred in the following order. In the FIX condition, participants had to fixate a fixation

point at eye level in front of them and the workspace was covered from vision using a

cardboard that was positioned below the participant's chin. In the FIXF condition, subjects

had to look at the same fixation position but they were able to see the workspace in their

lower visual field. In the FREE condition no fixation position was presented and the subjects

were told to directly look at the workspace in order to view the target area as well as their

moving hand above the table surface. For experiment 1, RW was tested in three sessions

taking place on three different days (15, 20, and 23 months post-stroke). Each target was

presented 4 times during the first two test sessions respectively and 8 times during the third

7

test session, yielding a total of 16 movements to each target in the fixation condition (FIX)

and a total of 12 trials for the FIXF and FREE condition for each target (recorded in the

second and third measurement only). Since the results from the three sessions were

comparable (Tab. 1, also see Fig. S1 in supplementary material), they were analysed together.

The targets were presented block-wise per hemi-field with a randomised order of the 3 targets

within a hemifield. The control subjects were tested once where each target occurred 8 times

per viewing condition. Between control subjects and between the different testing sessions of

RW the sequence of the hemifields as well as the sequence of the active hand was

randomised.

{please insert Table 1 about here}

In experiment 2, subjects were seated at the same table and the same array of targets as in

experiment 1. Wearing LCD shutter glasses (PLATO, Translucent Technologies), participants

pressed down a button with their index finger on the start position in the middle front of the

table prior to each trial. Each trial started when the LCD glasses turned from opaque to

transparent, revealing a yellow fixation point at the middle end of the miniature table and one

of the target positions marked by a red dot. Subjects were asked to reach to the target while

keeping their gaze on the fixation point. Again, as in experiment 1, participants were

encouraged to reach with a single movement and not to correct inaccurate reaches. Accurate

fixation in both experiments was controlled by the experimenter, and eye position was

additionally recorded using a digital video camera for later offline control. In the case of

fixation errors, trials were discarded and repeated. In the closed-loop (CL) condition, the

shutter glasses remained open for the full time of the trajectory. In the open-loop (OL)

condition, the shutter glasses turned opaque as soon as the movement was initiated, i.e. the

index finger released the start button. For experiment 2, RW (23 months post-stroke) and

8

control subjects were tested during one session and each target was presented 8 times per

viewing condition. As in experiment 1, the targets occurred in randomised order within

hemifields that were varied blockwise.

2.3 Analysis

Seven infrared light reflecting markers were attached to the right and left hand of the subject,

at each side of the wrist, half way of the os metacarpale II, the phalanx proximalis II, and to

the distal phalanxes of the thumb, index, and middle finger. Although the offline kinematic

analysis was based solely on the data of the marker at the distal phalanx of the index finger,

additional markers were included during motion tracking to improve the stability of the

motion capturing. The 3D position data of the movements were recorded with a sampling rate

of 200 Hz (Vicon Motion Systems, Oxford, UK). Data was analysed offline using custom

software based on Matlab 7.5 (Mathworks Inc., Sherborn, MA, USA). Raw data was

smoothed with an averaging window of 10 data points. Movement onset was defined from the

tangential speed of the wrist marker using a threshold of 50 mm/s. The end points of the

trajectories were determined manually at the time of contact with the table surface when the

velocity of the hand was minimal but at least below 50 mm/s. The same parameters were used

for RW as for the healthy controls. During all measurements RW produced swift and smooth

movements to the respective target positions and across all conditions we could not find any

substantial differences in trajectory kinematics despite a slightly later relative time point at

which the peak velocity was reached based on movement duration when reaching with her left

hand (Table S1, Supplementary Material). Errors of the end position were calculated as the

absolute difference, i.e. the deviation of the reached position in right-left (x) and anterior-

posterior (y) direction from the target point, and signed x- and y-errors separately. In x-

direction positive errors represent deviations in right direction and negative errors designate

9

deviations in left direction. In y-direction positive errors indicate an overshoot, while negative

errors represent an undershoot. For each target 95 % error ellipses were calculated to analyse

the variability of end points. To test within-subject differences between conditions and hands,

a 3 x 2 ANOVA was performed for the absolute error values of RW for experiment 1 and a 2

x 2 ANOVA for experiment 2. To test differences between the performance of patient RW

and the control group, significance tests for single-case statistics were performed

(SINGLIMS_ES.exe) (Crawford et al., 2010).

2.4 Anatomy

An MRI FLAIR image of patient RW was acquired 23 months post-stroke (Siemens

Magnetom Trio, Erlangen, Germany). The boundary of the lesion was manually delineated on

the FLAIR image for every transversal slice by one experimenter (M.H.) and verified by a

trained neuroradiologist (T.-K.H.) using MRIcron (http://www.cabiatl.com/mricro/). The

whole brain volume and, subsequently, the lesion mask were normalised to the T1-weighted

template from the Montreal Neurological Institute (MNI) into standard reference space using

the unified segmentation approach of SPM8 (Wellcome Department of Imaging

Neuroscience, London, UK) implemented in Matlab 7.9 (Mathworks Inc., Sherborn, MA,

USA). The lesion location was confirmed using the SPM Anatomy Toolbox (Eickhoff et al.,

2005; Bürgel et al., 2006). This atlas provides stereotaxic information on the location of

cortical areas in MNI reference space and being based on the cytoarchitectonic analysis of 10

post-mortem brains it serves as a measure of intersubject variability giving the respective

probabilities of Brodmann areas for each voxel.

3. Results

In the first experiment, we used a proprioceptive position matching paradigm to investigate

10

RW's ability to reach accurately to the felt position of the unseen index finger that was moved

passively to six possible locations (3 left of the body midsagittal line, 3 right of the body

midsagittal line, all invisible for the subject) by the experimenter below a black table surface

in front of her. The subject's accuracy was examined in three different conditions: the subjects

had to look straight ahead at a fixation position at eye level and the workspace was entirely

covered from vision (FIX), they had to look straight ahead at a fixation position at eye level

but were able to see the workspace above the table, i.e. their active hand (FIXF), or they were

instructed to directly look at the workspace above the table including their active hand

without fixation (FREE). The end positions of the reaching movements of patient RW in these

conditions are presented in Figure 2.

{please insert figure 2 about here}

Without visual feedback of either hand, RW showed a considerable leftward shift of

perceived locations when reaching with the right hand and an opposite rightward shift when

reaching with the left hand in the proprioceptive position matching task (Fig. 3 A, Tab. 2).

These errors improved significantly when vision of the active hand was allowed in conditions

FIXF and FREE as shown by a significant main effect for condition in the repeated measures

ANOVA with factors condition and hand (F[2, 68] = 19.87, p ≤ 0.001). However, not only

were position errors still significantly different from the controls if the contralesional hand

had to be localised but also if the presumably unaffected ipsilesional hand was the target

while the position of the contralesional, moving limb could be controlled visually (Fig. 3 A,

Tab. 2). Decomposing the absolute position errors into horizontal x- and vertical y-errors

shows that the absolute errors were mostly driven by errors in x-direction while the y-errors

were essentially in the normal range (Fig. 3 B and C, Tab. 3). We also found a clear hand

effect (F[1, 69] = 24.23, p ≤ 0.001) and a condition × hand interaction (F[2, 68] = 3.45, p =

11

0.037), meaning that RW showed higher absolute errors with the active left hand and

improved more with vision when moving the left hand. Additionally, the analysis of 95%

confidence ellipses showed a much higher variability of RW in comparison to controls in all

conditions with both hands (for both hands and all conditions t(11) > 4.594, p < 0.001 as

tested with the significance test for single-case statistics (Crawford et al., 2010)) (Fig. 4).

{please insert Table 2 and 3 about here}

{please insert figure 3 and 4 about here}

In a second experiment, we investigated the performance of patient RW when targets were

defined visually. The same array of six targets on the table surface and the same starting

position was used. A fixation point was shown at the middle top border of the table in order to

present targets with different eccentricities in the lower visual field. The reaching movements

were performed with each hand in two different conditions: in the open-loop condition (OL),

RW and healthy controls reached to the targets as soon as the target was visible but did not

receive visual feedback of their own movement and the target after movement onset. In the

closed-loop condition (CL), vision was allowed throughout the movement.

Reaching to visual targets in the closed-loop condition, RW was as accurate as controls with

either hand (Fig. 5). During OL reaching, however, she revealed significantly larger absolute

errors with both hands in comparison to healthy controls (Fig. 5, Tab. 2), providing further

evidence for a bilateral deficit. Interestingly, she showed a very stable tendency to overshoot

the targets with only small errors along the right-left axis (Fig. 6). This error pattern stands in

contrast to her performance in the proprioceptive task that was conducted in the same

workspace and required almost identical movement paths (Tab 4, Figure 6). As the target

array in principle represented a semi-circle around the starting position we also analysed the

12

same data encoded in a polar coordinate system. This complementary analysis supported our

abovementioned finding of higher directional errors in the proprioceptive conditions and

higher amplitude errors in the visual conditions. The respective absolute amplitude and

direction errors can be found in figure S2 of the supplementary material.

{please insert table 4 and figure 5 and 6 about here}

To analyse the anatomical structure of the lesion of RW, we used the probabilistic maps of the

SPM Anatomy Toolbox (Fig. 1). It was possible to localise 94.6 % of the lesion volume,

where 80 % was localised in the primary somatosensory cortex (39.9 % in Area 2, 19.7 % in

Area 3b, 15.9 % in Area 1, and 4.5 % in Area 3a). Less than 4 % of the voxels were also

found in Area 7 (3.9 %), Area 4 posterior (3.2 %), IPC (2.9 %), and Area 4 anterior (2.8 %).

The main fibre tracts – at least those that are already included in the probabilistic atlas – on

the contrary were spared (Fig. 1). Thus, the lesion analysis confirms with high probability that

the lesion of RW particularly covers the primary somatosensory areas.

4. Discussion

In summary, our results from both experiments suggest a bilateral proprioceptive deficit in

patient RW, which is caused by a localised lesion of early somatosensory areas of only the

right hemisphere. Surprisingly, we detected this highly significant impairment in the absence

of any notable deficits in typical clinical tests of proprioceptive function. This bilateral

impairment could be explained either by a representation of both hands in one hemisphere

that was damaged in RW or by the impact of the unilateral lesion on contralesional processing

through impaired interhemispheric connections. While these two alternatives could hardly be

disentangled based on our data presented here there is clear evidence for bilateral hand

representations in the postcentral gyrus from single-cell studies in monkeys, studies with

13

other patients and functional imaging studies with healthy controls.

Iwamura et al. (1994) reported bilateral receptive fields of hand digits in the somatosensory

cortex of macaques, specifically in the upper and middle part of Brodmann area 2 across the

border separating area 2 from areas 5 and 7 (Iwamura et al., 1994). The authors suggested that

the bilaterality of these neurons is conferred by callosal connections between homologous

areas and is necessary when information from both hands has to be integrated (Iwamura et al.,

1994). Bilateral representations in the postcentral gyrus were previously believed to be

restricted to body parts crossing the body midline, e.g. the mouth (Manzoni et al., 1989). This

assumption was formulated as the ‘midline fusion theory’, anatomical structures at the body

midline are dealt with by a fused bilateral representation. Iwamura (2000) proposed to extend

this concept also to bilateral representations of the hands, since both hands are often used

together and perform cooperative actions. Thereby our hands constitute kind of a virtual

midline through a high level of direct coordinated interactions and yield a unified image

(Iwamura, 2000). Sakata et al. (1973) investigated the characteristics of neurons in the

monkey BA 5 at the border to BA 2 posterior to the populations investigated by Iwamura et

al. (1994). They found ipsilateral or bilateral representations in 42% of these cells and

reported responses in these BA 5 neurons to multiple joint interactions as well as bilateral

interactions. Interestingly, they also found a subgroup of neurons ("matching cells") that

responded maximally when two separate body parts were brought into contact (Sakata et al.,

1973). Bioulac and Lamarre (1979) have recorded neurons from area 5 and S1 in awake

monkeys before and after deafferentation of the contralateral arm. Interestingly, in contrast to

neurons of the motor cortex and in area 5, neurons in S1 have shown no response to

movements after deafferentation, suggesting that the discharge of neurons in the postcentral

gyrus during arm movements corresponds to the processing of peripheral feedback from the

limb (Bioulac and Lamarre, 1979). These reported response characteristics are in good

14

agreement with the localisation deficit we have observed in the patient RW.

This evidence for ipsilateral somatosensory representations from studies with monkeys is

supported by several patient studies. Corkin et al. (1970) have reported ipsilateral

somatosensory deficits in patients with lesions of the postcentral gyrus. Several studies have

shown more somatosensory errors in patients with right cerebral lesions compared to patients

with left hemisphere lesions and significantly greater ipsilateral deficits after right cerebral

damage (Corkin et al., 1970; Fontenot and Benton, 1971; Boll, 1973). Brasil-Neto & de Lima

(2008) reported a decreased sensitivity to moving tactile stimuli at the ipsilateral hand for the

first time also in chronic stroke patients (Brasil-Neto and Lima, 2008). However, the reported

studies have utilised simple tactile recognition, somatosensory stimulus discrimination or

somatosensory stimulus localisation tasks but have not tested proprioceptive localisation in

these patients.

In agreement with the abovementioned observations in stroke patients and non-human

primates functional magnetic resonance imaging (fMRI) in healthy young subjects showed

that unilateral tactile stimulation of the fingers and lips of healthy subjects activates both

contralateral and ipsilateral S1 (Blatow et al., 2007). By delivering a vibrational stimulus

eliciting a movement illusion of either hand, Naito et al. (2005) found contralateral activation

in both hemispheres in BA 2 but also ipsilateral activation in right BA 2 during right hand

stimulation, arguing for a dominant role of the right hemisphere for hand proprioception

(Naito et al., 2005).

As we expected, visual feedback of the moving hand in the workspace helped RW to control

where the active hand was reaching. Interestingly, this effect was not exactly the same for

either hand as demonstrated by the interaction of the factors condition and active hand in the

analysis of the absolute errors (Fig. 3). This observation based on the absolute errors is

15

supported by a higher variability of end positions with the left hand in FIX and FIXF

condition but a contrary pattern for FREE with higher variability for the active right hand as

shown in the analysis of the confidence ellipses (Fig. 4). Thus, RW seemed to gain more from

visual feedback when moving the left hand to the concealed right hand than vice versa. Such a

pattern would be expected if the left hand was more affected than the right hand by the

damage to the right hand area despite of a general bilateral effect. In that case, RW would be

more confident about the static position of the concealed right hand than about the position of

the concealed left hand while the respective moving hand was more accurately guided

through vision in the condition FREE. In the other two conditions, larger errors for the

moving left hand might be expected based on the higher contribution of proprioceptive

feedback to online movement control (Bagesteiro et al. 2006).

The directional shift of proprioceptively perceived hand positions in experiment 1 suggests

spatially biased hand representations in the early somatosensory cortex itself or a spatial bias

in later processing areas caused by the unilateral damage. The spatial pattern resembles a

general rotation or lateral translation across the whole workspace and was not biased towards

fixation, since in the latter case directional errors should reverse across the midline. The

observed errors also do not indicate a clear reference to the shoulder positions. With RW's

overall shoulder width being 340 mm, such a reference of the proprioceptive bias towards one

or the other shoulder should have resulted in a reversal of error directions for the most

eccentric positions tested in experiment 1. Such a change was clearly not present in the data.

In contrast to the proprioceptive experiment 1, RW's performance in the closed loop condition

of experiment 2 was inconspicuous. RW obviously had no difficulties to encode the target

position correctly when it was defined visually in the first place. Furthermore, this

measurement showed that her motor control is generally intact. In the open loop condition,

16

however, she showed overshooting errors. These errors in movement amplitude stand in

contrast to the significant directional errors in experiment 1. We presume that this difference

is due to the fact that her impairment affects movement control in the two experiments in

somewhat different ways. While target positions for movement planning and control are

solely based on distorted proprioceptive information in experiment 1, she can encode the

target positions quite accurately based on the intact visual input in experiment 2 before the

start of any movement. As the control of movement distance seems to rely more on

proprioceptive input as the control of movement direction (Bagesteiro et al., 2006), it seems

plausible that RW produces predominantly movement amplitude errors in the open loop

condition of experiment 2 where no alternative visual feedback is available. This

interpretation would also be consistent with similar observations of predominant amplitude

errors in a patient with severe proprioceptive deficits due to peripheral neuropathy in a very

similar visuomotor task (Medina et al., 2010).

Our observations complement findings on proprioceptive functions reported by two

remarkable single-case studies in humans. Newport et al. (2001) reported the patient CT with

a left thalamic lesion who presented right-sided tactile extinction and astereognosis for her

right hand. Comparable to our patient RW the patient of Newport et al. (2001) also showed

normal power levels for both hands. In contrast to RW proprioceptive impairments could be

detected in this patient even in coarse clinical testing: she mirrored postures of her good arm

with the impaired arm with only "moderate success". Whereas RW's behavioural deficits

suggested some degree of bilateral impairment, CT demonstrated an isolated deficit of the

static position sense only for the right arm. Remarkably, in contrast to her bad performance in

localising her passively positioned right arm, patient CT was able to control her right arm

very accurately for guided movements without visual feedback. Considering the different

17

lesion locations in RW and CT, we might speculate that cortical disconnection of the early

proprioceptive representations results in some ipsilateral consequences due to callosal cortico-

cortical connections whereas a lesion of the thalamic relay nuclei exerts strictly unilateral

impairments. Blangero et al. (2007) reported proprioceptive deficits in two patients who

showed symptoms of optic ataxia after unilateral parietal damage inferior and posterior to the

lesion of RW. Both patients demonstrated poor performance in localisation matching tasks

similar to our first experiment without any visual feedback and with visual fixation.

Interestingly, in both patients the impairment was modulated by the visual hemifield of target

presentation, i.e. the deficit was more severe if the target hand was located in the

contralesional hemifield. Although we compared different accuracy measures between

hemifields in RW we could not find any evidence for such an effect of the visual field. The

damage in Blangero's patients (Blangero et al., 2007) affected regions posterior to the

postcentral gyrus in contrast to patient RW with a lesion that is essentially constrained to the

postcentral gyrus with a relatively small extent into the parietal cortex posterior to the

postcentral gyrus. Thus, the direct comparison between these cases supports the assumption

that an anterior to posterior proprioceptive to visual coding gradient might indeed be present

in the PPC, which also includes areas of combined coding schemes.

We are aware of the fact that interpretations based on the behaviour and anatomy of just a few

single cases are far from providing firm grounds for a concluding model of parietal

proprioceptive processing. We cannot rule out that a small overlap of the structural damage in

these individual patients, distant effects of dysfunctional connections, or undetected localised

hypometabolism in the parietal cortex produced or influenced the observed impairments in

RW, CT (Newport et al., 2001), OK, or CAN (Blangero et al. 2007). Furthermore, all these

patients were investigated in the chronic stage and thus individual changes of the functional

18

anatomy due to long-term brain plasticity cannot be ruled out without complementary

functional neuroimaging. However, with all these caveats in mind, our findings nevertheless

indicate that a unilateral localised lesion in the right postcentral gyrus affects proprioception

of both hands when asked to locate the spatial position of the opposite hand. This suggests

ipsilateral proprioceptive processing in early somatosensory areas, possibly mediated by

callosal connections. These findings highlight the demands for integrating proprioceptive

information from both hands already at an early processing stage in order to perform

coordinated actions.

19

5. Acknowledgments

We are indebted to RW for her patience and her motivation. This work was supported by the

European Union (ERC StG 211078). We are grateful to the Division of Neuropsychology

(Prof. Karnath) for the support in data acquisition.

20

6. References

Bagesteiro, L. B., Sarlegna, F. R., Sainburg, R. L. (2006). Differential influence of vision and

proprioception on control of movement distance. Exp Brain Res, 171(3): 358-70.

Bioulac, B., & Lamarre, Y. (1979). Activity of postcentral cortical neurons of the monkey

during conditioned movements of a deafferented limb. Brain Res, 172(3): 427-37.

Blangero, A., Ota, H., Delporte, L., Revol, P., Vindras, P., Rode, G., Boisson, D., Vighetto,

A., Rossetti, Y., & Pisella, L. (2007). Optic ataxia is not only 'optic': impaired spatial

integration of proprioceptive information. Neuroimage, 36 Suppl 2: T61-8.

Blatow, M., Nennig, E., Durst, A., Sartor, K., & Stippich, C. (2007). fMRI reflects functional

connectivity of human somatosensory cortex. Neuroimage, 37(3): 927-36.

Boll, T. J. (1973). Right and left cerebral hemisphere damage and tactile perception:

performance of the ipsilateral and contralateral sides of the body. Neuropsychologia, 12(2):

235-8.

Brasil-Neto, J. P., & de Lima, A. C. (2008). Sensory deficits in the unaffected hand of

hemiparetic stroke patients. Cog Behav Neurol, 21: 202-205.

Bürgel, U., Amunts, K., Hoemke, L., Mohlberg, H., Gilsbach, J. M., & Zilles, K. (2006).

White matter fiber tracts of the human brain: three-dimensional mapping at microscopic

resolution, topography and intersubject variability. Neuroimage, 29(4): 1092-105.

Corkin, S., Milner, B., & Rasmussen, T. (1970). Somatosensory thresholds--contrasting

effects of postcentral-gyrus and posterior parietal-lobe excisions. Arch Neurol, 23(1): 41-58.

Crawford, J. R., Garthwaite, P. H., & Porter, S. (2010). Point and interval estimates of effect

21

sizes for the case‑controls design in neuropsychology: Rationale, methods, implementations,

and proposed reporting standards. Cognitive Neuropsychology, 27: 245-260.

Cushing, H. (1909). A note upon the faradic stimulation of the postcentral gyrus in conscious

patients. Brain, 32: 44-53.

Eickhoff, S. B., Stephan, K. E., Mohlberg, H., Grefkes, C., Fink, G. R., Amunts, K., & Zilles,

K. (2005). A new SPM toolbox for combining probabilistic cytoarchitectonic maps and

functional imaging data. NeuroImage, 25(4): 1325-35.

Fontenot, D. J., & Benton, A. L. (1971). Tactile perception of direction in relation to

hemispheric locus of lesion. Neuropsychologia, 9(1): 83-8.

Gardner, E. P., & Kandel, E. R. (2000). Touch. In: E.R. Kandel et al., ed., Principles of

Neural Science (pp. 451-471). New York: McGraw-Hill.

Gonzalez, C. L., Gharbawie, O. A., Williams, P. T., Kleim, J. A., Kolb, B., & Whishaw, I. Q.

(2004). Evidence for bilateral control of skilled movements: ipsilateral skilled forelimb

reaching deficits and functional recovery in rats follow motor cortex and lateral frontal cortex

lesions. Eur J Neurosci, 20(12): 3442-52.

Hamster, W. (1980). Die Motorische Leistungsserie - MLS. Handanweisung. Mödling: Dr. G.

Schuhfried.

Inase, M., Mushiake, H., Shima, K., Aya, K., & Tanji, J. (1989). Activity of digital area

neurons of the primary somatosensory cortex in relation to sensorially triggered and self-

initiated digital movements of monkeys. Neurosci Res, 7(3): 219-34.

Iwamura, Y. (2000). Bilateral receptive field neurons and callosal connections in the

22

somatosensory cortex. Philos Trans R Soc Lond B Biol Sci, 355: 267-273.

Iwamura, Y., Iriki, A., & Tanaka, M. (1994). Bilateral hand representation in the postcentral

somatosensory cortex. Nature, 369: 554-556.

Jones, R. D., Donaldson, I. M., & Parkin, P. J. (1989). Impairment and recovery of ipsilateral

sensory-motor function following unilateral cerebral infarction. Brain, 112: 113-32.

Manzoni, T., Barbaresi, P., Conti, F., & Fabri, M. (1989). The callosal connections of the

primary somatosensory cortex and the neural bases of midline fusion. Exp Brain Res, 76: 251-

266.

Medina, J., Jax, S. A., Brown, M. J., Branch Coslett, H. (2010). Contributions of efference

copy to limb localization: Evidence from deafferentation. Brain Res, 1355: 104-11.

Naito, E., Roland, P. E., Grefkes, C., Choi, H. J., Eickhoff, S., Geyer, S., Zilles, K., &

Ehrsson, H. H. (2005). Dominance of the right hemisphere and role of area 2 in human

kinesthesia. J Neurophysiol, 93(2):1020-34.

Newport, R., Hindle, J. V., & Jackson, S. R. (2001). Links between vision and

somatosensation. Vision can improve the felt position of the unseen hand. Curr Biol, 11(12):

975-80.

Penfield, W., & Boldrey, E. (1937). Somatic motor and sensory representation in the cerebral

cortex of man as a studied by electrical stimulation. Brain, 60: 389-44.

Penfield, W., & Rasmussen, T. (1950). The cerebral cortex of man. A clinical study of

localization of function. New York: Macmillan.

Sakata, H., Takaoka, Y., Kawarasaki, A., & Shibutani, H. (1973). Somatosensory properties

23

of neurons in the superior parietal cortex (area 5) of the rhesus monkey. Brain Res, 64: 85-

102.

Schaefer, S. Y., Haaland, K. Y., & Sainburg, R. L. (2009). Hemispheric specialization and

functional impact of ipsilesional deficits in movement coordination and accuracy.

Neuropsychologia, 47(13): 2953-66.

Simões-Franklin, C., Whitaker, T. A., & Newell, F. N. (2011). Active and passive touch

differentially activate somatosensory cortex in texture perception. Hum Brain Mapp, 32(7):

1067-1080.

Warrington, E., James, M. (1991). Visual Object and Space Perception Battery. Bury St.

Edmunds: Thames Valley Test Company.

24

Figure Captions

Figure 1: MRI scan of patient RW. Top row: normalised FLAIR image of RW acquired 23

months post-stroke showing the lesion centred on the right postcentral gyrus. The horizontal

lines on the coronal image (y = -32) indicate the height of the axial slices (z = 46, z = 56, z =

66) shown in MNI coordinates. Middle row: an overlay of the lesion of patient RW (dark

blue) and an overlay of all areas of the SPM Anatomy Toolbox shown in different colors (S1

(Area 1, 2, 3a, 3b) in red, M1 (Area 4a, 4p) in green, SPL5 in violet, SPL7 in cyan, AIPS in

brown, IPC in yellow). Bottom row: an overlay of the lesion of patient RW (dark blue) with

all white matter fiber tracts of the SPM Anatomy Toolbox is shown in different colors

(corticospinal tract in cyan, callosal body in violet, optic radiation in green, cingulum in

yellow, fornix in light blue, superior longitudinal fascicle in red). As for the “maximum

probability maps” provided by the SPM Anatomy Toolbox (Eickhoff et al., 2005) the

thresholds for including a voxel into the map of all structures was set to be higher than 30 %,

i.e., an assigned voxel had a probability of ≥ 40% for being part of the defined region.

Figure 2: Proprioceptive reaching of RW. The end points of her reaching movements towards

the proprioceptively defined targets are shown when reaching with the active left hand

towards the passively moved right target hand (left column) and when reaching with the right

hand towards the left target hand (right column). The starting position of her active reaching

hand is in the middle front (small dot), while the end points for the six targets are shown in

different tokens. The behaviour of RW is shown for the three visual conditions: fixation

straight ahead and the reaching hand was occluded from vision (FIX), fixation straight ahead

but the reaching hand was visible in the periphery (FIXF), and free vision of the reaching

25

hand (FREE).

Figure 3: Proprioceptive reaching errors of RW and controls in mm. Mean errors are shown

for the control group with standard deviations. Errors are shown for the left (LH) and right

hand (RH) for all conditions: fixation straight ahead and the reaching hand was occluded from

vision (FIX), fixation straight ahead but the reaching hand was visible in the periphery

(FIXF), and free vision of the reaching hand (FREE). a) Absolute reaching errors. b) Positive

x-errors designate errors in right direction, whereas negative x-errors indicate errors in left

direction of the target. c) Positive y-errors represent an overshoot relative to the target

position, whereas negative y-errors indicate an undershoot-error.

Figure 4: Areas of 95 % error ellipses in mm2 of proprioceptive reaching end points of RW

and controls. Mean areas are shown for the control group with standard deviations. Areas are

shown for the left (LH) and right hand (RH) for all conditions.

Figure 5: Visual reaching errors of RW and controls in mm. Mean errors are shown for the

control group with standard deviations. Errors are shown for the left (LH) and right hand

(RH) for both conditions, closed-loop (CL) and open-loop (OL). a) Absolute reaching errors.

b) Positive x-errors designate errors in right direction, whereas negative x-errors indicate

errors in left direction of the target. c) Positive y-errors represent an overshoot relative to the

target position, whereas negative y-errors indicate an undershoot-error.

26

Figure 6: Visual reaching of RW. The end points of her reaching movements towards the

visually defined targets are shown when reaching with the left hand (left column) and when

reaching with the right hand (right column). The starting position of her reaching hand is in

the middle front (small dot), while the end points for the six targets are shown in different

tokens. The fixation dot was presented at position [0/200] in the middle back part of the table.

The behaviour of RW is shown for the two visual conditions: with visual feedback in the

periphery while reaching (CL), and when vision was occluded as soon as the movement was

started (OL).

27

Table 1: Mean absolute errors of RW for each condition for all three days of measurements

of experiment 1. The mean of the absolute errors of all six targets is given in mm. Condition

FIX was measured on three different days and 16 data points were acquired for each target,

while condition FIXF and FREE were measured on two different days with a total of 12 data

points for each of the six targets.

Left Hand Right Hand Measured

points per

target FIX FIXF FREE FIX FIXF FREE

Day 1 87.54 - - 59.97 - - 4

Day 2 63.97 78.19 50.49 68.20 55.31 64.40 4

Day 3 102.24 67.76 45.47 86.35 61.67 36.98 8

Mean 89.00 71.24 47.14 75.22 59.55 46.12 16/12/12

28

Table 2: Absolute errors of experiment 1 and 2. The mean of the absolute error of all six

targets is given in mm for the control group (with standard deviation) and patient RW for all

conditions and both hands. The results of the significance test are given as well as effect sizes

calculated with a significance test for single-case statistics (Crawford et al., 2010).

Control group Significance

test

Estimated % of

the control group

obtaining a

lower score than

RW

Estimated effect

size

Task Hand n Mean SD RW t p Point 95 % CI Point 95 % CI

Proprio

FIX

Left 12 38.72 15.90 89.00 3.038 0.006 99.44 95.84 -

100.00

3.162 1.73 -

4.57

Proprio

FIX

Right 12 33.03 14.60 75.22 2.777 0.009 99.10 94.07 -

100.00

2.89 1.56 -

4.20

Proprio

FIXF

Left 12 34.42 11.62 71.24 3.044 0.006 99.44 95.87 -

100.00

3.17 1.74 -

4.58

Proprio

FIXF

Right 12 32.71 11.88 59.55 2.171 0.026 97.36 87.63 -

99.96

2.26 1.16 -

3.34

Proprio

FREE

Left 12 24.83 9.37 47.14 2.288 0.021 97.85 89.17 -

99.98

2.38 1.24 -

3.50

Proprio

FREE

Right 12 27.25 9.97 46.12 1.818 0.048 95.18 81.98 -

99.78

1.89 0.92 -

2.84

Point

CL

Left 12 11.37 3.72 14.54 0.819 0.215 78.48 56.83 -

93.38

0.85 0.17 -

1.51

Point

CL

Right 12 12.14 3.42 14.73 0.728 0.241 75.90 53.90 -

91.78

0.76 0.10 -

1.39

Point

OL

Left 12 19.06 5.42 35.88 2.98 0.006 99.38 95.50 -

100.00

3.10 1.70 -

4.49

29

Point

OL

Right 12 19.42 8.22 36.05 1.944 0.039 96.1 84.17 -

99.87

2.023 1.00 -

3.02

30

Table 3: x- and y-errors of experiment 1. The means of the x- and y-errors of all six targets

are given in mm for the control group (with standard deviation) and patient RW for all

conditions and both hands. The results of the significance test are given as well as effect sizes

calculated with a significance test for single-case statistics (Crawford et al., 2010).

Control group Significance

test

Estimated % of

the control group

obtaining a

lower score than

RW

Estimated effect

size

Task Hand Error n Mean SD RW t p Point 95 % CI Point 95 % CI

FIX Left x 12 15.47 24.87 70.47 2.125 0.029 97.15 86.98 -

99.95

2.21 1.13 to

3.27

FIX Right x 12 -2.89 14.45 -63.26 -4.014 0.001 0.10 0.00 t-

0.92

-4.18 -5.98 - -

2.36

FIXF Left x 12 10.4 17.58 53.98 2.382 0.018 98.18 90.30 -

99.99

2.48 1.30 -

3.63 FIXF Right x 12 -1.41 12.79 -40.99 -2.973 0.006 0.63 0.00 -

4.56

-3.10 -4.48 - -

1.69

FREE Left x 12 7.42 12.65 33.23 1.960 0.038 96.21 84.44 -

99.88

2.04 1.01 -

3.04

FREE Right x 12 -10.54 13.25 -29.93 -1.406 0.094 9.37 1.14 -

26.73

-1.46 -2.28 - -

0.62

FIX Left y 12 -11.98 12.24 6.84 1.477 0.084 91.62 74.94 -

99.12

1.54 0.67 -

2.37

FIX Right y 12 -4.04 19.93 -19.85 -0.762 0.231 23.10 7.58 -

44.98

-0.79 -1.43 - -

0.13

FIXF Left y 12 -14.91 12.74 20.28 2.654 0.011 98.88 93.05 -

100.00

2.76 1.48 -

4.02

31

FIXF Right y 12 -8.97 20.03 -23.84 -0.713 0.245 24.53 8.49 -

46.57

-0.74 -1.37 - -

0.09

FREE Left y 12 2.34 13.94 20.82 1.274 0.115 88.55 69.97 -

98.20

1.33 0.52 -

2.10

FREE Right y 12 1.2 15.71 -24.09 -1.547 0.075 7.51 0.68 -

23.50

-1.61 -2.47 - -

0.72

32

Table 4: x- and y-errors of experiment 2. The means of the x- and y-errors of all six targets

are given in mm for the control group (with standard deviation) and patient R.W. for all

conditions and both hands. The results of the significance test are given as well as effect sizes

calculated with a significance test for single-case statistics (Crawford et al., 2010).

Control group Significance

test

Estimated % of

the control group

obtaining a

lower score than

RW

Estimated effect

size

Task Hand Error n Mean SD RW t p Point 95 % CI Point 95 % CI

CL Left x 12 1.14 2.15 1.15 0.004 0.498 50.17 28.73 -

71.58

0.01 -0.56 to

0.57

CL Right x 12 -1.52 2.97 -2.27 -0.243 0.406 40.64 20.54 -

62.86

-0.25 -0.82 -

0.33

OL Left x 12 4.85 6.61 -2.16 -1.019 0.165 16.51 3.91 -

37.06

-1.06 -1.76 - -

0.33

OL Right x 12 -4.77 4.81 -8.52 -0.749 0.235 23.48 7.82 -

45.40

-0.78 -1.42 - -

0.12

CL Left y 12 3.77 4.63 10.36 1.367 0.099 90.06 72.33 -

98.69

1.42 0.59 -

2.22

CL Right y 12 3.02 3.56 5.65 0.710 0.246 75.37 53.31 -

91.45

0.74 0.08 -

1.37

OL Left y 12 6.73 6.00 24.57 2.857 0.008 99.22 94.67 -

100.00

2.97 1.61 -

4.31

OL Right y 12 6.85 5.63 23.23 2.795 0.009 99.13 94.22 -

100.00

2.91 1.57 -

4.22

33

Figure Captions of the Supplementary Material

Figure S1: Proprioceptive reaching of RW during the three separate measurements. The end

points of her reaching movements towards the proprioceptively defined targets are shown

when reaching with the left hand towards the right target hand (left column) and when

reaching with the right hand towards the left target hand (right column). The starting position

of her active reaching hand is in the middle front (small dot), while the end points of the three

measurements are shown in different tokens. The behaviour of RW is shown for the three

visual conditions: fixation straight ahead and the reaching hand was occluded from vision

(FIX), fixation straight ahead but the reaching hand was visible in the periphery (FIXF), and

free vision of the reaching hand (FREE).

Figure S2: Direction and amplitude error values in a polar coordinate system of the

proprioceptive and visual reaching task. Mean errors are shown for RW and the control group

with standard deviations. Errors are shown for the left (LH) and right hand (RH) for all

conditions of the proprioceptive reaching task (fixation straight ahead and the reaching hand

was occluded from vision (FIX), fixation straight ahead but the reaching hand was visible in

the periphery (FIXF), and free vision of the reaching hand (FREE)) as well as for the visual

reaching task (closed-loop (CL) and open-loop (OL)). a) and c) Absolute direction errors in

degrees. b) Absolute amplitude errors in mm.

34

Supplementary Material

Table S1: Mean values of kinematic reaching parameters of RW and of controls (with

standard deviations (SD) across all experimental conditions). Trajectory time (ms) from start

position to the end position; PV: peak velocity (mm/s); %TPV: relative time point at which

PV was reached based on movement duration; velocity at determined end point of reaching

(mm/s); #VA: number of velocity adjustments; t-values and two-tailed p-values were

calculated with a significance test for single-case statistics (Crawford et al., 2010).

Trajectory

time (ms)

PV %TPV Velocity at

end

#VA

Left hand Controls

mean

894.93 439.4 65.89 17.77 2.03

Controls

SD

113.08 73.14 4.79 4.79 0.44

RW 821.81 510 77.98 24.79 1.92

t-value -0.621 0.927 2.425 1.408 -0.240

p-value 0.547 0.374 0.034 0.186 0.815

Right hand Controls

mean

933.1 440.79 73.37 16.61 2.14

Controls

SD

123.44 86.8 7.05 7.56 0.46

35

RW 867.98 441.08 70.12 13.74 2.3

t-value -0.507 0.003 -0.443 -0.365 0.334

p-value 0.622 0.998 0.666 0.722 0.745

Figure 1Click here to download high resolution image

Figure 2Click here to download high resolution image

Figure 3Click here to download high resolution image

Figure 4Click here to download high resolution image

Figure 5Click here to download high resolution image

Figure 6Click here to download high resolution image

Figure S1Click here to download high resolution image

Figure S2Click here to download high resolution image