Throwing is a fundamental human motor behavior; it was as important for the survival of our ancestors as walking (Roach et al., 2013). In modern times, throwing is regarded as one of the most important movements in sports such as baseball, handball, and javelin throw. Underscoring this point, many studies have focused on baseball throwing from the perspectives of performance improvement and injury prevention (DiGiovine et al., 1992, Hamilton et al., 1996, Jobe et al., 1983, Kotoshiba et al., 2021, Smidebush et al., 2019). The throwing motion can be divided into distinct temporal phases, and for each phase, the function of each individual upper limb muscle has been identified (DiGiovine et al., 1992, Jobe et al., 1983). Smidebush et al. (2019) compared muscle activity during fastball pitching and curveball pitching and reported that the former showed significantly greater activity in the upper and lower limb muscles than the latter. In a previous study investigating the relationship between shoulder joint instability and muscle activity, players with shoulder joint instability showed greater activity of the infraspinatus muscle during the acceleration phase than those without instability (Kotoshiba et al., 2021). Hamilton et al. (1996) showed that the activity of the wrist flexor muscles during throwing motion in baseball players with medial collateral ligament injury was significantly reduced compared with those without injury. As stated, investigating muscle activity during throwing motion is of great significance for understanding efficient throwing performance and the pathophysiology of throwing-related injuries, as well as for developing training and rehabilitation programs. While it is recognized that baseball throwing requires spatiotemporal coordination of many upper limb muscles, it remains unclear how such coordination is regulated by the central nervous system. Muscle synergy analysis is a technique for assessing such multimuscle coordination and has been successfully utilized to analyze muscle coordination in movement patterns such as walking, running, and reaching (Cappellini et al., 2006, d'Avella and Lacquaniti, 2013, Lacquaniti et al., 2012). Muscle synergy analysis decomposes the activation of a group of muscles into a muscle synergy vector, which refers to the relative weighting of each muscle within each synergy (motor module composition), and a synergy activation coefficient, which represents the temporal activation of each muscle synergy. For example, during walking, activated lower limb muscles can be grouped into four to five basic motor modules (Cappellini et al., 2006, Lacquaniti et al., 2012, Oliveira et al., 2014). Modular grouping has been shown to reduce the dimensionality of the muscle space and simplify the description of walking motion. Furthermore, muscle synergy analysis has revealed aberrant temporal patterns of muscle activation during walking among patients with neurological disorders, such as stroke and multiple sclerosis (Lencioni et al., 2016, Routson et al., 2013, Routson et al., 2014). Thus, muscle synergy analysis has proven to be a powerful tool for the assessment of the spatiotemporal coordination of muscle activation. However, the composition and temporal activation pattern of motor modules have not been determined for baseball throwing. The primary purpose of this study was to determine the composition and temporal activation pattern of motor modules for baseball throwing via muscle synergy analysis using the non-negative matrix factorization (NMF) algorithm.

The interindividual similarity of muscle synergies has been investigated. Although it is generally known that muscle synergies are common among individuals (d'Avella et al., 2006, Frere and Hug, 2012, Rimini et al., 2017), in some cases, there are individual-specific muscle synergies (Rimini et al., 2017, Torres-Oviedo and Ting, 2007, Yokoyama et al., 2016). However, it is unclear whether the spatiotemporal pattern of muscle synergy in throwing is specific to individuals or shared among them. Our second aim was to address this question using clustering techniques. We hypothesized that the proficient throwing motion acquired through years of practice would consist of shared muscle synergies among individuals. We believe that the findings related to the muscle synergies involved in throwing motion obtained in this study are potentially beneficial for training to improve throwing performance and for rehabilitation of throwing-related injuries.