As participation in wheelchair sports increases, the need of quantitative assessment of biomechanical performance indicators and of sports- and population-specific training protocols has become central. The present study focuses on junior wheelchair basketball and aims at (i) proposing a method to identify biomechanical performance indicators of wheelchair propulsion using an instrumented in-field test and (ii) developing a training program specific for the considered population and assessing its efficacy using the proposed method. Twelve athletes (10 M, 2 F, age = 17.1 ± 2.7 years, years of practice = 4.5 ± 1.8) equipped with wheelchair- and wrist-mounted inertial sensors performed a 20-metre sprint test. Biomechanical parameters related to propulsion timing, progression force, and coordination were estimated from the measured accelerations and used in a regression model where the time to complete the test was set as dependent variable. Force- and coordination-related parameters accounted for 80% of the dependent variable variance. Based on these results, a training program was designed and administered for three months to six of the athletes (the others acting as control group). The biomechanical indicators proved to be effective in providing additional information about the wheelchair propulsion technique with respect to the final test outcome and demonstrated the efficacy of the developed program.
The benefits of practicing sports and physical activities for individuals with disabilities are generally accepted and concern psychosocial health and functional ability as well as general quality of life [
As a consequence of the increase in participation, competitiveness has risen and the development of training protocols which are specific to the discipline and to the population has become more and more important. In this respect, quantitative biomechanical assessment of performance- and injury-related parameters provides information related not only to the overall outcome (product) of the analysed motor task, but also to the way this task is performed (process).
Extensive literature exists which analyzes biomechanical aspects of wheelchair basketball. In particular, the following areas have been widely investigated: wheelchair propulsion technique [
To this aim, wearable inertial measurement units (IMUs) have been proposed as a valid alternative to traditional laboratory-based instruments, allowing for in-the-field performance monitoring without neither constraining athletes’ movements nor significantly modifying the original wheelchair configuration. Few studies focused on the use of IMUs to assess biomechanical parameters in wheelchair sports, such as speed [
Concerning training methodologies, only few studies focused on the development and assessment of training programs specific for wheelchair users. One study [
In this framework, the purpose of the present study was twofold: (i) to propose a method based on inertial sensing technologies aimed at obtaining a set of biomechanical parameters able to provide performance-related information about in-the-field wheelchair propulsion and (ii) to develop a discipline- and population-specific training program for junior wheelchair basketball athletes and to assess its efficacy using the proposed biomechanical parameters.
Twelve junior wheelchair basketball players (Table
Participant characteristics.
Subject | Age [years] | Mass [kg] | Stature [m] |
|
Years of practice | Gender | Group |
---|---|---|---|---|---|---|---|
S1 | 20 | 66 | 1.70 | 2.5 | 7 | Male | CG |
S2 | 18 | 48 | 1.68 | 3 | 6 | Male | CG |
S3 | 15 | 42 | 1.66 | 2 | 4 | Female | CG |
S4 | 17 | 61 | 1.75 | 0.5 | 3 | Male | CG |
S5 | 12 | 46 | 1.45 | 2 | 2 | Male | CG |
S6 | 17 | 96 | 1.78 | 4.5 | 3 | Male | CG |
S7 | 20 | 48 | 1.55 | 2 | 7 | Male | EG |
S8 | 19 | 49 | 1.69 | 3 | 7 | Male | EG |
S9 | 16 | 42 | 1.66 | 2.5 | 5 | Female | EG |
S10 | 18 | 74 | 1.75 | 0.5 | 4 | Male | EG |
S11 | 13 | 45 | 1.45 | 2 | 3 | Male | EG |
S12 | 20 | 105 | 1.90 | 4.5 | 3 | Male | EG |
Descriptive statistics | 17.1 ± 2.7 | 60.2 ± 21.4 | 1.7 ± 0.1 | 2.25 ± 1 | 4.5 ± 1.8 |
Characteristics of all participants: descriptive statistics are reported in terms of mean ± standard deviation (SD), except for
To assess the efficacy of the proposed training program, participants were randomly divided into an experimental (EG) (
Three experimental sessions were carried out, during which an instrumented 20-metre sprint test was performed. The first session (ES1) was performed in October, at the beginning of the training season, and was used to fulfil the first aim of the study (i.e., the biomechanical characterisation). The results obtained at this stage were used as a guidance to design a discipline- and population-specific training program (see Section
The three experimental sessions, as well as the specific and the standard training program, were performed on the basketball court of the Santa Lucia Foundation and each athlete used the wheelchair commonly adopted during both training practice and competition. The same experimental protocol was followed for all the three sessions, as described in the following section.
During each experimental session, each athlete was equipped with three wearable IMUs (Opal, APDM Inc., Portland, Oregon, USA). These devices embed three-axial accelerometers (±6 g of full-range scale, 128 samples/s) providing the components of the vector sum of gravitational and inertial linear accelerations along the axes of a coordinate system fixed with the unit. Two IMUs were fixed on the right and left wrists using elastic bands, while the third unit was securely attached to the backrest of the wheelchair using double-sided tape (Figure
Wheelchair- and wrist-mounted IMUs with the relevant systems of reference.
Each athlete performed a 20-metre sprint test (20 mS). This test, focusing in particular on start-up and steady state velocities, was selected within those proposed for adult wheelchair basketball players [
In addition, before the beginning of the competitive season, each subject was medically examined and the peak power output (
To remove random noise, the measured accelerations were low-pass-filtered with a cut-off frequency of 12 Hz using a 4th-order zero-lag Butterworth filter [
The forward component of the acceleration measured by the wheelchair IMU was then used to identify the beginning of the steady state phase. To this aim, the signals were further low-pass-filtered with a cut-off frequency of 4 Hz and relative maxima were detected on the curve (Figure
(a) Forward component of the acceleration measured by the wheelchair IMU. Dots indicate relative maxima. The duration of a cycle is also indicated with a double arrow. (b) Trend of the push cycle duration over time. The beginning of the steady state phase is indicated.
Based on previous literature [
Based on the results of the above-mentioned instrumented 20 mS and on literature evidence related to the demands of wheelchair basketball athletes [
Strength exercises involved the following muscle groups: biceps, triceps, middle trapezius, and shoulder abductors and adductors [
Coordination exercises focused on the sport specific components of coordination skills and, in particular, on spatial orientation, kinaesthetic differentiation, reaction, adaptation, combination, and rhythm [
Following the guidelines proposed by Faigenbaum et al. [
The statistical analysis was performed using the IBM SPSS Statistics software (IBM Corp., Armonk, NY, USA). The alpha level of significance was set to 0.05 for all statistical tests.
For each subject and each trial, the normal distribution of the IMU-based estimated parameters was verified using the Shapiro-Wilk test of normality. As all parameters were not normally distributed, the Spearman (
For both ES2 and ES3, each subject, and each trial, the normal distribution of the IMU-based estimated parameters was verified using the Shapiro-Wilk test of normality. As all parameters were not normally distributed, a Wilcoxon signed-rank test was performed to investigate whether significant differences existed between ES2 and ES3 sessions, for both EG and CG. In addition, Mann-Whitney
Descriptive statistics of
Results of the biomechanical characterisation at ES1.
Parameter | Descriptive statistics | Correlation with | |
---|---|---|---|
|
|
||
|
6.4 ± 1.1 | ||
|
2.25 ± 1 | −0.675 |
|
|
1.46 ± 0.69 | −0.870 |
|
|
0.47 ± 0.06 | 0.757 |
|
|
113.93 ± 14.86 | −0.631 |
|
|
342.10 ± 240.37 | −0.242 |
|
|
48.77 ± 3.27 | −0.659 |
|
|
6.86 ± 6.67 | 0.548 |
|
|
11.09 ± 6.84 | 0.448 |
|
|
4.58 ± 3.16 | 0.520 |
|
Descriptive statistics and Spearman correlation coefficients of parameters obtained during ES1. Data are expressed as mean ± SD except for
The results of the multiple regression analysis are reported in Table
Results of the multiple regression model.
Parameter | Unstandardized coefficients |
|
|
|
||
---|---|---|---|---|---|---|
|
Std. error | Lower bound | Upper bound | |||
Constant | 7.344 | 0.225 | 32.593 |
|
6.896 | 7.792 |
|
−1.021 | 0.104 | −9.799 |
|
−1.229 | −0.814 |
|
0.100 | 0.020 | 4.998 |
|
0.060 | 0.140 |
|
0.031 | 0.009 | 3.634 |
|
0.014 | 0.048 |
|
−0.151 | 0.052 | −2.906 |
|
−0.254 | −0.048 |
Unstandardized
The mean and SD of
Results of the training program assessment (ES2 and ES3 results).
Parameter | Group | ES2 | ES3 | Wilcoxon test | |
---|---|---|---|---|---|
Mean ± SD | Mean ± SD |
|
| ||
|
CG | 6.1 ± 1.2 | 5.9 ± 1.1 | 0.248 | −1.156 |
EG | 6.1 ± 1.2 | 6.2 ± 1.1 | 0.080 | −1.753 | |
|
|||||
|
CG | 0.44 ± 0.05 | 0.45 ± 0.08 | 0.705 | −0.378 |
EG | 0.47 ± 0.09 | 0.44 ± 0.08 | 0.027 |
−2.207 | |
|
|||||
|
CG | 117.32 ± 14.62 | 118.30 ± 21.02 | 0.600 | −0.524 |
EG | 112.65 ± 18.71 | 119.97 ± 15.74 | 0.028 |
−2.201 | |
|
|||||
|
CG | 276.56 ± 148.16 | 289.57 ± 241.16 | 0.463 | −0.734 |
EG | 330.71 ± 191.10 | 588.32 ± 312.29 | 0.028 |
−2.201 | |
|
|||||
|
CG | 48.91 ± 3.82 | 47.86 ± 3.82 | 0.173 | −1.363 |
EG | 47.77 ± 3.56 | 48.62 ± 3.96 | 0.046 |
−1.992 | |
|
|||||
|
CG | 4.36 ± 3.64 | 6.65 ± 4.81 | 0.173 | −1.363 |
EG | 5.00 ± 3.38 | 4.12 ± 4.59 | 0.345 | −0.943 | |
|
|||||
|
CG | 12.00 ± 6.93 | 15.24 ± 10.47 | 0.249 | −1.153 |
EG | 12.16 ± 9.92 | 7.41 ± 3.12 | 0.249 | −1.153 | |
|
|||||
|
CG | 5.57 ± 4.38 | 4.56 ± 3.41 | 0.249 | −1.153 |
EG | 4.35 ± 2.74 | 4.17 ± 2.83 | 0.917 | −0.105 |
Descriptive statistics and Wilcoxon signed-rank test results (in terms of
This study presents a method for the quantitative characterisation of a wheelchair basketball field test (20-metre sprint test) based on the use of inertial measurement units (IMUs). A list of biomechanical parameters was defined as performance indicators, which allowed for the development of a sports- and population-specific training program. The method was applied to a team of junior wheelchair basketball athletes and was used to assess the efficacy of the developed training program which was administered for three months.
The proposed instrumented 20-metre sprint test (20 mS) allowed extracting a list of biomechanical indices associated with the performance of the test (
The results about
In terms of aerobic power, the present results are in agreement with previous findings about elite wheelchair athletes [
The results of the correlation analysis provide interesting insights into the importance of the biomechanical characterisation of the wheelchair propulsion during the 20 mS. All parameters display a statistically significant relationship with the performance index,
In particular, the peak power output (
Moderate positive correlations were obtained for all the computed coefficients of variation (
A significant negative relationship between
According to the results of the correlation analysis, the regression model identified four significant parameters able to predict
In the present study, an
On the other hand, when considering the IMU-based biomechanical parameters related either to strength or coordination, differences between ES2 and ES3 were obtained for the EG only. In particular, significant improvements were displayed for
These results indicate that although the administered training program does not influence the final outcome of the 20 mS, it affects the way this outcome is obtained by the athletes. This supports the hypothesis that biomechanical analysis can effectively provide additional performance indicators to the coaches relative to the way specific movements are performed, not limiting the analysis on the final product of the selected motor task.
The investigation included a relatively small sample size, particularly when considering the control and experimental groups separately (i.e., during ES2 and ES3). Also, the study involved participants with different pathologies and variable level of spinal cord lesions. It is plausible that these two aspects play an important role in determining the lack of significant differences between CG and EG after the training program administration. Still, significant differences for a list of biomechanical parameters were identified within each group when comparing the results obtained before and after the administration of the program. It is possible that between-group differences might have been pointed out if a larger cohort of athletes would have been admitted to the program.
Conversely, no statistical difference was detected when comparing the times to complete the 20 mS. This is probably related to the inadequate level of accuracy of the manual digital stopwatch, which could be improved by using photocells or a laser gun. Moreover, it is not excluded that differences in the test score could emerge from a program administration period greater than three months.
In the present study, the progression force was estimated by analysing the acceleration signals measured by the IMU located on the wheelchair.
This paper fills an existing gap in the field of junior wheelchair basketball. A methodology for the biomechanical assessment of the wheelchair propulsion was developed and a list of biomechanical indices associated with the performance of a 20-metre sprint test was obtained by means of wheelchair- and wrists-mounted inertial measurement units. These indices proved to correlate with the test performance and provide quantitative information about the way athletes obtained such performance. Therefore, they were used to define a sports- and population-specific training program focused on strength and coordination training. The proposed biomechanical methodology was then used to assess the efficacy of the defined training program after three months of administration. The estimated indices were effective in identifying both strength and coordination improvements following the training administration.
Both the biomechanical assessment method and the training program proved to be well perceived by the athletes and to be applicable in training conditions. Special attention, in fact, was paid to the organisational and practical aspects of the experimental protocol and of the program administration. It is worth underlining that the results of the present study were achieved thanks to the effective interaction within the multidisciplinary research group, which allowed addressing and answering the needs of both coaches and physiotherapists, through the complementary expertise of biomechanists, in terms that were valuable to the former professionals.
Inertial measurement unit
Control group
Experimental group
First, second, and third experimental sessions, respectively
20-metre sprint test
Functional classification score
Upper arms peak power output
Time to complete the 20-metre sprint test
Push cycle duration
Push cycle frequency
Peak progression force
Peak acceleration
Bilateral symmetry index
Coefficient of variation
Standard deviation
Interquartile range.
The authors declare no conflict of interests.
The authors would like to thank the athletes and their families for their voluntary participation in this study and the team coach, Djodji Damas Ntendarere, for his kind support. The contributions of Patrizio Farina to the training program administration and of Claudia Mazzà to the initial conception of the study are gratefully acknowledged.