Differential Adaptations of Muscle Architecture to High-Velocity Versus Traditional Strength Training in Cerebral Palsy

Introduction

Cerebral palsy (CP) is the most common motor disability originating in childhood. ¹ People with motor disabilities face even greater challenges in achieving adequate levels of physical activity as compared with the general population. It is well documented that youth with CP have decreased activity levels compared with typically developing youth.^2,3 Muscle weakness in CP is considered to be one of the primary impairments that causes activity limitation. In fact, weakness of the quadriceps has consistently been reported to be highly correlated with gross motor function.^4,5 The quadriceps are important for ambulation and other activities of daily living, such as stair climbing. During ambulation, the quadriceps muscle group contributes substantially to hip and knee extension during single limb support and is one of the largest contributors to the vertical acceleration of the center of mass.^6,7

The architecture of a muscle is the primary determinant of muscle function. ⁸ In normal muscle, physiological cross-sectional area (CSA) determines the capacity of the muscle to produce force and is indicative of the number of sarcomeres in parallel, although other measures of muscle size have also been shown to be predictive of force-generating capacity.^9,10 In children with CP, we have previously reported that anatomic CSA of the rectus femoris (RF) and muscle thickness of the vastus lateralis (VL) and RF are decreased and are predictive of voluntary peak torque production.^11,12 However, fiber length is considered to be the single most important architectural parameter. ¹³ Fascicle, or fiber length, is the primary determinant of the shortening velocity and excursion of a muscle and is indicative of the number of sarcomeres in series. We have also shown that RF fascicle length is decreased in youth with CP compared with typically developing youth. ¹¹ Muscle power, as the product of force and velocity, would be adversely affected by decreases in muscle size and fascicle length, respectively, as well as by other neural factors, such as voluntary activation deficits. Our preliminary data show that average muscle power during a knee extension isokinetic test at 60 deg/s is only 18% of an age-matched, typically developing comparison group. This decrement in quadriceps muscle power was much greater than muscle strength, which has been reported to be approximately 50% of typically developing youth.^14-16

Because one of the main goals of rehabilitation interventions in CP is to maintain upright ambulation, strength training (ST) of the knee and hip extensors is often advocated. However, despite modest increases in strength, there is no high-level evidence that lower extremity ST improves walking speed or other measures of functional walking capacity.^17,18 One possible reason for the lack of significant improvements in ambulation in youth with CP is the focus of resistance training programs solely on improving muscle strength. Because youth with CP have difficulty producing torque at higher velocities of movement ¹⁶ and have a decreased rate of force development, ¹⁵ it would appear that progressive resistance training focusing on greater movement velocities, and thus, muscle power, would be more beneficial than traditional ST. Second, several studies included in a recent meta-analysis were not intensive enough, in that the intervention did not provide an adequate load to effect changes at the muscle level. ¹⁷ This is not surprising, given that optimal dosing guidelines (frequency, intensity, and duration) for youth with CP have not been established.

Despite previously held beliefs that muscle changes in CP are irreversible, ¹⁹ we now know that muscles are one of the most plastic tissues in the body, adapting in very specific ways to both the amount and type of stimulation imposed on them. ²⁰ Traditional ST with high loads performed at slow and controlled movement speeds has been shown to increase muscle thickness, CSA, and fascicle angles in healthy adults.^21,22 In contrast, there have been reports of increased fascicle length, in addition to muscle thickness, in response to resistance training with the concentric portion performed at higher movement velocities.^{23 -25} Furthermore, evidence suggests that isokinetic training performed at high velocities is superior in inducing muscle hypertrophy as compared with low-velocity training.^25,26 These data support a specificity of training effect where the training stimulus was associated with increases in the corresponding muscle architectural parameters. However, it is unknown whether specificity of training principles hold true in CP.

Therefore, the purpose of this study was to investigate whether velocity training (VT), which included resistance training at increasingly higher velocities, would induce specific architectural and muscle performance adaptations not observed with traditional ST. We hypothesized that increases in muscle size and strength would occur in both the ST and VT groups in response to training. However, we hypothesized that fascicle length and muscle power would increase after VT only. A secondary aim was to determine the effects of ST and VT on walking speed and functional walking performance.

Methods

Outcome Measures

Muscle architecture

Muscle thickness, fascicle length, and fascicle angle of the RF and VL as well as RF CSA were examined with ultrasound imaging (GE Logiq i, GE Healthcare, Chalfont St. Giles, UK) in 2D B-Mode with a 6 to 12 MHz linear array transducer. Detailed methods have been described previously. ¹¹ Briefly, subjects were positioned supine with knees resting in full available extension. Images were taken at the midway point between the anterior superior iliac spine and the superior pole of the patella for the RF and at the midway point between the greater trochanter and lateral femoral condyle for the VL in the midsagittal plane of the muscle. These locations were marked clearly on the skin with a surgical pen. Probe orientation was transverse for calculation of CSA and longitudinal (in the plane of the fascicular path) for muscle thickness, fascicle angle, and fascicle length. Muscle thickness was measured as the perpendicular distance between the deep and superficial aponeurosis, and fascicle angle as the angle between the fascicle and the deep aponeurosis. To obtain fascicle length, the visible portion of the fascicle was measured within the image frame and the remainder was estimated as a linear continuation of the fascicle and aponeuroses in the proximal direction using digitizing software as illustrated in Figure 1 (NIH ImageJ version 1.44, National Institutes of Health, Bethesda, Maryland). To ensure accurate probe placement across repeated scans (baseline and post), acetate paper was placed over the thigh during the baseline assessment, and the probe placements, anatomic landmarks, and any other distinguishing features, such as scars or moles were marked. For the postassessment, anthropometric measurements were also taken and marked on the skin. The acetate paper was repositioned over the thigh to verify accuracy of measurements and probe placement. ²⁹ The knee joint angle was also measured with a goniometer on the postassessment to ensure that the knee was in the same position as the preassessment. A single, experienced examiner performed all of the ultrasound assessments. However, measurements from the obtained images were performed by a separate examiner to remove any potential bias.

OPEN IN VIEWER

Figure 1.

Ultrasound images of the rectus femoris preintervention (A) and postintervention (B) in one participant. Line AB represents the superficial aponeurosis. Line CD represents the deep aponeurosis. Line EF represents the measurement of muscle thickness as the perpendicular distance between lines AB and CD. Fascicle angle is measured as the angle between lines CD and AD. For fascicle length, a line of best fit was drawn along the length of a fascicle and joined lines AB and CD. In this example, fascicle length increased by 8 mm postintervention as illustrated visually in the upper right portion of the figure.

Muscle function

Each of the following assessments were administered on the Biodex System 4 Pro dynamometer with standardized positioning and set-up as described previously. ¹⁴ Two to 5 submaximal practice trials were given before each of the assessments. Knee extensor muscle strength was measured as the peak torque obtained over 3 trials during an isometric contraction with the knee flexed to 90°. Knee extensor strength was also measured isokinetically at 30 deg/s, 60 deg/s, 90 deg/s, and 120 deg/s using the passive concentric mode to test throughout the entire range of motion. Peak torque obtained over 5 repetitions per speed was used for analysis. Peak torque was gravity corrected using the Biodex system software. The change in torque–angle relationship was measured by calculating the angle of peak torque during maximal isokinetic knee extension contractions at 60 deg/s.

Peak power and peak velocity were measured in isotonic mode at a load equivalent to 20% of peak isometric torque. Isotonic contractions allow the assessment of muscle power without constraining the velocity of the movement and are believed to be more relevant to physical performance in activities of daily living. ³⁰ Instructions were to kick (extend) the leg as fast as possible against the load when given the start signal (reference). Power was calculated every 10 ms over 5 repetitions (velocity × torque) to determine peak power.

Results

Characteristics of the Study Population

Figure 2 describes the recruitment and randomization process in detail. Sixteen participants were randomized to the 2 intervention groups: ST group (n = 7; 11 legs) and VT group (n = 9; 13 legs). No significant differences in baseline characteristics were found between the groups (Table 1). The mean age of participants for the VT group and ST group was 13.9 ± 2.6 and 13.7 ± 4.3 years, respectively. Four participants with spastic diplegia were in each group. Five with hemiplegia were in the VT group, and 3 with hemiplegia were in the ST group with an even distribution across GMFCS levels based on stratification procedures (ie, no group differences >2 participants per individual GMFCS levels; Table 1).

OPEN IN VIEWER

Figure 2.

Consort flow diagram of recruitment.

OPEN IN VIEWER

Table 1.

Participant Characteristics

	Sex Male/Female	Mean Age, y (SD)	Weight, kg	Height, m	Leg Length, cm	Knee Extension, deg	Subdiagnosis	GMFCS Level
Velocity training (VT; n = 9)	4/5	13.9 (2.6)	45.5 (8.9)	1.51 (0.14)	81.5 (10.3)	−3.7 (3.7)	4 spastic diplegia; 5 hemiplegia	III = 4; I = 5
Strength training (ST; n = 7)	6/1	13.7 (4.3)	43.1 (17.1)	1.46 (0.16)	75.3 (11.6)	1.7 (2.8)	4 spastic diplegia; 3 hemiplegia	III = 2; II = 1; I = 4

Abbreviation: GMFCS, Gross Motor Function Classification System.

Outcome Measures

There were no significant differences between groups at baseline for any of the outcome measures.

Muscle architecture

A statistically significant difference between interventions was observed for RF fascicle length with a large effect size (d = 1.24, P = .012). RF fascicle length increased in the VT group (+1.23 cm; 95% confidence interval [CI] = 0.12-2.34, P = .044) and decreased in the ST group (−0.76 cm; 95% CI = −0.07 to −1.46, P = .049) after training as illustrated in Figure 3. Pre- and post- ultrasound images of a participant who increased fascicle length after VT are illustrated in Figure 1. Within-group comparisons revealed statistically significant increases in RF CSA for both groups (VT, +0.26 cm², 95% CI = 0.09-0.44, P = .029; ST, +0.57 cm², 95% CI = 0.17-0.97, P = .009). VL muscle thickness significantly increased after ST only (P = .011; Figure 3).

OPEN IN VIEWER

Figure 3.

Preintervention to postintervention changes in muscle architecture of the rectus femoris (RF) and vastus lateralis (VL) for velocity training (VT) and traditional strength training (ST) groups (CSA = cross-sectional area). All values are means ± standard error. ^†Significant between-group differences (P < .05). *Significant within-group differences (P < .05).

Muscle function

Preintervention and postintervention results for muscle function are illustrated in Figure 4. Peak velocity improved by 16.5% in the VT group (P < .001) and decreased by 2% in the ST group (P = .408). A medium between-group effect size for peak velocity was found (d = 0.84; P = .098) favoring the VT group. Furthermore, peak power significantly increased by 64% in the VT group only (P = .023). Both groups had significant increases in peak torque at all isokinetic velocities. In contrast, isometric peak torque increased after ST only (P = .007). A shift in the torque–angle curve was noted in the VT group only with an increase in the angle of peak torque to longer muscle lengths (VT—pre, 68.8° ± 3.8°; post, 75.9° ± 2.7°; P = .029 and ST—pre, 67.9° ± 4.7°; post, 71.7° ± 2.2°, P = .277).

OPEN IN VIEWER

Figure 4.

Preintervention to postintervention changes in knee extensor strength (peak torque), peak power; and peak velocity, for velocity training (VT) and traditional strength training (ST) groups. All values are means ± standard error. ^§Significant within-group differences (P < .01). *Significant within-group differences (P < .05).

Secondary Outcome Measures

Functional walking performance and speed

Baseline differences between groups were not statistically significant (P < .05). There were no significant improvements in the ST group. In contrast, self-selected and fast walking speeds, TUG, and 1-minute walk test all significantly improved in the VT group as illustrated in Table 2.

OPEN IN VIEWER

Table 2.

Pre and Post Means and SD for Gait and Functional Performance Variables

	VT			ST
	Pre	Post	P	Pre	Post	P
SS speed, m/s	0.98 (0.41)	1.08 (0.46)	.04	1.23 (0.23)	1.27 (0.26)	.49
Fast speed, m/s	1.34 (0.58)	1.46 (0.67)	.03	1.60 (0.34)	1.63 (0.34)	.65
TUG, seconds	13.04 (6.3)	11.52 (4.8)	.04	9.47 (3.8)	9.53 (3.8)	.87
1MWT, meters	84.6 (38.2)	92.5 (41.6)	.01	99.0 (22.8)	102.9 (24.1)	.14

Abbreviations: VT, velocity training; ST, strength training; SS speed, self-selected walking speed; fast speed, fast walking speed; TUG, timed up-and-go test; 1MWT, 1-minute walk test.

Discussion

Differential muscle architectural and performance adaptations were observed in response to 24 sessions of ST versus VT in youth with CP. To our knowledge, this is the first study to show increases in fascicle length in CP in response to a targeted exercise intervention, such as VT. These results highlight the enormous capability of muscles in CP for muscle plasticity. Not only did muscle architectural plasticity occur in both resistance training groups, but these results support the notion that muscles in CP are capable of responding differentially to specific training stimuli.

Fascicle length is the primary determinant of muscle shortening velocity. As hypothesized, training at higher shortening velocities resulted in increases in RF fascicle length and CSA whereas ST resulted in a decrease in RF fascicle length and muscle hypertrophy. Resistance training studies between 12 and 14 weeks in duration in young and older adults have shown increases in whole quadriceps or RF CSA between 3% and 10%^21,29,35 and VL muscle thickness between 7% and 12%.^23,36 Our results (+4% to 6%) are in line with these studies, given that our study consisted of 24 treatment sessions, whereas the aforementioned studies reported between 33 and 42 total training sessions. The 16% increase in RF fascicle length for the VT group in this study is also comparable with other studies that used high-velocity training protocols. These studies reported increases in fascicle length between 10% and 39% for the VL and RF.^23,24 We do not feel that these changes were related to a decrease in fascicle angle, as the change from pre to post for the VT group was only 1.5°, which is functionally insignificant. The ST group fascicle length decreased by 7%, which is consistent with other reports that have shown that resistance training at slow, controlled speeds results in either no change or a decrease in fascicle length.^21,22 However, evidence for contraction velocity as a stimulus for increasing fascicle length requires further investigation to substantiate this claim. Other proposed stimuli for increases in fascicle length are eccentric/lengthening contractions and muscle excursion range or stretch.^37-39 However, eccentric contractions were not performed in this study, and all participants exercised through their available range of motion. Therefore, the training velocity appears to be the primary factor that explains the differential adaptations between the 2 training groups.

The effect of an increase in fascicle or fiber length is an increase in the maximum contraction velocity of the muscle; and for an equivalent absolute velocity, a muscle with longer fibers will generate a higher force (see Lieber and Friden, ⁸ Figure 9B, for a review). These results suggest that increases in RF fascicle length and CSA after VT translated into significant improvements in peak movement velocity and peak torque. Therefore, muscle power, as the product of torque and velocity, significantly improved after VT only. Peak torque during isokinetic testing increased in both groups but only peak velocity increased after VT, suggesting that the increase in peak velocity was the primary factor responsible for the increase in muscle power after VT (Figure 4). Surprisingly, a specificity of training effect for peak torque was not observed. Rather, peak torque increased at all tested isokinetic velocities regardless of whether participants trained at slow (ST) or faster (VT) movement speeds. However, isometric torque significantly increased after ST only, highlighting the specificity of training effect of ST on muscle size.

Another effect of an increase in fascicle length is to increase the active muscle range of the length–tension relationship. ⁸ The increase in the angle of peak torque (ie, greater knee flexion) in the VT group and resultant shift in the torque angle curve to longer muscle lengths consistent with an increase in fascicle length. A limitation of noninvasive measures of fascicle length, such as obtained with ultrasound imaging, is the inability to measure actual sarcomere length. Therefore, an increase in fascicle length only infers that sarcomeres have increased in series.

Although muscle architectural adaptations were the subject of this investigation, we posit that the combined effect of muscle architectural and neural adaptations to resistance training contributed to the observed changes in muscle strength and power. Neural adaptations to resistance training include increases in motor unit recruitment and firing rates, among others. ⁴⁰ Although it is currently believed that strength gains in prepubertal and postpubertal children are primarily because of neural changes,^41,42 we show here that muscle architectural adaptations in these children with CP occur to a similar degree as young and older adults in response to resistance training.

Interestingly, there was a differential adaptation between muscle groups where only RF fascicle length was altered in response to training. A previous report showed that fascicle length was shorter in the VL as compared with the RF in TD children. ¹¹ This is also the case in this study, where VL fascicle length is shorter than RF fascicle length. One hypothesis is that in CP, the RF is predisposed to adapt to the velocity stimulus because of its longer length and greater excursion capacity. In fact, fascicle length of the RF adapted to both resistance training interventions but in opposite directions (increased in the VT group, decreased in the ST group).

Muscle power has been shown to be a more important muscle performance indicator than strength and a better predictor of physical function and mobility, particularly in older adults.^43,44 Therefore, it is currently recommended that resistance training programs designed for older adults should focus on improving muscle power, rather than muscle strength. ⁴³ Perhaps, the lack of improvement in functional walking performance and walking speed in the ST group was because of the inability of the training stimulus to specifically target muscle power. The lack of significant improvements in walking, activity, and participation as a result of ST interventions in CP has been previously well documented in recent systematic reviews.^17,45 Limitations of several studies reviewed include inadequate intensity and duration of the strength-training interventions. In our study, both groups trained at an appropriate intensity, volume, frequency, and duration according to the American Academy of Pediatrics and National Strength and Conditioning Association position statement papers for youth resistance training.^46,47 The only difference between the 2 interventions was the velocity of training. Training with higher velocity, maximum-effort muscle contractions in the VT group resulted not only in statistically significant, but also clinically meaningful changes in both self-selected and fast walking speeds that met or exceeded the minimal clinically important difference of 0.1 m/s for gait speed in CP ⁴⁸ (change [post − pre]: self-selected = 0.10 m/s; fast = 0.12 m/s). Furthermore, significant gains were made after VT in functional walking performance as measured by the TUG and 1-minute walk test. Similar to the results of the meta-analyses, traditional ST resulted in strength gains that did not translate into improvements in activity limitations related to walking.

Conclusions

Muscle architecture in CP is highly plastic and is capable of responding differentially to the type of training stimulus. As hypothesized, training at higher shortening velocities resulted in increases in fascicle length and CSA whereas ST resulted in increases in muscle size only. Furthermore, the relative magnitude of muscle architectural adaptations was similar to that in adults undergoing resistance training protocols. Collectively, these results suggest that VT is equally effective as traditional ST in improving isokinetic strength of the knee extensors but more effective in improving velocity of movement, muscle power, walking speed, and functional walking performance. Therefore, strengthening interventions involving higher velocity movements, such as VT or power training, should be incorporated into everyday clinical practice for youth with CP.