Keynote Lectures

Electrical Muscle Stimulation: What It Is, How It Works and How It Can Help the Sport Scientist
Nicola Maffiuletti, Schulthess Clinic, Switzerland, Switzerland

Shockwave Therapy in Sports Medicine
Amir Pakravan, European College of Sport and Exercise Physicians, United Kingdom, United Kingdom

Machine Learning Approaches Supporting and Not Supporting Sports Practice
Martin Lames, Technical University of Munich, Germany, Germany

Abstract
Sport scientists are more and more interested in the evaluation and improvement of neuromuscular function, likely because of its repercussions on sport-related performance. As such, technology-based tools such as transcutaneous electrical muscle stimulation are of potential interest for the sport science community. On one hand, this technique could help optimize neuromuscular function of various sportsmen, either as a post-exercise recovery strategy (to minimize the effects of fatigue) or as a training/rehabilitation modality (to improve strength/power). On the other hand, electrical muscle stimulation is increasingly used – in combination with portable dynamometry – to investigate neuromuscular (central and peripheral) fatigue induced by exercise/competition on the field, with implications for training plans, recovery interventions and injury prevention. This keynote presentation will focus on the methodological (“what it is”) and physiological (“how it works”) specificities of transcutaneous electrical muscle stimulation, as well as on the main evidence-based applications of this technological tool in sport science settings (“how it can help”).

Abstract
Extracorporeal Shockwave Therapy (ESWT) is a non-invasive treatment option used in the management of a variety of sports and musculoskeletal medicine presentations primarily to promote healing and improve pain, function and performance.
The modality uses high energy pulses of sound wave produced by any of the various methods, depending on the type of the machine, to trigger a cellular level biological response in the target tissue through a mechanism commonly known as mechanotransduction.
A few of the more common uses of ESWT are tendinopathies, plantar fasciitis and calcifications. However, there is growing evidence for using the modality in bone pathologies such as bony stress response, and potentially osteonecrosis.
Depending on the type of presentation and the target tissue, different protocols have been devised which may vary in duration as well as the number, frequency and force of the sound wave pulses. Despite an excellent safety record, there are certain precautions and contra-indications to using Shockwave Therapy which clinicians need to be aware of. 
ESWT is an example of successful collaboration between science, technology and clinical practice. The future direction of research in this field should be towards finding more specific and potentially personalised approach to treatment.

Abstract
In recent years, performance analysis (PA) has evolved more and more into a big data science. At least for the most prestigious competitions, such as leagues and championships in team sports with a high degree of professionalization, for each match a huge amount of information is recorded containing video images, action feeds and position data. Machine Learning (ML) is a bundle of different methods each designed for and capable of conducting analyses on large data sets.

As a consequence, ML may be considered as the king’s road for sports practice to unfold a greater impact and effectiveness. Among the dramatically increasing number of ML approaches in PA one may distinguish three different classes of machine learning applications that will be introduced with examples of representative studies each. First, we use acknowledged ML methods to solve standard ML problems as a tool embedded in PA applications. The most prominent example is maybe video-based position detection relying extensively on pattern recognition tools employing ML. Second, sports data is used in computer science as a show case for basic research in ML. For example, there are many ML papers on improving prediction methods making use of sports data, sometimes making explicitly use of the – for this purpose – very much desired property of sports matches of being essentially not predictable. Finally, we have ML applications that intend to give support to PA either by facilitating routine tasks or by creating innovative analyses that were not in reach before. An example for a routine task is automated game annotation, whereas new options arise from creating new and meaningful performance indicators.

Obviously, among these several applications there are also non-supportive ones to sports practice. These are sometimes not as easy to detect like, for example, the astrophysicist happening to be a football nerd and applying his fancy tools to football or the tactics coach employing a computer scientist as life belt against drowning in the data ocean.

Starting from the “real” needs of sports practice leads to a distinction between types of ML applications that are in general not supportive for sports practice and those with potential support. Among the latter ones it is distinguished between promising applications that are under way and ready-to-use ones.