1. Introduction
The stretch-shortening cycle (SSC) is a natural action involving the stretch, or eccentric lengthening, of an active skeletal muscle immediately prior to contraction [
1]. The primary role of the SSC is to optimize mechanical loading of the muscle-tendon complex, which may lead to a metabolically efficient and forceful muscle contraction [
2]. Neurophysiological mechanisms responsible for SSC enhancement of muscle performance are debated [
2], yet may include storage and recapture of elastic potential energy [
2,
3,
4,
5,
6], skeletal muscle pre-activation [
3,
4,
5,
6], increased active state [
2], residual force enhancement [
3,
4,
5,
6], pre-synaptic facilitation of alpha motor neurons from supraspinal drive [
7], and involuntary spinal reflex pathways [
3,
4,
5]. Given the role of the SSC in optimizing muscle efficiency and force, it is essential to identify measures that effectively differentiate SSC utilization during functional movement tasks.
Reactive strength was first introduced by Warren Young [
8] as a measure of lower-extremity SSC utilization in jumping. Specifically, Young [
8] defined reactive strength as “The ability to utilize stretching of the muscle and then change quickly from an eccentric to a concentric contraction”. Jumping movements that involve high stretch-loads and ground contact times (GCTs) ≤ 250 ms are categorized as fast SSC, with jumping movements involving GCTs ≥ 251 ms categorized as slow SSC [
9]. To measure slow SSC utilization, Young [
8] proposed taking the difference between jump heights (JHs) achieved using the countermovement (CMJ) and squat jumping (SJ) techniques (
Table 1). To measure fast SSC utilization, Young [
8] established a metric known presently as the Reactive Strength Index (RSI).
The RSI is computed from jumping movements comprising a distinct ground contact, or impact, phase by taking a ratio of JH to GCT (
Table 1). The RSI was introduced as a measure of depth jump (DJ) performance. DJ was introduced as the “shock method” of training by Yuri Verkhoshanksy [
10] and has persisted as a jumping movement commonly included in plyometric training programs targeting SSC enhancement [
11]. The DJ technique involves a maximal vertical jump performed immediately following landing impact from a self-initiated drop. DJ is considered a fast SSC movement, however, it is common for GCTs to exceed the 250 ms threshold when JH is emphasized through verbal instruction or when DJs are performed from a high drop [
12]. For example, Struzik et al. [
13] observed significantly longer
GCTs when DJs were performed with verbal instruction to maximize JH (GCT = 0.33–0.43 s) versus instruction to both maximize JH and minimize GCT (GCT = 0.23–0.28 s). Further, Addie et al. [
14] observed significantly longer GCTs in a sample of active young adults when DJs were performed from drop heights of 0.76 and 0.91 m versus drop heights ranging between 0.30 and 0.60 m.
The RSI was recently modified by Ebben and Petushek [
15] for application as an assessment of slow SSC utilization in CMJ. The RSI-modified (RSI-mod) is computed by replacing GCT with time to take-off (TTT;
Table 1), a temporal variable representing the duration between countermovement initiation and CMJ take-off.
The RSI and RSI-mod share similar nomenclature and both fit within Young’s [
8] original definition of reactive strength, yet their discrimination of lower-extremity reactive strength, or SSC utilization, may depend on jumping technique [
16,
17,
18]. For example, DJ and maximal repetitive jumping are the most common jumping techniques used to estimate the RSI. The 10/5 repeated jump test (10/5 RJT) is a maximal repetitive jumping technique wherein 10 bilateral jumps are performed immediately after an initial CMJ [
18]. Although DJ and the 10/5 RJT both comprise a distinct ground contact phase and involve maximal jumps for height subsequent to a landing impact, Stratford et al. [
17] observed only a moderate linear association (
R2 = 0.30) between RSI scores derived from DJ and the 10/5 RJT. While there is likely a multitude of factors that distinguish the 10/5 RJT from DJ, Stratford et al. [
17] comment on conceivable differences in the feedforward neuromotor control of landing impact, noting that the DJ technique allows for a motor response to be planned in advance of the self-initiated drop.
In contrast with CMJ, DJ tends to elicit shorter GCTs [
16], greater ground reaction force (GRF) magnitudes and greater rates of GRF development (RFD; [
19]). Additionally, DJ does not involve a prolonged unweighting phase and requires complex feedforward and feedback neuromotor control of a drop and landing phase whereas CMJ is performed entirely with the feet in contact with the ground [
20,
21,
22,
23]. Notably, the differences between DJ and CMJ provide a basis for questioning whether the RSI and RSI-mod are compatible measures of reactive strength. Recently, McMahon et al. [
16] observed a moderate linear association (
R2 = 0.22) between RSI-mod (CMJ) and RSI (0.30 m DJ) scores in a sample of professional male rugby athletes. The 22% shared variance reported by McMahon [
16] is evidence that the RSI and RSI-mod are somewhat distinct, yet it is also important to mention that results from linear regression do not necessarily reflect the extent of agreement between measures [
24,
25,
26].
To address the limitations of linear regression, Bland and Altman [
24] proposed an alternative analysis that provides meaningful discernment through informal interpretation of limits of agreement (95% CI) and measurement bias [
25,
26]. A comprehensive analysis of the association and agreement between RSI and RSI-mod scores is timely considering the recent and increasing interest of both measures in the literature (
Figure 1). Further, with the RSI and RSI-mod having been applied across diverse populations, a representative analysis of measure compatibility may be comprised of a mixed-sex sample of participants with varied reactive strength ability as opposed to a homogenous sample. Therefore, the purpose of the present investigation was to evaluate the association and agreement between RSI and RSI-mod scores acquired from a mixed-sex sample of National Collegiate Athletic Association (NCAA) division I basketball athletes and active young adults. NCAA athletes and recreationally active young adults are among the most common populations studied in the RSI and RSI-mod literature. The decision to include a mixed-population sample was made to provide representation of the literature and to strengthen the analysis of association and agreement of RSI and RSI-mod scores by means of measuring a broad range of jumping ability. We hypothesized that there would be a significant linear association between RSI and RSI-mod scores, yet the totality of evidence would not support application of the RSI and RSI-mod as interchangeable measures of reactive strength.
4. Discussion
Our hypothesis that the RSI and RSI-mod are associative but not interchangeable measures of reactive strength was supported by the results. RSI and RSI-mod scores in the present investigation were similar to values reported previously in the literature [
16,
38,
39] and, across all DJ conditions, RSI scores were substantially greater (+131–150%) than RSI-mod scores, a finding that was mostly attributable to longer TTT versus GCT (+115–120%).
DJ GCTs were significantly shorter than CMJ TTTs, yet they were also above the 250 ms threshold traditionally associated with a fast SSC action [
9]. For DJs, participants were instructed to maximize JH and minimize GCT. Using comparable verbal instruction, Struzik [
13] observed DJ GCTs below the 250 ms threshold for fast SSC action, while several authors have observed DJ GCTs that were similar to those reported in the present investigation [
16,
40,
41,
42]. To encourage a fast SSC action in DJ, it may be necessary to provide augmented feedback during familiarization or to emphasize jumping “as quickly as possible” without reference to jump height, which is observed to facilitate both shorter DJ GCTs [
43] and greater RSI-mod scores [
44]. In addition, we did not require participants to report a history of plyometric training. Reduced DJ GCTs [
45] are noted as a potential adaptation to plyometric training, thus limited prior exposure to performing the DJ may have contributed to the GCTs observed in the present investigation. Lastly, performing DJs from a drop height that exceeds an individuals’ reactive capacity is observed to result in prolonged GCTs [
14]. GCTs were not different between DJ conditions in the present investigation, which suggests that the drop heights were within the reactive strength capacity of participants.
Linear regression revealed significant positive associations between RSI and RSI-mod scores, however, the amount of shared variance (20–47%) returned from the models was moderate. This is largely consistent with the findings from McMahon et al. [
16] and suggests that the RSI and RSI-mod likely do not measure the same reactive strength characteristics. From a collective view, the shared variance between RSI-mod and RSI scores in the present investigation had a tendency to exceed the 22% reported by McMahon et al. [
16]. The present investigation comprised a larger sample of participants (
n = 47 vs. 21; [
16]) that were heterogeneous with respect to sex and athletic status. The heterogeneity of our sample likely resulted in a greater range of RSI and RSI-mod scores. Further, since sample size and predictor value range are two factors that can augment the shared variance returned from linear regression [
46], the heterogeneity of our sample may have contributed to the differences in association observed between the present investigation and McMahon et al. [
16].
The Bland–Altman analyses provide evidence of a poor agreement between RSI and RSI-mod scores. The Bland–Altman plots were consistent when RSI-mod scores were compared against RSI scores derived from 0.51, 0.66 and 0.81 m DJ. The measurement bias between RSI and RSI-mod scores (0.50–0.57) was greater than the mean for RSI-mod scores (0.42) and was statistically significant since the 95% CIs did not cross zero (95% CI = 0.40–0.68). In Bland–Altman analysis, two measures may be significantly biased yet retain strong agreement if the limits of agreement between measures is small. This was not the case in the present investigation as the ranges between upper and lower limits of agreement (1.27–1.51) were large and considerably greater than the mean values for both RSI (0.97–1.05) and RSI-mod (0.42) scores. For all Bland–Altman plots, linear regression revealed a significant performance-dependent effect on measurement bias, wherein the difference between RSI and RSI-mod scores was positively associated with the mean of RSI and RSI-mod scores. These effects suggest that there may be limited performance transfer between scores, whereby an increase in the RSI does not necessarily result in a similar increase in the RSI-mod. Further, these effects suggest that the agreement between RSI and RSI-mod scores may not be consistent when applied across populations with varied reactive strength ability.
As mentioned previously, DJ tends to elicit shorter GCT/TTT in conjunction with greater peak GRF and RFD when compared against
CMJ (
Figure 6). The DJ technique is also performed without a prolonged unweighting phase [
16,
26] and, as shown in
Figure 6, the total duration of a fast-SSC (<250 ms) ground contact in DJ can be shorter than the duration of the CMJ unweighting phase. Differences in the GRF profiles of DJ and CMJ also infer that a greater biomechanical demand is placed on the neuromuscular system during DJ. Consequently, the RSI-mod may be the more appropriate measure of reactive strength in populations that may not have the requisite strength needed to safely and skillfully accept the high stretch-loads that are applied to the muscle-tendon complex in DJ. The RSI-mod could be used to provide partial insight into reactive strength until sufficient tolerance to the biomechanical demands of DJ is realized, at which point the RSI may then be the preferred measure.
From a neuromotor control perspective, the self-initiated drop and landing impact phase of DJ may represent the most important distinction from CMJ. In DJ, the neuromotor system must develop a planned motor response to landing impact in anticipation of the timing and magnitude of GRF [
22]. This feedforward control strategy is executed during the drop phase, resulting in the pre-activation of lower extremity skeletal muscle prior to landing impact [
22]. Pre-activation increases the stiffness of skeletal muscle and, when impact forces are predicted correctly, facilitates a safe dispersion of stress through the muscle-tendon complex upon landing [
22]. Pre-activation may also enhance the SSC response to landing impact by preparing the muscle-tendon complex to store elastic energy and by modifying the short latency spinal reflex via input from supraspinal drive and alpha-gamma co-activation [
7,
22,
47,
48].
Several authors have acknowledged the pre-activation of skeletal muscle as a fundamental component of natural, or functional, lower-extremity SSC actions [
48,
49,
50]. For example, to evaluate functional lower-extremity SSC utilization, Nicol et al. [
49] recommend the performance of jumping techniques, such as DJ, that involve a rapid stretch of pre-activated skeletal muscle. From this perspective, DJ-derived RSI scores may be more representative of natural sport movements (e.g., running, sprinting, cutting and jumping) that evoke robust lower-extremity SSC actions in response to impact between the feet and the ground. In contrast, RSI-mod scores are derived from a controlled CMJ technique performed with the feet remaining in contact with the ground for the entire duration of the unweighting and vertical jump phases. RSI-mod scores are considered a valid and reliable assessment of slow lower-extremity SSC utilization and neuromuscular power [
51,
52,
53,
54], yet their application as a measure of functional lower-extremity reactive strength, or SSC utilization, may need to be reconsidered.
An informal search of the literature (
Figure 1) reveals a recent and increasing interest in the RSI and RSI-mod. Using the search term “reactive strength index”, Google Scholar (
http://scholar.google.com (accessed on 15 March 2020) returned 252 and 42 peer-reviewed manuscripts from 2000 to 2020 that feature the RSI or RSI-mod as a dependent measure, respectively (
Figure 1). Notably, among the 42 manuscripts that included the RSI-mod as a dependent measure, 8 (19%) incorrectly refer to the RSI-mod as the RSI. The results from the present investigation and McMahon et al. [
16] are evidence that the RSI and RSI-mod cannot be used interchangeably, yet, with the inconsistent application of the terms in the literature, it may be of practical value to consider further distinction between the measures. One example would be to revise the nomenclature of the RSI-mod to the Explosive Strength Index (ESI), which may better reflect the biomechanical and neuromotor demands of CMJ.
It is important to mention several limitations of the present investigation. First, we estimated the JH component of the RSI using a take-off velocity method. This approach required an estimation of DJ landing impact velocity taken from digitized 2-dimensional video. Video data were digitized in accordance with the de Leva [
29] anthropometric model which, when compared against criterion methods, yields valid estimates of whole-body CoM displacement [
55]. Estimating DJ take-off velocity through a combination of videography and force platform dynamometry is supported in the literature and may address the known threats to validity associated with estimating JH from flight time [
32,
33]. Regardless, there are several potential sources of measurement error that arise from the capture and digitization of 2-dimensional video, thus a recommendation for future investigation is to consider estimating DJ landing impact velocity using criterion methods, such as optical motion capture. Second, we instructed participants to jump as high and as quickly as possible. Our approach to verbal instruction likely contributed to DJ GCTs that exceeded the threshold for a fast SSC action. As such, the results of the present investigation should be considered in context with the specific verbal instructions provided to participants and DJ GCTs. Lastly, it is important to note that RSI scores were not different across DJs performed from varying drop heights. While this has been observed in prior literature [
27], it also brings into question the sensitivity of the RSI as a measure of reactive strength. For instance, achieving similar RSI scores in DJs performed from a low versus high drop may not account for differences in the amount of mechanical energy absorbed during landing impact. While the RSI is reliable and valid as a broad metric of DJ performance, there may be value in focusing future research on a more specific metric of reactive strength, which may include a direct analysis of the rate of absorption and production of mechanical energy.