Human sensitivity to vertical self-motion

Abstract

Perceiving vertical self-motion is crucial for maintaining balance as well as for controlling an aircraft. Whereas heave absolute thresholds have been exhaustively studied, little work has been done in investigating how vertical sensitivity depends on motion intensity (i.e., differential thresholds). Here we measure human sensitivity for 1-Hz sinusoidal accelerations for 10 participants in darkness. Absolute and differential thresholds are measured for upward and downward translations independently at 5 different peak amplitudes ranging from 0 to 2 m/s(2). Overall vertical differential thresholds are higher than horizontal differential thresholds found in the literature. Psychometric functions are fit in linear and logarithmic space, with goodness of fit being similar in both cases. Differential thresholds are higher for upward as compared to downward motion and increase with stimulus intensity following a trend best described by two power laws. The power laws' exponents of 0.60 and 0.42 for upward and downward motion, respectively, deviate from Weber's Law in that thresholds increase less than expected at high stimulus intensity. We speculate that increased sensitivity at high accelerations and greater sensitivity to downward than upward self-motion may reflect adaptations to avoid falling.

Fig. 1

Experimental setup

Fig. 2

Acceleration, velocity, and position traces for a baseline condition trial (a, b, c, respectively) and for a trial with 0.3 m/s² pedestal (d, e, f, respectively)

Fig. 3

Graphical representation of the peak amplitude of the comparison stimuli used. The gray dotted lines indicate the pedestal intensities around which the set of stimuli was selected for each condition. Each tick marks a possible acceleration value to be presented

Fig. 4

Psychometric functions fit (gray line) in log stimulus space to the data (black dots) obtained for one participant in the baseline condition (a) and for a pedestal of 1.6 m/s² (b). The probability of rating the comparison stimulus as stronger than the pedestal is on the y axis. The light gray line and the black dashed line represent the mean and the standard deviation of the fitted cumulative Gaussian, respectively. a The gray line represents the stimulus that corresponds to the participant’s absolute threshold. b The black dashed lines represent stimuli that are one standard deviation weaker (left dashed line) or stronger (right dashed line) than the pedestal. According to our definition of differential threshold, the region on the x axis between the black dotted lines encloses stimuli that cannot be distinguished from the pedestal

Fig. 5

Differential thresholds for upward movements (black triangles) are significantly higher than for downward movements (gray triangles). Error bars are SEM. Their relationship with the motion intensity is more consistent with a power function (black and gray continuous lines) as opposed to a linear fit (black and gray dashed lines)