Brain and muscles in archery: Zhang 2025 study

In 26 first-year university students taking an introductory course, the study recorded the brain's electrical activity (EEG) together with that of six muscles during shooting, before and after seven weeks of lessons, and estimated how much information passes from brain signals to muscle signals using an index called transfer entropy. The flow was stronger in the alpha and beta frequency bands than in gamma, with large differences (Cohen's d 0.91 between alpha and beta, 3.26 between alpha and gamma and 2.74 between beta and gamma) and no change between before and after the course.

After the course the flow from brain to muscles was lower over the frontal and central areas, while the flow in the opposite direction did not change. The link with shooting results changed sign: before the course the correlations between the index and the score were all positive (from 0.0929 to 0.2729 depending on the muscle), and afterwards all negative (from -0.0960 to -0.3649). In both tests the index fell as more arrows were shot, less so after the course. The authors read these changes as a better allocation of the brain's resources and as a sign that the movement is becoming automatic.

Recordings were made on the university's outdoor range with standard recurve bows and targets at 10 metres, in 7 women and 19 men with a mean age of 18.7, all right-handed. Each test consisted of 30 arrows, in three sets of ten with rest in between. The course involved one lesson a week of at least 100 minutes with a qualified instructor, covering posture, drawing, aiming and release, with no fixed progression.

The EEG had 32 channels; surface EMG recorded the deltoid and erector spinae on both sides and the superficial finger flexor and common finger extensor of the right arm. For each shot the authors took a window of 1.5 seconds either side of the peak in finger flexor activity, which they treat as the moment of release, and calculated transfer entropy with a 25 ms lag between the two signals, chosen because it gave the highest value. Frequency bands were tailored to each person on the basis of their resting alpha rhythm. Shots were sorted into three levels: missing the target, 6 to 8 and 9 to 10. Before and after were compared with paired t-tests, and shot levels with independent t-tests.

There is no control group, so the changes between the two tests may come from the course but also from getting used to being measured with a cap and electrodes on. The paper does not report whether scores had improved after seven weeks, and the three result levels leave out arrows scoring 1 to 5. In the correlation table, significance does not follow the size of the coefficients (0.0929 is marked significant, 0.2729 is not), and the text calls it Table 1. Before-and-after effects are of similar size over the parietal areas too, where the differences are reported as not significant, so the claim that the change mainly concerns the frontal and motor areas is fragile. With 30 arrows at 10 metres, talking about fatigue, as the abstract does, is an interpretation: fatigue was not measured.

The authors themselves point out that it is unclear whether transfer entropy measures the strength of the coupling between brain and muscles. Reading a lower flow as optimisation or as a sign of automatisation is an interpretive choice, in line with theories of motor learning but not demonstrated by these data, which are not publicly available.

For an archer the study offers nothing practical: it measures brain and muscle signals in beginners at 10 metres, and the paper does not say whether scores improved in the meantime. It does describe something consistent with the experience of learning: at first the movement seems to call for more top-down control, and after a few weeks of practice that control eases. This is the authors' interpretation of a debated index, not a measurement of automatisation.

It should also be kept apart from the question of target panic: a lower flow from brain to muscles in beginners says nothing about the de-automatisation hypothesis put forward by Matteo Campagna, which concerns experienced archers in whom the automatic movement breaks down. In the registry the study counts as an example of how the interaction between brain and muscles can be measured in shooting, with the limitations described above.