HOW OPEN- AND CLOSED SKILLS ATHLETES ADAPT DECISIONS TO AN UNCERTAIN ENVIRONMENT WITH AND WITHOUT A STRESSFUL CONDITION.

INTRODUCTION: Athletes must decide their actions under stressful situations. These abilities could also be transferred into a general do- main (1, 2). Thus, the present research analyzed the decision-making processes of athletes in an uncertain environment and how they acted when exposed to stressful situations. METHODS: Participants: 24 élite open-skills athletes (OSA; M age = 17.71 ± 1.02 y.o.), 13 élite closed-skill athletes (CSA; M age = 21.40 ± 3.55 y.o.) and 36 non- athletes ( (M age = 23.56 ± 1.57 y.o.) were recruited. Instruments: A reinforcement learning probabilistic task (3) with three levels of uncertainties (i.e., low, medium, and high) was employed. Participants had to score as many points as possible. The task was performed, in counterbalanced order across participants, two times: one in standard condition (SC) and one Under Pressure (UP). Heart rate was monitored. Several indexes, such as the Spatial Error (i.e., the ability to predict future events), the Gain (the ability to maximize the score), and the implicit decision confidence (through the bet on the decisions), were evaluated to examine the possible differences among the sport type (OSA vs CSA). The confounding factor of fluid intelligence was con- trolled (Raven-APM [4]) due to the possible relation between intelligence and decision-making. Data analysis: Linear mixed-effect regression for each index was performed. The independent variables were: sport type, condition, and level of uncer- tainty. Intelligence was set as a covariate variable. The subject was the random effect. Heart rate and mental effort analysed the effect of stressful condition. RESULTS: Mental effort and heart rate analysis reported that they were higher in UP compared to SC. Spatial Error results revealed an effect of intelligence (p < .001) only when the uncertainty was low and in SC (p < .05). Even if the triple interaction level of uncertainty x sport type x condition was significant ( p < .001), sport type did not influence the ability of predicting events (p > .05). Gain results revealed triple interaction level of uncertainty x sport type x condition was significant ( p < 0.01). However, the post-hoc analysis highlighted an improvement of the score from SC to UP condition in the low uncertainty environment for the control group; while in medium uncertainty for OSA. Confidence analysis highlighted triple interaction level of uncertainty x sport type x condition was significant (p < .001). Post-hoc results showed that only OSA modulated the bet between the two conditions when the uncertainty was low and high (p < .05). Conclusion Sport type did not influence the performance of this perceptual-decision task. However, the perception of the risk changed in UP. Specifi- cally, OSA seemed to behave conservatively when exposed to a stressful situation. Moreover, the results indicated that intelligence could help to predict events. 1Voss et al.(2010);2Chaddock et al.(2011);3Larsen&Coricelli(2017);4Raven(199)