3.1 Participant Enrollment
Study advertisement yielded six parties reaching out for enrollment. Six individuals were assessed for inclusion and exclusion criteria, and all were enrolled in the study. Two participants did not complete the study, one due to scheduling conflicts before post-testing could be initiated, and the other could not focus attention and follow instructions during Week 0 testing. We report the data on the four participants who completed at least the Week 7 assessment sessions. Of these, one did not participate in Week 10 assessments. The self-reported or parent-reported participant characteristics of the four individuals who completed the ballet technique course are presented in Table 1. In summary, four participants with mixed spastic and dyskinetic CP included in this case series participants had mean age = 12.5 years, SD = 6.9 years, three were female, and one was male. The achieved enrollment numbers allow for a case series report using individual within-subject statistical analyses of the quantitative outcome measures.
Table 1
Participant characteristics.
Personal Information |
Participant | 1 | 2 | 3 | 4 |
Age | 22 | 7 | 8 | 13 |
Gender | Female | Male | Female | Female |
Origins of CP | Congenital | Prenatal trimester 2: spontaneous blood flow in the parietal and frontal lobe, microcephaly and poly microgyria | Not reported | Prenatal trimester 2 |
GMFCS | 4 | 1 | 2 | 1 |
MACS | 4 | 3 | 1 | Left: 1, Right 3 |
Extremity Distribution | Quadriplegia and right-side neglect | Left hemiplegia | Mild Quadriplegia | Right hemiplegia and mild left hemiplegia |
Medication | Not reported | Not reported | Note reported | Not reported |
Frequency of Treatment | PT: once a week for 50 minutes, AT: twice per week for 15 minutes | PT: twice per week for 30 minutes | OT: 90 minutes per week, PT: 75 minutes per week, AT: 30 minutes per week, Writing at YMCA: 45 minutes per week | PT: once every other week for 50–60 minutes |
Botulinum Toxin History | None | Once on arms/thumb | Multiple injections from 2011 to 2015 | Multiple injections from 2007 to 2012, paired with serial casting |
Surgical History | Twice on lower limbs | None | Tethered cord release | Right wrist tendon release |
Other Health Condition | None | Seizures, attention deficit disorder | None | None |
Type of CP | Spastic unilateral CP, dystonic CP | Spastic unilateral CP, dystonic CP, choreo-athetotic CP | Spastic unilateral CP, dystonic CP, choreo-athetotic CP | Spastic unilateral CP, dystonic CP, choreo-athetotic CP |
Hypertonia | Spasticity, dystonia | Spasticity, dystonia | Spasticity, dystonia | Spasticity, dystonia |
Hypotonia | Trunk hypotonia | | Trunk hypotonia | |
Hyperkinesia | | Myoclonus (not spontaneous) | Myoclonus | Myoclonus |
Participant’s Ability and Lifestyle |
Physical Activity Level Compared to Same-age Peers | | | | |
Academic Performance |
Reading Skills Compared to Same-age Peers | Above average | Below average | Above average | Above average |
Mathematic Skills Compared to Same-age Peers | Average (keeps up) | Below average | Below average | Average (keeps up) |
Writing Skills Compared to Same-age Peers | Above average | Cannot write/unable to write a sentence with correct spelling | Below average | Above average |
Notes | All participants lived in Central Illinois. Once participant was lost due to scheduling conflicts and a second participant was excluded from the study in Week 0 due inability to focus attention and follow assessment instructions. (5). Demographic information from these two participants is excluded in Table 1. The classification of cerebral palsy follows the guideline of Surveillance of Cerebral Palsy in Europe [3]. CP: cerebral palsy; GMFCS: Gross Motor Function Classification System; MACS: Manual Ability Classification System; PT: physical therapy; AT: aquatic therapy; OT: occupational therapy. |
3. Quantitative Outcome Measures
Representative data of a single stretch of the elbow flexors, including EMG, angle, and angular velocity vs. time, are presented in Fig. 1(a). A representative example of angular velocity vs. angle is shown in Fig. 1(b). Not all stretches resulted in a stretch reflex response. The percentages of imposed stretches that elicited a DSRT response are included in the Supplementary Figure S1. Linear regression analysis of angular velocity of stretch vs. angle resulted in low correlation coefficients r (< 0.6452), and 69.7% of total trials had non-significant p-values for the slopes of the linear fits, indicating the high variability in the responses as expected for mixed CP in hypothesis 1. These results are reported in Supplementary Table S1.
Exploring further the behavior in the variability of the DSRT responses, all measured values of the DSRT angles for each individual participant in Weeks 0, 7, and 10 are reported in detail in Supplementary Figures S2, S3, S4, and S5. The corresponding box plots of the coefficients of variation of the DSRT angles for each testing week are plotted in Fig. 2 (left column), and the descriptive statistics values are included in the Supplementary Table S2. The results of Shapiro Wilk normality tests and equal variance tests are included in the Supplementary Table S3. The p-values of one-tailed student’s t-tests of each participant on the coefficients of variation of the DSRT angles within each bin of velocity of stretch in Weeks 0, 7, and 10 were calculated and are reported in Table 2. The significant results are indicated by asterisks in Fig. 2 (left column). In summary, reduced means were obtained after the intervention and at one-month follow-up (Weeks 7 and 10) for participant A (W0 vs. W7: p-value = 0.0112, CI = 1.030 to 12.34, df = 25; effect size = 0.947; W0 vs. W10: p-value = 0.0127, CI = 1.118 to 15.27, df = 20, effect size = 1.142) and for participant B (W0 vs. W10, p-value = 0.0469, CI = -1.024 to 11.26, df = 11, effect size = 1.137); no significant reductions in the mean of the coefficients of variation were obtained for participants C and D.
(INSERT FIGURE 1 HERE; Fig. 1 Caption below)
Figure 1. (a) Sample data from a single trial in session 5 of stretching elbow flexors from participant A. The raw and filtered EMG of biceps brachii long head are plotted in the top two panels. The third panel shows the change of angle as a function of time, and the bottom panel shows angular velocity vs. time. The DSRT angle is 65.75, and its corresponding angular velocity of stretch is 196.83/s. (b) Representative angular velocity vs. angle graph showing DSRTs for participant A in session 5. The plot shows that 20 DSRTs were evoked at various angular velocities in a single testing session. Similar angles are recorded for a large range of angular velocities, indicating the lack of velocity dependence of the stretch reflex, which would occur in spasticity. No significant linear relationship was found (p-value = 0.7514).
Addressing hypothesis 2, the frequency distribution plots for the DSRT angle corresponding to each bin of angular velocity of width = 10°/s for each participant were generated and are included in the Supplementary Figures S2, S3, S4, and S5. The corresponding box plots are presented in Figure (2) (right column) and the descriptive statistics values are included in the Supplementary Table S2. The results of Shapiro-Wilk normality tests and equal variance tests on the data on each participant are included in the Supplementary Table S3. In accordance with the results of the normality and equal variance tests presented in the in the Supplementary Table S3, we report the results of one-tailed student t-tests, one-tailed Welch tests, or two-sample Kolmogorov-Smirnov tests as appropriate- Significant results of the student t-tests, Welch test, and two-sample Kolmogorov-Smirnov test for differences between the distributions are shown with asterisks in Fig. 2 (right column). Details of the statistical tests are included in Table 2. Briefly, reduced angular velocity means were obtained after the intervention and at one-month follow-up (Weeks 7 and 10) for participant A (W0 vs. W7: p-value < 0.0001, CI = 22.16°/s to 59.68°/s, df = 116; effect size = 0.796; W0 vs. W10: p-value < 0.0001, CI = 17.34°/s to 53.43°/s, df = 98, effect size = 0.759); for participant B (W0 vs. W7: p-value < 0.0005, CI = 4.706°/s to 27.96°/s, df = 88; effect size = 0.616; W0 vs. W10, p-value = 0.0010, CI = 8.493°/s to 30.96°/s, df = 68, effect size = 0.887). For participant C, the Kolmogorov-Smirnoff test showed significant differences in the distributions from Week 0 to Week 7, with a smaller median angular velocity on Week 7. For participant D, the Week 0 distribution did not pass the Shapiro-Wilk normality tests. Therefore, the two-sample Kolmogorov -Smirnoff test was used to compare the distributions from Week 0 to Week 7 and Week 0 to Week 10. The two-sample Kolmogorov-Smirnoff test showed significant differences in the distributions from Week 0 to Week 7 (p-value = 0.0090; n0 = 60, n7 = 60) and Week 0 to Week 10 (p-value < 0.001; n0 = 60, n10 = 60), with progressively increasing median DSRT angular velocities from Week 0 to Weeks 7 and 10 (43).
In summary, all the distributions changed significantly. Participant D demonstrated an increasingly larger DRST angle of catch after participation, and the rest of the participants showed smaller angles of catch.
(INSERT FIGURE 2 HERE; Fig. 2 Caption below)
Figure 2. Box plots the coefficients of variation of the DSRT angles of catch (left column) and DSRT angular velocity responses (right column) for each participant. The box plots show the descriptive statistics values for the coefficient of variation of the DSRT angle and DSRT angular velocity for each participant across all trials on Weeks 0, 7, and 10, showing the minimum value, first quartile, median, third quartile, and maximum values measured. The normality and equal variance test results are presented in Supplementary Table S3. The corresponding frequency distribution plots are included in Supplementary Figures S2, S3, S4, and S5. The asterisks in the plots indicate significant differences in the corresponding statistical tests shown in Table 2. The asterisks in the graphs correspond to the following levels of significance: *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, ****p-value < 0.0001. For specific p-values, see Table 2.
Table 2
One-tailed student’s t-tests and Welch’s t-test were performed on the DSRT angular velocity data for participants A and B. The Kolmogorov-Smirnoff test was performed on the DSRT angular velocity data of participants C and D as a consequence of the results of the Shapiro-Wilk normality tests. The p-values are presented with the sample size for the corresponding week comparisons.
One-tailed Student’s t-tests on DSRT coefficient of variation |
Participant | W0 vs. W7 p-values | Effect size Cohen’s d | W0 vs. W10 p-values | Effect size Cohen’s d |
A | 0.0112* (n0 = 14, n7 = 13) | 0.947 | 0.0127* (n0 = 14, n10 = 8) | 1.142 |
B | 0.0558 (n0 = 8, n7 = 10) | - | 0.0469* (n0 = 8, n10 = 5) | 1.137 |
C | 0.3524 (n0 = 16, n7 = 13) | - | N/A | - |
D | 0.2961 (n0 = 10, n7 = 13) | - | 0.1272 (n0 = 10, n10 = 14) | - |
One-tailed tests on DSRT angular velocity |
| One-tailed Student’s t-tests on DSRT angular velocity W0 vs. W7 | One-tailed Welch’s t-test on DSRT angular velocity W0 vs. W10 |
Participant | p-value | Effect size Cohen’s d | p-value | Effect size Cohen’s d |
A | < 0.0001* (n0 = 60, n7 = 58) | 0.796 | < 0.0001* (n0 = 60, n10 = 40) | 0.759 |
B | 0.0005* (n0 = 30, n7 = 60) | 0.616 | 0.0010* (n0 = 30, n10 = 40) | 0.887 |
| Kolmogorov-Smirnov test on DSRT angular velocity |
| W0 vs. W7 | W0 vs. W10 |
C | < 0.0001* (n0 = 60, n7 = 55) | N/A |
D | 0.0090* (n0 = 60, n7 = 60) | < 0.0001* (n0 = 60, n7 = 60) |
3. 2 Clinical Measures
The QUEST, PBS, MTS, and DIS results are presented in Supplementary Table S4. The changes in QUEST showed that participants A and B improved beyond the MCID, while the other two participants remained stable throughout. Only participant A gained an MCID in the PBS, while the scores for the rest of the participants were stable (see Supplementary Table S4). Participants A, B, and D showed a larger range of motion in one joint before the angle of catch of the MTS, while participant C showed improvement bilaterally at the elbow joints. Participant D demonstrated an increased range of motion only at the elbow of the least affected side. Improvement in DIS larger than the SDD was observed in three participants, with ameliorated dystonic scores in participants A, B, and D reflecting dystonic movements of shorter duration and smaller amplitude during activity. Participant A also showed ameliorated dystonic symptoms during rest beyond the SDD. For the choreoathetosis subscale, an SDD amelioration was observed in participants A and B. Radar plots of the sub-scores of SCALE for Weeks 0, 7, and 10 are presented in the Supplementary Figure S6. These graphs demonstrate improvement in selective motor control in three participants (B, C, D). Two participants (B and D) demonstrated unilateral improvement of selective motor control, and one showed bilateral enhancement, with the most improvement observed at the hip and subtalar joint.