Visual Modulation of Human Responses to Support Surface Translation

Abstract

Vision is known to improve human postural responses to external perturbations. This study investigates the role of vision for the responses to continuous pseudorandom support surface translations in the body sagittal plane in three visual conditions: with the eyes closed (EC), in stroboscopic illumination (EO/SI; only visual position information) and with eyes open in continuous illumination (EO/CI; position and velocity information) with the room as static visual scene (or the interior of a moving cabin, in some of the trials). In the frequency spectrum of the translation stimulus we distinguished on the basis of the response patterns between a low-frequency, mid-frequency, and high-frequency range (LFR: 0.0165-0.14 Hz; MFR: 0.15-0.57 Hz; HFR: 0.58-2.46 Hz). With EC, subjects' mean sway response gain was very low in the LFR. On average it increased with EO/SI (although not to a significant degree p = 0.078) and more so with EO/CI (p < 10^-6). In contrast, the average gain in the MFR decreased from EC to EO/SI (although not to a significant degree, p = 0.548) and further to EO/CI (p = 0.0002). In the HFR, all three visual conditions produced, similarly, high gain levels. A single inverted pendulum (SIP) model controlling center of mass (COM) balancing about the ankle joints formally described the EC response as being strongly shaped by a resonance phenomenon arising primarily from the control's proprioceptive feedback loop. The effect of adding visual information in these simulations lies in a reduction of the resonance, similar as in the experiments. Extending the model to a double inverted pendulum (DIP) suggested in addition a biomechanical damping effective from trunk sway in the hip joints on the resonance.

Keywords: balance control; human posture control; modeling; support surface translation; vision.

FIGURE 1

Examples of stimulus, original response, and frequency response functions (FRFs). (Aa) PRTS translation stimulus (pp. 9.6 cm). (Ab) Evoked sway response of body COM around ankle joints (eyes closed). (B) Derived frequency response functions (FRFs) of gain, phase, and coherence. Vertical dashed lines separate low, middle, and high-frequency ranges of the stimulus; compare Figure 2). Gain represents the amplitude ratio between sway response amplitude and translation stimulus amplitude. Unity gain indicates a response of 1° sway amplitude per 1 cm of support translation and zero gain the absence of any evoked sway. Phase reflects the temporal relation between the translation stimulus and the sway response (0° indicating exactly in-phase response and –180° a lean counter to platform translation). Coherence is a measure of the signal to noise ratio of the stimulus-evoked sway (the coherence for a linear ideal system with no noise would be unity over the whole spectrum). To get an impression of the stimulus, see https://youtu.be/BUlcKQ67JCk.

FIGURE 2

Spectral characteristics of the PRTS (pseudo-random ternary sequence) translation stimulus. a, b, and c give the low, mid, and high frequency (and acceleration) ranges referred to in the text and the Results figures.

FIGURE 3

Results from the seven experiments reflecting the seven tested viewing conditions listed in Table 1. Shown are gain, phase, and coherence (in A–C with standard error values) in terms of gain, phase, and coherence over frequency. Dotted curves in panels (D,E) demonstrate discrepancy of obtained versus modeled data (see text). Curves in (F,G) represent “distorted” responses obtained with the cabin fixed on the translating body support.

FIGURE 4

Inverted pendulum model (simplified version of the model used by Hettich et al. (2014). The lower half (Ankle Module) shows the control of the body center of mass (COM) with negative feedback from proprioception and passive tissue (pas). The input is desired body COM orientation in space (BS!; 0°, upright). The controller C_A (proportional-derivative) provides after the lumped time delay Δt the ankle torque T_A that stabilizes at rest the upright body pose through the feedbacks from ankle proprioception (PROP_A) and passively from connective tissue (Ankle_pas). During support surface translations, deviations from the earth vertial body pose by the translation evoked inertial and gravitational ankle torque (T_A–inert; T_A–grav) are compensated for by a Disturbance Estimation and Compensation (DEC) mechanisms (gray box; details in Supplementary Appendix A). The role of the upper body stabilization on the hip joints for whole body COM stabilization follows comparabale principles and is adding to the sway of the body COM. In the context of the present experiments, we considered two modeling scenarios, one where the gain of the hip module was so much enhanced that essentially a SIP sway in the ankle joints resulted, and the other where a DIP scenario with additional sway in the hip joints resulted. In the latter case, a damping of the body sway responses from support translation in the ankle joints resulted, adding to sway damping resulting from visual input (details in Supplementary Appendix B). The figure emphasises that the balancing involves both, the ankle joint control and the hip joint control.

FIGURE 5

Model simulation results for COM balancing without vision (A; EC), with visual position information available (B; EO/SI), and with both visual position and velocity information available (C; EO/CI). Note that these results reproduce the main response features obtained in the human experiments (compare Figures 3A–C). The dash-dot coherence curve in (A) shows result of adding low-frequency noise to the control. It suggests that noise may explain to a large extent the difference in coherence seen between the simulations and the human data in the low-frequency range.

FIGURE 6

Robot simulation results for COM balancing without vision (A, EC for “eyes closes”) and with visual input corresponding to the continuous illumination condition (B, EO/CI).

FIGURE 7

SIP model simulations to test our hypothesis of a resonance effect. (A) Variations of loop gain. With a loop gain of unity, a peak occurs in the upper MFR (full line), but there is in addition also a tendency for a peak slightly below 0.2 Hz with reduced loop gain (dotted). Note also that the larger the peak is in the range of mid and high frequencies, the lower is the gain in the low-frequency range (by which we explain the absence of considerable sway in the low-frequency range with EC in our experiments; compare Figures 3A–C). Other parameter changes had different effects. For example, varying instead the time delay of the feedback loops increased, respectively, decreased the amplitude of the gain peak of the FRF (panel B with inset, simplified model), while increasing selectively the gain of the passive joint stiffness shifted primarily the gain peak toward higher frequencies (C).

FIGURE 8

(A,B) Further simulation results. (A) Effects of injecting low-frequency noise into the servo loop of the COM model (Figure 4). Dotted curves give simulation results obtained with noise alone. When adding a weak support translation stimulus, the response (full lines) developed a resonance peak and a coherence of almost unity in the upper MFR (arrow) with a phase lag of approximately 180°. (B) Results of simulations that tested the theoretical hypothesis of an antagonism between visual and proprioceptive effects. The dashed curves represent sway responses obtained with proprioceptive feedback alone, full lines those with continuous visual feedback alone, and the dotted curves give the results obtained with the combination of both (note its resemblance to Figure 3C).