Influence du mode d’accès proprioceptif sur un principe de contrôle

Influence du mode d’accès proprioceptif sur un principe de contrôle

DISCUSSION

The aim of the Experiment was to determine to what extent deafferentation affected the human ability to intercept moving balls depending on visual sources of information contained in the environment. We asked two groups of participants (Middle-Aged, Patient GL) to intercept balls that travelled toward them obliquely in a virtual environment by manipulating a joystick allowing them to control their velocity. More precisely, we set-out an experiment in which different visual information specifying the direction of displacement were drained form the visual environment, leading finally participants to rely on proprioceptive information. In the Full Environment, the visual scene contained two visual information related to the direction of displacement: the FoE and a visual egocentric frame of reference. In the Ground Environment, only the FoE was available. In the Landmark Environment, only a visual egocentric frame of reference was available. In the Empty environment, the visual scene do not allowed to determine one’s the direction of displacement on the basis of visual signals. Analyses of performance (AE) revealed that the Patient GL achieved worse score than her healthy counterparts (Middle-Aged Participants). Moreover, whereas the performances of Middle-aged participants were damaged as the environment was drained from its visual content (Full, Ground, Landmark and Empty), the Patient GL was able to keep constant its performances in all Environments. Middle-Aged Participants produced a better AE than the Patient GL in three of the four environment conditions (i.e., Full, Landmark and Empty conditions). Kinematics analyses showed that the Patient GL exhibited jerkier velocity profiles than Middle-Aged participants. Moreover, whereas the velocity profiles performed by Middle-Aged participants do not differed between Environment conditions, the Patient GL exhibited jerkier velocity profiles in particular when the three types of perceptual signals (i.e., visual allocentric, visual egocentric, proprioceptive egocentric) were available (Full condition). The CBA model failed to explain the behavior observed by the Middle-Aged participant and the Patient GL. Interestingly however, adding adjusted perceptual thresholds in the numerical simulations allowed the ‘Bounded-CBA’ model to provide a good account of the behavior produced by the three groups of participants in all environment conditions.

The influence of deafferentation

The core issue of the present work is the evaluation of the interceptive performance when the egocentric reference system is greater impaired by the absence of proprioceptive signals. The three levels of analysis (Performance, Kinematics and Perceptual-Motor strategy) provide complementary pieces of answer. First of all, in comparison with Middle-Aged participants the Patient GL produced a lower performance in three out of the four environment conditions (Full, Landmark and Empty). While this result was expected in the Empty condition, we expected the Patient GL to be as accurate as Middle-Aged participants in the presence of visual information, i.e., in the other three conditions. Moreover, contrary to our expectations, the Patient GL reached the same level of performance when visual information was lacking (i.e., Empty condition) and when the environment was visually enriched. The kinematic analyses performed on displacement kinematics provide some insights into these unexpected results. These analyses confirm previous studies; the displacement velocity profiles produced by both Middle-Aged and Deafferented patients are highly jerky (Riviere & Thakor, 1996). These jerky velocity profiles are even more pronounced for the Patient GL, in particular in the Full condition, i.e., when visual information is available through both optic flow and retinal signals. Finally, the analyses on the perceptual-motor strategies enable us to clarifying the picture one step further. The initial version of the Constant Bearing Angle model failed to account for the regulation behavior produced by the Patient GL. Conversely, this study confirms the need for neuro-physiologically grounded architecture of law of control. As found by Francois et al. (in press), the ‘Bounded-CBA’ model allowed accounting for jerky velocity profiles performed by Middle-Aged participants, revealing thus that perceptual threshold for perceiving θ & drove the control of displacement. Moreover, our study also revealed that the ‘Bounded-CBA’ model allowed accounting for velocity profiles performed by the Patient GL.Consequently, the sudden and steep slope in the displacement adaptations could express the patient’s difficulties to detect small angular changes. This could have led her to ‘bounce’ from the upper part of the threshold to the lower part of the threshold. Interestingly the perceptual thresholds found for the Patient GL differed across the environment conditions (2.3, 2.6, 2.8 and 2.8 °/s for the Full, Ground, Landmark and Empty conditions respectively). In the Full condition, the patient’s threshold is not only minimal, but is also of the same magnitude as the threshold found for Middle-Aged participants. This result suggests that, when available, optic flow and retinal signals compensate for the lack of extra-retinal signals in detecting the rate of change in bearing angle. The reason why this condition did not give rise to an increase in performance is probably related to the constraints imposed to the participant and in particular to the impossibility to see both the hand and the joystick. Surprisingly the Patient GL’s performance did not decrease when both optic flow and retinal (i.e. the body-centered cross on the screen) information were removed. This result was unexpected given that the Patient GL suffers from a lack of proprioception from the neck muscles and that this information greatly contributes to determine object position and motion relative to the body (Biguer, Donaldson, Hein, & Jeannerod, 1988; Taylor & McCloskey, 1991). Although head position and muscular activity were not recorded in the present experiment, we could clearly notice that the Patient GL kept her head directed towards the ball, possibly by stiffening of her trunk and neck muscles. Freezing body segments is a common strategy of patients and Middle-Aged individuals with sensory impairments (Benjuya, Melzer, & Kaplanski, 2004; Bloem, Allum, Carpenter, Verschuuren, & Honegger,2002; Lajoie et al., 1992). Having both the head and gaze directed towards the ball, the Patient GL may have compensated perceptible changes in gaze direction (i.e., in bearing angle) by accelerating or decelerating accordingly. Within this framework, sensorimotor signals originating from the extra-ocular muscles (Gauthier, Nommay, & Vercher, 1990) would have an important role to detect the rate of change in the bearing angle. Relying essentially on these signals, the Patient GL would be able to perform the task with a reasonable accuracy in comparison with the other environment conditions.

CONCLUSION

This study confirms the need for neuro-physiologically grounded architecture of law of control but do not jeopardize the status of the Constant Bearing Angle strategy as a perceptual-motor principle being able to account for the regulation behavior of participants. More precisely, perceptual constraints added in numerical simulations of the ‘Bounded-CBA’ model perfectly fit with the ‘perception actuation’ level of analysis suggested by Bootsma (Bootsma, 1998). This study also reveals the perceptual problem encountered by Deafferented patients, whose perceptual systems allow them to be able to exploit redundant visual perceptual variables and switch from visual allocentric variables to egocentric ones. However, this study revealed the importance of proprioceptive signal for the control of interception in impoverished visual environments, providing thus converging results with previous study (Bastin et al., 2006a).