Muscle coordination in archers: Zhang 2021 study

Twenty-eight Chinese archers at three levels, eight from the national team, twelve from provincial squads and eight from city clubs, each shot twelve arrows at 30 metres with thirteen surface electrodes on the trunk and arms. The analysis looks for muscle synergies: the few activation patterns which, combined, reconstruct the activity of all the recorded muscles.

The main result is that the number of patterns does not change with level: three in all three groups, with 2.9 plus or minus 0.4 in the elite archers, 2.6 plus or minus 0.6 in the intermediates and 2.9 plus or minus 0.4 in the beginners, and p equal to 0.33. The first pattern corresponds to the final moment of the draw and involves the arm and shoulder muscles on the bow side plus the middle trapezius; the second to aiming, with the elbow extensors of the bow arm and the shoulder extensors of the drawing arm, which are the muscles that hold the anchor; the third to the period after release, with the biceps and the finger extensor of the drawing arm.

What changes is when the third one starts. Its centre of activity falls at 91.4% of the cycle plus or minus 3.7 in the elite archers, at 87.9% plus or minus 5.7 in the intermediates and at 71.9% plus or minus 29.5 in the beginners, with p equal to 0.03 against the elite group and p equal to 0.00 against the intermediates. Starting earlier, the release pattern in the beginners overlaps more often with the aiming one, and the average frequency of overlaps is higher, peaking midway through the phase that follows the release. The width of the patterns, by contrast, does not differ between groups.

There are then two differences on individual muscles. In the beginners the middle deltoid contributes less to the aiming pattern than in the intermediates (p equal to 0.01) and the elite archers (p equal to 0.04). In the elite archers the flexor digitorum superficialis contributes less to the release pattern than in the intermediates (p equal to 0.01) and the beginners (p equal to 0.02).

The three groups were eight archers from the Chinese national team, twelve from a provincial squad and eight from clubs in Tianjin and Shijiazhuang. Muscle activity was recorded with a thirteen-channel wireless surface electromyography system (Trigno, Delsys) at 2,000 hertz. On the drawing side: middle trapezius, latissimus dorsi, middle deltoid, posterior deltoid, long head of triceps, biceps, extensor digitorum and flexor digitorum superficialis. On the bow side: latissimus dorsi, long head of triceps, lateral head of triceps, biceps and anterior deltoid. The skin was cleaned with alcohol and the hair removed.

Two reflective markers, one on the right elbow and one at the middle of the stabiliser, were tracked by seven Qualisys cameras at 200 hertz, with a reference video at 95 hertz, all synchronised by an analogue signal.

The recordings were made indoors, with each archer's own bow and arrows, on a target at 30 metres. After the usual warm-up and five consecutive shots with thirty seconds between them, each archer shot two recorded rounds of six arrows. The shot cycle was divided into three phases: the draw, from the bow's highest point to the moment the right elbow stops (velocity below 0.25 millimetres per second); the aim, from there to the release; and the phase after the release, fixed at half a second for everyone.

The signal was band-pass filtered between 20 and 500 hertz, rectified and low-pass filtered at 20 hertz with a fourth-order Butterworth, then normalised to each muscle's maximum across all trials. Each phase was resampled to 100 points, which is 300 per cycle, and the twelve cycles were joined into a 13 by 3,600 matrix for each archer. From there the patterns were extracted by non-negative matrix factorisation, with a sampling adequacy index of 0.86 plus or minus 0.03. From the activation curves the centre of activity was computed by circular statistics, along with the full width at half maximum and the average frequency of overlaps, compared between groups with a one-way analysis of variance and the LSD test.

The paper gives no age, no sex, no years of practice and no draw weight for any of the three groups, so what separates a beginner from an intermediate here is the squad they belong to and nothing else. The groups are eight, twelve and eight people.

The result the paper rests on has a spread that needs stating: in the beginner group the centre of activity of the third pattern has a standard deviation of 29.5 against a mean of 71.9, which means that group did not do one thing, it did several, and that the mean describes nobody in particular. Beside that comparison the table reports a statistical power of 0.21, an unusual figure to find next to a p below 0.01.

The authors list their measurement limits: the rotator cuff muscles cannot be reached with surface electrodes and were not recorded, the antagonists of the drawing shoulder were not recorded, and the surface signal over the latissimus dorsi is not reliable. The phase after release was defined as a fixed half second for everyone, rather than by anything the archers did.

Above all, no score was measured and no arrow's impact point was recorded. So none of these differences is connected to accuracy: level stands in for performance, and the stand-in here is membership of a squad. The authors' closing hypothesis, that elite archers might show more patterns under harder conditions such as wind, is stated as a hypothesis and was not tested.

What this study adds is a fine distinction. Between a national-team archer and a club archer, the difference is not in how many activation patterns the body uses, since there are three in all three groups, but in when the last one starts. In the beginners the pattern associated with the release begins earlier, while the aiming pattern is still running, and the two overlap more often.

Said without the laboratory vocabulary: in the less experienced archers the arm starts doing what it does after the release while the arrow is still on the string. It is the first time in this registry that an instrumented measure says something about the end of the shot rather than about the draw or the aim.

Nothing operational follows from it, though. Nobody shot for score, so the study cannot say that an early release costs rings, and it did not compare two ways of finishing the shot: it compared people of different levels. The recommendation the authors draw, to work on the timing of the follow-through, is their own inference from the data and they write it as such, leaning on what coaches say about the follow-through needing to flow rather than stop abruptly.

The second useful observation concerns the hand. In the national-team archers the finger flexor contributed less to the release pattern than in the other two groups, which runs in the same direction as an older finding on the forearms: the best archers let the string go by relaxing, not by opening.