Small lights were attached to the backs of walkers at eight main points of
their bodies, i.e., the shoulders. hips, knees and heels. The walkers. four
men and four women aged 18 to 62, walked on a moving belt at a speed of 1.
0 to 3.0 km/h. Their gaits, outlined by point-lights, were videotaped and p
rojected on a screen. A total of 420 participants, 212 female and 208 male,
were asked to judge the gender and age of the point-light walkers (PLWs).
Our findings were that (1), there were 72% correct gender-identifications w
ith the figure bring 85% or more for older walkers; (2), the gaits of two y
oung PLWs were perceived as those of old persons. The reason for this appea
red to be correlative changes between the movements of the individual ports
of the body, especially of shoulders and heels: (3), signal detectability
(d') for gender seemed to be a useful measure of gender-identification. The
above findings suggest that both back-view and side-view observations of t
he PLWs provide valuable information for gait perception.