When we want to know which way is up, we can use any of the following sources of information.Non-visual information
1. The otolith organs on each side of the head contain small crystals attached to hairs. When the head tilts, the hairs bend and generate nerve impulses that indicate the direction of head tilt.
2. Touch and pressure sense organs in the skin indicate the direction of pressure arising from the support surface.
3. Kinaesthetic sense organs in muscles and joints indicate the direction in which gravity pulls the limbs.
4. The body axis. Normally, the feet are down and the head is up. Astronauts floating in space often feel down to be where the feet are.
Visual information
1. The visual frame. Surfaces such as walls and floors are normally horizontal or vertical. Also, many lines, such as the horizon and the corners of rooms, are horizontal or vertical. Such surfaces and lines are called the visual frame. A simple visual frame indicates horizontal and vertical but does not indicate which way is up.
2. Intrinsic visual polarity. Objects such as houses, people, and trees have a recognizable top and bottom and usually have a consistent orientation to gravity.
3. Extrinsic visual polarity. Spatial relationships between objects can indicate the direction of gravity. For example, an object supported on another object, such as a box on a shelf, a falling object, and an object hanging on a string, possess extrinsic polarity.
Witkin pioneered the study of the effects of tilting the visual frame on the perception of the vertical. He found that a luminous square frame in dark surroundings tilted 28 degrees in the frontal plane causes a vertical rod to appear tilted by 6 degrees on average in the opposite direction (Witkin 1949). This is known as the rod and frame effect. Howard and Childerson (1994) asked observers to set a rod to the apparent vertical as they sat erect in an empty cubic room inclined to various angles about the line of sight (the roll axis). Settings were accurate whenever an axis of symmetry of the room was aligned with the observer's body axis, which occurred at 45-degree intervals. At intermediate angles, the rod was set on average 4 degrees in the direction of tilt of the nearest axis of symmetry. Thus a stationary tilted frame has only a limited effect on the perceived vertical. The non-visual senses restrain the effects produced by a tilted frame.
But what happens when we observe a rotating scene? We measured the pure effects of motion by sitting people at the centre of a 3-metre-diameter sphere lined with dots. Continuous rotation of the sphere at 30 degrees about the roll axis induced an illusion of continuous self-rotation accompanied by a paradoxical sensation of being tilted by up to 20 degrees (Howard and Childerson 1994). Since a sphere contains no visual information about the direction of gravity these illusions must have been induced by visual motion alone. We then added a visual frame by placing people in a rotating cubic room. This also produced illusory self-rotation but people felt upright every time an axis of symmetry became aligned with the subject's body axis.
We then asked what happens when we add visual polarity. Wertheimer (1912) reported that a furnished room seen in a mirror so that it was optically tilted 45 degrees appeared gradually to right itself. Asch and Witkin (1948) had subjects look into a small furnished room tilted 22 degrees about the roll axis. A vertical rod appeared tilted about 15 degrees in the opposite direction. We obtained similar limited illusory tilts of a rod, or of the self, when we placed people in a fully furnished room containing a table, chairs, shelves with objects on them, hanging pictures, and a multitude of other objects with intrinsic and/or extrinsic polarity, as shown in Fig. 1. Thus, a tilted polarized scene induces larger displacements of the vertical than does a simple visual frame.
Finally, we asked what happens when people are placed in a fully furnished room that rotates about the roll or pitch axis. In a fairground device built in Los Angeles towards the end of the last century, observers sat in a stationary gondola suspended in a furnished room. When the room rocked to and fro, the people in the stationary gondola felt that the room was stationary and that they and the gondola had rocked. In Germany in 1937, Kleint placed people in a furnished room that rotated completely about the roll axis. Some subjects reported sensations of total self-rotation but no quantitative data was presented. We found that 80 per cent of observers experienced complete head-over-heels illusory tumbling of the self about the roll or pitch body axis when sitting upright in our furnished room that rotated completely about the roll or pitch axis (Howard and Hu 2001).
We concluded that the non-visual senses are overwhelmed by vision only when a richly polarized room rotates around the observer. We agreed with other investigators that a stationary tilted visual scene, even when it contains a rich variety of information about the direction of gravity, has only a limited capacity to overcome the non-visual senses. But the experiments we conducted next showed that we were wrong.
In the experiments mentioned so far, the room was tilted with respect to both gravity and the observer's body axis. We asked what happens if the observer and the room are tilted together so that the room remains aligned with the body axis. We slowly rotated the furnished room and the observer through 90 degrees so that the observer ended up supine looking up at the same wall that had been vertically in front before the motion started. For many subjects something remarkable happened. Many observers felt that neither they nor the room had moved. At the end of the rotation, they felt upright in an upright room. When they moved their arms out from the body they felt as if they were weightless. An object hanging down from the wall above them looked as if it were magically suspended in space. We had a professional magician who thought that we had suspended the object by a jet of air! We tested nine NASA astronauts and the five that experienced the illusion reported that it felt just like being in zero gravity. We call this the 'levitation illusion'. We then rotated the room and the observer together through 180 degrees. Many observers reported that they were upright in an upright room when in fact they were upside down in an upside-down room. These effects are due entirely to the presence of familiar polarized objects. There is no motion and the visual frame (the corners and surfaces of the room) is not tilted. Thus, for many people, the visual polarity of a stationary scene can completely overcome conflicting information about the direction of gravity arising from the non-visual senses. This happens when the body axis and the up–down polarity axis of the room are aligned.
We then constructed a full-scale replica of NASA's Space Lab, as shown in Fig. 2. Space Lab is where astronauts conducted experiments in the weightless conditions of space before the advent of Space Station. Space Lab contained virtually no familiar polarized objects. We found that observers did not experience the levitation illusion in this environment. Space Station is also devoid of familiar polarized objects. Thus, the visual environments in which astronauts work do not provide a stable framework for maintaining a consistent sense of orientation. In space, the non-visual senses do not provide information about orientation and the Space Station does not contain a stable visual substitute for gravity. It is no wonder that astronauts experience disorientation.
We found some strong age effects. The illusion of self-tilt in the rotating sphere and the head-over-heels illusion of self-rotation induced by the rotating furnished room were experienced by most subjects of all ages. It is known that the effects of visual motion on the perception of body orientation develop in very young children and are served by primitive postural reflexes. On the other hand, the proportion of observers experiencing the levitation illusion increased with age (Howard, Jenkin, and Hu 2000). Only 20 per cent of a group of 12-year-olds experienced the illusion but 80 per cent of 70-year-olds experienced it. There are two reasons for these age effects. First, the otolith organs become less sensitive as we get older. Secondly, the levitation illusion depends entirely on our familiarity with the normal orientation of objects (visual polarity). Older people come to rely on this knowledge as their otolith organs lose their sensitivity. Old people are prone to fall and our results predict that they would be especially prone to fall in a place like a stairwell where surfaces are not horizontal and where there is an absence of familiar objects. Our work was funded by NASA and we are recommending that astronaut disorientation would be less severe if the walls of Space Station contained pictures of familiar objects.
Fig. 1. A veiw of the interior of a furnished room.
Fig. 2. Full-scale replica of NASA'S Space Lab.
(Published 2004)
— Ian P. Howard
- Bibliography
- Asch, S. E., and Witkin, H. A. (1948). 'Studies in space orientation: II. Perception of the upright with displaced visual fields and with body tilted'. Journal of Experimental Psychology, 38.
- Howard, I. P., and Childerson, L. (1994). 'The contribution of motion, the visual frame and visual polarity to sensations of body tilt'. Perception, 23.
- — — and Hu. G. (2001). 'Visually induced reorientation illusions'. Perception, 30.
- — — Jenkin, H. L., and Hu, G. (2000). 'Visually induced reorientation illusions as a function of age'. Aviation, Space and Environmental Medicine, 71.
- Kleint, H. (1937). 'Versuche über die Wahrnehmung II. Über Bewegung'. Zeitschrift für Psychologie, 141.
- Wertheimer, M. (1912). 'Experimentelle Studien über das Sehen von Bewegung'. Zeitschrift für Psychologie und Physiologie des Sinnesorgane, 61.
- Witkin, H. A. (1949). 'Perception of body position and the position of the visual field'. Psychological Monographs, 6 (whole 7).
