Effet de l’âge sur un principe de contrôle

Performance

The panel A of the Figure 50 depicts the frequency distributions of participant’s final Ypositions (i.e., participant’s positions along the Y-axis at time t=8 s, when the ball crossed the participant’s displacement axis) cumulated across trials for the three offset conditions as compared to the position of the Interception Point (IP equal to 5.5, 8.2 and 10.5 m for the -2.5, +0.2 and +2.5 m Offset conditions). For all Offset conditions, the frequency distributions of Young participant’s final Y-positions display sharpen peaks, spreading over 0.5 m forward and backward to the IP, whereas the middle aged distributions of final Y-positions were relatively flat and were spread up to 2 m forward and backward the IP. Distributions of final Y-positions show that, in average, Young participants slightly less overshot the IP (final Ypositions equal to 5.60, 8.32 and 10.6 m) than did Middle-aged participant’s (final Y-positions equal to 5.69, 8.38 and 10.63 m). Such overshoot occurred more often for Young participants than for Middle-aged participant’s (71.63 vs. 58.29 % of trials). More generally, such distributions of trials, in which the IP was sometimes overshot and sometimes undershot (especially for Middle-aged participants) do no indicate a systematic bias toward the IP. The panel B of Figure 50 displays the absolute errors (AE) computed in two ways for the two groups in the different Environment conditions. We first considered absolute errors as the Euclidian distance between the agent and the ball at the moment at which the ball crossed the participant’s displacement axis (dotted bars). This criterion indicated that Young participants were able to intercept the targets with their head at time t = 8 s as instructed (mean AE equal to 0.14 m) whereas Middle-aged participants were only able to virtually catch the targets with their arms (mean AE equal to 0.43 m). To control that Middle-aged participants succeeded in the task by overcoming the instructions and by only attempting to intercept targets at any moment of the trials, we also computed absolute error as the minimum Euclidian distance between the agent and the ball at any moment during the overall trial course (plain bars). The ANOVA (2 Groups × 4 Environments) performed on the AE mean values (computed with the latter definition) revealed a significant main effects of Group – 134 – (F(1, 12) = 134.41, p<.05, η²p = 0.92) and Environment (F(3, 36)= 11.57, p<.05, η²p = 0.49). The Environment × Group (F(3, 36)= 6.25, p<.05, η²p = 0.34) interaction was also significant. Posthoc analysis revealed that Young participants were more accurate than Middle-aged participants (0.10 ± 0.03 m vs. 0.26 ± 0.03 m) showing thus that, because the latter errors were appreciatively similar to a head diameter, Middle-aged participants succeeded in the task. A posteriori comparisons also revealed that, while the performance of Young participants did not significant vary between the different environments (p>.05), Middle-aged participants were more accurate in the Full environment than in the other environments and were the least accurate in the Empty environment (0.21 m ± 0.03, 0.26 m ± 0.01, 0.26 m ± 0.04 and 0.31 m ± 0.05 m for the Full, Ground, Landmark and Empty conditions, respectively; p<.05).

Kinematics

Three-ways repeated measures ANOVAs (4 Environments × 3 Offsets × 16 Time Intervals) on velocity profiles (Figure 51A) were performed separately for both groups of participants (cf. Table 1). Young participants. Analyses performed on individual mean velocity profiles revealed significant main effects of Offset (p<.05, η²p = 0.99), Environment (p<.05, η²p = 0.53) and Time Interval (p<.05, η²p = 0.46). Moreover, both the Offset × Time Intervals (p<.05, η²p = 0.72) and Environment × Time Intervals (p<.05, η²p = 0.46) interactions were also significant. A posteriori comparison revealed that significantly different velocity profiles were produced in the three Offset conditions during the last 6 seconds of the trial (p<.05) (Figure 51A). The velocity changes were in accordance with the task requirements, with positive offset conditions giving rise to a higher overall displacement velocity in comparison with the negative offset conditions, and with intermediate offset conditions giving rise to intermediate – 136 – displacement velocity profiles (p<.05). Moreover, Young participants produced similar displacement velocity profiles in all environment conditions (p>.05) except during the very last Time interval (i.e., close to ball contact) in which the velocity produced in the Empty condition was higher than that produced in the other conditions. Middle-aged participants. Analyses performed on individual mean velocity profiles revealed significant main effects of Offset (p<.05, η²p = 0.99) and Time Intervals (p<.05, η²p = 0.35) but no significant main effect of Environment (p>.05, η²p = 0.46). Moreover, Offset × Time Intervals (p<.05, η²p = 0.48) interaction was also significant. A posteriori comparisons revealed that significantly different velocity profiles were produced in the three Offset conditions during the last 6 seconds of the trial (p<.05) (Figure 51A). Once again the velocity changes were in accordance with the task requirements. Moreover it is worth noting that contrary to the other participants, the Middle-aged participants decelerated systematically at the very beginning of the trial, whatever the offset conditions, before producing adaptive velocity changes. Taken together, these results show that the velocity profiles exhibited by both Young and Middle-aged groups of participants are highly affected by the Offset but only marginally by the Environment. At a more descriptive level, it is also worth noting that the velocity adaptations produced by Young participants are very smooth, contrary to those produced by Middle-aged participants. One can finally notice that the Middle-aged participants sometimes even produced backward displacements in the negative offset condition which was never produced by Young participants. We further analyzed the displacement kinematics by counting the number of zero crossings ( Y ZC && ) exhibited in the acceleration profiles. The number of Y ZC && is indicative of whether the displacement adaptations are gradual (very few Y ZC && ) or conversely nonlinear (numerous Y ZC && ) (Figure 51B). Two-way ANOVA (4 Environments × 2 Groups) with repeated measures on the Environment factor performed on the individual mean number of Y ZC && performed by Young and Middle-aged groups of participants revealed a significant main effect of Group (F (1, 12) = 5.20, p<.05, η²p = 0.99), but no significant main effect of the Environment factor (F (3, 36) = 0.3, p>.05, η²p = 0.02). Young participants produced less Y ZC && than the Middle-aged ones (2.60 ± 0.63 vs. 2.99 ± 0.39).