Biomedical Engineering Reference
In-Depth Information
along the future path (as in Fig.
4.4
a). The distance to the point of convergence and
the amount of time remaining until the gap closed was manipulated across trials such
that gaps ranged from easily passable to impassable. The cylinders disappeared 1 s
after they began moving and within 1.2 s of the onset of cylinder motion, subjects
had to press one of two buttons on the remote mouse to indicate whether they could
have safely passed through the gap before it closed.
According to the hypothesis introduced in the previous section, judging the pass-
ability of a shrinking gap requires detecting information about the minimum loco-
motor speed needed to safely pass through the gap
, which in turn requires
the visual system to factor out the influence of self-motion. Therefore, manipula-
tions of visual and/or non-visual self-motion information should affect judgments of
passability.
To investigate the influence of non-visual self-motion information (e.g., proprio-
ception), [
8
] took advantage of the fact that such information, to be useful, must be
continually recalibrated. This is because the relation between non-visual self-motion
information and self-motion is dependent upon factors such as surface compliance
and load that can vary from situation to situation. For such information to be useful
for the purposes of perceiving self-motion, it must be possible to recalibrate when
conditions change. In a VE, recalibration can be brought about by increasing or
decreasing the speed with which subjects move through the VE relative to the physical
world (i.e., the visual gain). Fajen and Matthis increased visual gain to 1
(ν
min
)
, which
means that subjects moved through the VE 50 % faster than they moved through
the physical world. The manipulation of visual gain affects the relation between
non-visual self-motion information and self-motion in the VE. Therefore, as sub-
jects moved around the VE with the increased visual gain, non-visual self-motion
information became recalibrated—that is, subjects learn to attribute more optic flow
to their own actively generated self-motion. If subjects rely on non-visual self-motion
information to factor out the influence of self-motion, then the component of optic
flow that is attributed to self-motion should be greater when subjects are calibrated to
the faster-than-normal visual gain (see Fig.
4.6
). The remaining component (i.e., the
component that is attributed to object motion) should be less. Therefore, subjects
should perceive that the obstacles will converge toward a point that is farther away
along the locomotor axis and should be less likely to perceive that the gap is pass-
able. This prediction was tested by comparing passability judgments when subjects
were recalibrated to the faster-than-normal visual gain versus when they were cali-
brated to a normal visual gain. As predicted, subjects were less likely to perceive the
gap as passable when they were calibrated to a faster-than-normal visual gain. Such
findings indicate that non-visual self-motion information plays a role in recovering
the object-motion component of optic flow, which is required to detect information
about
.
5
×
ν
min
. In a follow-up experiment, Fajen and Matthis (in press) demonstrated
that visual self-motion information also plays a role.
To summarize, when people move in the presence of other moving objects,
the optic flow field is the vector sum of the self-motion and object-motion com-
ponents. Visual information that is relevant to perceiving affordances such as the
passability of a shrinking gap is found in the object-motion component of optic
