Brain networks while aiming in archery: Gu 2022

The study compared the electrical activity of the brain in two groups of Chinese archers while they aimed at 70 metres: 14 from the national team, with experience of international competition (the "elite" group), and 14 from the Beijing team, who had shot national but not international events (the "expert" group). In the test the mean score per arrow was 8.7 in the first group and 6.8 in the second.

The authors looked at the three seconds before release, split into three one-second windows, and calculated how far the signals from the 30 electrodes were in phase with each other, using an index called WPLI. Differences between the groups appeared only in the last second and only in the two beta bands: there the national team archers showed stronger connections, mainly between frontal and central regions, and no connection was stronger in the other group. Treating these connections as a network, the national team had a higher mean clustering coefficient, global efficiency and local efficiency and a shorter characteristic path length, which in graph theory terms means a more integrated network.

The link with score, on the other hand, was clearer in the Beijing group, with more connections correlated with the mean arrow score, while few remained in the national team. The authors read this as performance that depends less on the state of the brain at that moment.

Thirty-one archers were recruited, 16 from the national team and 15 from the Beijing team, all right-handed, with a mean age of 23 and on average 8.1 and 6.1 years of training. Three were excluded because of artefacts in the signal, leaving the 14 per group who were analysed. Testing took place on the national team's outdoor range: each archer shot 35 arrows at their own pace, with their own bow, at a target 70 metres away, with an infrared sensor marking the moment of each shot and a spotter calling the score after every arrow. The target is described as a sheet 52 by 52 centimetres, with a 10 ring 10 centimetres across and scores from 6 to 10.

EEG was recorded with a portable 30-channel amplifier at 500 Hz, referenced to the mastoids. After cleaning, 856 trials remained, about 30 per archer. Frequency bands were set from each archer's individual alpha frequency, and the interval from 4 to 3 seconds before release served as the baseline. The authors first picked the window and bands with the largest difference between the groups, then examined connections, their change from baseline and network measures within those, using non-parametric tests, FDR correction and Spearman correlations with the mean of 30 arrows.

The groups are small and differ in many ways besides level: years of training, score, experience of international events and perhaps equipment. The study is cross-sectional, so it shows how two established groups differ and cannot say whether those EEG features help archers get better or come from being better; the title asks exactly that question, but nobody was followed over time.

The window and bands were chosen because they gave the largest difference between the groups, and all the other analyses were then run on them, a procedure that tends to inflate results. To compare two independent groups the text says it used the Wilcoxon signed-rank test, which is meant for paired data, and in the results it describes the scores as non-normal with a p above 0.05, which would indicate the opposite. The correlations between the network measures and score, with 14 archers per group, have absolute r values between 0.54 and 0.58 and p between 0.03 and 0.05, and the authors admit they cannot explain the positive and negative correlations that appear side by side. The abstract states that eigenvector centrality was higher in the Beijing group, while the results report no significant difference on that measure. A 52-centimetre target at 70 metres is not the regulation face for the distance, and the paper does not explain it. Finally, 30-channel scalp EEG picks up activity from large areas of cortex: assigning functions such as attention or spatial awareness to particular electrodes is the authors' interpretation.

For anyone who shoots, the study changes nothing in practical terms. It shows that in the last second before release, archers of international and of national standard have differently organised brain activity, and that in the stronger archers this activity is less tied to the score of the individual arrow. This fits the idea that more experienced archers depend less on the state of the moment, but it gives no indication of how to get there.

The authors envisage using these measures to select athletes and for neurofeedback. These are prospects, since neither use was tested here, and with 28 archers and a cross-sectional design any selection based on EEG would have no foundation.