Results for static stimuli (control conditions) showed no significant variations in postural response (BSA and vRMS), regardless of the area (size and location) of visual field stimulated. These findings indicate that in an ecological environment stimulating peripheral or central visual field does not generate a significant difference in postural control for stance condition. However, other authors (Amblard & Carblanc,
1980; Berencsi et al.,
2005) stated that peripheral vision contributes more than central vision for postural control in static conditions. Nougier et al. (
1998) also used a static stimulus and concluded that peripheral and central vision had different and complementary functional roles. These studies, however, used non-ecological static stimuli with no depth cues.
The results of the present study suggest that when stimulation is quasi-homogenous across the full visual field, the contribution of the visual system in postural control is invariant, regardless of the part of the visual field stimulated. Roles of central and peripheral vision for the control of upright stance are not significantly different when in an ecological static environment. This is consistent with the findings of Straube et al. (
1994), who used static stimuli that took into account the cortical magnification factor.
When observing a static visual scene, the perceived motion of the environment is due to body sway. Amblard and Carblanc (
1980) and Kelly, Loomis, and Beall (
2005) postulated that this relative perceived motion provides visual information for postural control. However, observing a moving environment in which amplitude and velocity are greater than those produced by spontaneous body sway creates the illusion of moving through the environment or that the environment itself is moving. This illusion of self-motion is called vection. Vection elicits a correlated compensatory postural response intended to reduce changes in the optic flow (Kawakita et al.,
2000; Kuno, Kawakita, Kawakami, Miyake, & Watanabe,
1999; Lestienne, Soechting, & Berthoz,
1977; van Asten et al.,
1988).
When stimulating the same area of the visual field, BSA was significantly greater with dynamic than with static stimuli, except for C4 and C7 visual field conditions. This suggests that there was little or no vection induced by dynamic stimuli in the central 7° of the visual field. With respect to stimuli presented in the central part of the visual field, we found that vection was induced in the C15 and C30 conditions. The BSA increase for C15 and C30 between static and dynamic condition was similar to that observed for peripheral and FF conditions (
Figure 5). Thus, the relation between vision and postural response suggests that central vision corresponds to an area that includes at least the central 7° but is smaller than the central 15° of the visual field. These findings indicate that peripheral vision plays a greater role than central vision for inducing vection with a tunnel design stimulus; this is consistent with the conclusions of van Asten et al. (
1988) and Lestienne et al. (
1977).
The results of the current study also suggest that peripheral and central vision are used differently in the perception of the optic flow experienced with linear motion. In these conditions, it is noticeable that apparent velocity of the stimulus (texture elements) decreases from the periphery to the center (Johnston & Wright,
1986). Despite the fact that central vision is more sensitive for motion detection than peripheral vision (Finlay,
1982; McKee & Nakayama,
1984; Orban, Van Calenbergh, De Bruyn, & Maes,
1985) stimulation of central vision (area including the central 7° of the visual field) with a dynamic stimulus failed to induce a postural response of significantly greater amplitude as compared with a static stimulus.
Central dynamic stimulation failed to induce vection whereas peripheral dynamic stimulations produced vection. This is inconsistent with the results of Andersen and Braunstein (
1985). They claimed to induce greater vection by centrally stimulating only 7.5° of the visual field compared to 21.2°, using a radially structured stimulus moving along the line of sight. The kind of stimulus used in the present study had the same central structure as the stimulus used by Andersen and Braunstein (
1985). Thus, if differences between our findings and theirs seem to be unrelated to the structure of the stimulus, they may be due to the fact that these researchers used more subjective measures or less salient visual textures.
vRMS is similar in both dynamic and static environments suggesting that postural stability is not significantly affected by a dynamic visual stimulation for young adult. Thus, although an optic flow induces a postural reactivity to reduce vection between the body and virtual environment, postural stability is preserved. However, it appeared that instability slightly increased as a function of dynamic stimulation of peripheral visual field. Berencsi et al. (
2005) failed to modify the postural response by introducing spatiotemporal changes in their stimulus. They hypothesized that motion direction rather than temporal changes may have a crucial role in postural control and suggested the use of optic flow to further examine this issue. Our findings concerning BSA are consistent with this hypothesis (induced vection) but the dynamic stimulation does not significantly increase postural instability in this population.
Our results show a decrease of BSA with the stimulus frequency, regardless of the area of visual field stimulated. This is in agreement with previous studies (Lestienne et al.,
1977; van Asten et al.,
1988). Our data corroborate the conclusions of Mergner, Schweigart, Maurer, and Blümle (
2005) supporting that postural reactivity to a dynamic visual stimulation depicts both frequency and displacement (sway amplitude) effects. Indeed, BSA tend to decrease for high temporal frequencies (
Figure 6) highlighting a frequency dependant trend. BSA also depicts a plateau for low temporal frequencies described as a displacement saturation effect.
With respect to visual field effect on postural control, the findings of the present study show no significant differences of BSA between C4 and C7 conditions. BSA measured in C15 and C30 conditions is quite similar to the BSA measured in peripheral and FF conditions since paired comparisons showed no differences between C15, C30, P4, P7, and P30 nor between C30, P4, P7, P15, P30, and FF conditions. Significant differences between C4 and each of C30, P4, P7, P15, P30, and FF conditions as well as between C7 and each of C15, C30, P4, P7, P15, P30, and FF conditions were found. This confirms that a functional central vision definition for postural control should subtend an area between the central 7° and 15°, whereas peripheral vision should subtend the surrounding area. Central vision should correspond to an area that includes at least the central 7° but is smaller than the central 15° of the visual field. Since BSA tended to increase while the central area of visual field increased there may be a transition between central and peripheral vision for postural control. These results follow neuro-anatomical findings relative to cone and rod distributions (Jonas, Schneider, & Naumann,
1992), which are not homogenous and different for both across the retina. They are also consistent with neuro-anatomical findings related to visual field defects due to occipital cortex lesions (Horton & Hoyt,
1991; McFadzean, Brosnahan, Hadley, & Mutlukan,
1994), which state that the central 10° of the visual field correspond approximately to 50–60% of the striate cortex but cannot separate accurately central and peripheral vision. Isolating purely central or peripheral vision would therefore prove to be very difficult.
Concerning vRMS, no significant difference was found between C4 and C7, the same for C15, C30, P4, P7, P15, P30, and FF conditions. Nevertheless, vRMS in each of C4 and C7 conditions were significantly different from the C15, C30, P4, P7, P15, P30, and FF conditions. These results confirm what was observed with BSA.
With dynamic stimuli, postural responses (BSA and vRMS) are significantly greater when stimulating peripheral rather than central vision (central 7° of the visual field). This is consistent with previous studies that used optic flow stimuli and stated that peripheral vision plays the primary role in maintenance of posture (Kawakita et al.,
2000; Lestienne et al.,
1977; Previc & Neel,
1995). Other studies also came to similar conclusions although different methodologies were used. Amblard and Carblanc (
1980) and Berencsi et al. (
2005) used only 2D static stimuli. Berencsi et al. (
2005) also tested a dynamic condition that consisted of two alternating random-dot patterns (i.e., a non-directional spatial/temporal perturbation), but this condition had no effect on postural control. Optic flow differs from this pattern because it represents an entity easily assimilated to ecological conditions.
Since cortical representations of central and peripheral parts of the stimuli can be considered similar, the only difference between stimuli for peripheral and central vision is the stimulus distance, in which nearer elements stimulate peripheral vision. This leads to a purely geometrical explanation. Considering elements of equal real size, the closer they are, the greater their apparent size and velocity (
Figure 8). Therefore, postural response to a dynamic stimulus seems to depend partly on the distance to the stimulus and, so, on its apparent velocity. Furthermore, it would be interesting to examine this point by using a spherical optic flow structure similar to the one described by Gibson (
1979), with the eyes of the subject being placed at the center of the sphere. In addition, we can note that the shape of the optic flow experienced during anterior–posterior motion, as described by Gibson (
1979), resembles the retinal shape. The retinal portion on which the central part of a moving visual scene is projected (like the one used in this study) can be regarded as a plane that is perpendicular to the direction of motion. The peripheral retina, which is stimulated by lateral parts of the visual scene, can be considered as a series of parallel planes. When an observer is moving forward in a linear fashion, the central part of the retinal image expands radially from the fovea. Expansion velocity increases with retinal eccentricity (Johnston & Wright,
1986). As for the peripheral part of the retinal image, it expands minimally; it drifts on the retina, parallel to motion direction. Drift velocity also increases with eccentricity. Therefore, optic flow is perceived as a radial expansion in its central part and a longitudinal drift in the periphery of the visual scene.
The greatest changes in the optic flow field involve peripheral vision. Furthermore, vection induces a compensatory postural response intended to reduce these changes occurring in the optic flow (Kawakita et al.,
2000; Kuno et al.,
1999; Lestienne et al.,
1977; van Asten et al.,
1988). Therefore, when the stimulus is moving, it is coherent to record induced postural responses of greater amplitude when stimulating peripheral vision than when stimulating central vision.
With respect to the texture of the stimuli used in this study, we can assume that the stimulations of central and peripheral retina are similar for spatial frequencies and that cortical representation of the different parts of the stimuli are almost equivalent. Motion, in spite of the velocity gradient in the moving visual scene, appears homogenous (Johnston & Wright,
1986). This suggests that, when the visual system is stimulated with an optic flow, central and peripheral vision are used in a different way for processing visual information and for control of stance than in a static environment.
Therefore, respective contributions of central and peripheral vision seem to be related to relative magnitude of changes of the optic flow. While peripheral vision contributes to compensatory sway (Kawakita et al.,
2000; Kuno et al.,
1999; Lestienne et al.,
1977; van Asten et al.,
1988), central vision seems to contribute to postural stability. When comparing postural responses (BSA) for peripheral and full visual field conditions, there are no significant differences. If central vision had a more important role in postural reactivity BSA should have been lower during FF condition. Such results were observed by Previc and Neel (
1995); they found postural responses to be of greater amplitude when stimulating peripheral vision as compared with full field stimulation.
When moving linearly, peripheral vision is important for updating the structure of the environment for spatial location, while central vision is important for orientation (Turano, Yu, Hao, & Hicks,
2005). Central and peripheral vision also represent inputs of equal importance in the perception of optic flow (Habak, Casanova, & Faubert,
2002). When immersed in optic flow, visual information may be processed as follows for the control of stance: central vision serves to orientate the response direction, while peripheral vision processes visual information on location and velocity, allowing an adapted postural response to the perceived perturbation.