Effects of motion paradigm on human perception of tilt and translation

Abstract

The effect of varying sinusoidal linear acceleration on perception of human motion was examined using 4 motion paradigms: off-vertical axis rotation, variable radius centrifugation, linear lateral translation, and rotation about an earth-horizontal axis. The motion profiles for each paradigm included 6 frequencies (0.01-0.6 Hz) and 5 tilt amplitudes (5°-20°). Subjects verbally reported the perceived angle of their whole-body tilt and the peak-to-peak translation of their head in space and used a joystick capable of recording 2-axis motion in the sagittal and transversal planes to indicate the phase between the perceived and actual motions. The amplitudes of perceived tilt and translation were expressed in terms of gain, i.e., the ratio of perceived tilt to equivalent tilt angle, and the ratio of perceived translation to equivalent linear displacement. Tilt perception gain decreased, whereas translation perception gain increased, with increasing frequency. During off-vertical axis rotation, the phase of tilt perception and of translation perception did not vary across stimulus frequencies. These motion paradigms elicited similar responses in roll tilt and interaural perception of translation, with differences likely due to the influence of naso-occipital linear accelerations and input to the semicircular canals that varied across motion paradigms.

Figure 1

The 4 motion paradigms used in this study. (A) Off-vertical axis rotation (OVAR) generated linear acceleration along both the interaural axis and the naso-occipital axis without input to the semicircular canals. (B) Variable radius centrifuge (VRC) generated both a centripetal acceleration along the interaural axis and a Coriolis linear acceleration along the naso-occipital axis without input to the semicircular canals. (C) Linear lateral acceleration (LLT) generated linear acceleration along the interaural axis at low frequency (due to the limited sled length without naso-occipital linear acceleration and without input to the semicircular canals. (D) Earth horizontal axis rotation (EHAR) generated a linear acceleration along the interaural axis without naso-occipital linear acceleration, and input to the semicircular canals declined at low frequency of rotation.

Figure 2

Modulations of linear acceleration along the interaural axis (solid) and the naso-occipital axis (dashed) during the 4 motion paradigms generating a 10° tilt magnitude. (A) Off-vertical axis rotation (OVAR) was consistent across all frequencies (velocities). (B) Variable radius centrifuge (VRC) generated a Coriolis linear acceleration that increased with stimulus frequency. (C) Linear lateral acceleration (LLT) generated pure linear acceleration along the interaural axis with no naso-occipital linear acceleration. (D) Earth horizontal axis rotation (EHAR) generated a linear acceleration along the interaural axis due to roll tilt relative to gravity with no naso-occipital linear acceleration.

Figure 3

Mean ± standard error of tilt gain and translation gain (from 12 subjects) as a function of frequency during each of the 4 motion paradigms generating an equivalent tilt of 10°. Tilt gain is the ratio of perceived tilt angle and equivalent (i.e., the actual or the resultant tilt from linear acceleration and gravitation acceleration stimuli) tilt angle. Translation gain is the ratio of perceived translation displacement and the equivalent linear displacement.

Figure 4

Data from Fig. 3 superimposed in a single diagram. The closed symbols and the solid fit lines show the tilt gains across all motion paradigms. The open symbols and the dashed fit lines show the translation gain across all motion paradigms. Curve fittings are linear curve fits. The crossover frequency between tilt gain and translation gain occurred between 0.15 Hz and 0.6 Hz depending on the motion paradigms.

Figure 5

Mean ± standard error of the phase of the joystick tilts and translation displacements as a function of motion frequency during off-vertical axis rotation (OVAR) in 12 subjects.

Figure 6

Mean ± standard error of perceived amplitudes of tilt and perceived amplitudes of translation as a function of stimulus tilt angle during the 4 motion paradigms at 0.15 Hz and at 0.6 Hz. Refer to Table 1 for the clarification of trials conducted at each angle and frequency.