Vection refers to the perception of self-motion induced by visual stimuli. An example of this sensation is that if one is seated in a train parked at the station while the train next to it pulls out from the station, one may have a sustained sensation that one is moving in the opposite direction. A possible close relationship between presence and vection will be introduced in Section 3.3.3. In view of this relationship, prior work on vection measures is of special interest. This section borrows from .
The simplest types of vection are circular and linear. To induce circular vection, participants may be seated in a chair surrounded by a cylinder (often painted with stripes) which rotates around the participant. Linear vection is typically induced by a display in which objects seem to be approaching or receding.
There is a good discussion of vection measures in Carpenter-Smith et al. . Most previous vection studies have been based on a magnitude estimation measure, in which a participant is requested to assign numbers or joystick positions to perceptions. The deficiencies of magnitude estimation were discussed above.
There is a small but interesting literature concerning other measures for vection. This literature can be thought of as dividing into threshold [111,12] and nulling [112,48,49,50] studies. In threshold studies, one investigates how visual stimuli affect the minimally detectable magnitude of inertial motion (or conversely, how inertial motion affects the onset of vection). Young et al.  examined the interaction of visual and inertial rotation cues, by placing participants on a rotatable chair surrounded by a stripe pattern rotating at constant angular velocity. Among their findings were higher thresholds for the detection of inertial acceleration when the inertial cues conflict with the vection cues. Berthoz et al.  placed participants in a cart which moved linearly and induced vection by providing moving images in the lateral field. Their report includes vection onset thresholds in the range of the threshold for visual motion detection, indicating the importance of vision to the perception of motion. A disadvantage of threshold studies is possible variance due to participants adopting different confidence criteria for reporting threshold.
In nulling studies, one set of stimuli are opposed by another and participants are asked to determine the point at which the two stimuli counterbalance each other. Zacharias and Young  set up a circular vection nulling experiment. Participants were asked to maintain a stationary position by adjusting their inertial rotation, in the presence of a rotating visual surround and inertial disturbance. Other research using visual-inertial nulling is described by Huang and Young: for yaw rotation ; lateral motion ; and roll and pitch . These papers developed models for visual-inertial interaction.
Related but distinct is the research by Carpenter-Smith et al. . Prone subjects were translated along their head x-axis (fore-aft). In the presence of various inertial and visual surround conditions, participants were asked to report their direction of motion. By running many trials for each participant in each condition, a point of subjective equality (PSE), at which participants would think themselves as at rest, could be determined mathematically for each condition. Shifts in PSE as a result of changes in the visual surround were used to develop, for the first time, a scale for linear vection. (This is not to say that a scale for vection cannot be developed using the more traditional techniques, only that it had not been previously done.)
The presence measurement experiments described in Chapter 4 are an extension and refinement of the procedure reported by Carpenter-Smith and his colleagues. An attempt is also made to relate visually-induced self motion, indicated by the visual-inertial nulling procedure, to self-reported presence.