The response characteristics of seated human subjects exposed to fore-aft ( x -axis) and lateral ( y -axis) vibration are investigated through measurements of dynamic interactions between the seated body and the seat pan, and the upper body and the seat backrest. The experiments involved: (i) three different back support conditions (no back support, and upper body supported against a vertical and an inclined backrest); (ii) three different seat pan heights (425, 390 and 350 mm); and three different magnitudes (0.25, 0.5 and 1.0 m/s² rms acceleration) of band limited random excitations in the 0.5-10 Hz frequency range, applied independently along the fore-aft and lateral directions in an uncoupled manner. The body force responses, measured at the seat pan and the backrest along the direction of motion, are applied to characterize the total body apparent mass (APMS) reflected on the seat pan, and those of the upper body reflected on the backrest. Unlike the widely reported responses of seated occupants under vertical vibration, the responses to horizontal vibration show strong effect of excitation magnitude. The large displacements at lower frequencies cause considerable rotations of the upper body, and the knees and ankles, particularly when seated without a back support, which encouraged the occupants to continually shift larger portion of the body weight towards their feet. This together with the strong dependence on the excitation magnitude resulted in considerable inter-subject variability of the data. The addition of a back support causes stiffening of the body to limit the low frequency rocking motion of the upper body under x -axis motion, while considerable dynamic interactions with the backrest occur. The mean apparent mass (APMS) responses measured at the seat pan and the backrest suggest strong contributions due to the back support condition, and the direction and magnitude of horizontal vibration, while the role of seat height is important only in the vicinity of the resonant frequencies. In the absence of a back support, the seat pan responses predominate at a lower frequency (near 0.7 Hz) under both directions of motion, while two secondary peaks in the magnitude also occur at relatively higher frequencies. The addition of back support causes the seat pan response to converge mostly to a single primary peak, resulting in a single-degree-of-freedom like behavior, with peak occurring in the 2.7-5.4 Hz range under x -axis, and 0.9-2.1 Hz range under y -axis motions, depending upon the excitation magnitude and the back support condition. This can be attributed to the stiffening of the body in the presence of the constraints imposed by the backrest. A relaxed posture with an inclined backrest, however, causes a softening effect, when compared to an erect posture with a vertical backrest. The backrest, however, serves as another source of vibration to the seated occupant, which tends to cause considerably higher magnitude responses. The considerable magnitudes of the apparent mass response measured at the seat back under fore-aft motions suggest strong interactions with the backrest. Such interactions along the side-to-side motions, however, are relatively small. The results suggest that the biodynamic characterization of seated occupants exposed to horizontal vibration requires appropriate considerations of the interactions with the backrest.