Archery training and attention on EEG: Hussain 2025

Twenty adults who had never shot were randomly allocated to two groups of ten: one did six weeks of archery, the other nothing. Before and after, with an electroencephalogram on their heads, all of them performed two computer tasks: a Stroop test (press the key for the colour a word is printed in, ignoring the word itself) and a choice reaction time test (press the key for the square in which a cross appears).

Reaction times came down in both groups. On the Stroop test with word and colour matching, the archery group went from 840 to 726 milliseconds and the controls from 910 to 811; with word and colour mismatched, from 950 to 827 and from 984 to 866. In both cases the effect of time is significant (p 0.028 and p 0.012) but the interaction between time and group is nowhere near it (p 0.865 and p 0.947): the two groups improved by much the same amount. On the choice reaction task the archery group went from 526 to 520 milliseconds and the controls from 579 to 533, and this time not even the effect of time reaches significance (p 0.074). Put another way, on the simpler task the group that did not shoot improved more than the one that did, and the authors write that the differences observed did not come from the intervention alone.

On the electrical signals the picture is more mixed. During the Stroop task alpha power changes over time (p below 0.001, partial eta squared 0.38) and there is a time by group interaction (p 0.04, partial eta squared 0.22); when it comes to where the difference sits, though, the archery group had higher values than the controls mainly before the intervention, at five electrodes, and after it at a single one. During the choice reaction task alpha power does not change over time (p 0.92), nor does beta (p 0.27); differences between electrodes remain, along with some interactions involving group. Theta does not change over time in either task.

Twenty-four healthy volunteers were recruited with no shooting experience and no history of attention or neurological disorders (8 men and 16 women, mean age 23.3 years), all right-handed by the Edinburgh inventory. One withdrew and three were excluded for technical problems with the EEG recording, so the data analysed come from 20 people, ten per group, four men and six women in each. A power calculation indicated a minimum of 16 participants for a medium effect.

The intervention ran for six weeks, three times a week, 30 minutes a session, led by instructors from the Malaysian traditional archery organisation: basic technique and coordination, 25 to 35 arrows per session at three metres, with at least two days of rest between sessions and the equipment left with participants so they could practise in between. The control group did neither regular physical activity nor anything else.

The tasks were run in a quiet, dimly lit room, seated with elbows at 90 degrees in front of the screen. The session lasted twenty minutes: five of resting recording with eyes open on a central cross, three of Stroop (40 trials, half of them matching), five of choice reaction (8 practice and 150 test trials, with the interval between stimuli varied at random between one and three seconds), with a one-minute break between blocks. The EEG was recorded with a commercial 14-electrode headset sampling at 128 Hz; signals were filtered between 1 and 40 Hz with a 50 Hz notch, segments beyond 70 microvolts were rejected, and power was computed with a Fourier transform and then expressed as a percentage change from baseline. The bands stated in the methods are theta 4 to 7 Hz, alpha 8 to 12 and beta 13 to 30. The analysis is a repeated-measures analysis of variance with Greenhouse-Geisser correction and follow-up comparisons corrected by the Bonferroni method.

The sample analysed is twenty people, ten per group, and the controls did nothing: the difference between the groups is not between archery and another activity, but between archery and no activity at all, with everything that brings with it (leaving the house three times a week, an instructor, a group). There is no blinding: whoever ran the tasks knew who had been shooting. There is no follow-up.

The stated conclusion and what the tables show do not match. The abstract speaks of notable improvements in reaction times on both tasks and of robust evidence for an effect of archery; in the data, times improve in both groups, the time by group interaction is never significant, and on the choice reaction task not even the effect of time is. The abstract places the increase in alpha power in both tasks: during choice reaction there is no effect of time on alpha at all. The abstract gives the alpha band as 9 to 12 Hz, the methods as 8 to 12.

Errors remain in the degrees of freedom: for the interactions with group the text reports F(5.913, 18), F(6.558, 18) and F(6.56, 18), where the second number should be of the order of a hundred, as it is in the lines beside them. The participant characteristics give a mean height of 1.63 m with a standard deviation of 0.73 m, which cannot be right as printed. It is not known how much participants shot between sessions, given that the equipment was left with them: the dose actually taken is not measured.

Finally, the study measures nothing about shooting: there is no score, no measure of accuracy, and the participants are not archers. It concerns archery as an activity for beginners, not performance on the shooting line.

For anyone who shoots, this study says nothing about performance: the participants were adults handling a bow for the first time, they shot at three metres and nobody counted a point. It concerns the different question of whether six weeks of archery change the attention of people who are not archers.

On the substance, the answer the data support is more cautious than the one the paper gives. Those who shot responded faster at the end of the six weeks, but so did those who had not shot, and the comparison between groups does not separate the two: part of the improvement, perhaps all of it, is simply doing the same task a second time. On the brain signal side there is a change in alpha power during the Stroop task, but the point at which the groups differ most is the baseline measurement, and that weakens any reading of a training effect.

In the registry the study counts as weak evidence on the attention line in beginners, and as a reminder about reading abstracts: here the abstract asserts a result that the tables of the same paper do not support. Anyone teaching groups of beginners can take from it that six weeks at three metres did not dull anyone's attention; beyond that, the paper does not reach.