Imagery training in elite archers: Kim 2020 study

Six Korean national team archers, three men and three women aged 23 to 36 with 12 to 22 years of shooting, did eight 30-minute sessions of mental imagery training in the laboratory. With eyes closed, listening to audio recorded at a World Cup event, they imagined the five minutes before the start of competition and the three arrows of the first end. EEG recorded beta activity over the motor areas, and a small motor vibrated whenever the activity exceeded a personal threshold.

The authors report that beta activity fell in four of the six archers, rose in one and stayed almost unchanged in the last, with differences they themselves call minuscule: for instance from 0.000186 to 0.000157 in one archer and from 0.000376 to 0.000393 in the one whose activity rose, with no units given. Every difference comes out significant, but the tests are run on individual signal samples, more than 500 per second. In the arrow-by-arrow analysis, vibration decreased from the first to the second and third arrows in only two archers.

On the questionnaires, competitive anxiety changed little, falling in three archers, barely changing in one and rising in the other two, very slightly in one and somewhat more in the other, with almost identical group means (cognitive anxiety 1.88 to 1.85, somatic anxiety 1.70 to 1.605, trait anxiety 16.33 to 16). Imagery ability on the Sport Imagery Questionnaire rose on average in all five areas, most of all kinaesthetic (8.5 to 12.166) and visual (8.166 to 11).

Before the main study, a pilot with 16 electrodes had shown beta activity during imagery at C3, Cz and C4 in all six archers, so the main study used only these three sites over the motor areas (Laxtha WEEG-32, impedance below 5 kΩ, filters at 0.5 and 100 Hz). At the start of each session a one-minute resting recording set the personal threshold, at the eighth stanine, roughly the 89th percentile; during imagery a LabVIEW programme made a small motor vibrate whenever beta exceeded it. The audio, recorded at a real competition, reproduced in real time the five-minute call, the signals and the start of shooting, and each imagined arrow was analysed over a 20-second window.

Before and after the eight sessions the archers completed the CSAI-2 for state anxiety and confidence, the SCAT for trait anxiety, and the Sport Imagery Questionnaire, which rates the visual, auditory and kinaesthetic vividness of images, the mood that goes with them and control over them. Pre and post EEG data were compared with separate t-tests for each archer; for the questionnaires there are only means and case-by-case descriptions. The archers were also given cards for practising imagery at home.

There is no measure of shooting at all: no scores, no competitions, no arrows shot. There is no control group and no condition without vibration, so the effect of neurofeedback cannot be separated from that of imagery itself, of repeating the sessions, or of the home practice cards, which according to the authors some archers used a lot and others little.

The EEG analysis is the weakest part. The t-tests treat tens of thousands of signal samples as independent observations, so any difference, however tiny, becomes significant. The text presents as the mean for all archers values that the table shows belong to the first archer, and the column labelled vibration frequency actually holds a difference between signal and threshold, including negative values, which makes it hard to tell what changed. It is not clearly stated which sessions make up pre and post.

The questionnaire figures also contain arithmetic slips: the change in cognitive anxiety is given as -0.33, but 1.88 to 1.85 is -0.03, and the kinaesthetic change is given as +3.833 when 8.5 to 12.166 is +3.666. Finally, the abstract says beta stabilised and vibration decreased in all archers, whereas the results describe one archer whose beta rose and one with no clear difference.

For an archer, the study describes a training idea more than a result: imagining the start of a competition with audio recorded at the venue, and getting a signal when brain activity rises above one's resting level. The archers reported more vivid images, especially in bodily sensations, and some reported slightly less anxiety, but without a comparison group this may simply reflect having spent eight sessions on imagery.

The study does not show that neurofeedback improves shooting, because shooting was not measured, and the EEG data, as analysed, do not allow the conclusion that the archers' brain activity changed in any meaningful way. Anyone weighing up a neurofeedback system for training will find here an example of how it might be used with elite athletes, not evidence that it works.