Impairment of Multi-Finger Force Control in Adult Patients with Dyskinetic Cerebral Palsy
Article information
Abstract
PURPOSE
This study aimed to investigate the impact of cerebral palsy (CP) on motor control strategies, with a focus on understanding the altered capability of multi-finger force control in adult patients with dyskinetic CP.
METHODS
Eight adults with dyskinetic CP and ten age- and sex-matched healthy controls participated in this study. The participants performed three force production tasks using four fingers of the dominant hand under isometric conditions: maximum voluntary contraction (MVC) to assess maximum force production, a ramp task to evaluate finger interdependency (enslaving), and a pulse task to examine the ability to rapidly and precisely adjust force. The co-contraction index (CCI) of finger flexor and extensor was also determined during steady state and pulse force production.
RESULTS
Patients with dyskinetic CP demonstrated significant impairments in finger force control compared to healthy controls. Specifically, the accuracy and consistency of force production were reduced in the CP group, and they exhibited a longer time to peak pulse force. The CP group also showed increased finger interdependency and elevated CCI during pulse force production, suggesting a greater reliance on co-contraction and less efficient motor control strategies.
CONCLUSIONS
This study highlights the distinct motor control deficits in individuals with dyskinetic CP, particularly in tasks requiring rapid and precise finger force adjustment. These results have important implications for the functional rehabilitation of patients with CP. Therapies aimed at reducing excessive co-contraction and decreasing finger interdependency may be beneficial for improving muscle coordination and overall motor function.
INTRODUCTION
Cerebral palsy (CP) is a neurological disorder that occurs in fetuses and collectively refers to a non-progressive condition caused by permanent damage to the immature brain [1]. CP is one of the most common neuromotor disorders, affecting 0.2-0.3% of all fetuses [2–4]. The specific motor function disorders that arise depend on the area of the brain that is damaged. Spastic CP, the most common form, is caused by damage to the corticospinal tract, whereas dyskinetic CP is associated with hypoxic damage to the basal ganglia and thalamus [5]. The bodily functions most affected by damage to the basal ganglia in dyskinetic CP include movement functions related to posture and the fine motor control of the fingers [6].
While symptoms of spastic CP have been measured using various quantitative methods, such as clinical rating scales [7], torquevelocity relationships [8], and velocity-electromyography relationships [9], very few studies have quantified the changes in the behavioral patterns and neuromuscular function in dyskinetic CP. Additionally, although CP is classified as a non-progressive disorder and research has predominantly focused on children, the increasing survival rate of CP patients highlights the necessity of investigation on adult CP patients [10]. Therefore, the current study focused on quantifying motor function characteristics of adult patients with dyskinetic CP.
The hand is a representative organ that exemplifies the complex motor control characteristics of the human body, and many previous studies have observed changes in human motor control characteristics due to neurological lesions by analyzing the hand/finger system [11–15]. Building on these findings, previous studies on CP patients have provided valuable insights into the deficits in hand function associated with the disorder. Specifically, the patients with CP have been reported to exhibit reduced finger force production capabilities [16], along with impaired ability to regulate force in a stable and precise manner due to altered muscle coordination patterns [17]. Additionally, increased finger interdependency, which reflects reduced dexterity in hand use, has been identi-fied as a contributing factor to diminished hand function in these individuals [13]. However, because these findings are based on studies of spastic CP patients, they cannot be directly extrapolated to dyskinetic CP patients, who are affected by different pathophysiological mechanisms. Given the crucial role of the basal ganglia in motor planning and execution, damage to this area in dyskinetic CP patients may result in a distinct pattern of motor impairment.
Therefore, this study designed an experimental task to quantify an individual’s ability to generate and regulate force using multiple fingers. This study aimed to investigate the altered finger force control capabilities in adult patients with dyskinetic CP by comparing them with healthy individuals, thereby identifying the distinctive characteristics that emerge in their motor behavior. Following a set of hypotheses was formulated. First, the ability to perform finger force production tasks (ability of maximum force production; and accuracy and consistency of force production) of CP patients will be smaller than that of the healthy control group. Second, the finger interdependency of CP patients will be greater than those of the healthy control group. Third, the forearm muscle activation patterns (co-contraction of finger flexor and extensor) of CP patients will be different from those of the healthy control group. This approach would provide valuable insights into the altered behavioral features of dyskinetic CP patients, offering important considerations for their rehabilitation.
METHODS
1. Participants
Eight patients diagnosed with dyskinetic CP with diplegia by a pedi-atric orthopedic surgeon were classified into the CP group (eight females; 45.7±6.4 years old; right-handed). Those selected as the subjects in the CP group were able to carry out daily life activities in terms of hand function such as eating and writing [13]. Before the experiment, they were confirmed to have no sensory defects of both hands and fingers by a monofilament test (Semmes-Weinstein Monofilament Test, North Coast Medical Inc., USA), and the sensory threshold was set to filament number 3.61 (0.4 g) [18]. Sex and age-matched ten healthy adults (ten females; 46.1±10.1 years old; right-handed) participated in the study as a healthy control group. This study was conducted with the approval of the Institutional Review Board (IRB) of Seoul National University Bundang Hospital (IRB No. B-1707-408-302).
2. Apparatus
Four unidirectional piezoelectric transducers (Model 208A03, PCB Piezotronics Inc., Depew, NY, USA) were used to measure the flexion and extension forces of each finger under isometric conditions. Each force transducer was attached vertically to a frame made of aluminum, and a U-shaped groove was placed for each sensor to allow isometric contraction with the distal phalanx of each finger inserted (Fig. 1A). The vertical direction (y-axis) distance of each sensor was fixed at 3 cm, and the position in the mediolateral direction (x-axis) and the angle in the anteroposterior direction (z-axis) were adjusted according to the morphological characteristic of the subject's fingers. A cylindrical handle was fixed to be positioned on the palm of the subject to maintain the constant configuration of the hand and fingers (Fig. 1A). Signals from each sensor were collected at a frequency of 500 Hz using the LabVIEW soft-ware (National Instrument, Austin, TX, USA). The surface electromyo-gram (sEMG) system (Trigno IM, Delsys Inc., Boston, MA, USA) was used to measure the level of muscle excitation of the finger flexors and extensors during the all experimental tasks (Fig. 1B). EMG electrodes were attached to the three flexor muscles (FDS, flexor digitorum superficialis; FCU, flexor carpi ulnaris; FCR, flexor carpi radialis) and the three extensor muscles (EI, extensor indicis; ED, extensor digitorum; ECU, extensor carpi ulnaris) to collect signals at a frequency of 2,000 Hz.
The illustrations of the experimental equipment set up. (A) Four unidirectional force transducers (PCB Piezotronics Inc., Depew, NY, USA) attached to the customized aluminum panel to measure four finger forces. The four fingers were naturally flexed the proximal interphalangeal joint for about 10-20º and inserted in corresponding U-shaped grooves. (B) The surface EMG sensors (Delsys Inc., Boston, MA, USA) were attached to six muscles including finger flexors and extensors on the forearm.
3. Experimental procedures
The experimental tasks include three divided tasks as follows: (1) maximum voluntary contraction task (MVC task), (2) ramp task, and (3) pulse task. The subject’s dominant hand, their right hand, was used in all experimental tasks. During the task, the subject was asked to maintain a 45° flexion of the shoulders and elbows, and the height of the chair was adjusted to fit the body. Throughout the experiment, the sum of four finger forces (FTOT) values were displayed as real-time feedback on a monitor fixed at eye level, 50 cm away from the front (27-inches; 1,920×1,080 resolution at 60 Hz).
1) MVC task
The MVC was a task that involves all four fingers to create the maximum force in the flexion or extension direction for five seconds (Fig. 2). The maximum finger forces in the flexion and extension directions were measured, and the trials with a high FTOT were selected among the two MVC trials. The measured maximum force of each finger (MVCi, index [I], middle [M], ring [R], little [L], and maximum FTOT [MVCTOT]) were used to set the target force for the other two tasks.
The illustrations of the feedback screen in three experimental tasks. Maximum voluntary contraction task (MVC task), ramp task, and pulse task.
2) Ramp task
The ramp task was a task to measure the interdependency between each finger force, and it is a task to gradually increase the force by one finger along the trajectory provided as feedback (Fig. 2). The trajectory provided as feedback was set to 5% MVCi (first steady-state phase) for 4 seconds, 5-40% MVCi (increase phase) for 12 seconds, and 40% MVCi (second steady-state phase) for 4 seconds. It was measured in the flexion and extension directions for each finger, and forces that were unintentionally generated by the fingers (non-task fingers) other than the finger performing the task (task finger) were simultaneously recorded, but not provided as feedback. The subjects were encouraged to focus only on the task fingers. A detailed description of the ramp task protocol can be found in [19].
3) Pulse task
The last task was a pulse task to generate a short pulse force in the flexion or extension direction simultaneously using four fingers (Fig. 2). The subjects were asked to produce a constant force (5% of the average flexion or extension MVCTOT) for about 5 seconds and produce a fast pulse force (20% of the average of flexion and extension MVCTOT) by changing the direction of the force at a self-paced timing. The task was performed under two conditions as follows: the flexion to extension (FE) condition, where the flexion force was maintained and a pulse force was produced in the direction of extension, and the extension to flexion (EF) condition, which was opposite to the former condition. During the task, the FTOT produced by the subject was displayed in the flexion (-) and extension (+) directions based on the vertical axis of the graph on the monitor in real-time. A total of 50 trials were performed, 25 times for each condition (FE and EF), and about 10 seconds break (or more if necessary) was taken between every two trials to prevent muscle fatigue. A detailed description of the pulse task protocol can be found in [13].
4. Data analysis
The collected data were analyzed by the universal program (MAT-LAB, MathWorks, Natick, MA, USA). The collected force values of each finger were processed using a low-pass filter (10 Hz cut-off with zero-lag, 4th-order Butterworth filter). The collected EMG data were processed using a notch filter (60 Hz) to remove external electrical noise and then rectified to apply a band-pass filter (50 Hz lower cut-off and 450 Hz upper cut-off).
1) Indices of task performance
The pulse task data were used to quantify the performance indices. The accuracy index (ACI) of the task was a value indicating how far the peak value of the pulse force produced in the trial was from the target force value, calculated by Equation 1. The precision index (PRI) was a value indicating how far the peak value of the pulse force produced in the trial was from the average of the peak values of the pulse force generated in each trial, calculated by Equation 2 [20].
where i denotes each trial; Fi denotes the peak value of the pulse force produced in each trial; Ftarget denotes the target force; and Fmean =the average of the pulse force peak values produced in each trial. Further, the time from the moment when the change in steady finger force begins (pulse onset) to the moment when the peak pulse occurs was quantified as the time to peak force, serving as an indicator of the rate of pulse force development.
2) Index of finger interdependency (finger enslaving)
The ramp task data were used to calculate the enslaving matrix and the enslaving index, and two enslaving matrix and index values were calculated for each subject according to the conditions (i.e., two force directions). The enslaving matrix (E) was expressed by the equation below (Equation 3).
where k denotes flexion or extension; and the enslaving matrix (Ek) represents the ratio of forces unintentionally generated by the remaining fingers (off-diagonal elements) while each finger i (the main diagonal element) performs the task (gradual force production). Each element was obtained by calculating the coefficient of the regression equation (e i,j,k) between each finger force by multiple linear regression. The enslaving index was calculated as the average of the off-diagonal elements of the Ek matrix. A detailed description for the calculation of the enslaving index can be found in [21].
3) Co-contraction indices (CCI) of finger flexor and extensor muscles
To calculate the CCI, the EMG signal for 500 ms in the steady-state force production phase (before pulse onset) was integrated (i EMG) and normalized as the integral EMG of the MVC task for each condition. CCI refers to the relative level of antagonist muscle excitation (i EMGANT) compared to the total level of muscle excitation (i EMGTOT), calculated by the following equation (Equation 4) [13]. The CCI of the pulse phase (from pulse onset to peak) was also calculated using the same method.
where k denotes FE or EF; i EMGANT is defined depending on the direction, as the sum of the extensor i EMGs for the EF and the sum of the flexor i EMGs for the FE; and i EMGTOT is defined as the sum of all six muscles.
5. Statistics
Descriptive statistics and two-way mixed ANOVA were used to identify the influence of the independent variables on the major dependent variables (ACI, PRI, time to peak force, MVC force, enslaving, CCI). Factors included in the independent variable were as follows: Group (two levels: CP and Control), Direction (two levels: FE and EF). The sphericity assumption was confirmed by Mauchly’s sphericity test. If sphericity was not assumed, the corrected p-value corrected by the Greenhouse-Geisser correction was presented. According to a previous study conducted using the same method and equipment, the results of identifying the intra-class correlation coefficient to verify the reliability of repeatedly measured forces and EMG values all showed a high value of 0.9 or higher [13]. For post-hoc analysis, the Mann-Whitney U test was performed. The significance level of all statistical tests was set as α=0.05.
RESULTS
1. Performance indices (ACI, PRI, and time to peak force)
The ACI and PRI (Fig. 3A and B, respectively) showed higher values in the CP group than in the healthy control group in both directions, and relatively large in the EF condition compared to the FE condition. These results were supported by two-way mixed design ANOVA with the factors Group and Direction, which showed significant main effects on Direction (ACI: F[1,14] =16.36, p =.002, ηp² =0.60; PRI: F[1,14] =5.76, p =.042, ηp² =0.30) and Group (ACI: F[1,14] =11.64, p =.006, ηp² =0.51; PRI: F[1,14] =22.37, p =.001, ηp² =0.67) without factor interactions.
Results of (A) accuracy index (ACI), (B) precision index (PRI), and (C) time to peak force during the pulse force production task. filled bars and empty bars indicate CP group and control group, respectively. The Flx-Ext represents pulse force direction from flexion to extension, and Ext-Flx indicates vice versa. The asterisks indicate significant differences between groups. Values are means±standard errors across the subjects.
In the time to peak force (Fig. 3C), the CP group also showed a relatively larger value than the healthy controls, and showed the main effect on the Group factor (F[1,14] =19.35, p =.001, ηp² =0.64). However, the FE condition showed a relatively large value compared to the EF condition, which could be supported by the main effect on the Direction (F[1,14] = 6.90, p =.024, ηp² =0.39).
2. Maximal finger forces (MVC force) and enslaving index
The capacity of digit MVC force production of the CP group was lower than that of the healthy controls for the both flexion and extension direction (Fig. 4A). In addition, the flexion direction showed larger MVC force than the extension direction for both the CP and control group. These results were supported by two-way mixed design ANOVA with the factors Group and Direction, which showed significant main effects on Direction (F[1,16] =167.95, p <.001, ηp² =0.91), and Group (F[1,16] = 9.16, p <.001, ηp² =0.35).
(A) Maximal voluntary contraction (MVC) forces, and (B) the index of finger force enslaving during flexion and extension effort for the CP and control group (CP: filled bars, Control: empty bars). The asterisks indicate significant differences between groups. Values are means±standard errors across the subjects.
In the case of the enslaving index (Fig. 4B), the CP group showed a larger value than the healthy controls in the all directions, and a relatively large value in the extension condition compared to the flexion condition. These results supported by the main effect on the Direction (F[1,16] = 17.41, p =.001, ηp² =0.51) and the Group (F[1,16] =38.50, p <.001, ηp² =0.68).
3. CCI of finger flexor and extensor muscles
In the case of the CCI between the flexor and extensors of the fingers, there was no significant difference between groups in the steady-state phase (Fig. 5A CCISS). However, in the pulse phase, it was confirmed that the CP group showed significantly larger values than the healthy controls (Fig. 5B CCIpeak). In other words, the healthy controls showed a large decrease in CCI in the pulse phase compared to the SS phase, while the CP group still showed a large CCI value in the pulse phase. These results were supported by two-way mixed design ANOVA with the factors Group, and Direction, which showed significant main effects on Group (F[1,14] =5.42, p =.035, ηp² =0.28) in the CCIpeak.
The co-contraction index (CCI) of the (A) steady-state phase and the (B) pulse phase for each direction condition (Flx-Ext, Ext-Flx). The value was calculated by defining the finger flexor as the antagonist muscle in the Flx-Ext condition and the finger extensor in the Ext-Flx condition. The asterisks indicate significant differences between groups. Values are means±standard errors across the subjects.
DISCUSSION
This study aimed to investigate the impact of neurological changes caused by dyskinetic CP on human motor control mechanisms, particularly in the context of multi-finger force control. Participants performed three force production tasks using four fingers of dominant hand. The MVC task was designed to assess their maximum finger force production capability, while the ramp task evaluated finger interdependency, which is related to hand dexterity. The pulse task was conducted to examine the ability to rapidly and accurately adjust force. The results confirmed all the hypotheses mentioned in the introduction, and it provide important insights into the altered motor control strategies in individuals with dyskinetic CP.
The reduced accuracy and consistency in force production observed in the CP group during the pulse task (Fig. 3A and B) can be attributed to the inherent challenges associated with dyskinetic CP. Unlike spastic CP, where motor impairments are primarily due to increased muscle tone and rigidity [22], dyskinetic CP is characterized by involuntary muscle contractions and fluctuating muscle tone, leading to unpredictable and uncontrolled movements [23]. These characteristics likely contribute to the difficulty dyskinetic CP patients experience in rapidly and precisely adjusting their finger force to match target levels. The longer time to peak pulse force observed in these patients further underscores the impact of dyskinetic CP on their motor control efficiency (Fig. 3C).
This inefficiency in control strategies was also evident in the results of the enslaving index, which reflects the increased interdependency between fingers. The enslaving index, which measures the unintended force production by non-task fingers, was notably higher in dyskinetic CP patients compared to healthy controls (Fig. 4B). This suggests that the motor control strategy in dyskinetic CP is less efficient, with a greater degree of involuntary force spillover between fingers [24]. Such inefficiency likely contributes to the reduced dexterity and precision observed in these patients [25]. The higher enslaving index may also indicate a compensatory mechanism, where the nervous system attempts to stabilize movements by recruiting additional fingers, albeit at the cost of precision and agility [21]. Previous studies have shown that individuals with spastic CP also experience significant impairments in precise control of finger movements [13,26]. However, the mechanisms underlying these impairments differ between spastic and dyskinetic CP. In spastic CP, the primary issue is hyperactivation of muscles and reduced flexibility, resulting in stiff, jerky movements. In contrast, dyskinetic CP involves impairments in the basal ganglia, causing patients to struggle with the coordination and timing of muscle contractions.
The co-contraction of agonist and antagonist muscles observed in dyskinetic CP patients further complicates their ability to generate precise finger forces. Co-contraction, which is the simultaneous activation of opposing muscle groups, typically serves to stabilize joints in healthy individuals. However, in dyskinetic CP, it may contribute to increased muscle stiffness and resistance to movement, ultimately impairing fine motor control. This finding aligns with previous research that has high-lighted the role of co-contraction in motor deficits [27–29]. In our study, the CCI was calculated separately for the steady-state finger force production phase and the pulse force production phase. Interestingly, while there was no significant difference in CCI between groups during the steady-state phase (Fig. 5A), the CCI was significantly higher in the CP group during the pulse force production phase (Fig. 5B). This suggests that dyskinetic CP patients may struggle more with rapid force adjustments rather than maintaining a steady force. The elevated CCI during the pulse phase indicates that when a quick and precise adjustment in force is required, dyskinetic CP patients may rely more heavily on co-contraction, potentially as a compensatory mechanism to stabilize the fingers during the dynamic phase of force production [30]. However, this increased reliance on co-contraction likely contributes to the reduced accuracy and efficiency observed in their force control, further impairing their fine motor performance.
Interestingly, while dyskinetic CP patients showed significant impairments in force control, they were still able to perform the tasks, albeit with reduced efficiency. This suggests that despite the profound motor deficits, there is some preserved ability to generate and regulate force, which could be targeted in therapeutic interventions. Rehabilitation strategies that focus on improving coordination and reducing involuntary movements may help enhance the motor control capabilities of dyskinetic CP patients. For instance, therapies that reduce muscle co-contraction and finger interdependency could potentially improve the precision of finger movements and overall hand function. However, the current study focused only on hand function, which may not fully capture the broader motor challenges faced by dyskinetic CP patients. Future research should extend these findings to gross motor functions, such as gait and posture control, to provide a more comprehensive understanding of the motor impairments in dyskinetic CP and to develop targeted interventions that address both fine and gross motor deficits.
CONCLUSION
This study evaluated motor control deficits in adult patients with dyskinetic CP, focusing on tasks requiring rapid and precise finger force control. The findings demonstrated that CP patients exhibit significant impairments in precise finger force modulation and rapid force adjustment. The high interdependency between fingers and increased co-contraction of finger muscles may contribute to these deficits and should be considered key targets in physical therapy.
Notes
CONFLICT OF INTEREST
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
AUTHOR CONTRIBUTIONS
Conceptualization: J Song, K Kim; Data curation: J Song, K Kim; Formal analysis: J Song, K Kim; Visualization: J Song, K Kim; Writing – original draft: J Song, K Kim; Writing – review & editing: J Song, K Kim.