Game Development Reference
the Sciences of Human Movement), these techniques constitute a significant advance-
ment in the feasibility of the study of the working of the human subject (Cruz-Neira,
1998; Gaggioli, 2003), in his abilities to process information, motor control and cog-
nitive processing (sensorimotor determinants of cognition and cognitive determinants
of activity control). The introduction of virtual reality techniques is thus in line with
an “interactive'' approach of the “perception-action'' looping and of the “in real time''
study of the sensorimotor transformation.
If we refer to the approach of Loomis et al. (1999), in the field of Behavioural
Sciences we can follow the historical idea of the search for a compromise between the
ecological validity of a laboratory study and the experimental control required for the
scientific and objective study of a behaviour. In other words, we still oscillate between
the necessary scientific reductionism and the desire to study a situation as close as
possible to reality (“we must bring the real world into the laboratory'', wrote Gibson
(1979)). If we consider a “precomputer'' period (before 1950) as the starting point, we
can cite the tachistoscope as an example of a device that makes it possible to present
single images for very brief periods. The advantage of the device was the maximum
control that the experimenter had on the stimulation presented to the subjects of the
experiment, thus making it possible to draw conclusions as regards the processing of the
visual information by the subjects. However, the tachistoscope enables the presentation
of still, two-dimensional images. The ecological validity of the conclusions that it helps
to achieve as regards visual perception is thus limited. In this “precomputer'' period,
“on-the-field'' experimentation was most often limited to observations and interviews.
Under such conditions, the ecological validity of the behaviour studied is high,
as the “subject'' was in a natural situation, interacting with the real world. Thus, the
problem is that the experimental control was close to zero.
With respect to this “precomputer'' period, the development of computer tools
has made it possible to bend this linear relation, in the sense of an experimental con-
trol becoming possible under relatively ecological conditions. It can be noted that the
progress of computer graphics has made it possible to present the subject with sensory
stimulations that are much more complex (and kinematic) than those with the tachisto-
scope, under controlled conditions. However, it must be noted that we are most often
at the level of a psychophysical approach of perception, in which the subject remains
a “passive perceiver''. We can also observe that the progress of the computerised sys-
tems for recording behaviour has made it possible to go past the stage of a subjective
behaviour evaluation. However, the experimental control has remained limited.
The use of virtual reality thus makes it possible to get closer to an ideal objec-
tive in which experimental control co-exists with the ecological validity of the studied
behaviour. It is therefore about bringing the real world into the laboratory, while
emphasising the multi-modal (and controllable) character of the sensory stimulations
(visual, sound, haptic, vestibular, etc.) and the interaction. For example, this is exactly
what flight and driving simulators do. The experimental situation is perfectly con-
trolled by the designers of the device, and the subject is placed in a situation of
interaction with this virtual environment. It can be said that we are getting closer
to the “psychophysics of the activity'' (Flach, 1990), in which the “perception-action''
coupling is respected in the experimental approach.
An example of such an approach is represented by the studies on visual control
of locomotion (Buekers et al., 1999), in which a moving carpet is connected to a