The phenomenon of exaggerated motor overflow is well documented in
stroke survivors with spasticity. However, the mechanism underlying the
abnormal motor overflow remains unclear. In this study, we aimed to
investigate the possible mechanisms behind abnormal motor overflow and
its possible relations with post-stroke spasticity. 11 stroke patients
(63.6 ± 6.4 yrs; 4 women) and 11 healthy subjects (31.18 ± 6.18 yrs; 2
women) were recruited. All of them were asked to perform unilateral
isometric elbow flexion at submaximal levels (10, 30, and 60% of maximum
voluntary contraction). Electromyogram (EMG) was measured from the
contracting biceps (iBiceps) muscle and resting contralateral biceps
(cBiceps), ipsilateral flexor digitorum superficialis (iFDS), and
contralateral FDS (cFDS) muscles. Motor overflow was quantified as the
normalized EMG of the resting muscles. The severity of motor impairment
was quantified through reflex torque (spasticity) and weakness. EMG-EMG
coherence was calculated between the contracting muscle and each of the
resting muscles. During elbow flexion on the impaired side, stroke
subjects exhibited significant higher motor overflow to the iFDS muscle
compared with healthy subjects (ipsilateral or intralimb motor
overflow). Stroke subjects exhibited significantly higher motor overflow
to the contralateral spastic muscles (cBiceps and cFDS) during elbow
flexion on the non-impaired side (contralateral or interlimb motor
overflow), compared with healthy subjects. Moreover, there was
significantly high EMG-EMG coherence in the alpha band (6–12 Hz) between
the contracting muscle and all other resting muscles during elbow
flexion on the non-impaired side. Our results of diffuse ipsilateral and
contralateral motor overflow with EMG-EMG coherence in the alpha band
suggest subcortical origins of motor overflow. Furthermore, correlation
between contralateral motor overflow to contralateral spastic elbow and
finger flexors and their spasticity was consistently at moderate to high
levels. A high correlation suggests that diffuse motor overflow to the
impaired side and spasticity likely share a common pathophysiological
process. Possible mechanisms are discussed.
Introduction
When a stroke survivor with spastic hemiplegia is asked
to squeeze the hand or flex the elbow joint on the non-impaired side as
shown in Figure 1, there is involuntary activation of spastic finger and elbow flexors on the impaired side (Figures 1A, B). This phenomenon of involuntary activation of spastic muscles can occur in about 30% of hemiplegic stroke (1). It is often referred as motor overflow or associated reaction (1–8).
Other terms, such as mirror movement, global synkinesis, are sometimes
used interchangeably for the same clinical observation (8).
Motor overflow is one form of the spastic muscle overactivity. Other
types of muscle overactivity are also seen clinically, such as spastic
dystonia, co-contraction (9, 10).
Motor overflow is commonly
observed in the contralateral homologous resting muscle(s). It can also
be seen from proximal muscles to distal muscles in a form of abnormal
synergy (11, 12), and between limbs on the impaired side through interlimb coupling (13). As demonstrated in Figures 1C,D,
motor overflow to the contralateral spastic finger and elbow flexors
occurs during voluntary finger extension on the non-impaired side. These
clinical presentations indicate that motor overflow to the spastic
muscles is non-selective, diffuse, and concomitantly with voluntary
activation of other muscles. In contrast, motor overflow seen in
neurologically intact adults is mainly in contralateral homologous
muscles in the context of extreme effort or fatigue [see review (14)].
Therefore, motor overflow in stroke survivors is likely mediated by
different mechanisms than in healthy adults. However, the underlying
mechanisms for motor overflow after stroke are poorly understood.
A number of methods have been used in the literature to
evaluate motor overflow after neurological impairments, including
surface EMG, goniometry, dynamometry, electrogoniometry, and clinician
rating form. Surface EMG is the most commonly used laboratory-based
method (8). In our recent studies (15, 16),
involuntary EMG activities of the contralateral resting muscles were
used to quantify the extent of motor overflow during unilateral
voluntary elbow flexion tasks. Using quantitative assessment, the level
of motor overflow is found to be graded by the effort of the
non-impaired muscles (3).
Furthermore, EMG-EMG coherence analysis between EMG signals from the
contracting muscle and the contralateral resting muscles could provide
potential sources of motor overflow. Coherence analysis is based on the
cross-correlation between two separate signals in the frequency domain.
Coherence values fall between 0 and 1. Commonly studied frequency bands
include 6–12 Hz (alpha band), 13–30 Hz (beta band), and 30–60 Hz (gamma
band). It is well accepted that both beta and gamma bands have cortical
origins (17–20). Coherence in the alpha band is believed to have subcortical influences, may be related to the reticulospinal drive (21).
For example, EMG signals were recorded from bilateral homologous
muscles, such as biceps muscles during motoric responses of acoustic
startle reflex and during similar voluntary movements in healthy
subjects. EMG-EMG coherence in the alpha band was significantly greater
during startle reflex responses than during voluntary movement,
suggestive of a reticulospinal origin of such coherence in the alpha
band (21).
Motor overflow is often seen and elicited in stroke
survivors with spasticity. Its relation with post-stroke spasticity
remains controversial. Motor overflow is found to be associated with
spasticity in some studies (2, 3, 6), but not in others (1, 4).
In all these studies, spasticity was assessed using clinical scales,
such as modified Ashworth scale or Tardieu scale. Quantitative
assessment is likely to provide better insights into this relationship.
Based on the velocity-dependent increase in resistance feature of
spasticity, a quantitative assessment with computerized control of
external stretch was developed (22, 23).
In this approach, a joint is stretched by a motorized device at a
controlled, constant speed. Resistance torque is obtained to quantify
responses from spastic muscles. Reflex torque is quantified objectively
by subtracting passive resistance at a very slow speed of stretch, e.g.,
5°/s from that at a fast speed, e.g., 100°/s. Reflex torque is
attributed primarily to underlying neural mechanisms of spasticity. In a
previous study (24),
we have demonstrated that reflex torque was velocity-dependent at the
same wrist position (muscle length), and changed with various wrist
positions at the same speed of stretch. This biomechanical
quantification of spasticity is also sensitive to quantify reflex and
non-reflex responses from spastic elbow flexors in response to
controlled cold exposure (25).
In the present study, the specific aim was to examine the
possible mechanisms mediating the phenomenon of motor overflow in
chronic stroke. Stroke survivors and healthy controls were instructed to
flex the elbow joint voluntarily at submaximal levels. Surface EMG
signals were recorded from bilateral elbow flexors and finger flexors to
quantify motor overflow. Within-limb and between-limb EMG-EMG coherence
analyses were performed. Elbow flexor spasticity was quantified using
our established biomechanical approach. Since motor overflow is commonly
seen in stroke survivors with spasticity, they may share the same
underlying pathophysiology. We hypothesized that there is greater motor
overflow to the spastic elbow and finger flexors and that greater motor
overflow is highly correlated with spasticity, as compared to the
control group. Furthermore, post-stroke spasticity is primarily
attributed to reticulospinal hyperexcitability and has separate
underlying mechanisms for weakness (26, 27).
between-limb intermuscular EMG signals were hypothesized to have
significant EMG-EMG coherence in the alpha band to reflect
reticulospinal hyperexcitability. Motor overflow was further
hypothesized to correlate with spasticity (reflex torque), but not
weakness.
Yen-Ting Chen
Shengai Li
Elaine Magat1,2, 
Ping Zhou
Sheng Li