Project topics and materials
Get Complete Project Material File(s) Now! »
Determining the orientation of common objects
Using a same or mirror image discrimination task, Shepard and Metzler (1971) found the time to indicate two images of three dimensional novel asymmetrical objects were the same or not increased as a linear function of the angular disparity between the two views, regardless of the axis of rotation27. In explanation of this phenomenon, Shepard and Metzler reported their participants could “imagine the rotation around whichever axis was required with equal ease” (p. 702), and that their participants claimed to “imagine one object as rotated into the same orientation as the other” (p. 701). Shepard and Metzler coined the act of rotating an imagined object in an analogue fashion “mental rotation”, but they noted “the response times necessarily include any time taken by the subjects to decide how to process the pictures” (p. 703, emphasis added).
In a “same-different” mental rotation task, such as that created by Shepard and Metzler (1971), at least two criteria must be met in order for mental rotation to occur. First, as noted by Shepard and Metzler (1971), the two objects must not differ in a way that would allow their difference to be determined independently of their orientation. Using an object and its mirror image or two identical objects ensures there are no such differences (Cooper & Shepard, 1973b; Shepard & Metzler, 1971). If two objects differed by a feature or features that could be detected independently of their orientation (e.g., very different colours) then no effect of orientation would be found for a “different” response. Second, before either of the objects can be mentally rotated into the view of the other, the relative orientations of both objects must be determined. Primary support for this claim comes from the mental rotation function itself. The response time increase as a linear function of the difference in orientation (as found by Shepard & Metzler, 1971) indicates that the objects are mentally rotated into alignment via the shortest angular distance. However, unless the orientations of both objects has been determined, the direction of the shortest angular distance could not be known (But see Ullman, 1996). In other words, it does not appear that the objects are mentally rotated in order to determine their orientation, as has been suggested at times (De Caro & Reeves, 2000; Rock, 1973). Without knowledge of the orientations, objects would be rotated in a random direction until they were aligned. This would result in a flat response time function across 27 “Rotation” refers to dynamic movement in space, while “orientation” refers to static position in space. orientation, as half of the objects would be randomly rotated the short direction, and half
rotated the long direction.
The aim of the following three experiments was to determine if in fact the orientation of objects is determined independently of their orientation and, therefore, without the aid of any normalisation28 process such as mental rotation. The results of two previous studies provide some support for this claim. The first study was by Corballis, Zbrodoff, Shetzer and Butler (1978), who instructed participants to respond “yes” only to alphanumeric stimuli at a “target” orientation and “no” to alphanumeric stimuli at all other orientations. Averaged across the various target orientations, Corballis et al. found that responses to orientations which were further from upright (120º, 180º and 240º) were not slower than responses to the more upright orientations (300º, 0º and 60º). The second study was by Maki (1986). Maki presented common objects at 4 orientations and instructed participants to move a joystick in the direction of the tops of these objects. In her Experiment 3, the objects were presented at 0º, 90º, 180º, or 270º, and in Experiment 4, at 45º, 135º, 225º, or 315º. While Maki did not directly test the difference in response times to objects at 0º and 180º, based on her Figure 3 (p, 378), the difference appears to be approximately 25ms which is much smaller than the smallest significant difference Maki found in Experiment 3 (78ms). More conclusive evidence, however, comes from her Experiment 4 in which Maki found that responses to objects that were more upright (at 45º or 315º) were not faster than responses to objects that were more upside down (at 135º or 225º).
The effect of inverted noise on signal identification
The following series of experiments was designed to further test the notion that the identification of letters and “basic” shapes is independent of their orientation. The methodology employed is that of the flanker compatibility task originally developed by Eriksen and Eriksen (1974). The flanker compatibility task was created as a visual search task but with the location of the target and distracting stimuli (distracters / noise) kept constant.
By keeping the stimuli locations constant the extent to which responses to the target were influenced by the similarity between the target and distracters (e.g., shared response or imilar shapes) could be measured. In the task, a target usually appears in the middle (at the location of focused attention) and is flanked by distracters. Participants are instructed to ignore the flanking stimuli and respond only to the central target. Eriksen and Eriksen (1974) used letters as targets and distracters. In their experiments, two letters required a left response and two a right response. The distracters were the same letter as the target (identical condition), a different letter but with the same response as the target (response compatible condition), a letter with a different response (response incompatible condition), a similar letter but not associated with a response (similar neutral condition), or a dissimilar letter again not associated with a response (dissimilar neutral condition). They also varied the spacing between the letters. They found response times were fastest when all the letters were identical, slowest when the flanking letters were associated with the other response (response incompatible condition), and intermediate for the other three conditions.
In relation to visual similarity, Eriksen and Eriksen (1974) found response times to the target letters were faster when flanked by letters that were similar to the targets (similar neutral condition) than when flanked by letters that were dissimilar. However, this difference only occurred when targets to noise letter distance was 0.5o 17 and 1o from the target. At 0.06o, however, there was no difference in response times to targets when flanked by noise similar and noise dissimilar letters. However, fewer mistakes were made in the similar neutral condition than dissimilar neutral condition. This issue was followed up by Yeh and Eriksen (1984) who found that visual similarity between the targets and the flanking letters (flankers) influenced responses more than name similarity. Yeh and Eriksen (1984) used upper and lower case letters that could be presented with upper or lower case flankers that were of 17 Here visual angle is used to indicate the distance in visual space (e.g., the distance between stimuli on a computer screen). 1o of visual angle equates to 1cm of distance when viewed at a distance of 57cm. shape.
Chapter 1: General Introduction
Overview
Templates
Detection Theory
Implications of this History on the Current Research
Evidence of Object Categorisation in the Brain
Evidence of Object Categorisation in Responses
Mental Rotation
Models of Matching External with Internal Objects
The Current Research
Chapter 2: The effect of inverting noise on signal identification
Introduction
Experiment 2.1
Experiment 2.2
Experiment 2.
General Discussio
Chapter 3: Determining the orientation of common objects
Introduction
Experiment 3.1
Experiment 3.2
Experiment 3.3
Introducing FAT
General Discussion
Chapter 4: The “Which way?” experiments
Introduction
Experiment 4.1
Experiment 4.2
Experiment 4.3
General Discussion
Chapter 5: Evidence of a digital brain
References
