Is there a contextual interference effect for sub-elite alpine ski racers learning complex skills?

Abstract

Scientific understanding of the contextual interference effect stems mainly from studies on unskilled participants learning artificial laboratory tasks. Although one goal of such studies is to extrapolate the findings to include real-world learning situations such as sports, this generalization is not straightforward. This study tested the contextual interference effect with 66 sub-elite, competitive alpine ski racers who learned a new movement pattern−the pumping technique to increase velocity in slalom−by practicing this skill in three different slalom courses over a 3-day training period. The interleaved group practiced all three courses each day in a semi-random order. In contrast, the blocked group practiced only one course each day, which was randomized and counterbalanced across the participants in this group. A retention test was delivered 72 h after the last practice day. In contrast to our hypothesis, the interleaved group did not display significantly better retention than the blocked group. The interleaved group’s performance was also not significantly attenuated during skill learning compared to the blocked group. Our results underscore the importance of conducting motor learning experiments in natural environments to understand the conditions that facilitate learning beyond the laboratory environment.

Reinforcement learning enhances training and performance in skilled alpine skiers compared to traditional coaching instruction

“…The main slalom course was used in all sessions, except during the transfer test, and featured a 10 m distance and a 1.9 m offset. The course distance aligned with our previous study [42], but we opted for a slightly larger offset to better suit the skill level of our skiers. The transfer test evaluated how well the skiers transferred their learning to a new slalom course more realistic to a typical alpine ski race course.…”

“…Finding the best strategy requires a delicate balance between exploiting the strategy known to yield the best payoff and exploring alternative strategies that may offer superior benefits. In our study, the options consisted of four strategies that skiers could employ to improve their race times on flat slopes in slalom, grounded in physics-based coaching manuals for alpine ski racing [40, 41, 47, 48], biomechanical research on elite skiers [43, 44, 49] and common strategies used by coaches. The four strategies were named “stand against”, “rock skis forward”, “extend”, and “extend with rock skis forward (see Fig.…”

“…The radial component of the ground reaction force moving the center of mass closer to the center of rotation increases velocity, in accord with the principle of mechanical energy conservation [38]. Simulation studies of individuals extending their body in rollers or during carved slalom turns have shown that this movement can increase speed [39, 40], and the effect has been observed in various ways with elite skiers during training and competition [13, 42, 73].…”

“…In their model, the gain in rotational kinetic energy from this motion is proportional to the amount of work the skier does against the centrifugal force (from the skier’s frame of reference); therefore, a larger extension movement will accomplish a greater increase in rotational kinetic energy. Simulation studies of individuals extending their body in rollers or during carved slalom turns have shown that this movement can increase speed [41, 42], and the effect has been observed in various ways with elite skiers during training and competition [44, 49, 77].…”

“…2b). In the supervised (target skill) learning group, we recruited ski coaches to instruct skiers to select the strategy that we defined as the theoretically best strategy based on computational modeling [38][39][40] and field observations of elite skiers [41,42]. This group serves as a benchmark for the upper limit of performance achievable through optimal strategy choices and thus offers an upward comparison for the performance yielded through reinforcement learning of strategy selection.…”