Sensory reweighting with translational visual stimuli in young and elderly adults: the role of state-dependent noise

The properties of sensory reweighting for control of human upright stance have primarily been investigated through experimental techniques such as sinusoidal driving of postural sway. However, other forms of visual inputs that are commonly encountered, such as translation, may produce different adaptive responses. We directly compared sinusoidal and translatory inputs at stimulus parameters that made stimulus velocity comparable with each type of stimulus. Young healthy individuals were compared with healthy elderly and elderly designated as “fall-prone” to investigate whether the hypothesized basis for poor balance control in the “fall-prone” elderly is related to their ability to reweight sensory inputs appropriately. Standing subjects were presented with visual displays which moved in the medial–lateral direction either by (1) oscillating at different amplitudes or (2) simultaneously oscillating and translating at different speeds. All three subject groups showed that increasing the amplitude of the oscillations led to a decrease in gain. Increasing translation speed led to decreases in gain only at speeds above 1 cm/s. This suggests that the nervous system is processing more than just stimulus velocity to determine the postural response. A model implementing “state-dependent noise”, in which visual stimulus noise increases with relative speed, was developed to account for the difference between translation and oscillation. The weak group effects question the common view that the fall-prone elderly are deficient in sensory reweighting. One explanation for the apparent discrepancy is that the slow, small-amplitude visual stimuli used in this study probe the asymptotic dynamics of the postural response. If given enough time, even the fall-prone elderly are able to adapt to a new sensory environment appropriately. However, the asymptotic adaptive response may not be functional in terms of preventing falls.