Arm tremor and archery training: Shinohara 2023 study

Twenty male university students with no archery experience were split by lottery into two groups: one did a side-bridge exercise with the trunk inclined at 40 degrees for four weeks, the other did nothing. The choice of 40 degrees comes from the first part of the study, in which eleven participants held the side bridge at 25, 40, 55 and 70 degrees for fifteen seconds: arm tremor measured at the elbow was 0.00022 ± 0.00018 at 25 degrees, 0.00048 ± 0.00047 at 40, 0.00215 ± 0.00182 at 55 and 0.00161 ± 0.00167 at 70, with 25 and 40 degrees indistinguishable from each other and different from the other two. Forty degrees is therefore the greatest inclination that produces no tremor.

After four weeks, in the group that had trained, bow tremor during aiming fell from 8.68 ± 4.90 to 6.83 ± 4.84 (p = 0.047, d = 0.72) and the median power frequency of the middle deltoid's EMG fell from 76.27 ± 13.64 to 68.04 ± 9.92 hertz (p = 0.003, d = 1.30). The other three measures did not change: upper trapezius from 72.42 to 71.37 (p = 0.850), lower trapezius from 64.03 to 58.14 (p = 0.076), serratus anterior from 51.83 to 53.78 (p = 0.439). In the group that did not train nothing moved, tremor included, which went from 7.69 ± 5.24 to 8.19 ± 9.09 (p = 0.854).

The authors read the fall in median frequency as a greater share of slow muscle fibres in the middle deltoid, and conclude that tremor during the sighting phase can be reduced by a side bridge held at the inclination that does not provoke it.

The participants are twenty male students aged 20.8 ± 0.3 years, between 20 and 23, all right-handed and all without archery experience. Anyone with a history of shoulder dislocation or surgery, cervical symptoms, documented injury to the shoulder complex or regular training habits was excluded; allocation to the two groups was by lottery, and at the start the groups did not differ on age, height, weight, body mass index, tremor or any of the four median frequencies.

The reference task is not a shot. All used a 9-kilogram bow, that is twenty pounds, with the bow shoulder abducted to 90 degrees and the drawing shoulder in 45 degrees of horizontal flexion; as soon as the thumb of the drawing hand touched the chin a signal was given and the aiming position was held for fifteen seconds. There are no arrows, no target and no score.

Tremor was measured with accelerometers taped to the bow handle and to the lateral epicondyle of the humerus, sampling at a thousand hertz. Of the fifteen seconds recorded, the middle 8.192 seconds were used and transformed with a fast Fourier transform; the tremor value is the summed power in the 3 to 18 hertz band, chosen because it covers the bands reported in the literature for several kinds of postural tremor. Surface EMG, with electrodes twenty millimetres apart, recorded the middle deltoid, upper trapezius, lower trapezius and serratus anterior on the bow side, and the median power frequency was computed for each.

The training was five one-minute sets of the side bridge at 40 degrees, with one minute of rest between sets, five days a week for four weeks: twenty sessions in all, about ten minutes a day, always in the same laboratory, with completion confirmed by email. Before-and-after comparisons were made with a paired t-test within each group, and the four inclinations compared with non-parametric methods.

The limitation the authors declare is also the most important: the participants were not archers, so there is no telling whether the exercise changes anything in the movement of someone who shoots. Add to that the fact that not a single arrow was shot: the task is an aiming position held for fifteen seconds with a twenty-pound bow, with no target and no score, so the study says nothing about the relation between tremor and accuracy.

The analysis does not compare the two groups with each other. The tests are run within each group, before and after, and the conclusion rests on a difference being significant in one group and not in the other, which is not a test of the difference between groups. With ten people per group, moreover, tremor only just passes the threshold (p = 0.047) and the lower trapezius only just misses it (p = 0.076): with numbers that small either could easily swap places.

The comparison group did nothing, not a sham exercise, and nobody was blinded. The authors themselves note that the exercise uses almost the same position as the final task, so part of the result may be simple familiarity with that position. The participants were also selected partly for having no regular training habits, and in untrained people four weeks of any exercise produces changes that are easy to come by.

The interpretation in terms of slow muscle fibres is an inference from the EMG median frequency, not a measurement of the fibres, and the authors acknowledge this by writing that the physiological changes remain unknown. In the conclusions the exercise is also proposed for beginners and for people with the yips, but nobody with that problem was studied, and the injury prevention mentioned in the abstract was not measured.

The study is not about shooting but about bow tremor during aiming, and it measures that in complete beginners holding a position for fifteen seconds. Within that boundary it says something precise: trunk inclination in the side bridge decides whether the exercise provokes tremor, and at 55 and 70 degrees tremor appears while at 25 and 40 it does not. After four weeks of training at the inclination that does not provoke it, bow tremor during aiming went down, while in those who did not train it stayed where it was.

For an archer the interesting part is the idea of dosing behind it: not the number of sets, but the fact that the load of an isometric exercise can be chosen by watching whether the arm shakes or not. It is the authors' criterion, and in this study it worked on people who do not shoot.

What cannot be taken from it is just as clear. There is no score, so there is no telling whether a steadier bow during aiming puts arrows closer together; there are no archers, so there is no telling whether someone with years of shooting behind them has any margin left on that tremor; and the comparison between groups, as the study sets it up, does not allow anyone to say how much of the effect comes from the exercise and how much from familiarity with the position. The authors mention the yips as a possible field of future application: that is a suggestion written in the conclusions, not a result of the study.