This article was downloaded by: [130.132.123.28] On: 29 December 2014, At: 15:23 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
The Journal of General Psychology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/vgen20
Recognition of Altered Shapes by Hearing and HearingImpaired Subjects a
Kathleen C. Chen & Mou-Ta Chen
b
a
Behavioral Science Division: Psychology , Rochester Institute of Technology , USA b
Department of Mathematics , State University of New York , College at Brockport, USA Published online: 06 Jul 2010.
To cite this article: Kathleen C. Chen & Mou-Ta Chen (1990) Recognition of Altered Shapes by Hearing and Hearing-Impaired Subjects, The Journal of General Psychology, 117:1, 27-37, DOI: 10.1080/00221309.1990.9917770 To link to this article: http://dx.doi.org/10.1080/00221309.1990.9917770
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the Content.
Downloaded by [] at 15:23 29 December 2014
This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
The Journal ofGeneral Psychology, 117(1), 27-37
Recognition of Altered Shapes by Hearing and Hearing-Impaired Subjects
Downloaded by [] at 15:23 29 December 2014
KATHLEEN C. CHEN Behavioral Science Division: Psychology Rochester Institute of Technology MOU-TACHEN Department ofMathematics State University ofNew York, College at Brockport
ABSTRACT. In the present study, we explored whether mental transformation or feature detection is involved in visual processing of shape recognition; another purpose was to compare the visual processing of hearing subjects with that of hearingimpaired subjects. Deprivation of hearing, according to Myklebust (1964), impedes perceptual functioning in some respects and enhances such functioning in other respects. An experiment was conducted using two-dimensional random shapes under similarity transformations in Euclidean space. Reaction time served as the behavioral measure. The analysis of variance results showed significant main effects of orientations, isometries, and shapes, but not of groups and sizes. There were significant linear trends for orientations and sizes. Both groups demonstrated mental isometric transformation.
HUMAN BEINGS ARE CONTINUALLY FACED with the problem of recognizing shapes or objects in varying orientations. Milner (1974) suggested that recognition of shapes might be achieved by the extraction of features that are independent of orientation, spatial location, and mirror reflections. Similar to this view is the idea proposed by Corballis (1988) that recognition of a familiar shape independent of its orientation may be accomplished by extracting a description of the shape that is frame-independent or independent of any coordinate system. Such a description is usually sufficient to find the stored representation of the shape in long-term memory.
Requests for reprints should be sent to Kathleen C. Chen, College of Liberal Arts. Rochester Institute ofTechnology, One Lomb Memorial Drive, Rochester, NY 14623. 27
Downloaded by [] at 15:23 29 December 2014
28
The Journal of General Psychology
Palmer (1983) proposed another view, stating that shape constancy is accomplished by using a reference-frame hypothesis while focusing on properties that vary over the transformations of a group. He assumed that the effects of transformations were neutralized by imposing an intrinsic frame of reference, thereby achieving shape constancy. The intrinsic frame corresponds with the structure of the figure. He believed that when intrinsic frames are appropriately chosen, they can compensate for structural changes induced by similarity transformations: translations, rotations, reflections, dilations, and their composites. Palmer chose similarity transformations in Euclidean space in his analysis of perceptual organization because similarities are preferred by the human visual system to other transformations. A type of mental transformation was demonstrated by Cooper and Shepard (1973). They provided empirical evidence that mental rotation is involved in judgments of disoriented shapes. Subjects were timed as they decided whether rotated alphanumeric characters were normal or mirror reversed; reaction times increased with the angular departure of the characters from the normal upright orientation. Similar results were obtained by Cooper (1975) using random two-dimensional shapes. Those results have been taken as evidence that mental rotation is an analogue process and that mentally rotated shape representation is at some level the same as that of a shape that has been physically rotated. Identification of shape independent of size has long been considered a problem (Bundesen & Larsen, 1975; Kohler, 1929). One type of explanation of this ability involves the assumption that we transform a mental image of an object of a given size to an image of an object of the same shape but of another size (Bundesen, Larsen, & Farrell, 1981; Posner, 1969). Another type of explanation is based on direct extraction of size-invariant features from the sensory input (Dodwell, 1970; Gibson, 1969). The purpose of the present study was to determine whether mental transformation or feature detection is involved in the visual processing of shape recognition and to compare the perceptual processing of hearing and hearingimpaired subjects. According to Myklebust (1964), deprivation of hearing impedes perceptual functioning in some respects and enhances such functioning in other respects. Hearing-impaired subjects' visual processing may be enhanced because of auditory deprivation.
Method The present investigation used two-dimensional random forms under similarity transformations in Euclidean space. A transformation of a plane is a oneto-one map of the points of the plane onto themselves. An isometry is a transformation that preserves distance between two points. Because of its distancepreserving property, an isometry also preserves the shape and the size of a
Chen & Chen
29
figure. A plane isometry is a translation, a rotation, a reflection, or a glide reflection, which is a composition of a reflection and a translation. A dilation is an enlargement or a contraction of a ratio about a point. A similarity is one of the following: an isometry, a dilation, or a composition of an isometry and a dilation (Martin, 1982). Mental isometric transformation will lead to increase in reaction time as the angular departure of shape from identity is increased or as the size changes from its standard size.
Downloaded by [] at 15:23 29 December 2014
Subjects Eight hearing students of the Rochester Institute of Technology and eight hearing-impaired students of the National Technical Institute for the Deaf, Rochester Institute of Technology, served in the experiment. There were equal numbers of male and female subjects in both groups, and all subjects had 201 20 or corrected 20/20 vision.
Stimulus Materials Stimuli were computer generated. Three random shapes of 6, 12, and 24 points generated by Attneave and Amoult (1956) were used (Figure 1). The three shapes and their reflected images represented three levels of rated complexity and three sizes of linear scales: 1.6,0.8, and 0.4. The isometries used were rotations and reflections. Each shape of a scale had four orientations: 0°, 60°, 120°, and 180° (clockwise angular departures from the standard shape for rotation) and 0°, 30°, 60°, and 90° (angular departures of the mirrors for reflection). The angular departure of the position of a reflected image was twice as much as the angular departure of its mirror. With three shapes in three sizes, two isometries, and four angular orientations, there were a total of 72 shapes (3 x 3 x 2 x 4). The smallest display subtended a visual angle of approximately 3° 54'; the medium display, an angle of 7° 48'; and the largest display, an angle of 15° 36'.
Procedure Both hearing and hearing-impaired groups followed the same procedure. The hearing-impaired subjects were given the same written instructions on the screen as the hearing subjects. Extra instruction was given to subjects by written verbal communications or sign language to ensure that they understood the instructions before the test session. A sign-language interpreter served as an assistant when needed. The procedure was programmed in Pascal language for a Gigi digital terminal and a Barco CD33 screen. Each subject was tested individually. During three training trials, subjects studied drawings of three standard random shapes of 6, 12, and 24 points of the 1.6 scale for 2 min each.
Downloaded by [] at 15:23 29 December 2014
30
The Journal of General Psychology
FIGURE 1. Standard forms.
They were instructed that forms would be presented to them on the screen one at a time and that they were to press the left-hand key if the displayed form was a rotated image of one of the standard forms. If the image was identical to one of the standard forms, they were to consider it a 0° rotated image of the standard form and to press the left-hand key accordingly. If the displayed form was a reflected image of one of the standard forms, they were to press the right-hand key. They were also informed that the forms displayed could differ in size from the standard forms and that they did not have to indicate what the size of the form was. They were instructed that if the size of the stimulus form was different from the size of one of the standard forms, they should continue to judge whether the form with size variation was a rotated or reflected image of one of the standard forms by pressing the leftor right-hand key. A visual display of two colored keyboards indicating all
Downloaded by [] at 15:23 29 December 2014
Chen & Chen
31
possible left- or right-hand keys that could be used was shown after the instructions. Each form appeared for 0.5 s, centered within a circular field with a black surround. Subjects' key-presses terminated the stimulus display. Following a key-press, a I-s cyan warning field preceded the presentation of the next stimulus. During training trials, subjects' responses were followed by the words "right" or "wrong" displayed on the screen. A I-hr test session that consisted of four trials immediately followed the training period. Of the 72 test figures given in each trial, there were three forms each in three sizes, two isometries, and four angular orientations. The order of presentation for each trial was randomized by the computer; reaction time (RT) and errors were recorded by the computer also. Test figures on which errors were made were presented again at the end of each trial so that we would obtain a complete set of error-free data for each subject. For each subject, the complete set of data consisted of 288 errorless RTs.
Results A six-way analysis of variance (ANOVA) (Subjects x Groups x Forms x Sizes x Isometries x Orientations) was performed. The analysis indicated no significant between-groups effect, F(l, 14) = .23, p > .05; a significant effect offonns, F(2, 28) = 6.18, p < .01; no significant effect of sizes, F(2, 28) = 2.16, p > .05; a highly significant isometry effect, F(l, 14) = 81.54, p < .001; and a highly significant orientation effect, F(3, 42) = 20.62,p < .001. The interaction between sizes and orientations was marginally significant, F(6, 84) = 2.88, P < .05, as was the interaction between isometries and orientations, F(3, 42) = 3.01, P < .05. Other two- and three-way interactions between the main variables (groups, forms, sizes, orientations, and isometries) were not statistically significant. Alteration of Forms The mean RT as a function of forms for both hearing and hearing-impaired groups is shown in Figure 2. A trend analysis showed a significant quadratic effect, F(l, 28) = 11.75, P < .01, but no significant linear trend, F < 1. Comparing the results of hearing and hearing-impaired groups, we found similar effects for both groups, with significant quadratic effects, F(l, 14) = 4.68, P < .05, for the hearing group and F(l, 14) = 9.97, P < .01, for the hearing-impaired group. Alteration of Sizes The combined data of hearing and hearing-impaired groups used to compute the RT function for sizes are illustrated in Figure 3. A trend analysis revealed
32
The Journal of General Psychology
I
,
... .- ..., /
/ I
,
Downloaded by [] at 15:23 29 December 2014
\
I
\
,
,
~
I
1100 1600
I
,
\
I
1800
'", ,
I
\
~
\
I
1
\
2
,
Forms
FIGURE 2. Mean RT as a function of form with 6 points (1), 12 points (2), and 24 points (3).
a significant linear trend, F(l, 28) = 4.30, P < .05, and a nonsignificant deviation from this trend, F < 1. Separate trend analysis of sizes for the hearing group still showed a significant linear trend, F(l, 14) = 6.27, P < .05, but that for the hearing-impaired group did not approach statistical significance, F < 1. The hearing-impaired group tended to react with longer RTs and lower error rates than the hearing group. However, the ANOVA showed no significant difference between RTs, and a t test showed no significant difference in error rates. Size effect became more pronounced for the hearing group when reflection was included with rotation in the analysis of total subjects. This is shown in the significant difference (by t test, p < .05) in RTs of rotated and reflected forms in the standard size with the 1.6 scale used in training. A similar analysis for the hearing-impaired group did not approach the significance level of .05. Orientations
The linearity of the RT function of orientations for the combined hearing and hearing-impaired groups is shown in Figure 4. Regression for orientations
Chen & Chen
33
1850
1800
Downloaded by [] at 15:23 29 December 2014
1750
1'100
·2
.,
Size
FIGURE 3. Mean RT as a function of size with scales 1.6 (SI)' 0.8 (8J, and 0.4 (83) ,
yielded a linear trend that was highly significant, FO, 42) = 61.57, p < .001, but no significant quadratic or higher order effects, F < I. When hearing and hearing-impaired groups were compared, there were similar highly significant linear trends for the hearing group, FO, 21) = 27.44, P < .001, and the hearing-impaired group, FO, 21) = 34.90, p < .001. Again there were no significant higher order interactions in either group, F < 1.
RotatedForms The hearing and hearing-impaired subjects' mean RTs (averaged over subjects, forms, and sizes) for correctly determining the rotated test forms are also illustrated in Figure 4. RTs for errors are not included in the mean RTs in Figure 4. The most striking feature of the data presented in Figure 4 is the linearity of the increase in RT with angular departure of the test stimulus from the trained orientation. This result supports previous studies using alphanumeric stimuli (Cooper & Shepard, 1973) and random two-dimensional shapes (Cooper, 1975). The error ranges of rotation were between 3% and 20% and 5% and 16% for hearing and hearing-impaired groups, respectively.
34
The Journal of General Psychology
"..
2200 2100
....... 2 1900 1800 v;-
g
",
1700
Eo-
1600
;j
1500
.'" "
'"
" '"'"
'"
." --- ---
0::
Downloaded by [] at 15:23 29 December 2014
G)
::E
--_--..--
1400 1)00
~
Reflection Reflection and Rotation Rotation
1200
-e 600 Angular departure from trained orientation (degrees)
FIGURE 4. Mean RT in isometries as a function of angular departure of test shape from trained orientation.
Reflected Forms
Hearing and hearing-impaired groups' RT functions (averaged over subjects, forms, and sizes) for correctly judging the reflected forms are also plotted in Figure 4. The striking linearity depicted for rotation is also apparent for reflection. The hearing group showed a tendency toward higher error rates than did the hearing-impaired group; however, the differences were not significant (t tests, p > .05) either for orientation in terms of error rates or for RT. The range of error rates in reflection for the hearing group was between 12% and 33% and for the hearing-impaired group, between 2% and 15%. Discussion Mental isometric transformation was demonstrated by the significant main effects of orientation and isometry (rotation/reflection) in ANOVAs for both groups of subjects. Both hearing and hearing-impaired subjects were alike in demonstrating linearity of RT functions for orientation, which has implica-
Downloaded by [] at 15:23 29 December 2014
Chen & Chen
35
tions for the nature of the internal processes involving rotated-reflected discrimination with size variation. Hearing-impaired subjects and hearing subjects demonstrated mental rotation. These results replicate the findings of Cooper and Shepard (1973), Cooper (1975), and Cooper and Shepard (1978). RTs for reflected images were longer than for rotated images; they replicate the findings of Corballis, Zbrodoff, Shetzer, & Butler (1978) that latencies for identifying reflected characters were longer than those for rotated characters. Our results are at odds with the feature detection theories of shape recognition proposed by Deutsch (1962) and Milner (1974) because these theories predict invariance with respect to both angular orientation and mirror reflection. Another tenable explanation for the present results is that there is a process of mirror-image generalization, perhaps involving homotopic recording and transfer of memory traces from one to the other side of the brain so that form registration leaves mirror traces both for the form itself and for its mirror-reflected image (Achim & Corballis, 1977; Corballis & Beale, 1976). Such a process serves as an adaptive function in our everyday world, as it allows us to recognize patterns regardless of left-right orientation. The subjects in our study were exposed in the training session to the standard forms in the version we consider normal. Hence the traces under mirror reflection that depended on interhemispheric transfer were weaker than the normal ones of direct input. Our results may be explained by the reference frame theory, as mental rotation is involved in returning the disoriented form to uprightness with respect to its frame. The identity 0° orientation is invariably the easiest to recognize. Consistent with Palmer, Rosch, and Chase (1981), recognition latency decreases as the form becomes more canonical or upright. Hinton and Parsons (1981) proposed that perception of a shape involves the assignment of an intrinsic frame of reference to shape, and the frame can be either leftor right-handed and can be decided only in an analogue way. Even Corballis' frame-independent position recognizes the role of mental rotation as necessary in mirror-image discrimination (1988). Our procedure involved mirrorimage discrimination in differentiating the two isometries (rotation/reflection). Template theories of visual pattern recognition assume the operation of preprocessing to deal with irrelevant information, such as stimulus size discrepancies. One form of normalizing the stimulus is mental transformation of size. Current literature shows widespread acceptance of the size transformation explanation (Bundesen & Larsen, 1975; Bundesen et al., 1981; Cooper & Shepard, 1978; Larsen & Bundesen, 1978; Shepard, 1981). In the present experiment, trend analysis showed a significant linear function of RT and size for combined hearing and hearing-impaired groups (Figure 2). The size effect was not significant by ANOVA. Separate analysis
Downloaded by [] at 15:23 29 December 2014
36
The Journal of General Psychology
of the hearing group showed a significant linear trend and also a significant difference by t test (p < .05) of RT of rotated plus reflected forms on the 1.6, 0.8, and 0.4 scales and rotated plus reflected forms on the 1.6 (training) scale. These results suggest that some hearing subjects mentally transformed size to match the form against the internal template on the 1.6 scale. There was no pronounced size transformation effect for the hearing-impaired group. Nevertheless, the combined groups still showed a significant linear trend. This demonstrates some similarity transformation. Some subjects mentally rotated, flipped, and transformed size to match an internal template. It is possible, however, to discriminate forms of different sizes without size transformation. The procedure in this experiment differs from that in the size transformation studies cited above. In those studies, a matching task was performed while both stimulus forms were present in the visual field, whereas our study required the subject to match the stimulus with an internal memory image. Other outcomes of this experiment were the significant quadratic effects between complexity of the forms and RT for both hearing and hearingimpaired groups and a significant main effect of forms by ANOVA. These results are different from Cooper's (1975) findings. One possible explanation of the complexity variations in the forms to produce RT differences relates to the nature of the internal representation or memory image, which is highly schematic in comparison with the rich detail of the form itself. A different explanation is based on the subjects' reports that some achieved the correct response by imagining only a few distinctive features of the test form. It is quite obvious that the l2-point form is the most difficult form. Investigation of the visual functioning of hearing-impaired subjects may have important implications for the psychology of sensory deprivation. We concluded that hearing deprivation did not affect these subjects' visual form perception. Both hearing and hearing-impaired groups performed mental isometric transformations. We also found a relationship between complexity of forms and rate of mental transformation. The ease of recognition depended not on the complexity of forms, but on the distinctiveness of the features of forms.
REFERENCES Achim, A., & Corballis, M. C. (1977). Mirror-image equivalence and the anterior commissure. Neuropsychologia, 15, 475-478. Attneave, F., & Arnoult, M. D. (1956). The quantitative studyof shapeand pattern. Perception, 53, 452-471.
Bundesen, C,; & Larsen, A. (1975). Visual transformations of size. Journal of Experimental Psychology: Human Perception and Performance, 1, 214-220.
Bundesen, c., Larsen, A., & Farrell, J. E. (1981). Visual transformations of size and orientation: Spatial imagery and perceptual processes. In A. Baddeley & J.
Downloaded by [] at 15:23 29 December 2014
Chen & Chen
37
Long (Eds.), Attention and performance IX (pp. 279-294). Hillsdale, NJ: Erlbaum. Cooper, L. A. (1975). Mental rotation of random two-dimensional shapes. Cognitive Psychology. 7. 20-43. Cooper, L. A., & Shepard, R. N. (1973). Chronometric studies of rotation of mental images. In W. G. Chase (Ed.), Visualinformation processing (pp. 75-176). New York: Academic Press. Cooper, L. A., & Shepard, R. N. (1978). Transformations on representations of objects in space. In E. G. Carterette & M. P. Friedman (Eds.), Handbook ofperception (Vol. 8, pp. 104-146). New York: Academic Press. Corballis, M. C. (1988). Recognition of disoriented shapes. Psychological Review. 95(1), 115-123. Corballis, M. C., & Beale, I. L. (1976). Thepsychology ofleft and right. Hillsdale, NJ: Erlbaum. Corballis, M. c.. Zbrodoff, N. J., Shetzer, L. I., & Butler, P. B. (1978). Decision about identity and orientation of related letters and digits. Memory and Cognition. 6(2), 98-107. Deutsch, J. A. (1962). A system of shape recognition. Psychological Review. 69. 491-500. Dodwell, P. C. (1970). Visual pattern recognition. New York: Holt. Gibson, E. J. (1969). Principles ofperceptual learning and development. New York: Appleton-Century-Crofts. Hinton, G. E., & Parsons, L. M. (1981). Frames of reference and mental imagery. In A. D. Baddeley & J. Long (Eds.), Attention andperformance (Vol. 9, pp. 261278). Hillsdale, NJ: Erlbaum. Kohler, W. (1929). Gestaltpsychology. New York: Live-Right. Larsen, A., & Bundesen, C. (1978). Size scaling in visual pattern matching. Journal ofExperimental Psychology: HumanPerception and Performance. 4. 1-20. Martin, G. E. (1982). Transformation geometry. New York: Springer-Verlag. Milner, P. M. (1974). A model for visual shape recognition. Psychological Review. 81. 521-535. Myklebust, H. R. (1964). The psychology of deafness (2nd 00.). New York: Grone and Stratton. Palmer, S. (1983). The psychology of perceptual organization. In J. Beck, B. Hope & A. Rosenfield (Eds.), Human and machine vision (pp. 269-339). New York: Academic Press. Palmer, S., Rosch, E., & Chase, P. (1981). Canonical perspective and the perception of objects. In A. C. Baddeley & J. Long (Eds.), Attentionand performance (Vol. 9, pp. 135-151). Hillsdale, NJ: Erlbaum. Posner, M. I. (1969). Abstraction and the process of recognition. In G. H. Bowen & J. J. Spence (Eds.), Thepsychology oflearning and motivation (Vol. 5, pp. 441(0). New York: Academic Press. Shepard, R. N. (1981). Psychological complementary. In M. Kubovy & J. R. Pomerantz (Eds.), Perceptual organization (pp. 279-341). Hillsdale, NJ: Erlbaum.
ReceivedApril 27, 1989
The Journal of General Psychology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/vgen20
Recognition of Altered Shapes by Hearing and HearingImpaired Subjects a
Kathleen C. Chen & Mou-Ta Chen
b
a
Behavioral Science Division: Psychology , Rochester Institute of Technology , USA b
Department of Mathematics , State University of New York , College at Brockport, USA Published online: 06 Jul 2010.
To cite this article: Kathleen C. Chen & Mou-Ta Chen (1990) Recognition of Altered Shapes by Hearing and Hearing-Impaired Subjects, The Journal of General Psychology, 117:1, 27-37, DOI: 10.1080/00221309.1990.9917770 To link to this article: http://dx.doi.org/10.1080/00221309.1990.9917770
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the Content.
Downloaded by [] at 15:23 29 December 2014
This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
The Journal ofGeneral Psychology, 117(1), 27-37
Recognition of Altered Shapes by Hearing and Hearing-Impaired Subjects
Downloaded by [] at 15:23 29 December 2014
KATHLEEN C. CHEN Behavioral Science Division: Psychology Rochester Institute of Technology MOU-TACHEN Department ofMathematics State University ofNew York, College at Brockport
ABSTRACT. In the present study, we explored whether mental transformation or feature detection is involved in visual processing of shape recognition; another purpose was to compare the visual processing of hearing subjects with that of hearingimpaired subjects. Deprivation of hearing, according to Myklebust (1964), impedes perceptual functioning in some respects and enhances such functioning in other respects. An experiment was conducted using two-dimensional random shapes under similarity transformations in Euclidean space. Reaction time served as the behavioral measure. The analysis of variance results showed significant main effects of orientations, isometries, and shapes, but not of groups and sizes. There were significant linear trends for orientations and sizes. Both groups demonstrated mental isometric transformation.
HUMAN BEINGS ARE CONTINUALLY FACED with the problem of recognizing shapes or objects in varying orientations. Milner (1974) suggested that recognition of shapes might be achieved by the extraction of features that are independent of orientation, spatial location, and mirror reflections. Similar to this view is the idea proposed by Corballis (1988) that recognition of a familiar shape independent of its orientation may be accomplished by extracting a description of the shape that is frame-independent or independent of any coordinate system. Such a description is usually sufficient to find the stored representation of the shape in long-term memory.
Requests for reprints should be sent to Kathleen C. Chen, College of Liberal Arts. Rochester Institute ofTechnology, One Lomb Memorial Drive, Rochester, NY 14623. 27
Downloaded by [] at 15:23 29 December 2014
28
The Journal of General Psychology
Palmer (1983) proposed another view, stating that shape constancy is accomplished by using a reference-frame hypothesis while focusing on properties that vary over the transformations of a group. He assumed that the effects of transformations were neutralized by imposing an intrinsic frame of reference, thereby achieving shape constancy. The intrinsic frame corresponds with the structure of the figure. He believed that when intrinsic frames are appropriately chosen, they can compensate for structural changes induced by similarity transformations: translations, rotations, reflections, dilations, and their composites. Palmer chose similarity transformations in Euclidean space in his analysis of perceptual organization because similarities are preferred by the human visual system to other transformations. A type of mental transformation was demonstrated by Cooper and Shepard (1973). They provided empirical evidence that mental rotation is involved in judgments of disoriented shapes. Subjects were timed as they decided whether rotated alphanumeric characters were normal or mirror reversed; reaction times increased with the angular departure of the characters from the normal upright orientation. Similar results were obtained by Cooper (1975) using random two-dimensional shapes. Those results have been taken as evidence that mental rotation is an analogue process and that mentally rotated shape representation is at some level the same as that of a shape that has been physically rotated. Identification of shape independent of size has long been considered a problem (Bundesen & Larsen, 1975; Kohler, 1929). One type of explanation of this ability involves the assumption that we transform a mental image of an object of a given size to an image of an object of the same shape but of another size (Bundesen, Larsen, & Farrell, 1981; Posner, 1969). Another type of explanation is based on direct extraction of size-invariant features from the sensory input (Dodwell, 1970; Gibson, 1969). The purpose of the present study was to determine whether mental transformation or feature detection is involved in the visual processing of shape recognition and to compare the perceptual processing of hearing and hearingimpaired subjects. According to Myklebust (1964), deprivation of hearing impedes perceptual functioning in some respects and enhances such functioning in other respects. Hearing-impaired subjects' visual processing may be enhanced because of auditory deprivation.
Method The present investigation used two-dimensional random forms under similarity transformations in Euclidean space. A transformation of a plane is a oneto-one map of the points of the plane onto themselves. An isometry is a transformation that preserves distance between two points. Because of its distancepreserving property, an isometry also preserves the shape and the size of a
Chen & Chen
29
figure. A plane isometry is a translation, a rotation, a reflection, or a glide reflection, which is a composition of a reflection and a translation. A dilation is an enlargement or a contraction of a ratio about a point. A similarity is one of the following: an isometry, a dilation, or a composition of an isometry and a dilation (Martin, 1982). Mental isometric transformation will lead to increase in reaction time as the angular departure of shape from identity is increased or as the size changes from its standard size.
Downloaded by [] at 15:23 29 December 2014
Subjects Eight hearing students of the Rochester Institute of Technology and eight hearing-impaired students of the National Technical Institute for the Deaf, Rochester Institute of Technology, served in the experiment. There were equal numbers of male and female subjects in both groups, and all subjects had 201 20 or corrected 20/20 vision.
Stimulus Materials Stimuli were computer generated. Three random shapes of 6, 12, and 24 points generated by Attneave and Amoult (1956) were used (Figure 1). The three shapes and their reflected images represented three levels of rated complexity and three sizes of linear scales: 1.6,0.8, and 0.4. The isometries used were rotations and reflections. Each shape of a scale had four orientations: 0°, 60°, 120°, and 180° (clockwise angular departures from the standard shape for rotation) and 0°, 30°, 60°, and 90° (angular departures of the mirrors for reflection). The angular departure of the position of a reflected image was twice as much as the angular departure of its mirror. With three shapes in three sizes, two isometries, and four angular orientations, there were a total of 72 shapes (3 x 3 x 2 x 4). The smallest display subtended a visual angle of approximately 3° 54'; the medium display, an angle of 7° 48'; and the largest display, an angle of 15° 36'.
Procedure Both hearing and hearing-impaired groups followed the same procedure. The hearing-impaired subjects were given the same written instructions on the screen as the hearing subjects. Extra instruction was given to subjects by written verbal communications or sign language to ensure that they understood the instructions before the test session. A sign-language interpreter served as an assistant when needed. The procedure was programmed in Pascal language for a Gigi digital terminal and a Barco CD33 screen. Each subject was tested individually. During three training trials, subjects studied drawings of three standard random shapes of 6, 12, and 24 points of the 1.6 scale for 2 min each.
Downloaded by [] at 15:23 29 December 2014
30
The Journal of General Psychology
FIGURE 1. Standard forms.
They were instructed that forms would be presented to them on the screen one at a time and that they were to press the left-hand key if the displayed form was a rotated image of one of the standard forms. If the image was identical to one of the standard forms, they were to consider it a 0° rotated image of the standard form and to press the left-hand key accordingly. If the displayed form was a reflected image of one of the standard forms, they were to press the right-hand key. They were also informed that the forms displayed could differ in size from the standard forms and that they did not have to indicate what the size of the form was. They were instructed that if the size of the stimulus form was different from the size of one of the standard forms, they should continue to judge whether the form with size variation was a rotated or reflected image of one of the standard forms by pressing the leftor right-hand key. A visual display of two colored keyboards indicating all
Downloaded by [] at 15:23 29 December 2014
Chen & Chen
31
possible left- or right-hand keys that could be used was shown after the instructions. Each form appeared for 0.5 s, centered within a circular field with a black surround. Subjects' key-presses terminated the stimulus display. Following a key-press, a I-s cyan warning field preceded the presentation of the next stimulus. During training trials, subjects' responses were followed by the words "right" or "wrong" displayed on the screen. A I-hr test session that consisted of four trials immediately followed the training period. Of the 72 test figures given in each trial, there were three forms each in three sizes, two isometries, and four angular orientations. The order of presentation for each trial was randomized by the computer; reaction time (RT) and errors were recorded by the computer also. Test figures on which errors were made were presented again at the end of each trial so that we would obtain a complete set of error-free data for each subject. For each subject, the complete set of data consisted of 288 errorless RTs.
Results A six-way analysis of variance (ANOVA) (Subjects x Groups x Forms x Sizes x Isometries x Orientations) was performed. The analysis indicated no significant between-groups effect, F(l, 14) = .23, p > .05; a significant effect offonns, F(2, 28) = 6.18, p < .01; no significant effect of sizes, F(2, 28) = 2.16, p > .05; a highly significant isometry effect, F(l, 14) = 81.54, p < .001; and a highly significant orientation effect, F(3, 42) = 20.62,p < .001. The interaction between sizes and orientations was marginally significant, F(6, 84) = 2.88, P < .05, as was the interaction between isometries and orientations, F(3, 42) = 3.01, P < .05. Other two- and three-way interactions between the main variables (groups, forms, sizes, orientations, and isometries) were not statistically significant. Alteration of Forms The mean RT as a function of forms for both hearing and hearing-impaired groups is shown in Figure 2. A trend analysis showed a significant quadratic effect, F(l, 28) = 11.75, P < .01, but no significant linear trend, F < 1. Comparing the results of hearing and hearing-impaired groups, we found similar effects for both groups, with significant quadratic effects, F(l, 14) = 4.68, P < .05, for the hearing group and F(l, 14) = 9.97, P < .01, for the hearing-impaired group. Alteration of Sizes The combined data of hearing and hearing-impaired groups used to compute the RT function for sizes are illustrated in Figure 3. A trend analysis revealed
32
The Journal of General Psychology
I
,
... .- ..., /
/ I
,
Downloaded by [] at 15:23 29 December 2014
\
I
\
,
,
~
I
1100 1600
I
,
\
I
1800
'", ,
I
\
~
\
I
1
\
2
,
Forms
FIGURE 2. Mean RT as a function of form with 6 points (1), 12 points (2), and 24 points (3).
a significant linear trend, F(l, 28) = 4.30, P < .05, and a nonsignificant deviation from this trend, F < 1. Separate trend analysis of sizes for the hearing group still showed a significant linear trend, F(l, 14) = 6.27, P < .05, but that for the hearing-impaired group did not approach statistical significance, F < 1. The hearing-impaired group tended to react with longer RTs and lower error rates than the hearing group. However, the ANOVA showed no significant difference between RTs, and a t test showed no significant difference in error rates. Size effect became more pronounced for the hearing group when reflection was included with rotation in the analysis of total subjects. This is shown in the significant difference (by t test, p < .05) in RTs of rotated and reflected forms in the standard size with the 1.6 scale used in training. A similar analysis for the hearing-impaired group did not approach the significance level of .05. Orientations
The linearity of the RT function of orientations for the combined hearing and hearing-impaired groups is shown in Figure 4. Regression for orientations
Chen & Chen
33
1850
1800
Downloaded by [] at 15:23 29 December 2014
1750
1'100
·2
.,
Size
FIGURE 3. Mean RT as a function of size with scales 1.6 (SI)' 0.8 (8J, and 0.4 (83) ,
yielded a linear trend that was highly significant, FO, 42) = 61.57, p < .001, but no significant quadratic or higher order effects, F < I. When hearing and hearing-impaired groups were compared, there were similar highly significant linear trends for the hearing group, FO, 21) = 27.44, P < .001, and the hearing-impaired group, FO, 21) = 34.90, p < .001. Again there were no significant higher order interactions in either group, F < 1.
RotatedForms The hearing and hearing-impaired subjects' mean RTs (averaged over subjects, forms, and sizes) for correctly determining the rotated test forms are also illustrated in Figure 4. RTs for errors are not included in the mean RTs in Figure 4. The most striking feature of the data presented in Figure 4 is the linearity of the increase in RT with angular departure of the test stimulus from the trained orientation. This result supports previous studies using alphanumeric stimuli (Cooper & Shepard, 1973) and random two-dimensional shapes (Cooper, 1975). The error ranges of rotation were between 3% and 20% and 5% and 16% for hearing and hearing-impaired groups, respectively.
34
The Journal of General Psychology
"..
2200 2100
....... 2 1900 1800 v;-
g
",
1700
Eo-
1600
;j
1500
.'" "
'"
" '"'"
'"
." --- ---
0::
Downloaded by [] at 15:23 29 December 2014
G)
::E
--_--..--
1400 1)00
~
Reflection Reflection and Rotation Rotation
1200
-e 600 Angular departure from trained orientation (degrees)
FIGURE 4. Mean RT in isometries as a function of angular departure of test shape from trained orientation.
Reflected Forms
Hearing and hearing-impaired groups' RT functions (averaged over subjects, forms, and sizes) for correctly judging the reflected forms are also plotted in Figure 4. The striking linearity depicted for rotation is also apparent for reflection. The hearing group showed a tendency toward higher error rates than did the hearing-impaired group; however, the differences were not significant (t tests, p > .05) either for orientation in terms of error rates or for RT. The range of error rates in reflection for the hearing group was between 12% and 33% and for the hearing-impaired group, between 2% and 15%. Discussion Mental isometric transformation was demonstrated by the significant main effects of orientation and isometry (rotation/reflection) in ANOVAs for both groups of subjects. Both hearing and hearing-impaired subjects were alike in demonstrating linearity of RT functions for orientation, which has implica-
Downloaded by [] at 15:23 29 December 2014
Chen & Chen
35
tions for the nature of the internal processes involving rotated-reflected discrimination with size variation. Hearing-impaired subjects and hearing subjects demonstrated mental rotation. These results replicate the findings of Cooper and Shepard (1973), Cooper (1975), and Cooper and Shepard (1978). RTs for reflected images were longer than for rotated images; they replicate the findings of Corballis, Zbrodoff, Shetzer, & Butler (1978) that latencies for identifying reflected characters were longer than those for rotated characters. Our results are at odds with the feature detection theories of shape recognition proposed by Deutsch (1962) and Milner (1974) because these theories predict invariance with respect to both angular orientation and mirror reflection. Another tenable explanation for the present results is that there is a process of mirror-image generalization, perhaps involving homotopic recording and transfer of memory traces from one to the other side of the brain so that form registration leaves mirror traces both for the form itself and for its mirror-reflected image (Achim & Corballis, 1977; Corballis & Beale, 1976). Such a process serves as an adaptive function in our everyday world, as it allows us to recognize patterns regardless of left-right orientation. The subjects in our study were exposed in the training session to the standard forms in the version we consider normal. Hence the traces under mirror reflection that depended on interhemispheric transfer were weaker than the normal ones of direct input. Our results may be explained by the reference frame theory, as mental rotation is involved in returning the disoriented form to uprightness with respect to its frame. The identity 0° orientation is invariably the easiest to recognize. Consistent with Palmer, Rosch, and Chase (1981), recognition latency decreases as the form becomes more canonical or upright. Hinton and Parsons (1981) proposed that perception of a shape involves the assignment of an intrinsic frame of reference to shape, and the frame can be either leftor right-handed and can be decided only in an analogue way. Even Corballis' frame-independent position recognizes the role of mental rotation as necessary in mirror-image discrimination (1988). Our procedure involved mirrorimage discrimination in differentiating the two isometries (rotation/reflection). Template theories of visual pattern recognition assume the operation of preprocessing to deal with irrelevant information, such as stimulus size discrepancies. One form of normalizing the stimulus is mental transformation of size. Current literature shows widespread acceptance of the size transformation explanation (Bundesen & Larsen, 1975; Bundesen et al., 1981; Cooper & Shepard, 1978; Larsen & Bundesen, 1978; Shepard, 1981). In the present experiment, trend analysis showed a significant linear function of RT and size for combined hearing and hearing-impaired groups (Figure 2). The size effect was not significant by ANOVA. Separate analysis
Downloaded by [] at 15:23 29 December 2014
36
The Journal of General Psychology
of the hearing group showed a significant linear trend and also a significant difference by t test (p < .05) of RT of rotated plus reflected forms on the 1.6, 0.8, and 0.4 scales and rotated plus reflected forms on the 1.6 (training) scale. These results suggest that some hearing subjects mentally transformed size to match the form against the internal template on the 1.6 scale. There was no pronounced size transformation effect for the hearing-impaired group. Nevertheless, the combined groups still showed a significant linear trend. This demonstrates some similarity transformation. Some subjects mentally rotated, flipped, and transformed size to match an internal template. It is possible, however, to discriminate forms of different sizes without size transformation. The procedure in this experiment differs from that in the size transformation studies cited above. In those studies, a matching task was performed while both stimulus forms were present in the visual field, whereas our study required the subject to match the stimulus with an internal memory image. Other outcomes of this experiment were the significant quadratic effects between complexity of the forms and RT for both hearing and hearingimpaired groups and a significant main effect of forms by ANOVA. These results are different from Cooper's (1975) findings. One possible explanation of the complexity variations in the forms to produce RT differences relates to the nature of the internal representation or memory image, which is highly schematic in comparison with the rich detail of the form itself. A different explanation is based on the subjects' reports that some achieved the correct response by imagining only a few distinctive features of the test form. It is quite obvious that the l2-point form is the most difficult form. Investigation of the visual functioning of hearing-impaired subjects may have important implications for the psychology of sensory deprivation. We concluded that hearing deprivation did not affect these subjects' visual form perception. Both hearing and hearing-impaired groups performed mental isometric transformations. We also found a relationship between complexity of forms and rate of mental transformation. The ease of recognition depended not on the complexity of forms, but on the distinctiveness of the features of forms.
REFERENCES Achim, A., & Corballis, M. C. (1977). Mirror-image equivalence and the anterior commissure. Neuropsychologia, 15, 475-478. Attneave, F., & Arnoult, M. D. (1956). The quantitative studyof shapeand pattern. Perception, 53, 452-471.
Bundesen, C,; & Larsen, A. (1975). Visual transformations of size. Journal of Experimental Psychology: Human Perception and Performance, 1, 214-220.
Bundesen, c., Larsen, A., & Farrell, J. E. (1981). Visual transformations of size and orientation: Spatial imagery and perceptual processes. In A. Baddeley & J.
Downloaded by [] at 15:23 29 December 2014
Chen & Chen
37
Long (Eds.), Attention and performance IX (pp. 279-294). Hillsdale, NJ: Erlbaum. Cooper, L. A. (1975). Mental rotation of random two-dimensional shapes. Cognitive Psychology. 7. 20-43. Cooper, L. A., & Shepard, R. N. (1973). Chronometric studies of rotation of mental images. In W. G. Chase (Ed.), Visualinformation processing (pp. 75-176). New York: Academic Press. Cooper, L. A., & Shepard, R. N. (1978). Transformations on representations of objects in space. In E. G. Carterette & M. P. Friedman (Eds.), Handbook ofperception (Vol. 8, pp. 104-146). New York: Academic Press. Corballis, M. C. (1988). Recognition of disoriented shapes. Psychological Review. 95(1), 115-123. Corballis, M. C., & Beale, I. L. (1976). Thepsychology ofleft and right. Hillsdale, NJ: Erlbaum. Corballis, M. c.. Zbrodoff, N. J., Shetzer, L. I., & Butler, P. B. (1978). Decision about identity and orientation of related letters and digits. Memory and Cognition. 6(2), 98-107. Deutsch, J. A. (1962). A system of shape recognition. Psychological Review. 69. 491-500. Dodwell, P. C. (1970). Visual pattern recognition. New York: Holt. Gibson, E. J. (1969). Principles ofperceptual learning and development. New York: Appleton-Century-Crofts. Hinton, G. E., & Parsons, L. M. (1981). Frames of reference and mental imagery. In A. D. Baddeley & J. Long (Eds.), Attention andperformance (Vol. 9, pp. 261278). Hillsdale, NJ: Erlbaum. Kohler, W. (1929). Gestaltpsychology. New York: Live-Right. Larsen, A., & Bundesen, C. (1978). Size scaling in visual pattern matching. Journal ofExperimental Psychology: HumanPerception and Performance. 4. 1-20. Martin, G. E. (1982). Transformation geometry. New York: Springer-Verlag. Milner, P. M. (1974). A model for visual shape recognition. Psychological Review. 81. 521-535. Myklebust, H. R. (1964). The psychology of deafness (2nd 00.). New York: Grone and Stratton. Palmer, S. (1983). The psychology of perceptual organization. In J. Beck, B. Hope & A. Rosenfield (Eds.), Human and machine vision (pp. 269-339). New York: Academic Press. Palmer, S., Rosch, E., & Chase, P. (1981). Canonical perspective and the perception of objects. In A. C. Baddeley & J. Long (Eds.), Attentionand performance (Vol. 9, pp. 135-151). Hillsdale, NJ: Erlbaum. Posner, M. I. (1969). Abstraction and the process of recognition. In G. H. Bowen & J. J. Spence (Eds.), Thepsychology oflearning and motivation (Vol. 5, pp. 441(0). New York: Academic Press. Shepard, R. N. (1981). Psychological complementary. In M. Kubovy & J. R. Pomerantz (Eds.), Perceptual organization (pp. 279-341). Hillsdale, NJ: Erlbaum.
ReceivedApril 27, 1989
