Exp Brain Res (2009) 199:59–70
DOI 10.1007/s00221-009-1972-5
RESEARCH ARTICLE
Palmar arch modulation in patients with hemiparesis after
a stroke
Archana P. Sangole Æ Mindy F. Levin
Received: 26 April 2009 / Accepted: 29 July 2009 / Published online: 19 August 2009
Ó Springer-Verlag 2009
Abstract Hand shape modulation has traditionally been
studied within the framework of reach-to-grasp tasks by
examining the control of arm transport, grip aperture
scaling, and finger joint excursions. However, global
parameters characterizing arm and hand movement can be
enhanced by additional knowledge of biomechanical
changes in the hand. We previously examined palmar arch
modulation during grasping in healthy subjects by identi-
fying thenar and hypothenar displacement. This method
was used to characterize hand shape modulation in 10
stroke survivors with mild hand paresis, as assessed by the
Chedoke-McMaster clinical scale, during two types of
grasps (spherical, cylindrical). Palmar arch modulation was
examined during the three phases of prehensile movement:
transport shaping (P1), preshaping (P2), and contact
shaping (P3). Compared to the control group, the stroke
survivors showed significant differences (spherical:
F2,18 = 12.025, P \ 0.001; cylindrical: F2,18 = 9.054,
A. P. Sangole (&)
Department of Mechanical Engineering,
Ecole Polytechnique, 2900 Edouard-Montpetit,
Montreal, QC H3T 1J4, Canada
e-mail: archana.sangole@polymtl.ca
A. P. Sangole
Sainte-Justine University Hospital Center,
Montreal, QC, Canada
M. F. Levin
Center for Interdisciplinary Research in Rehabilitation,
Montreal, QC, Canada
e-mail: mindy.levin@mcgill.ca
M. F. Levin
School of Physical and Occupational Therapy,
McGill University, 3654 Promenade Sir-William-Osler,
Montreal, QC H3G 1Y5, Canada
P \ 0.001) in palmar arch modulation particularly during
P3 wherein fine adjustments are made to the grip in prep-
aration for object manipulation. While control subjects
completed most of hand shape modulation early in the task,
stroke survivors took longer to complete each phase. Fur-
thermore, stroke survivors started with a flatter hand which
required relatively more arch modulation during the latter
part of the task, thereby reflecting a temporal and spatial
concurrency between the phases. Stroke survivors with
well-recovered hand grasping ability tended to incorporate
compensations/adaptations in hand posture during specific
grasping phases. Palmar arch analysis provides us with a
more complete understanding about how hand biome-
chanics, specifically palmar concavity articulation, is
changed post-stroke. This will allow us to better identify
the motor compensations used for grasping and to re-focus
rehabilitation interventions to reduce compensations and
improve functional motor recovery.
Keywords Prehension
Grasp
Hand posture

Preshaping
Palmar arch
Hand biomechanics

Reaching
Stroke
Introduction
Reach-to-grasp tasks have been primarily described by how
the hand is transported to the object (arm transport) and
how arm–hand movement is modulated in preparation for
and during grasping (hand shape modulation; for review,
see Sangole and Levin 2007). From the perspective of
understanding the control of hand function, the classical
theory explains prehension as a system of two independent
visuomotor channels comprised of a transport component
and a grasp component (Jeannerod 1984, 1999).
123
60
Alternative perspectives are that prehensile movement is a
complex or higher order motor behavior consisting of
elementary actions that occur concurrently in a given time
window (Smeets and Brenner 1999; van Bergen et al. 2007;
Sangole and Levin 2008b) or as a result of a change from
one frame of reference to another (Pilon et al. 2007;
Feldman et al. 2007; Ma and Feldman 1995). The ele-
mentary motor actions occurring concurrently involve arm-
transport, grasp, hand rotation, or orientation (Soechting
and Flanders 1993; Stelmach et al. 1994; Desmurget et al.
1995, 1996, 1998; Desmurget and Prablanc 1997; Rand and
Stelmach 2005; Paulignan et al. 1990; Mamassian 1997,
van Bergen et al. 2007) and hand-shape modulation
(Sangole and Levin 2008a, b). Much of the understanding
of hand function has been derived from examining how the
grip aperture changes during arm transport (Santello and
Soechting 1997, 1998; Gentilucci et al. 2003; Saling et al.
1996) or the angular excursion of finger joints during
grasping (Kamper et al. 2003; Santello et al. 1998, 2002;
Mason et al. 2001; Ansuini et al. 2006).
While temporal and spatial relationships between the
elementary motor actions (Rosenbaum et al. 1995; Rossi
et al. 2002), e.g., arm–hand and trunk–arm coordination
have been well-described, there is still limited knowledge
about how stroke impacts each action. A better under-
standing of the motor deficits in post-stroke patients may be
gained by characterizing changes in each elementary action
of grasping within the framework of the whole task. Stroke
survivors with upper limb hemiparesis often have difficulty
manipulating objects during functional tasks (Brunnstrom
1970; Fries et al. 1993; Lai et al. 2002; Muellbacher et al.
2002). Grasping deficits have been described in terms of
altered hand orientation as the hand approaches the object
which may be related to compensatory movement strategies
of the arm and trunk (Roby-Brami et al. 2003; Michaelsen
et al. 2004). While altered movement strategies and their
compensations for reaching and grasping have been descri-
bed such as increased shoulder elevation, reduced elbow
extension, and shoulder horizontal adduction accompanied
by excessive trunk movement (Roby-Brami et al. 2003;
Cirstea and Levin 2000), there is less evidence associating
these strategies with deficits in the ability to modulate hand
configuration or aperture (e.g., Michaelsen et al. 2004).
The grasp component of reaching is composed of at
least two phases, one that adjusts the hand configuration
and aperture according to object shape and orientation
(opening phase) and another that closes the hand and fin-
gers on the object (closure phase). As the hand reaches to
grasp an object, the grip aperture attains a maximal width,
and then begins to establish contact with and encompass
the object (Jeannerod 1984). The time at which maximal
grip aperture is attained is a fundamental kinematic land-
mark commonly used to describe hand shape modulation
123
Exp Brain Res (2009) 199:59–70
and has been extensively used as a descriptor in reach-to-
grasp related studies (Castiello 2005). However, by the
time of maximal grip aperture, hand shape modulation has
already occurred and all that remains is the action of hand
closure on the object. Previous studies have shown that in
stroke survivors, grip aperture scaling to object size is
preserved and despite longer reaching and grasping times,
the relative time to maximal aperture occurs earlier than in
healthy controls (Michaelsen et al. 2009). On the other
hand, tasks performed using compensatory strategies to
facilitate hand opening (such as using the object to open the
hand) can be characterized by a maximal grip aperture
which is similar in magnitude to healthy subjects because it
is dictated by the size of the object. From the standpoint of
designing effective treatment interventions, not only is it
important to know where there is a deficit, e.g., differences
in grip aperture, but also it is necessary to know why there
is a difference, e.g., weakness in the thenar or hypothenar
muscles causing either flattening of the hand or unstable
grip. While the measure of maximal grip aperture can be
used to describe deficits in the coordination of reaching and
grasping and the ability to preshape the hand to object
dimensions and orientation, it does not provide insights
into the underlying associated biomechanical changes of
the hand, which are likely to exist concomitantly. Char-
acterizing grasping deficits may require descriptors more
specific to the hand which may not be apparent in measures
related to a global characterization of arm or finger
movement.
One such descriptor that may provide important addi-
tional biomechanical information about hand shape mod-
ulation in subjects with stroke is that of the palmar arch
(Sangole and Levin 2008a, b). The palmar concavity plays
an essential role in hand shape modulation to ensure a
secure or stable grasp conforming to the intended use of the
object (Kapandji 1982; Lewis 1977; Landsmeer 1962).
Palmar dynamics during two types of power grasping have
been previously described in healthy subjects. Prehensile
movement was divided into three components of hand
shape modulation: transport shaping, preshaping, and
contact shaping. Transport shaping reflects hand shape
modulation as the arm extends to reach for the object
(during the transport component). The other two phases
characterize the grasping component: preshaping is the
modulation of the hand as it begins to encompass the object
and contact shaping reflects the fine adjustments to the
grasp during the initial stages of the intended use of the
object. We hypothesized that post-stroke patients would
have deficits in the ability to conform the palm to the object
which would lead to problems in grasping and object
manipulation. In addition, a deficit in the ability to ensure a
stable grasping interface may lead to the adoption of
movement compensations at the hand or arm level,
Exp Brain Res (2009) 199:59–70 61

ultimately interfering with motor recovery (Alaverdashvili
et al. 2008). Therefore, in this study, we characterized hand
shape modulation during object manipulation involving the
whole hand in individuals with post-stroke hemiparesis. In
addition, hand opening and closing was divided into three
phases in order to examine which phase is most affected
and the associated impact on the biomechanics of palmar
concavity formation which cannot be derived from discrete
parameters such as maximal grip aperture.
Materials and methods
Subjects
Study participants included ten stroke survivors (4 women,
mean age 65 ± 9 years, range 51–79) with hemiparesis,
see Table 1 and eight healthy adult volunteers (4 women,
mean age 55 ± 10 years, range 41–68). The control group
participants were right-handed and had no history of injury
to the hand or arm. Stroke survivors had a diagnosis of
unilateral stroke and were at least 4 months post-stroke at
the time of the study. Those with some functional arm and
hand movement [Chedoke-McMaster Hand Assessment
Scale (Gowland et al. 1993) C 3] were included. All par-
ticipants were informed of the aim of the study and signed
consent forms approved by the local Ethics Committee.
Tasks were performed by the affected side in the stroke
survivors and the dominant side in the control group.
Experimental procedure
Hand shape modulation was examined during two natural
power grasps—spherical (ball: 10.5 cm diameter) and
cylindrical (plastic glass: 8.5 cm diameter, 14 cm high).
Participants performed 10 trials in separate blocks for each
grasp type. Object transport and manipulation components
were included in order to replicate a functional task so as to
encourage a more natural reaching and grasping movement
pattern (Trombly 1992; Ada et al. 1994; Chen et al. 2008).
Subjects were seated in a comfortable chair with a back
support. In the initial position, the hand was placed on an
adjustable platform with the arm positioned at approxi-
mately 45° shoulder abduction, 90° elbow flexion, and the
forearm in 90° pronation (Fig. 1). Objects were placed at a
distance of 90% of the subject’s arm length in the sagittal
plane in front of the ipsilateral shoulder on an adjustable
table, approximately 5 cm above the initial hand position
(Mark et al. 1997; Levin et al. 2002). Arm length was
measured as the distance between the medial border of the
axilla and the wrist crease. The participants grasped the
object and transported it medially to a surface, approxi-
mately 10 cm higher than the initial support surface, also
located at 90% of the arm’s length just beyond the midline
before returning the hand to the initial position. The initial
and final hand and object positions were standardized
across subjects and trials for both grasping tasks. Move-
ment was initiated in response to a verbal cue and subjects
were asked to perform the activity at a comfortable speed.
Data recording and analysis
A six-camera ViconÒ motion analysis system was used to
record three-dimensional movement of the arm and hand.
Reflective markers (5-mm) were placed on the radial and
ulnar styloids of the wrist, with two additional markers
placed on the forearm. Hemispherical reflective markers
(3-mm) were attached to the dorsum of the hand and the
finger joints. The marker configuration is illustrated in
Fig. 2a. The two markers at the base of the 2nd and 5th
metacarpal bones (MCP) constituted the carpal markers
and the respective landmark for the palmar arch formula-
tion was calculated by bisecting the line segment con-
necting these two markers. Data were recorded at 120-Hz
for 8 s and filtered using a second order Butterworth filter
with a cut-off frequency of 10 Hz.

Table 1 Clinical and

Subject Sex Age Side of Time since Chedoke-McMaster Purdue test Dominant
demographic data of the stroke
(yrs) hemiparesis injury (yrs) score arm/hand LA/A side side
survivors
S1 F 55 Left 9 4/3 10/3 R
S2 F 60 Left 2 5/4 9/5 R
S3 F 67 Left 3.2 7/7 13/10 R
S4 M 68 Left 1.9 5/6 13/2 L
S5 M 77 Right 6 5/6 9/4 L
S6 F 79 Left 4.2 4/5 11/3 R
The maximal score for both the S7 M 59 Right 2.1 7/6 13/9 R
arm and hand on the Chedoke- S8 M 67 Right 3.3 5/5 11/6 L
McMaster scale is 7 S9 M 70 Right 3.1 4/3 10/1 R
LA less affected, A affected, S10 M 51 Left 1.4 7/3 10/0 R
R Right, L Left

123
62 Exp Brain Res (2009) 199:59–70

included the initial portion of the object manipulation

Adjustable
height
Adjustable
height
Adjustable
height
Fig. 1 Schematic illustration of the experimental set-up and the two
grasp types (spherical and cylindrical)
The prehensile movement was divided into three phases
(P1: transport shaping, P2: preshaping, and P3: contact
shaping) using the tangential velocity trace of the ulnar
marker of the wrist (Fig. 2b). Phase 1 (P1) involved hand
movement from onset to the time of maximal wrist
velocity; phase 2 (P2) extended from the end of P1 to the
time when object transport began, i.e., the first minimum of
the tangential velocity trace; and phase 3 (P3) extended
from the end of P2 to the time of maximal velocity of
object transport, i.e., second peak in the wrist tangential
velocity trace, also considered to be the end of task (END).
While ‘preshaping’ in phase 2 reflects initial hand–object
contact, movements in phase 3 reflect the fine adjustments
in grip posture that occur during object stabilization at the
time of object transport, hence the term ‘contact shaping’.
Current methods generally define the end of grasp at the
completion of arm transport (at the end of phase ‘2’). We
phase since fine adjustments occur to the grip as object
manipulation begins, i.e., phase 3 as identified in our
method. The same has been shown in our earlier study
(Sangole and Levin 2008a, b). The tangential velocity was
computed by numerical differentiation of the x, y, and z
coordinates. Movement onset was defined as the time at
which the tangential velocity of the ulnar marker exceeded
and remained above 5% of the peak wrist velocity. This
time was referred to as zero-ms for subsequent data anal-
ysis. The end of the movement (end of task) was defined as
the instance when maximal tangential velocity was attained
during object transport. This definition of task phase
includes fine adjustments in hand shape modulation up
until the initial stages of object transport (Sangole and
Levin 2008a, b). Two parameters were measured at the
beginning of each phase and at the end of the task: palmar
arch position and time.
We defined the palmar arch as being composed of two
components (thenar and hypothenar) articulating with a
rigid middle plane formed by the carpus-index MCP-mid-
dle MCP. This formulation was considered reasonable
because there is little or no relative movement between the
respective carpo-metacarpal joints of the index and middle
fingers (Levangie and Norkin 2001). The thenar and
hypothenar components are henceforth referred to as the
thenar arch and the hypothenar arch, respectively. The
thenar arch was defined using the carpus, thumb MCP, and
index finger MCP, while, the hypothenar arch was defined

Fig. 2 a Marker configuration (A)

on the hand and wrist (dorsal
view), and b Phases of the
reach-to-grasp movement (P1:
transport shaping, P2:
preshaping, P3: contact shaping;
separated by vertical dotted
lines) identified in the tangential
velocity trace of the wrist ulnar
marker for a control subject (left
panel) and a representative trial
of a stroke patient (right panel)

(B) P1 P2 P3 P1 P2 P3
500
Tangential velocity (mm/s)

400

400
300
300
200
200

100
100

0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.00 0.50 1.0 1.50 2.00 2.50
Time (s) Time (s)

123
Exp Brain Res (2009) 199:59–70
Fig. 3 Biomechanical (A)
formulation of the palmar arch:
a dorsal view showing the
markers used for the arch
definition and the respective
planes and b transverse view at
the metacarpal joint (MCP) middle
level plane
thenar
plane
using the carpus, middle finger MCP, and little finger MCP.
Thus, the biomechanical formulation of the palmar con-
cavity combined the articulations of the thenar and hypo-
thenar arches (Fig. 3a, b). A smaller angle reflects more
concavity. This characterization helped to identify thenar
and hypothenar contributions to the palmar arch formation
and is therefore more descriptive of hand function.
Movements of the thumb are reflected in the thenar arch
and those of the 5th metacarpal are captured in the hypo-
thenar arch. Additional details are provided in Sangole and
Levin (2008a, b).
In order to examine phase-wise differences in grasping
patterns between the two groups, the change in palmar arch
angle and in thenar–hypothenar contributions during each
phase was calculated by expressing the change in angle in
each phase (P1, P2, P3) as a percentage of the total change
from the beginning of P1 to END (Sangole and Levin
2008b). The total task time and time taken to complete
each phase were also compared. For each subject, these
values were averaged across trials. Coordination of the
index and middle fingers was investigated by examining
the linearity between the angular displacements at the
metacarpal (MCP) and proximal phalangeal (PIP) joints of
each finger. Those trials in which the trajectories could not
be successfully reconstructed due to missing markers were
excluded (approximately 20%).
In a power grasp, the thumb, index, and middle fingers
play an essential role (Napier 1956). Since our biome-
chanical formulation of the palmar arch takes into account
the MP joint of the thumb, its modulation would be
reflected in that of the thenar arch, while the same is not
true for the index and middle fingers. Therefore, in the
analysis, the relative motions of the MCP and PIP joints of
the index and middle fingers were examined for linearity.
From related studies (Kamper et al. 2003), it was hypoth-
esized that there is a linear relationship between the
movements of the MCP and PIP joints of the index and
middle fingers.
The analysis thus included separate comparisons of (a)
palmar arch modulation (arch angle at the beginning of
63
(B)
Ring Middle
MCP MCP Index
Little MCP
hypothenar MCP
Thumb
plane
Hypothenar
MCP
arch
Thenar
arch
Palmar arch = thenar arch +
hypothenar arch
each phase and phase-wise change in arch angle), thenar–
hypothenar contributions, (b) total task time and the time
required to complete each phase, (c) and MCP-PIP linearity
for each grasp type. A mixed design ANOVA with factors
group (stroke, control) and phase (P1, P2, P3, End position)
was used. Comparisons between grasp types were done
only for the total task times and the three phase completion
times. Post-hoc comparisons were adjusted using Bonfer-
roni’s correction. Correlations between palmar arch mea-
sures and clinical scores of hand motor impairment and
function [Chedoke-McMaster and Purdue scores (Backman
et al. 1992)] were computed with Pearson statistics. An
initial significance level of P \ 0.05 was used for all tests.
Results
Phase-wise changes in palmar arch modulation
and thenar–hypothenar contribution
For each task, palmar arch measures at the beginning of
each phase (P1, P2, P3) and at the end of the task were
compared. The values at the initial position (P1) and at the
end of the task were similar between groups (Table 2) but
the hand was flatter (greater arch angle) in the stroke group
prior to object contact (beginning of P3) for both the
spherical (F2,18 = 4.097, P \ 0.05; P3: P \ 0.001) and the
cylindrical grasp (F2,18 = 18.7, P \ 0.001; P3:
P \ 0.001).
Figure 4a illustrates phase-wise arch modulation as a
percentage of the total, e.g., P1 as a percentage of
P1 ? P2 ? P3. Arch modulation, throughout the three
phases of grasp for each grasp type differed between
groups (spherical: F1,2 = 12.025, P \ 0.001; cylindrical:
F1,2 = 9.054, P \ 0.001). The stars indicate the phases in
which significant differences were observed between
groups. Most of the hand modulation occurred in P1
(P \ 0.002, spherical: control 45% ± 15, stroke
40% ± 15; cylindrical: control 63% ± 14, stroke
43% ± 14). In the cylindrical grasp, the stroke subjects
123
64 Exp Brain Res (2009) 199:59–70

Table 2 Measures of palmar arch angle in initial and final positions (palmar arch = thenar ? hypothenar) at the beginning of each phase (P1,
P2, P3) and at the end of the task (End) in the reach-to-grasp task in control (C) subjects and in subjects with stroke (S)
Subject Spherical Cylindrical
P1 (deg) P2 (deg) P3 (deg) End (deg) P1 (deg) P2 (deg) P3 (deg) End (deg)

C1 123 ± 8 118 ± 4 109 ± 2 108 ± 4 130 ± 7 115 ± 3 103 ± 6 104 ± 2

C2 128 ± 8 100 ± 7 103 ± 3 90 ± 7 120 ± 5 102 ± 2 105 ± 1 100 ± 1
C3 124 ± 5 107 ± 9 97 ± 3 92 ± 7 120 ± 6 108 ± 2 106 ± 2 105 ± 3
C4 120 ± 7 114 ± 1 110 ± 1 107 ± 6 119 ± 5 108 ± 1 110 ± 2 105 ± 2
C5 138 ± 10 119 ± 2 108 ± 2 107 ± 2 130 ± 11 111 ± 2 105 ± 1 105 ± 1
C6 108 ± 5 102 ± 4 99 ± 1 93 ± 10 106 ± 3 95 ± 9 97 ± 8 99 ± 2
C7 124 ± 3 120 ± 3 105 ± 2 104 ± 6 124 ± 4 114 ± 2 104 ± 2 107 ± 14
C8 123 ± 2 112 ± 2 116 ± 4 114 ± 4 126 ± 2 109 ± 1 110 ± 2 111 ± 3
Avg ± SD 124 ± 8 112 ± 8 106 ± 6* 102 ± 8 122 ± 8 108 ± 6 105 ± 4* 105 ± 4
S1 125 ± 3 123 ± 5 114 ± 4 118 ± 5 121 ± 4 116 ± 3 109 ± 3 108 ± 4
S2 131 ± 13 123 ± 6 120 ± 7 120 ± 4 137 ± 1 120 ± 17 121 ± 6 112 ± 21
S3 138 ± 5 108 ± 15 113 ± 6 97 ± 11 111 ± 34 94 ± 31 97 ± 12 86 ± 16
S4 139 ± 8 122 ± 3 112 ± 4 106 ± 2 141 ± 6 127 ± 12 118 ± 13 112 ± 4
S5 134 ± 8 127 ± 2 114 ± 1 112 ± 4 129 ± 5 114 ± 5 107 ± 4 98 ± 3
S6 126 ± 3 109 ± 24 100 ± 16 98 ± 19 119 ± 5 117 ± 5 107 ± 7 100 ± 16
S7 122 ± 3 111 ± 1 102 ± 2 100 ± 5 121 ± 1 108 ± 5 109 ± 14 102 ± 2
S8 120 ± 3 117 ± 1 109 ± 4 102 ± 2 120 ± 9 117 ± 2 115 ± 4 101 ± 11
S9 138 ± 9 139 ± 5 128 ± 6 124 ± 4 136 ± 8 130 ± 8 119 ± 11 101 ± 1
S10 135 ± 4 88 ± 14 103 ± 10 89 ± 3 133 ± 2 n/a n/a 98.6
Avg ± SD 131 ± 7 117 ± 14 112 ± 9 107 ± 10 127 ± 10 116 ± 10* 111 ± 7* 102 ± 8
* Difference between groups (P \ 0.001)

modulated hand shape more than the control subjects in P2
(P \ 0.01, control 23% ± 14, stroke 32% ± 12) and P3
(P \ 0.001, control 14% ± 9, stroke 25% ± 11). The
pattern of relative phase-wise change differed in each
group (Fig. 4b, c bottom panels). In the stroke group,
palmar arch shape changed less in P1 and more in P2 and
P3 compared to the control group for both grasp types
(P \ 0.01). Figure 4b and c (top panels) shows the amount
of change of arch angle in degrees. Since the hand in the
stroke subjects was several degrees flatter in the initial
position, hand shape in each phase in this group was 5°–7°
more than the control group and was more evident in P2
and P3, particularly for the cylindrical grasp.
Figure 5a illustrates the proportion of phase-wise thenar
and hypothenar arch modulation relative to the total arch
modulation. The stars indicate areas of significant differ-
ences between groups. When considering thenar and
hypothenar angles, smaller angles reflect greater arch
concavity and therefore increased contribution to the arch
formation. Thus, the shorter bars in the histogram shown in
Fig. 5, indicate that the thenar part of the hand contributed
more to arch formation than the hypothenar part for both
grasps in both groups (approximately 15°; spherical:
F2,18 = 4.711, P \ 0.014; cylindrical: F2,18 = 5.153,
P \ 0.01; Fig. 5a). Phase-wise thenar and hypothenar
contributions differed for each grasp type similarly in both
groups. For the spherical grasp, there was more thenar
contribution to the total arch modulation (palmar concav-
ity) during P2 (control: thenar 49° ± 9, hypothenar
63° ± 5; stroke: thenar 48° ± 13, hypothenar 68° ± 5;
P \ 0.01) and P3 (control: thenar 47° ± 7, hypothenar:
59° ± 5; stroke: thenar 47° ± 10, hypothenar 59° ± 5;
P \ 0.003) compared to P1 and more hypothenar contri-
bution during P3 (P \ 0.005), Fig. 5b. In the cylindrical
grasp there was more thenar contribution only during P3
(P \ 0.002), Fig. 5c, whereas the hypothenar contribution
was the same across all phases.
Total task time and phase completion times
Participants with stroke took approximately 60% longer to
complete the task for the spherical (control: 871 ± 280 ms,
stroke: 1,443 ± 455 ms) and cylindrical (control:
929 ± 264 ms, stroke: 1,953 ± 740 ms) objects, respec-
tively, compared to the control group. The relative time
spent in each phase varied. Overall, there were significant
differences between groups (F1,1,2 = 47.84, P \ 0.001),
between grasp types (F1,1,2 = 6.05, P \ 0.02), and a
phase-wise interaction for the cylindrical grasp
(F218 = 3.811, P = 0.006). The time spent in completing

123
Exp Brain Res (2009) 199:59–70 65

Fig. 4 a Phase-wise percent of Control Stroke

arch modulation; Phase-wise (A)100 p<0.002
interaction in arch modulation 90 p<0.002
(°; top panels) and percent of 80 p<0.01

Arch modulation (%)

arch modulation (bottom 70 p<0.001
panels) for b spherical grasp 60
type and c cylindrical grasp type 50 Contact shaping (P3)
40 Pre-shaping (P2)
30 Transport shaping (P1)
20
10
0
Spherical Cylindrical Spherical Cylindrical
grasp grasp grasp grasp

(B) Spherical grasp (C) Cylindrical grasp

Estimated marginal means (°)

16 16
Control
14 14
Patients
12 12

10 10

8 8

6 6

4 4
1 2 3 1 2 3
Estimated marginal means (%)

50
60

40 50

30 40

30
20
20

10 10
1 2 3 1 2 3
Phase Phase

each grasp phase in stroke subjects was longer compared to
the control subjects for all three phases, particularly for the
cylindrical grasp [spherical grasp: P1 (control:
302 ± 59 ms, stroke: 437 ± 138 ms, P \ 0.02) and P3
(control: 248 ± 88 ms, stroke: 520 ± 178 ms, P \ 0.004);
cylindrical grasp: P1 (control: 284 ± 45 ms, stroke:
482 ± 117 ms, P \ 0.001), P2 (control: 364 ± 184 ms,
stroke: 767 ± 345 ms, P \ 0.006), and P3 (control:
281 ± 73 ms, stroke: 704 ± 342 ms, P \ 0.002)], Fig. 6a.
When the two tasks were compared within groups, the
control group took longer to complete P3 (P \ 0.04) in the
cylindrical grasp as compared to the spherical grasp. The
stroke patients, however, took longer to complete the pre-
shaping (P2: control: 364 ± 184 ms, stroke: 767 ±
345 ms, P \ 0.01) and the contact shaping phases (P3:
control: 281 ± 73 ms, stroke: 704 ± 342 ms, P \ 0.053),
see Fig. 6b, particularly for the cylindrical grasp type.
Patients who had relatively high hand function scores (C5
on Chedoke-McMaster Scale and close to normal hand
function on the Purdue Pegboard test) required relatively
less time to complete phases 2 and 3 of the spherical grasp
as compared to their performance on the cylindrical grasp.
The total grasp time was negatively correlated with the
Chedoke-McMaster scale scores (spherical r = -0.59;
cylindrical r = -0.90) indicating that stroke survivors
with better hand function took less time to complete the
tasks. This was also true for each phase of the reaching
task, especially for the cylindrical object with correlations
ranging from -0.71 to -0.91 for Chedoke scores vs. times
spent in phases 1–3 and from -0.53 to -0.71 for Purdue
Pegboard Scores) versus each phase.
Proximal interphalangeal (PIP)
and metacarpo-phalangeal (MCP) joint linearity
To better characterize hand-related deficits due to stroke,
aside from palmar modulation, finger joint coordination
was also investigated by examining phase-wise linearity

123
66 Exp Brain Res (2009) 199:59–70

Fig. 5 a Phase-wise thenar and (A)

hypothenar contribution for Contact shaping (P3) Pre-shaping (P2) Transport shaping (P1)
each grasp type and respective
error bars. A smaller angle Control Stroke
200
indicates greater palmar
concavity and the stars indicate

Arch modulation (°)

phases in which the differences 150
between groups were significant
(P \ 0.01). Phase-wise
interaction in thenar (top panels) 100
and hypothenar (bottom panels)
contribution for b spherical and,
c cylindrical grasp type. In the 50
spherical grasp, there was more
thenar contribution to palmar
concavity formation during P2 0

thenar

hypothenar

thenar

hypothenar

thenar

hypothenar

thenar

hypothenar
and P3 and, more hypothenar
contribution during P3. In the
cylindrical grasp, there was
more thenar contribution only Spherical grasp Spherical grasp
Cylindrical grasp Cylindrical grasp
during P3, while hypothenar
contribution was the same (B) Spherical grasp (C) Cylindrical grasp
across all phases
Control
Estimated marginal means (°)

72 72
Patients
68 68
64 64
60 60
56 56
52 52
48 48
44 44
Thenar Thenar
40 40
1 2 3 1 2 3
Estimated marginal means (°)

72 72
68 68
64 64
60 60
56 56
52 52
48 48
44 44
40 Hypothenar 40 Hypothenar

1 2 3 1 2 3
Phase Phase

between MCP and PIP joints in the index and middle fin- The same pattern was not observed for the cylindrical grasp
gers. Although there were no overall phase-wise differ- (Fig. 7b). While there was still a tendency for the most
ences between groups in the correlation between the linearity to occur in P1 and the least linearly to occur in P2
linearity of PIP and MCP movement, there were distinct in the control group, there was only a significant difference
PIP–MCP interaction effects (F2,14 = 4.763, P \ 0.05) between groups in P2 for the middle finger.
between the two groups for each grasp type (Fig. 7a, b).
For the spherical grasp (Fig. 7a), the pattern of index and
middle PIP–MCP movement was more linear in P1 in the Discussion
control subjects compared to the patient group. In addition,
the linearity was greater in the patient group at P2 for the Most studies of reach-to-grasp movements have high-
index and in the control group at P3 for the middle finger. lighted differences in reaching and grasping movement

123
Exp Brain Res (2009) 199:59–70 67

2500 descriptions, the relationship between these deficits and the

(A)
Cylindrical grasp decreased use of the post-stroke upper limb, even in well-
2000
recovered patients remains elusive (Lai et al. 2002; Mayo
Spherical grasp
1500 p<0.001
et al. 2002). This relationship may be more completely
Time (ms)

p<0.004 understood by including the description of the biome-

1000 p<0.006 chanical limitations in the ability of the hand to conform to
the object to be grasped. To better understand how post-
500 p<0.02 p<0.002
stroke hemiparesis impacts hand movement, we formulated
0 a measure to quantify palmar arch modulation and identi-
Control Stroke Control Stroke
2500
fied three phases characterizing prehensile movement. Our
(B) earlier studies validated these methods in healthy individ-
Stroke
2000 uals with no history of hand dysfunction (Sangole and
Control p<0.053 Levin 2007, 2008a, b). In this article, we extended the same
Time (ms)

1500
methods to characterize modulations in hand movement
1000 p<0.04
p<0.01
during functional reaching tasks in individuals following
stroke.
500 Our findings showed that both groups exhibited com-
parable performance in phase-wise absolute palmar arch
0
Spherical Cylindrical Spherical Cylindrical modulation for the spherical grasp. However, for cylin-
grasp grasp grasp grasp
drical grasping there were differences during the preshap-
Contact shaping (P3) Pre-shaping (P2) Transport shaping (P1)
ing (phase 2: establishing object contact) and contact
Fig. 6 The time (ms) to complete each phase of the reaching shaping phases (phase 3: making fine adjustments to the
movement and the total reaching time a for the same grasp type grip at object manipulation). These differences may be
between groups; b for the two grasp types within each group attributed to the fact that in a spherical grasp, there are
support reaction elements that stabilize the object as the
patterns between stroke survivors and healthy subjects with hand establishes contact with it. For example, the direction
a main focus on hand aperture and grip forces (for a recent of the hand movement while establishing contact with the
review, see Nowak 2008). Yet, despite detailed spherical object is normal to the supporting surface
Fig. 7 Group interaction in (A) (B)
phase-wise PIP–MCP linearity a
spherical grasp (left panels) and 1.0 1.0 Control
Estimated marginal means

b cylindrical grasp (right Patients

0.9 0.9
panels). The y-axis is the linear
regression coefficient value 0.8 0.8

0.7 0.7

0.6 0.6
p<0.05
0.5 0.5
E

0.4 Indexfinger 0.4 Indexfinger

1 2 3 1 2 3

1.0 1.0
Estimated marginal means

0.9 0.9
08
0.8 0.8
0.7 0.7
0.6 0.6
0.5 p<0.05
0.5
0.4 p<0.05 0.4
0.3 0 3 Middle finger
0.3
Middle finger

1 2 3 1 2 3
Phase Phase

123
68
whereas this is not so for the cylindrical object. While
grasping a spherical object does not necessitate object
stabilization, a cylindrical object may tip over if the hand is
not aligned appropriately. Tipping of the object is often
avoided by using the other hand for stabilization. In the
current experimental protocol, however, subjects could not
avail themselves of assistance from the unaffected side.
Furthermore, cylindrical grasping requires a re-orientation
of the hand at the wrist joint. These aspects impact the
hand–object dynamics, particularly when object interaction
occurs which requires grip adjustment for overall prehen-
sile stability during object manipulation (phases 2 and 3).
The results thus suggest that although both grip types are
power grasps, grasping a cylindrical object requires addi-
tional hand re-orientation at the time of establishing object
contact and adjustment of the grip for overall prehensile
stability during object manipulation. This explains why
upper limb assistive technology efforts in the design of
doorknobs, utensils etc. minimize the need to re-orient the
hand. Assistive technology is essentially designed to
maximize the use of gravity or body weight in order to
manipulate objects. For example, using a lever-shaped
doorknob (e.g., LeveronÒ handle door knob adaptors), door
opening can be done by leaning on the knob with the arm
or hand instead of supinating the hand which is required to
turn a standard spherical doorknob.
In the healthy subjects, most of hand shape modulation
was accomplished during transport shaping (P1), with fine
adjustments occurring during the P2 and P3 phases. Thus,
hand shape modulation was accomplished early in the
reach-to-grasp task. In contrast, the stroke patients started
with a much flatter hand (greater palmar arch angle) and
thus required more arch modulation during the three pha-
ses. They exhibited relatively more arch modulation during
P2 and P3 as compared to the healthy subjects. This not
only reflects the temporal concurrency between the three
phases (Sangole and Levin 2008b) but also suggests a
spatial concurrency. Phase-wise times differed in the stroke
subjects and were related to hand motor impairment and
function scores (Chedoke-McMaster Hand Scale, Perdue
Pegboard). The only moderate correlation between the
times and the Chedoke-McMaster scores may be attributed
to a ceiling effect of the scale for fine hand control.
From a motor control standpoint, our findings show that
modulation of hand shaping at the level of the palm occurs
at the same time as changes in the finger aperture during
reach-to-grasp tasks. This suggests that the nervous system
controls these actions together in an anticipatory way prior
to object contact which is likely an important feature of
successful grasping. Hand functional retraining in patients
with post-stroke hand paresis may be more successful if
focused on the coordination of these two elements of hand
preshaping.
123
Exp Brain Res (2009) 199:59–70
The phase-wise PIP–MCP linearity pattern revealed an
interaction between the groups (see Fig. 6). The interaction
effect in the spherical grasp may be attributed to PIP–MCP
linearity during transport shaping and preshaping (phases 1
and 2). The PIP–MCP linear relationship explains the rel-
ative coordination between the two finger joints. If they are
synchronous with each other, then both PIP and MCP joint
angles during a grasping phase are either extending or
flexing together. If they are not synchronous then one fin-
ger is extending while the other is flexing or vice versa.
The interaction effect in Fig. 6a suggests that the stroke
survivors exhibit different, perhaps opposite, characteris-
tics of PIP–MCP phase-wise linearity during hand opening
and closing compared to the healthy individuals thus
reflecting an altered finger-joint coordination pattern. This
would impact their ability to use the hand for fine manip-
ulative tasks and provides a more detailed level of
description of the hand impairment. Further studies of more
complex grasping tasks primarily focusing on finger joint
angular excursions are required to examine the PIP–MCP
coordination for other grasping tasks.
Conclusions
Previous studies of reaching and grasping in subjects with
post-stroke hemiparesis have focused on arm movement to
explain hand-related deficits by characterizing hand ori-
entation relative to the arm and grip aperture scaling. What
is not clear from studies of arm motor compensations
during reach to grasp tasks is whether the compensations
are related to deficits in arm trajectory formation per se or
to deficits in the ability of the hand and fingers to conform
to the object and develop a stable grasp. This study is the
first to demonstrate that deficits in palmar shape modula-
tion occur concurrently with deficits in coordinated finger
movement. This new information about palmar shape
modulation may help to explain why even patients who are
considered to be well-recovered on standard clinical testing
still have deficits in performing activities of daily living
(Mayo et al. 2002).
Our findings provide additional insights into how stroke
impacts hand configuration during the performance of
functional hand tasks. Palmar arch modulation was most
affected in latter segments of the prehensile movement, i.e.,
phases 2 and 3, which involve hand–object contact and fine
grip adjustment. This indicates that discrete parameters
such as maximal grip aperture cannot characterize hand-
related deficits because maximal grip aperture is already
attained by the time of phase 2. Although both of the tasks
studied (spherical and cylindrical grasps) are power grips,
our findings showed that stroke survivors with mild hand
impairment had more difficulty with the cylindrical grasp
Exp Brain Res (2009) 199:59–70
possibly because the stability of the grip is threatened
during hand–object interaction. If the same task were to be
executed using a bimanual strategy, wherein the other hand
stabilizes the object during contact, we presume that the
performance would be comparable to the hand–object
dynamics during a spherical grasp.
Interpretation of the findings is limited by the small
sample size and the fact that a majority of the patients had
relatively moderate hand function capabilities. Neverthe-
less, the findings reveal distinct differences of how hand
configuration is affected by stroke. Our findings suggest
that combined with other measures of hand function such
as grip forces during object manipulation (e.g., Nowak
2008), hand shape modulation may form the basis of a
standardized assessment of deficient manual dexterity for
better diagnosis and rehabilitation of the hand post-stroke.
In particular, rehabilitation interventions for patients with
grasping deficits stressing anticipatory hand preshaping in
addition to wrist and finger joint movement may result in
better outcomes.
Acknowledgments APS was supported by a Fonds de la Recherche
en Santé du Quebec (FRSQ) fellowship. The research was conducted
at the Jewish Rehabilitation Hospital, during her postdoctoral term at
McGill University. The authors express sincere gratitude to the par-
ticipants, to the students—Liza Azeff, Mi Yun, Olivier Vourantonis,
and Luis Alberto Knaut who assisted on various aspects of data
recording as well as to Dr. Valeri Goussev for his assistance with data
analysis. MFL holds a Canada Research Chair in Motor Recovery and
Rehabilitation.
References
Ada L, Canning CG, Carr JH, Kilbreath SL, Shepherd RB (1994)
Task-specific training of reaching and manipulation. In: Stel-
mach GE, Vroon PA (eds) Insights into the reach to grasp
movement, vol 105. North-Holland, The Netherlands; Elsevier
Science B.V., pp 239–265
Alaverdashvili M, Foroud A, Lim DH, Whishaw IQ (2008) ‘‘Learned
baduse’’ limits recovery of skilled reaching for food after
forelimb motor cortex stroke in rates: a new analysis of the effect
of gestures on success. Behav Brain Res 188:281–290
Ansuini C, Santello M, Massaccesi S, Castiello U (2006) Effects of
end-goal on hand shaping. J Neurophysiol 95:2456–2465
Backman C, Cork S, Gibson D, Parsons J (1992) Assessment of hand
function: the relation between pegboard dexterity and applied
dexterity. Can J Occup Therapy 59:208–213
Brunnstrom S (1970) Movement therapy in hemiplegia: a neuro-
physiological approach. Harper & Row Publishers, New York
Castiello U (2005) The neuroscience of grasping. Nat Rev: Neurosci
6:726–736
Chen HC, Lin KC, Chen CL, Wu CY (2008) The beneficial effects of
a functional task target on reaching and postural balance in
patients with right cerebral vascular accidents. Motor Control
12(2):122–135
Cirstea MC, Levin MF (2000) Compensatory strategies for reaching
in stroke. Brain 123(5):940–953
Desmurget M, Prablanc C (1997) Postural control of three dimen-
sional prehension movements. J Neurophysiol 77:452–464
69
Desmurget M, Prablanc C, Rossetti Y, Arzi M, Paulignan Y, Urquizar
C, Mignot JC (1995) Postural and synergic control for three-
dimensional movements of reaching and grasping. J Neurophys-
iol 74:905–910
Desmurget M, Prablanc C, Arzi M, Rossetti Y, Paulignan Y,
Urquizar C (1996) Integrated control of hand transport and
orientation during prehension movements. Exp Brain Res
110:265–278
Desmurget M, Grea H, Prablanc C (1998) Final posture of the upper
limb depends on the initial position of the hand during
prehension movements. Exp Brain Res 119(4):511–516
Feldman AG, Goussev V, Sangole AP, Levin MF (2007) Threshold
position control and the principle of minimal interaction in motor
actions. Exp Brain Res 165:267–281
Fries W, Danek A, Scheidtmann K, Hamburger C (1993) Motor
recovery following capsular stroke. Brain 116:369–382
Gentilucci M, Caselli L, Secchi C (2003) Finger control in the tripod
grasp. Exp Brain Res 149:351–360
Gowland C, Stratford P, Ward M, Moreland J, Torresin W, Van
Hullenaar S et al (1993) Measuring physical impairment and
disability with the Chedoke-McMaster Stroke Assessment.
Stroke 24(1):58–63
Jeannerod M (1984) The timing of natural prehension movements.
J Mot Behav 16:235–254
Jeannerod M (1999) Visuomotor channels: their integration in goal-
directed prehension. Hum Mov Science 18:201–218
Kamper DG, Cruz EG, Siegel MP (2003) Stereotypical fingertip
trajectories during grasp. J Neurophysiol 90(6):3702–3710
Kapandji IA (1982) Physiology of the joints—upper extremities, vol
1. Churchill Livingstone, New York
Lai SM, Studenski S, Duncan PW, Perera S (2002) Persisting
consequences of stroke measured by the Stroke Impact Scale.
Stroke 33:1840–1844
Landsmeer JMF (1962) Power grip and precision handling. Ann
Rheum Dis 21:164–170
Levangie PK, Norkin CC (2001) The wrist and hand complex. In:
Davis FA Joint structure and function, 3rd edn. Philadelphia,
pp 251–289
Levin MF, Michaelsen S, Cirstea C, Roby-Brami A (2002) Use of the
trunk for reaching targets placed within and beyond the reach in
adult hemiparesis. Exp Brain Res 143:171–180
Lewis OJ (1977) Joints remodeling and the evolution of the human
hand. J Anat 123:157–201
Ma S, Feldman AG (1995) Two functionally different synergies
during arm reaching movements involving the trunk. J Neuro-
physiol 73(5):2120–2122
Mamassian P (1997) Prehension of objects oriented in three-
dimensional space. Exp Brain Res 114(2):235–245
Mark LS, Nemeth K, Gardner D, Dainoff MJ, Paasche J, Duffy M,
Grandt K (1997) Postural dynamics and the preferred critical
boundary for visually guided reaching. J Exp Psych: Hum
Percept Perform 23:1365–1379
Mason CR, Gomez JE, Ebner TJ (2001) Hand synergies during reach-
to-grasp. J Neurophysiol 86(6):2896–2910
Mayo NE, Wood-Dauphinee S, Cote R, Durcan L, Carlton J (2002)
Activity, participation, and quality of life 6 months poststroke.
Arch Phys Med Rehabil 83:1035–1042
Michaelsen SM, Jacobs S, Roby-Brami A, Levin MF (2004)
Compensation for distal impairments of grasping in adults with
hemiparesis. Exp Brain Res 157(2):162–173
Michaelsen SM, Magdalon EC, Levin MF (2009) Grip aperture
scaling to object size in chronic stroke. Motor Control
13(2):197–217
Muellbacher W, Richards C, Ziemann U, Wittenberg G, Weltz D,
Boroojerdi B, Cohen L, Hallett M (2002) Improving hand
function in chronic stroke. Arch Neurol 59:1278–1282
123
70
Napier JR (1956) The prehensile movements of the human hand.
J Bone Joint Surg 38B(4):902–913
Nowak DA (2008) The impact of stroke on the performance of
grasping: usefulness of kinetic and kinematic motion analysis.
Neurosci Biobehav Rev 32(8):1439–1450
Paulignan Y, MacKenzie C, Marteniuk R, Jeannerod M (1990) The
coupling of arm and finger movements during prehension. Exp
Brain Res 79:431–435
Pilon JF, DeSerres SJ, Feldman AG (2007) Threshold position control
of arm movement with anticipatory increase in grip force. Exp
Brain Res 181(1):49–67
Rand MK, Stelmach GE (2005) Effect of orienting the finger
opposition space in the control of reach-to-grasp. J Mot Behav
37(1):65–78
Roby-Brami A, Feydy A, Combeaud M, Biryukova EV, Bussel B,
Levin MF (2003) Motor compensation and recovery for reaching
in stroke patients. Acta Neurol Scand 107(5):369–381
Rosenbaum DA, Loukopoulos LD, Meulenbroek RGJ, Vaughan F,
Engelbrecht SE (1995) Planning reaches by evaluating stored
postures. Psychol Rev 102:28–67
Rossi E, Mitniski A, Feldman AG (2002) Sequential control signals
determine arm and trunk contributions to hand transport during
reaching in humans. J Physiol 538:659–671
Saling M, Mescheriakov S, Molokanova E, Stelmach GE, Berger M
(1996) Grip reorganization during wrist transport: the influence
of an altered aperture. Exp Brain Res 108:493–500
Sangole AP, Levin MF (2007) A new perspective in the understand-
ing of hand dysfunction following neurological injury. Top
Stroke Rehab 14(3):80–94
123
Exp Brain Res (2009) 199:59–70
Sangole AP, Levin MF (2008a) Arches of the hand in reach to grasp.
J Biomech. doi:10.1016/j.jbiomech.2007.11.006
Sangole AP, Levin MF (2008b) Palmar arch dynamics during reach-
to-grasp tasks. Exp Brain Res 190:443–452
Santello M, Soechting JF (1997) Matching object size by controlling
finger span and hand shape. Somatosens Motor Res 14(3):203–
212
Santello M, Soechting JF (1998) Gradual molding of the hand to
object contours. J Neurophysiol 79(3):1307–1320
Santello M, Flanders M, Soechting JF (1998) Postural hand synergies
for tool use. J Neurosci 18(23):10105–10115
Santello M, Flanders M, Soechting JF (2002) Patterns of hand
motion during grasping and the influence of sensory guidance.
J Neurosci 22(4):1426–1435
Smeets JB, Brenner E (1999) A new view on grasping. Motor Control
3(3):237–271
Soechting JF, Flanders M (1993) Parallel, interdependent channels for
location and orientation in sensorimotor transformations for
reaching and grasping. J Neurophysiol 70(3):1137–1150
Stelmach GE, Castiello U, Jeannerod M (1994) Orienting the finger
opposition space during prehension movements. J Mot Behav
26(2):178–186
Trombly CA (1992) Deficits of reaching in subjects with left
hemiparesis: a pilot study. Am J Occup Therapy 46(10):887–897
van Bergen E, Swieten LM, Williams JHG, Mon-Williams M (2007)
The effect of orientation on prehension movement time. Exp
Brain Res 178:180–193