INT'L. J. AGING AND HUMAN DEVELOPMENT,Vol. 31(4) 267-277, 1990
VISUAL FIELD DEPENDENCE IN ELDERLY FALLERS AND NON-FALLERS
STEPHEN R. LORD, BSC, MA IAN W. WEBSTER, MD, FRACP. HEAD School of Community Medicine University of New South Wales
ABSTRACT
Two tests of visual field dependence (a measure of reliance upon the spatial framework provided by vision in the perception of the uprighthroll vection and the rod and frame test-were administered to 136 participants aged fifty-nine to ninety-seven years. It was found that the fifty-nine participants who had experienced one or more falls in the past year were significantly more visually field dependent in both tests compared with the seventy-seven participants who had not fallen. Mean error in perception of the true vertical in the rod and frame test was 20.7 degrees for the fallers and 17.2 degrees for the non-fallers. Mean error in perception of the true vertical in the roll vection test was 6.6 degrees for the fallers and 3.6 degrees for the non-fallers. The test of roll vection was the better discriminator between fallers and non-fallers, which may be due in part to less misunderstanding of the required task by the participant. The results suggest that tilted or rolling visual stimuli may be factors leading to postural instability and falls in the elderly. The findings support the claim that greater dependence on visual information shown by fallers may be the result of reduced proprioceptive and vestibular function resulting from increased age and chronic health problems.
One of the major problems associated with aging is the increased susceptibility to falling. Community studies estimate that approximately one in three persons over the age of sixty-five experience one or more falls per year [l-41.Campbell et al., analyzed a stratified population sample of 533 participants aged sixty-five years and over and found that one in three experienced one or more falls per year [l]. Similarly, in a study of 2,793 participants aged sixty-five years and over, Prudham and Grimley Evans estimated an annual incidence for accidental falls of 28 267
0 1990, Baywood Publishing Co.,Inc.
doi: 10.2190/38MH-2EF1-E36Q-75T2 http://baywood.com
268 I LORDANDWEBSTER
percent [2]. Tinetti et al. [3], examined 336 participants aged seventy-five years and over in a prospective study and found that 32 percent fell in a period of one year, while in a community survey, Blake et al., found that 35 percent of 1042 participants aged sixty-five years and over reported one or more falls in the preceding year [4]. Falls have serious consequences for the elderly. In a prospective study, Lucht found that the incidence rate for persons who require acute care in hospital because of falls increased from 1percent per annum to in the age group sixty to sixty-five years to 10.4 percent per annum in the age group ninety years and over [5]. Injuries that result from falls include soft-tissue damage and fractures of the radius, humerus, and femoral neck. Falls can also lead to decreased mobility, which often results in increased dependency on others and hence an increased probability of being admitted to an institution. Smallegan found that falls are cited as a contributing factor in 40 percent of admissions to nursing homes [6]. A number of investigators have suggested that impaired vision may be a predisposingfactor for falls among the elderly [7-81. It has been shown that one of the functions of vision-the provision of information about the position and movements of body parts in relation to each other and the external environmentis an important source of information for postural control [7]. It has been found that many visual processes including visual acuity, contrast sensitivity, glare sensitivity, dark adaption, accommodation,and depth perception decline with age (especially beyond 40 years) [9]. In spite of the demonstrated deterioration of vision with age, it has been hypothesized that old people may rely largely on the spatial framework provided by vision in an attempt to compensate for reduced postural sensation. Orientation in space involves the use of spatial information provided by a number of senses. Postural (proprioceptive, vestibular, tactile) stimulation provides information about the gravitational upright and visual stimulation, which provides information about the main contours of a framework-the axes of which are in the horizontal and vertical planes. The interaction of postural and visual information produces a unitary experience, not two separate experiences of the upright. This is because visually upright objects are typically gravitationally upright as well, and that tilting the body changes both postural and visual information but leaves the relationship between the two “senses” unchanged [lo, 111. Vision and proprioception have been placed in conflict by a number of experimental techniques. Witkin et al. had participants seated on a tilt chair attempt to align themselves to the true vertical while viewing a tilted visual field (a tilted room) [12]. In another experiment they tested participants’ abilities to set a rod to the true vertical when it was enclosed in a tilted frame. They found that some people were “field dependent” in that they relied greatly on the spatial framework provided by vision, and that others were able to disregard the visual frame of reference and align themselves or the rod near the true vertical. Wide
VISUAL FIELD DEPENDENCE / 269
individual differences were found in all conflict situations, but individuals tended to be consistent from situation to situation. Over suggested that if excessive reliance is placed on vision, difficulties in maintaining a stable upright body position may arise when minimal, ambiguous, or misleading spatial information is provided by vision; if body position is wrongly determined by the visual framework, a fall may result [ll]. He suggested that old people with a history of falls (drop attacks) would be more field dependent (a response to a testing procedure that places visual and other postural cues in “cunflict”) than peopIe of the same age who have never fallen. Recently, Tobis et al., tested this hypothesis using a portable rod and frame apparatus [13]. They found that sixty-eight elderly fallers were significantly more field dependent than a control group of 131 elderly participants. Moving visual fields also have strong destabilizing effects on posture, particularly when visually perceived motion does not correspond to the body shift sensed by the vestibular and somatosensory systems [7, 14-17]. Roll vection was included in this study since it was thought that moving visual stimuli may also be a factor leading to postural instability and falls in the elderly. For instance, individuals with poor proprioception and other non-visual postural control systems when exposed to “misleading” moving visual stimuli may be at risk of fa11ing. In this article we examine whether visual field dependence as measured by two tests (a stationary tilted visual field and a rotating visual field) is associated with falls. METHODS
Description of the Sample The sample was comprised of 136 persons from the Anglican Retirement Villages in Sydney, Australia. The study was undertaken in two stages. In stage one, a sample of forty-one volunteers from the main village complex undertook the tests (26 participants resided in self-care units and 15 in hostels). Thirty-seven of these participants undertook the tests a second time six months after initial testing so that test-retest reliability could be assessed. In stage two, ninety-five residents from one hostel took part in the study. This hostel housed 124 residents. Of the twenty-nine non-participants,four were sick, five were absent (on holidays, etc.) and twenty declined. The participation rate of available residents (excluding those sick and absent) was 82.6 percent. The participants were between fiity-nine and ninety-seven years of a g e t h e age and gender distributions are shown in Table 1. Classification of participants into fallers or non-fallers was based on whether the participant had one or more falls in the past year. Falls status was primarily
270 I LORD AND WEBSTER
Table 1. Age and Gender Distribution of the Sample . Age Group
Men
59-69
0
70-79 80-89
8
90-97
Total
Mean (SD)
15
Women 7 41 52 10
Total
7 49
67
-
13
3 26
110
136
81.7 (6.2)
80.8 (7.1)
80.9 (7.0)
determined by self-reports. For residents with memory impairments (who may have fallen but have no recall of doing so) a secondary check was made using reports from nursing staff and accident records. Fifty-six women reported having no falls in the past year while fifty-four women reported one or more falls. Of the twenty-six men who undertook the tests, five had fallen. The mean age of the fallers was 81.9 years, and non-fallers 80.3 years, (t = 1.31, N S ) .
Apparatus and Procedures Visualfield dependence was measured using two tests. “Static”fie1d dependence - The static field test was measured using a portable rod and frame apparatus designed by Oltman, consisting of a rectangular enclosure tilted to the right at an angle of 28 degrees [MI. A rod (28 cm x 9.6 mm) was visible at one end of the enclosure, and a headrest was constructed at the opposite end. The rod, with its central fixation point positioned in the center of the end panel (30 cm2), could be rotated through 360 degrees by the experimenter. A protractor located behind the enclosure (i.e., behind the rod) enabled the experimenter to assess the angle of the rod. The sides of the enclosure were made of 30 cm x 60 cm x 3.2 mm translucent white acrylic sheets that blocked visual access to external vertical or horizontal reference cues. The tilted visual frame was enhanced by black plastic tape on the inside edges of the frame. The participant’s view was restricted to the interior of the enclosure by means of a curved shield attached to the headrest. The participant viewed a tilted visual field for a period of three minutes, and then the experimenter moved the rod (initially aligned at an angle of 28 degrees from the vertical) towards the vertical plane until the participant stated that the rod appeared upright (relative to gravity). Six trials: three with the rod inclined with the tilt and three with the rod inclined against the tilt were performed (see Figures l a and lb). Roll vection (or field dependence lo moving stimuli) - Roll vection was measured using a rotating disk and adjustable straight edge based on a design by
VISUAL FIELD DEPENDENCE / 271
H a. The subject viewing the rod and tilted field from an opening in one end of the apparatus.
I
b. The rod and frame as seen by the subject.
Figure 1. Diagram of rod and frame test apparatus.
I
272 I LORD AND WEBSTER
I
Figure 2. Diagram of the roll vection test. The subject attempts to align the black rod on the small disc with the upright while the lar e disk is rotated around her line of sight at 96 degrees sec- (16 rpm).
P
VISUAL FIELD DEPENDENCE / 273
Dichgans et al. [16]. In this test, the participant viewed a disk, which restricted his view to 130 degrees of visual angle. The disk was rotated about the participant's line-of-sight and contained low spatial frequency information (16 x 22.5 degree black and white segments). A target disk (which shared the same central fHation point as the large disk) was positioned 3 centimeters in front of the large disk. The target disk subtended 32 degrees of visual angle and was marked with a straight edge (20 cm x 1 cm). The rotation of the target disk was controlled by the participant. The participant was asked to position the target disk so that the straight edge was in the vertical plane before the test commenced (before the disk was rotated), and then the rod was positioned at 28 degrees from the vertical in the direction of the roll. The participant viewed the visual field which was rotated at 96 degrees sec-' for 30 seconds and then while the large disk was still rotating attempted to position the target disk so that the straight edge was aligned with the vertical plane. Six trails were performed-three with the straight edge initially tilted in the direction of the roll and three trials with the straight edge tilted in the direction opposite to the roll. Any error in positioning the straight edge in the vertical plane was measured in degrees (see Figure 2). The tests were administered by one experimenter (S.L.) who was blind to faller status.
Statistical Analysis Student t-tests were performed to examine whether there were differences in age and test performance between the fallers and non-fallers. It was found that the raw data for the roll vection test were positively skewed so log scores were examined. Chi2 tests for contingency tables were used to examine differences in test scores when performance was categorized as normal or abnormal.
Test Reliability The two tests of field dependence were re-administered to thirty-seven participants six months after the initial testing. The extent of the roll vection had a reliability coefficient (test-retest correlation) of 0.93, while rod and frame field dependence had a reliability coefficient of 0.66 (both significant at p < 0.01). All of the tests were administered by the one investigator on both occasions. RESULTS Mean scores plus standard deviations of the test results for the fallers, nonfallers, and total sample are shown in Table 2. The extent of the roll vection and rod and frame field dependence were significantly higher in the group who reported falls in the past year than in the group who reported no falls (t = 3 . 0 2 , ~= 0.003 and t = 2.59, p = 0.011 respectively). Mean error in perception of the true vertical in the rod and frame test was 20.7 degrees for the fallers and 17.2 degrees
274 /
LORDANDWEBSER
Table 2.Mean Error in Perception of the True Vertical in the Rod-and-Frame and Roll Vection Tests Faller/Non-Faller Comparisons
Test
Fallers (N = 59) Mean (S.D.)
Non-Fallers (N = 77) Mean (S.D.)
Total (N = 136) Mean (S.D.)
Rod and Frame a Roll Vection *
20.66 (6.60) 6.62 (6.22)
17.24 (8.52) 3.63 (4.86)
18.70 (7.91) 4.93 (5.67)
a Significant difference between fallers and non-fallers at p c 0.02. Significant difference between fallers and non-fallers at p c 0.01.
Table 3.Participants with Abnormal Responses in the Rod-and-Frame and Roll Vection Tests. Fallermon-Faller Comparisons Test
Rod and Frame a ROII vection
Fallers N (Percent)
Non-Fallers N (Percent)
47 (83.9) 31 (53.4)
55 (73.3) 25 (33.3)
a Set rod more than 12 degrees from the vertical, definition of abnormality from Tobis et al. [13]. Chi2 = 2.09; df = 1, NS. Set rod more than 4 degrees from the vertical, new definition of abnormality. Chi2 = 5.43, df= 1, p c 0.02.
for the non-fallers. Mean error in perception of the true vertical in the roll vection test was 6.6 degrees for the fallers and 3.6 degrees for the non-fallers. Table 3 shows the number and proportion of fallers, non-fallers, and the total sample who recorded abnormal results in the tests. Twelve degrees from the vertical was used as the criterion for the rod-and-frame test as this value was reported by Tobis et al., to discriminate between fallers and non-fallers [13]. The value of 4 degrees for the roll vection test was developed after examining frequency distributions. When 12 degrees from the vertical is used as the criterion for an abnormal response in the rod and frame test, it was found that most fallers (84%) and non-fallers (73%)were classified as abnormal. In the roll vection test, 4 degrees was used as a criterion for an abnormal response. Using this criterion, 53 percent of fallers were classified as abnormal with only 33 percent of the non-fallers. Pearson correlation coefficients were computed to examine the relationships between performance in the roll vection and rod and frame tests for the fallers, non-fallers, and total sample. The correlation coefficients (r) were 0.128 (p = 0.177) for the fallers group, 0.421 (p = 0.000) for the non-fallers group, and 0.379 (p = 0.000) for the total sample.
VISUAL FIELD DEPENDENCE / 275
DISCUSSION
It has been clearly demonstrated that tilted and moving visual fields can affect the perceived upright. This apparent displacement of the postural vertical appears to be the result of a central recomputation that results from an attempt to weigh the contradictory sensory inputs of the vertical [17]. The most extensive series of experiments on visual field dependence using stationary tilted fields were undertaken by Witkin et al., in the 1950s [12]. They found wide individual differences in all conflict situations, but that individuals tended to be consistent from situation to situation. Wide individual differences have also been found in the extent of roll vection [16]. The present study found that in both the rod and frame and roll vection tests, fallers aligned the rod at greater deviations from the vertical compared with non-fallers. As in the study by Tobis et al. [13], a significant association was found between falls and performance on the rod and frame test. The main difference being that in the present study, higher proportions of both fallers and non-fallers aligned the rod at substantial deviations from the vertical. Tobis et al., found that an average error in excess of 12 degrees in the rod and frame test may predict serious falls among independently functioning older adults. This difference in the proportion of the participants who aligned the rod at large angles from the upright in the two studies may be the result of sample differences. Tobis et al.’s sample was drawn from three groups: working seniors, senior citizens center attenders, and affluent seniors; all of whom resided in the community. In contrast, the present sample was drawn from residents of a retirement village-mostly hostel residents. The incidence of falls in this population (43%) is higher than that reported in community studies, and it is likely that this population is made up of frailer, more dependent persons. The mean age of the present sample (80.9 years) is also greater than in the study by Tobis et al. (72.5 years). These factors, increased age and dependency, may be associated with increased visual field dependence. Another reason for the high proportion of participants who set the rod at substantial deviations from the vertical could be that the rod and frame test procedure is open to misunderstanding by the participant, no matter how carefully it is explained. The test of roll vection has an advantage over the rod and frame test in that the participant can be asked to position the rod in the vertical plane before the test commences. This ensures that the participant understands the test procedure. In the rod and frame test, however, this cannot be done because the positioning of the rod in the vertical plane prior to the test provides the participant with cues as to how to position the rod-it simply “gives the game away.” With elderly participants, especially those with sub-clinical confusional states, understanding of the task requirements cannot be assured. Performance in the two tests was significantly associated in the non-faller’s group, however, there was only a weak association in the faller’s group. This lack of association appears to support the
276 / LORD AND WEBSTER
suggestion that some fallers may not understand the test procedure for the rod and frame test. The test of roll vection was also found to be a better discriminator between fallers and non-fallers than the rod and frame test. This may be because the participant undertakes the test standing (rather than sitting) and because the rotating disk provides a powerful visual stimulus. The effect of the rotating field was clearly disorienting and unpleasant for certain elderly fallers. The results of the present study suggest that elderly fallers have a greater dependence on visual information than elderly non-fallers. These findings support the hypothesis that increased visual dependence shown by fallers may develop in response to impaired proprioceptive and vestibular systems as a result of age and chronic health problems. The study examined elderly participants under artificial situations, and the findings may only partly generalize to real world situations-although the implications are that tilted or rolling visual stimuli may be related to factors that lead to postural instability and falls in the elderly. For instance, individuals with poor proprioception and vestibular sense, when exposed to “misleading” tilted or moving visual stimuli such as tilted forms and landmarks, moving clouds, vehicles, or structures may be at increased risk of falling.
REFERENCES 1. A. J. Campbell, J. Reinken, B. C. Allan, and G. S. Martinez, Falls in Old Age: A Study of Frequency and Related Clinical Factors,Age andAgeing, 10, pp. 264-270,1981. 2. D. Prudham and E. J. Grimley, Factors Associated with Falls in the Elderly: A Community Study,Age andAgeing, 10, pp. 141-146,1981. 3. M. E. Tinetti, M. Speechley, and S. F. Ginter, Risk Factors for Falls among Elderly Persons Living in the Community, The New England Journal of Medicine, 319, pp. 1701-1707, 1988. 4. A. J. Blake, K. Morgan, M. J. Bendall, H. Dallosso, S. B. J. Ebrahim, T. H. D. Arie, P. H. Fentem, and E. J. Bassey, Falls by Elderly People at Home: Prevalence and Associated Factors, Age and Ageing, 17, pp. 365-372, 1988. 5. U.Lucht, A Prospective Study of Accidental Falls and Resulting Injuries in the Home among Elderly People, Acta Sociomedica Scandinavica, 2, pp. 105-120,1971. 6. M. Smallegan, How Families Decide on Nursing Home Admission, Geriatric Consultations, 1 5 , pp. 21-24,1983. 7. D. N. Lee and J. R. Lishman, Visual Proprioceptive Control of Stance, Journal of Human Movement Studies, 1, pp. 87-95,1975. 8. C. Owsley, R. Sekuler, and D. Siemsen, Contrast Sensitivity throughout Adulthood, Vision Research, 23, pp. 689-699,1983. 9. D. G. Pitts, The Effects of Aging on Selected Visual Functions, in Aging in Human Visual Functions, R. Sekuler, D. W. Wine, and K. Dismukes (eds.), Liss, New York, 1982. 10. I. P. Howard and W. B. Templeton, Human Spatial Orientation, John Wiley and Sons, London, 1966.
VISUAL FIELD DEPENDENCE / 277
11. R. Over, Possible Visual Factors in Falls by Old People, Gerontologist,6, pp. 212-214, 1966. 12. H. A. Witkin, H. B. Lewis, M. Hertzrnan, K. Machover, P. B. Meissner, and S . Wapner, Personality through Perception, Harper, New York, 1954. 13. J. S. Tobis, S. Reinsch. et al., Visual Perception Dominance of Fallers among Community Dwelling Older Adults, Journal of the American Geriairic Society, 33, pp. 330333,1985, 14. D. N. Lee and J. R. Lishrnan, Vision-The Most Effective Source of Proprioceptive Information for Balance Control,Agressologie, 18a pp. 83-94,1977. 15. W.Bless, T. S. Kapteyn, and G. De Wit. Effects of Visual-Vestibular Interaction on Human Posture, Advances in Oto-Rhino-Laryngology,22, pp. 111-118,1977. 16. J. Dichgans, R. Held, L. R. Young, and T. Brandt, Moving Visual Scenes Influence the Apparent Direction of Gravity, Science, 178, pp. 1217-1219,1972. 17. T. Brandt, W. Paulus, and A. Straube, Vision and Posture, in Disorders ofPosture and Gait, W . Bles and T.Brandt (eds.), Elsevier, Amsterdam, 1986. 18. P. K. Oltman, A Portable Rod-and-Frame Apparatus, Perception and Motor Skills, 267, pp. 503-506,1968.
Direct reprint requests to: Stephen R. Lord School of Community Medicine University of New South Wales P.O. Box 1 Kensington, N.S.W.2033 AUSTRALIA
VISUAL FIELD DEPENDENCE IN ELDERLY FALLERS AND NON-FALLERS
STEPHEN R. LORD, BSC, MA IAN W. WEBSTER, MD, FRACP. HEAD School of Community Medicine University of New South Wales
ABSTRACT
Two tests of visual field dependence (a measure of reliance upon the spatial framework provided by vision in the perception of the uprighthroll vection and the rod and frame test-were administered to 136 participants aged fifty-nine to ninety-seven years. It was found that the fifty-nine participants who had experienced one or more falls in the past year were significantly more visually field dependent in both tests compared with the seventy-seven participants who had not fallen. Mean error in perception of the true vertical in the rod and frame test was 20.7 degrees for the fallers and 17.2 degrees for the non-fallers. Mean error in perception of the true vertical in the roll vection test was 6.6 degrees for the fallers and 3.6 degrees for the non-fallers. The test of roll vection was the better discriminator between fallers and non-fallers, which may be due in part to less misunderstanding of the required task by the participant. The results suggest that tilted or rolling visual stimuli may be factors leading to postural instability and falls in the elderly. The findings support the claim that greater dependence on visual information shown by fallers may be the result of reduced proprioceptive and vestibular function resulting from increased age and chronic health problems.
One of the major problems associated with aging is the increased susceptibility to falling. Community studies estimate that approximately one in three persons over the age of sixty-five experience one or more falls per year [l-41.Campbell et al., analyzed a stratified population sample of 533 participants aged sixty-five years and over and found that one in three experienced one or more falls per year [l]. Similarly, in a study of 2,793 participants aged sixty-five years and over, Prudham and Grimley Evans estimated an annual incidence for accidental falls of 28 267
0 1990, Baywood Publishing Co.,Inc.
doi: 10.2190/38MH-2EF1-E36Q-75T2 http://baywood.com
268 I LORDANDWEBSTER
percent [2]. Tinetti et al. [3], examined 336 participants aged seventy-five years and over in a prospective study and found that 32 percent fell in a period of one year, while in a community survey, Blake et al., found that 35 percent of 1042 participants aged sixty-five years and over reported one or more falls in the preceding year [4]. Falls have serious consequences for the elderly. In a prospective study, Lucht found that the incidence rate for persons who require acute care in hospital because of falls increased from 1percent per annum to in the age group sixty to sixty-five years to 10.4 percent per annum in the age group ninety years and over [5]. Injuries that result from falls include soft-tissue damage and fractures of the radius, humerus, and femoral neck. Falls can also lead to decreased mobility, which often results in increased dependency on others and hence an increased probability of being admitted to an institution. Smallegan found that falls are cited as a contributing factor in 40 percent of admissions to nursing homes [6]. A number of investigators have suggested that impaired vision may be a predisposingfactor for falls among the elderly [7-81. It has been shown that one of the functions of vision-the provision of information about the position and movements of body parts in relation to each other and the external environmentis an important source of information for postural control [7]. It has been found that many visual processes including visual acuity, contrast sensitivity, glare sensitivity, dark adaption, accommodation,and depth perception decline with age (especially beyond 40 years) [9]. In spite of the demonstrated deterioration of vision with age, it has been hypothesized that old people may rely largely on the spatial framework provided by vision in an attempt to compensate for reduced postural sensation. Orientation in space involves the use of spatial information provided by a number of senses. Postural (proprioceptive, vestibular, tactile) stimulation provides information about the gravitational upright and visual stimulation, which provides information about the main contours of a framework-the axes of which are in the horizontal and vertical planes. The interaction of postural and visual information produces a unitary experience, not two separate experiences of the upright. This is because visually upright objects are typically gravitationally upright as well, and that tilting the body changes both postural and visual information but leaves the relationship between the two “senses” unchanged [lo, 111. Vision and proprioception have been placed in conflict by a number of experimental techniques. Witkin et al. had participants seated on a tilt chair attempt to align themselves to the true vertical while viewing a tilted visual field (a tilted room) [12]. In another experiment they tested participants’ abilities to set a rod to the true vertical when it was enclosed in a tilted frame. They found that some people were “field dependent” in that they relied greatly on the spatial framework provided by vision, and that others were able to disregard the visual frame of reference and align themselves or the rod near the true vertical. Wide
VISUAL FIELD DEPENDENCE / 269
individual differences were found in all conflict situations, but individuals tended to be consistent from situation to situation. Over suggested that if excessive reliance is placed on vision, difficulties in maintaining a stable upright body position may arise when minimal, ambiguous, or misleading spatial information is provided by vision; if body position is wrongly determined by the visual framework, a fall may result [ll]. He suggested that old people with a history of falls (drop attacks) would be more field dependent (a response to a testing procedure that places visual and other postural cues in “cunflict”) than peopIe of the same age who have never fallen. Recently, Tobis et al., tested this hypothesis using a portable rod and frame apparatus [13]. They found that sixty-eight elderly fallers were significantly more field dependent than a control group of 131 elderly participants. Moving visual fields also have strong destabilizing effects on posture, particularly when visually perceived motion does not correspond to the body shift sensed by the vestibular and somatosensory systems [7, 14-17]. Roll vection was included in this study since it was thought that moving visual stimuli may also be a factor leading to postural instability and falls in the elderly. For instance, individuals with poor proprioception and other non-visual postural control systems when exposed to “misleading” moving visual stimuli may be at risk of fa11ing. In this article we examine whether visual field dependence as measured by two tests (a stationary tilted visual field and a rotating visual field) is associated with falls. METHODS
Description of the Sample The sample was comprised of 136 persons from the Anglican Retirement Villages in Sydney, Australia. The study was undertaken in two stages. In stage one, a sample of forty-one volunteers from the main village complex undertook the tests (26 participants resided in self-care units and 15 in hostels). Thirty-seven of these participants undertook the tests a second time six months after initial testing so that test-retest reliability could be assessed. In stage two, ninety-five residents from one hostel took part in the study. This hostel housed 124 residents. Of the twenty-nine non-participants,four were sick, five were absent (on holidays, etc.) and twenty declined. The participation rate of available residents (excluding those sick and absent) was 82.6 percent. The participants were between fiity-nine and ninety-seven years of a g e t h e age and gender distributions are shown in Table 1. Classification of participants into fallers or non-fallers was based on whether the participant had one or more falls in the past year. Falls status was primarily
270 I LORD AND WEBSTER
Table 1. Age and Gender Distribution of the Sample . Age Group
Men
59-69
0
70-79 80-89
8
90-97
Total
Mean (SD)
15
Women 7 41 52 10
Total
7 49
67
-
13
3 26
110
136
81.7 (6.2)
80.8 (7.1)
80.9 (7.0)
determined by self-reports. For residents with memory impairments (who may have fallen but have no recall of doing so) a secondary check was made using reports from nursing staff and accident records. Fifty-six women reported having no falls in the past year while fifty-four women reported one or more falls. Of the twenty-six men who undertook the tests, five had fallen. The mean age of the fallers was 81.9 years, and non-fallers 80.3 years, (t = 1.31, N S ) .
Apparatus and Procedures Visualfield dependence was measured using two tests. “Static”fie1d dependence - The static field test was measured using a portable rod and frame apparatus designed by Oltman, consisting of a rectangular enclosure tilted to the right at an angle of 28 degrees [MI. A rod (28 cm x 9.6 mm) was visible at one end of the enclosure, and a headrest was constructed at the opposite end. The rod, with its central fixation point positioned in the center of the end panel (30 cm2), could be rotated through 360 degrees by the experimenter. A protractor located behind the enclosure (i.e., behind the rod) enabled the experimenter to assess the angle of the rod. The sides of the enclosure were made of 30 cm x 60 cm x 3.2 mm translucent white acrylic sheets that blocked visual access to external vertical or horizontal reference cues. The tilted visual frame was enhanced by black plastic tape on the inside edges of the frame. The participant’s view was restricted to the interior of the enclosure by means of a curved shield attached to the headrest. The participant viewed a tilted visual field for a period of three minutes, and then the experimenter moved the rod (initially aligned at an angle of 28 degrees from the vertical) towards the vertical plane until the participant stated that the rod appeared upright (relative to gravity). Six trials: three with the rod inclined with the tilt and three with the rod inclined against the tilt were performed (see Figures l a and lb). Roll vection (or field dependence lo moving stimuli) - Roll vection was measured using a rotating disk and adjustable straight edge based on a design by
VISUAL FIELD DEPENDENCE / 271
H a. The subject viewing the rod and tilted field from an opening in one end of the apparatus.
I
b. The rod and frame as seen by the subject.
Figure 1. Diagram of rod and frame test apparatus.
I
272 I LORD AND WEBSTER
I
Figure 2. Diagram of the roll vection test. The subject attempts to align the black rod on the small disc with the upright while the lar e disk is rotated around her line of sight at 96 degrees sec- (16 rpm).
P
VISUAL FIELD DEPENDENCE / 273
Dichgans et al. [16]. In this test, the participant viewed a disk, which restricted his view to 130 degrees of visual angle. The disk was rotated about the participant's line-of-sight and contained low spatial frequency information (16 x 22.5 degree black and white segments). A target disk (which shared the same central fHation point as the large disk) was positioned 3 centimeters in front of the large disk. The target disk subtended 32 degrees of visual angle and was marked with a straight edge (20 cm x 1 cm). The rotation of the target disk was controlled by the participant. The participant was asked to position the target disk so that the straight edge was in the vertical plane before the test commenced (before the disk was rotated), and then the rod was positioned at 28 degrees from the vertical in the direction of the roll. The participant viewed the visual field which was rotated at 96 degrees sec-' for 30 seconds and then while the large disk was still rotating attempted to position the target disk so that the straight edge was aligned with the vertical plane. Six trails were performed-three with the straight edge initially tilted in the direction of the roll and three trials with the straight edge tilted in the direction opposite to the roll. Any error in positioning the straight edge in the vertical plane was measured in degrees (see Figure 2). The tests were administered by one experimenter (S.L.) who was blind to faller status.
Statistical Analysis Student t-tests were performed to examine whether there were differences in age and test performance between the fallers and non-fallers. It was found that the raw data for the roll vection test were positively skewed so log scores were examined. Chi2 tests for contingency tables were used to examine differences in test scores when performance was categorized as normal or abnormal.
Test Reliability The two tests of field dependence were re-administered to thirty-seven participants six months after the initial testing. The extent of the roll vection had a reliability coefficient (test-retest correlation) of 0.93, while rod and frame field dependence had a reliability coefficient of 0.66 (both significant at p < 0.01). All of the tests were administered by the one investigator on both occasions. RESULTS Mean scores plus standard deviations of the test results for the fallers, nonfallers, and total sample are shown in Table 2. The extent of the roll vection and rod and frame field dependence were significantly higher in the group who reported falls in the past year than in the group who reported no falls (t = 3 . 0 2 , ~= 0.003 and t = 2.59, p = 0.011 respectively). Mean error in perception of the true vertical in the rod and frame test was 20.7 degrees for the fallers and 17.2 degrees
274 /
LORDANDWEBSER
Table 2.Mean Error in Perception of the True Vertical in the Rod-and-Frame and Roll Vection Tests Faller/Non-Faller Comparisons
Test
Fallers (N = 59) Mean (S.D.)
Non-Fallers (N = 77) Mean (S.D.)
Total (N = 136) Mean (S.D.)
Rod and Frame a Roll Vection *
20.66 (6.60) 6.62 (6.22)
17.24 (8.52) 3.63 (4.86)
18.70 (7.91) 4.93 (5.67)
a Significant difference between fallers and non-fallers at p c 0.02. Significant difference between fallers and non-fallers at p c 0.01.
Table 3.Participants with Abnormal Responses in the Rod-and-Frame and Roll Vection Tests. Fallermon-Faller Comparisons Test
Rod and Frame a ROII vection
Fallers N (Percent)
Non-Fallers N (Percent)
47 (83.9) 31 (53.4)
55 (73.3) 25 (33.3)
a Set rod more than 12 degrees from the vertical, definition of abnormality from Tobis et al. [13]. Chi2 = 2.09; df = 1, NS. Set rod more than 4 degrees from the vertical, new definition of abnormality. Chi2 = 5.43, df= 1, p c 0.02.
for the non-fallers. Mean error in perception of the true vertical in the roll vection test was 6.6 degrees for the fallers and 3.6 degrees for the non-fallers. Table 3 shows the number and proportion of fallers, non-fallers, and the total sample who recorded abnormal results in the tests. Twelve degrees from the vertical was used as the criterion for the rod-and-frame test as this value was reported by Tobis et al., to discriminate between fallers and non-fallers [13]. The value of 4 degrees for the roll vection test was developed after examining frequency distributions. When 12 degrees from the vertical is used as the criterion for an abnormal response in the rod and frame test, it was found that most fallers (84%) and non-fallers (73%)were classified as abnormal. In the roll vection test, 4 degrees was used as a criterion for an abnormal response. Using this criterion, 53 percent of fallers were classified as abnormal with only 33 percent of the non-fallers. Pearson correlation coefficients were computed to examine the relationships between performance in the roll vection and rod and frame tests for the fallers, non-fallers, and total sample. The correlation coefficients (r) were 0.128 (p = 0.177) for the fallers group, 0.421 (p = 0.000) for the non-fallers group, and 0.379 (p = 0.000) for the total sample.
VISUAL FIELD DEPENDENCE / 275
DISCUSSION
It has been clearly demonstrated that tilted and moving visual fields can affect the perceived upright. This apparent displacement of the postural vertical appears to be the result of a central recomputation that results from an attempt to weigh the contradictory sensory inputs of the vertical [17]. The most extensive series of experiments on visual field dependence using stationary tilted fields were undertaken by Witkin et al., in the 1950s [12]. They found wide individual differences in all conflict situations, but that individuals tended to be consistent from situation to situation. Wide individual differences have also been found in the extent of roll vection [16]. The present study found that in both the rod and frame and roll vection tests, fallers aligned the rod at greater deviations from the vertical compared with non-fallers. As in the study by Tobis et al. [13], a significant association was found between falls and performance on the rod and frame test. The main difference being that in the present study, higher proportions of both fallers and non-fallers aligned the rod at substantial deviations from the vertical. Tobis et al., found that an average error in excess of 12 degrees in the rod and frame test may predict serious falls among independently functioning older adults. This difference in the proportion of the participants who aligned the rod at large angles from the upright in the two studies may be the result of sample differences. Tobis et al.’s sample was drawn from three groups: working seniors, senior citizens center attenders, and affluent seniors; all of whom resided in the community. In contrast, the present sample was drawn from residents of a retirement village-mostly hostel residents. The incidence of falls in this population (43%) is higher than that reported in community studies, and it is likely that this population is made up of frailer, more dependent persons. The mean age of the present sample (80.9 years) is also greater than in the study by Tobis et al. (72.5 years). These factors, increased age and dependency, may be associated with increased visual field dependence. Another reason for the high proportion of participants who set the rod at substantial deviations from the vertical could be that the rod and frame test procedure is open to misunderstanding by the participant, no matter how carefully it is explained. The test of roll vection has an advantage over the rod and frame test in that the participant can be asked to position the rod in the vertical plane before the test commences. This ensures that the participant understands the test procedure. In the rod and frame test, however, this cannot be done because the positioning of the rod in the vertical plane prior to the test provides the participant with cues as to how to position the rod-it simply “gives the game away.” With elderly participants, especially those with sub-clinical confusional states, understanding of the task requirements cannot be assured. Performance in the two tests was significantly associated in the non-faller’s group, however, there was only a weak association in the faller’s group. This lack of association appears to support the
276 / LORD AND WEBSTER
suggestion that some fallers may not understand the test procedure for the rod and frame test. The test of roll vection was also found to be a better discriminator between fallers and non-fallers than the rod and frame test. This may be because the participant undertakes the test standing (rather than sitting) and because the rotating disk provides a powerful visual stimulus. The effect of the rotating field was clearly disorienting and unpleasant for certain elderly fallers. The results of the present study suggest that elderly fallers have a greater dependence on visual information than elderly non-fallers. These findings support the hypothesis that increased visual dependence shown by fallers may develop in response to impaired proprioceptive and vestibular systems as a result of age and chronic health problems. The study examined elderly participants under artificial situations, and the findings may only partly generalize to real world situations-although the implications are that tilted or rolling visual stimuli may be related to factors that lead to postural instability and falls in the elderly. For instance, individuals with poor proprioception and vestibular sense, when exposed to “misleading” tilted or moving visual stimuli such as tilted forms and landmarks, moving clouds, vehicles, or structures may be at increased risk of falling.
REFERENCES 1. A. J. Campbell, J. Reinken, B. C. Allan, and G. S. Martinez, Falls in Old Age: A Study of Frequency and Related Clinical Factors,Age andAgeing, 10, pp. 264-270,1981. 2. D. Prudham and E. J. Grimley, Factors Associated with Falls in the Elderly: A Community Study,Age andAgeing, 10, pp. 141-146,1981. 3. M. E. Tinetti, M. Speechley, and S. F. Ginter, Risk Factors for Falls among Elderly Persons Living in the Community, The New England Journal of Medicine, 319, pp. 1701-1707, 1988. 4. A. J. Blake, K. Morgan, M. J. Bendall, H. Dallosso, S. B. J. Ebrahim, T. H. D. Arie, P. H. Fentem, and E. J. Bassey, Falls by Elderly People at Home: Prevalence and Associated Factors, Age and Ageing, 17, pp. 365-372, 1988. 5. U.Lucht, A Prospective Study of Accidental Falls and Resulting Injuries in the Home among Elderly People, Acta Sociomedica Scandinavica, 2, pp. 105-120,1971. 6. M. Smallegan, How Families Decide on Nursing Home Admission, Geriatric Consultations, 1 5 , pp. 21-24,1983. 7. D. N. Lee and J. R. Lishman, Visual Proprioceptive Control of Stance, Journal of Human Movement Studies, 1, pp. 87-95,1975. 8. C. Owsley, R. Sekuler, and D. Siemsen, Contrast Sensitivity throughout Adulthood, Vision Research, 23, pp. 689-699,1983. 9. D. G. Pitts, The Effects of Aging on Selected Visual Functions, in Aging in Human Visual Functions, R. Sekuler, D. W. Wine, and K. Dismukes (eds.), Liss, New York, 1982. 10. I. P. Howard and W. B. Templeton, Human Spatial Orientation, John Wiley and Sons, London, 1966.
VISUAL FIELD DEPENDENCE / 277
11. R. Over, Possible Visual Factors in Falls by Old People, Gerontologist,6, pp. 212-214, 1966. 12. H. A. Witkin, H. B. Lewis, M. Hertzrnan, K. Machover, P. B. Meissner, and S . Wapner, Personality through Perception, Harper, New York, 1954. 13. J. S. Tobis, S. Reinsch. et al., Visual Perception Dominance of Fallers among Community Dwelling Older Adults, Journal of the American Geriairic Society, 33, pp. 330333,1985, 14. D. N. Lee and J. R. Lishrnan, Vision-The Most Effective Source of Proprioceptive Information for Balance Control,Agressologie, 18a pp. 83-94,1977. 15. W.Bless, T. S. Kapteyn, and G. De Wit. Effects of Visual-Vestibular Interaction on Human Posture, Advances in Oto-Rhino-Laryngology,22, pp. 111-118,1977. 16. J. Dichgans, R. Held, L. R. Young, and T. Brandt, Moving Visual Scenes Influence the Apparent Direction of Gravity, Science, 178, pp. 1217-1219,1972. 17. T. Brandt, W. Paulus, and A. Straube, Vision and Posture, in Disorders ofPosture and Gait, W . Bles and T.Brandt (eds.), Elsevier, Amsterdam, 1986. 18. P. K. Oltman, A Portable Rod-and-Frame Apparatus, Perception and Motor Skills, 267, pp. 503-506,1968.
Direct reprint requests to: Stephen R. Lord School of Community Medicine University of New South Wales P.O. Box 1 Kensington, N.S.W.2033 AUSTRALIA
