- Email: [email protected]
Age related effects of restricted head movements on the useful field of view of drivers
Accid. Anal. and Prev., Vol. 29, No. 6, pp. 793-801, 1997 0 1997 Elsevier Science Ltd All rights reserved. Printed in Great Britain OOOI-4575/97 $17.00 + 0.00
Pergamon PII: s0001-4575(97)00048-1
AGE RELATED EFFECTS OF RESTRICTED HEAD MOVEMENTS ON THE USEFUL FIELD OF VIEW OF DRIVERS ROBERT B. ISLER, BARRY S. PARSONSON and GLENN J. HANSSON Universityof Waikato,PsychologyDepartment,Trafficand Road Safety Research Group, Private Bag 3 105, Hamilton, New Zealand
Abstract-Eighty drivers in a sample of four groups of 20 (10 males, 10 females), aged under 30 years (‘young’), 40-59 (‘middle aged’), 60-69 (‘older’), and 70 years and over (‘oldest’), participated in tests of head rotation and of several visual functions relevant to safe driving. Head rotation data showed that the oldest drivers had lost about l/3 of movement and that the loss tended to be more evident in males. Second attempts almost always produced slightly better results. All participants had at least 20/40 binocular vision, however, tests of monocular visual acuity, stereovision, and horizontal peripheral vision revealed that the poorest performers were aged 60 years and over and that the degree of decrement increased with age. Many of the older and oldest drivers in the sample were both severely restricted in their ability to turn their head and evidenced a loss of visual functioning. Analysis of the effect of reduced head movement on the useful field of view indicated that, for the drivers aged 60 years and over, there was an evident restriction on the distances at which approaching traffic could be brought into the central, stationary field, so that even at maximum head rotation plus one saccade (lY), approaching vehicles would not be clearly perceived beyond a distance of 50 m. The findings are discussed in relation to older drivers’ involvement in intersection accidents. 0 1997 Elsevier Science Ltd. Keywords-Traffic
accident,
Older driver, Head movement,
INTRODUCTION
Useful field of view, Intersection
Driving safely through intersections requires gathering distal and proximal visual information from many different angles, normally received through eye and/or head movements and, for visual areas behind the drivers, through the use of a vehicle’s mirrors. Blind angles cover areas in the traffic scene from which the drivers cannot extract information. Such angles are determined to some extent by external physical factors such as the vehicle’s design (especially A, B, and C pillar location and thickness, and rear window area, shape, and slope) and the quality, size, and location of the mirrors or barriers such as buildings, fences, trees, or land contour features. However, to a much larger degree, blind angles vary according to the physical and perceptual capacities of the drivers. They are minimized if drivers have good head turning and eye movement capabilities and are readily able to move their head and eyes towards peripherally located road scenes. In order to perceive important information in detail, such as traffic signs and relevant traffic, it needs to be first brought into central vision, that small area of the visual field (less than three degrees) in which visual acuity is highest. Eye movement
The number of drivers on the road in developed countries who are aged 60 years and older is increasing, and there also is evidence of an increasing involvement in road crashes among such drivers (Barr, 1991; Drummond and Yeo, 1992; Waller, 1991). International traffic accident statistics indicate that a major contributing driving error is failure to yield right of way at intersections (Evans, 1991a), with older drivers over represented in such accidents. Similarly, Garber and Srinivasan ( 1991) found that the probability of older drivers committing violations at an intersection or during change of lane manoeuvres was much higher than for younger drivers. Intersections are particularly demanding traffic environments for drivers, requiring a series of complex visual perception and decision processes. When drivers approach intersections, visual scanning using eye and head movements occurs (Kito et al., 1989), providing data for directional and ‘Go/No Go’ decisions. The directional decisions follow detection and interpretation of road signs and lane markings while the ‘Go/No Go’ decisions are based on judgements of the speed and distance of approaching traffic. 793
194
R. B. ISLER et al.
studies in real traffic situations show that drivers continually scan the road scene for useful information (e.g. Mourant and Rockwell, 1972). When peripheral information enters the visual field, attentional processes determine which information needs to be further inspected in detail using central vision. For some roading information, its appearance in the visual field does not necessarily mean that it will be perceived. For example, in some ‘looked but failed to see’ incidents, the drivers cannot perceive information in their visual fields because of attentional deficits, or because the information was presented to a particular area of the retina which is functionally blind. A number of visual deficits occur naturally with aging, including, cataracts, macular degeneration, open-angle glaucoma, and diabetic retinopathy, and are likely to affect driving performance (Klein, 199 1; Wood and Troutbeck, 1992). Visual and behavioural deficits (e.g. hemiagnosias) that affect driving performance can result from strokes. Drivers in the older age groups are the most likely candidates for these types of impairment and it seems that they often are not aware of such visual loss because of gradual onset or may not reveal it for fear of loss of their licence. Keeping the high accident rate of older drivers at intersections and during change of lane manoeuvres in mind, along with the changes in visual acuity and head movement ability, it seems reasonable to assume that the range, in degrees, of blind angles increases with age. Many intersection and change of lane accidents involving older drivers may, therefore, be caused by their not perceiving some of the important information essential for making safe decisions in those traffic environments. Recent research by Owsley et al. (1991) found that, in a sample of 53 older drivers (57-83 years), the size of useful field of view (UFOV) correlated well with accident frequency. The size of the full UFOV is smaller than the visual field as determined by clinical perimetry and does not cover the whole area of visual sensitivity (Ball et al., 1990). It consists of a small binocular stationary field and a larger eye field (binocular in more central areas) and refers to the total visual field area within which particular information can be acquired without eye and head movements, i.e. within one eye fixation (Sanders, 1993). According to Sanders, two stimuli within a small angular separation can be processed as a single one in the stationary field, while in the eye field subjects can obtain only an hypothesis from a peripheral stimulus before centrally fixating. Clearly defined eye movements can then be initiated in order to inspect peripheral stimuli in detail. Such eye movements are found to be much faster than for a new percept without an hypothesis.
Important boundaries lie between the eye field, which is part of the UFOV, and the two laterally extended monocular eye fields and head fields, from which no information can be obtained while attending to a central task. Appropriate eye movements may give access to stimuli in the extended eye fields, while head movements and, possibly, more eye movements would be necessary in order to detect peripheral stimuli in the head fields (Sanders, 1993). The boundaries between the eye fields and the extended eye fields are not clearly defined because they vary substantially across individual drivers and situations. Research has shown that increase of central cognitive load, stress, noise, and aging are factors, which may result in a generalized shrinkage of the (peripheral) eye field (see, for example, Williams, 1982; Ikedo and Takeuchi, 1975; Leibowitz and Appelle, 1969; Mackworth, 1965; Scialfa et al., 1987). In such situations, effective eye and head movements are particularly important as a ‘narrower beam of vision’ needs to be moved around faster to get the same information from the driving environment. It is a common observation that many older drivers appear to find it increasingly difficult to turn their head beyond certain angles owing to physical neck conditions (e.g., arthritis, lower muscle tone, neurological disorders). In severe cases, where head rotation is strictly limited, very little information from the head field is accessible, as body movements, which normally would serve to compensate for impaired head movements, are likely to be restricted by the car seat and seat belt. Another option would be to compensate head movement restrictions with larger eye movements. However, as Bahill et al. (1975) report, most of the naturally occurring eye movements (saccades) have an amplitude of only 15” or less and larger eye movements would put considerable strain on the saccadic system. Additionally, Carter et al. (1983) found that older people have longer latencies of saccadic eye movements. Surprisingly, while much attention has been given to research on older drivers’ visual deficits, there is no similar body of research available which examines the effects of restricted head movements on older drivers’ driving performance. As head movement restrictions seem predominantly to be a problem of older drivers and as these drivers are over-represented in intersection and manoeuvring accidents, where the accessibility to the extended eye fields and head fields of the field of view is particularly important, research in this area seems pertinent. This study aimed to assess the degree of head movement restriction in drivers in four different age
Age related
effects of restricted
groups and to test several visual functions relevant to UFOV in order to examine the relation between age, head movement, and UFOV.
METHOD Participants
The sample consisted of 80 drivers, comprising 10 male and 10 female participants in each of four age groups, ‘young’ (under 30 years, mean age M= 23.4), ‘middle aged’ (40 to 59 years, M=49.9); ‘older’ (60 to 69 years, M=65.3), and ‘oldest’ (70 years and over, M=76.1). In addition to equal numbers of male and female participants, an effort was made to include participants in the broader community, from various ethnic backgrounds and occupations. Participants were recruited as volunteers (with travel costs reimbursed) through an advertisement in a local newspaper, through contacts with Age Concern (a self help organization of older people), and by word of mouth. Apparatus
A brief driver behaviour questionnaire sought information on participants’ self-reported number of accidents over the last five years. Additionally, there was a question regarding participants’ self-judged driving ability with three response categories; below average, average, and above average. A consent form was used which outlined what was required of the participants, assured confidentiality, and acknowledged other ethical obligations. Most vision testing (binocular and monocular, visual acuity and stereovision) was conducted with the aid of a vision screener (Keystone VS-II, model 1135A from Keystone View, Division of Mast/Keyestone, Inc., NV, USA). A standard perimeter (Lafayette Instrument Company, Inc., IN, USA) was used to measure horizontal peripheral vision. Different coloured target disks (red, yellow, green, blue and white, 5 mm diameter) served as targets, however, only the data obtained with the white disk were analysed for this study. In order to measure the degree of maximum head turning, a head turning measurement device was developed. It consisted of a cycle helmet with a stylus pointer protruding from the centre front, parallel with the centre of the wearer’s line of vision when the eyes were looking directly ahead. A plumb bob suspended on the cord from the stylus allowed a reading to be taken of head rotation (in degrees of arc). The 210” arc, marked in five-degree segments from zero at the front centre to 105” on either side, was made of transparent plastic and mounted on a shoulder harness.
795
head movements
Procedure
After the participants had signed the consent form and completed the questionnaire, they performed the vision tests, followed by the test which measured their maximal degree of head movement. Visual acuity. Far point (6 metres) monocular and binocular visual acuity testing was performed. Seven ratings, Snellen values from 20/20 (highest visual acuity) to 20/200, were possible. Stereovision (depth perception). Stereovisual performance defines the extent to which objects may be discriminated in three dimensions. In this test, the participants identified which of five different symbols appeared to stand out from the others. Five levels of stereovision were tested, with the first being the easiest (10% stereovision), followed by three intermediary levels (30, 60, and 75%) and the final being the most difficult (85% stereovision). Seventy-five percent stereovision (level 4) or greater, was considered acceptable according to criteria set by the Keystone VS-II Vision Screener standards guide. Horizontal peripheral vision. The degree of peripheral vision was determined with five different coloured target stimuli (red, yellow, green, blue and white). Starting always from one or other peripheral side (90° from centre), each of the three target stimuli were moved along the perimeter towards the centre until the participant was able to perceive and name the colour of the test stimulus. Only the values for the white target stimulus were considered for the analysis. Maximum head movements. The head turning measurement device was used to determine the range (in degrees) of head movements, separately to the left and the right side, for each participant. Once the participant had turned their head to the maximum of articulation on either side, a record of the degree of turn was made. Data were obtained from two trials on each side.
RESULTS The results from the questionnaire on accident frequency during the last five years and self-judged driving ability revealed that the eighty participants reported a total of 23 accidents during that period (Table 1). Eight (35% of the sample) involved the youngest group (~30 years), of these, five of the 10 young male drivers reported having at least one accident during the last five years. Eight drivers (40%) from the 40 to 59 year-old age group each reported a single accident over that time span. The older and oldest participants (60-69 years and 70 + years of age) reported 7 (33%) accidents, however 25 drivers
196
R. B.
Table 1. Results from the Driver Behaviour Vision Testing
Questionnaire
and the
Age groups Total N=80 N
<30 n=20
40-59 n=20
60-69 n=20
70+ n=20
40
8
40
3
15
4
20
45 55
7 13
35 65
10 14
50 50
5 5
25 25
0 55 45
0 9 11
0 45 55
0 10 10
0 50 50
0 12 8
0 60 40
0
1
5
4
20
7
35
10
1
5
3
15
10
50
15
5
25
9
45
13
65
n%n%n%n%
Number of accidents last five years 23 8 Distance travelled per year 41 9 < 10,000 km 11 > 10,000 km 39 Self reported driving ability below average 0 0 average 42 11 9 above average 38 Monocular visual acuity less than 20/40 12 0 Stereovision 2 less than 75% 16 Horizontal peripheral vision 3 less than 30” 29
(63%) from these groups estimated their travel distance to be less than 10,000 km per year. The question regarding participants’ self-judged driving ability revealed that not one person in the sample of 80 drivers considered him or herself as a ‘below average’ driver. Eight of the 10 young male drivers rated themselves as above average (not shown on Table l), while almost all young female drivers identified themselves as average drivers. About equal numbers in the remaining age groups rated themselves either as average or above average drivers. There was no clear gender effect on the two eldest age groups. Table 1 also shows the frequency distribution of visual acuities tested for both eyes (binocular) and separately for the left and right eye (monocular) by age. All drivers had binocular visual acuity above the minimal standard of 20/40 required for driving in most countries. However, 12 drivers had severely impaired visual acuity in one of their eyes. Four of these were older drivers and seven were in the oldest age group. Of the latter, four had such a severe visual acuity impairment in one eye (20/200 or less) that they could be considered as functionally one-eyed drivers. Table 1 further shows the frequency distribution for stereovision levels (expressed in percentages of full stereovision) by age. Sixteen participants (20% of the whole sample) had less than 75% stereovision (unacceptable according to criteria set for the Keystone VS-II Screener). Ten ( 13%) of these drivers were from the oldest age group. The results of the horizontal peripheral vision task also revealed an age-related loss of such vision. As Table 1 shows, the percentage of drivers with less than 30” of peripheral
ISLER
et
al.
vision increased for the older (45%) and oldest (65%) age groups. Figure 1 shows mean degrees of maximum head movement, for the first and second attempt, by age and gender. A preliminary analysis (t-test) did not reveal a difference between the degrees of maximum left and right head movements pooled for all age groups, so the data for these two conditions were combined. The graph shows a clear age effect, with a gradual decline in mean degrees of maximum head movement across the age groups. The youngest participants had a mean maximum head movement of 86”, which is 14” more than the corresponding mean value of the 40 to 59 year-old participants, 19” more than the 60 to 69 year-olds, and 27” more than the 70 year-old and above participants. Except for those in the 60 to 69 year-old group, the female participants in each group were normally able to turn their head further than their male counterparts. Second attempts typically produced higher degrees of head movement (mean value of improvement taken over all groups: M= lY), with the middle-aged female group (40-59 years) as the only exception. A 4 x 2 x 2 (age groups x gende x trial, first versus second attempt) analysis of variance on the degrees of maximum head movement with repeated measures on the factor Trial confirmed main effects for Age, F(3, 72) =24.3, ~~0.01, Trial F(1, 72)=10.3, ~~0.01, but not for Gender F( 1, 72) =2.6, p>O.O5. There were no significant interactions. Scheffe’s post hoc tests revealed that the youngest age group had significantly greater maximum head movement compared with the middle, older and oldest age groups (all ps ~0.05). There were no significant differences between the maximum head movement when comparing either the middle age group with the older age group or the older age group with the oldest age group. However, the middle aged group produced significantly greater head movements than the oldest age group (~~0.05). Figure 2 shows a scatterplot for the whole sample with their degrees of maximum head movement in the right plane on the x-axis and the individual size of their right horizontal peripheral visual field as measured with a standard perimeter, on the y-axis. The graph reveals that in most cases the older people with severely restricted head movements (e.g. 60” or less) also had small degrees of horizontal peripheral vision. A significant Pearson’s product-moment correlation between these two variables was obtained, r=0.34, p ~0.05. Most young drivers (< 30 years) had a maximum head movement of more than 70” while the majority of the older and oldest drivers had maximum head movement of less than 70”. Three in the sample of oldest drivers had less than 50” of head
Age related
m -30
f <30
effects of restricted
m40-59
Gender Fig. 1. Mean degrees
40
50
60
70
of maximum
80
90
f 40-59
(m=male;
head movement,
100
110
Maximum Head Movement (deg) Fig. 2. Scatterplot and regression line with 95% confidence interval of individual degrees of maximum head movement plotted against sizes of right horizontal peripheral visual field.
movement and two of these drivers had also less than 20” horizontal peripheral vision. A further analysis was attempted in order to evaluate the potential effects of restricted head movements on the visual field of the sample of drivers assuming that they were arriving at the decision point of a T-intersection with an assumed road width of 8 m. Cohen (1996) showed that such a road width would be most adequate for rural roads. At the decision point drivers would need to judge the distance and speed of the oncoming traffic from the right in order to find a safe gap to cross the intersection and merge with the traffic on the left lane (drivers in New Zealand keep to the left side of the road).
797
head movements
m60-69
f=female)
f 60-69
m70+
f 70+
J
and Age (years)
for the first and second attempt,
by age and gender.
First, when looking straight ahead, for all drivers the oncoming traffic would appear almost perpendicular to their line of sight, in their extended lateral monocular eye or head field and no distance and speed judgements would be possible. Normally, much head movement, possibly supported by some eye movements, would then be initiated in order to bring the traffic into central vision to allow speed and distance judgements. In cases of restricted head movement, however, larger performed horizontal eye movement angles would be needed in order to compensate for the head movement loss. Figure 3 shows for each participant in each age group their maximum achieved head movement angle and their calculated critical detection distance, Dcrit, by which oncoming traffic would be brought into central vision assuming that no additional eye movements were performed (round symbols). Dcrit was calculated as
Dcrit = Ld/tan ~1 where Ld=lateral distance (about 4 m) from participant to oncoming car if road width was 8, c1=maximum achieved head movement angle. Figure 3 further shows Dcrit for each driver assuming that they would support their maximum head movement with a 15” horizontal eye movement (squared symbols). Such a movement corresponds to one natural saccade (Bahill et al., 1975). Finally, Fig. 3 also shows Dcrit values (triangular symbols) assuming that drivers’ maximum head movement was supported by 30” of horizontal eye movement (two natural saccades). The data in Fig. 3 show that the maximum achieved head movement angles of the older drivers would not be sufficient to bring oncoming traffic in central vision at distances exceeding 20 m without additional eye movements to the right. In contrast,
798
R. B. ISLER et al.
Drivers
50
<30 Years
60 Maximum
Fig. 3. Calculated movement (HM), 30” eye movement those on the right
and 40-59
70 Head
Movement
Years
60
Dnvers
90
100
110
(deg)
60-69
50
Years
60 Maximum
70 Head
critical distance by which oncoming traffic would be brought into central vision by maximum head movement and additional 15” eye movement (HM+ 1.5), and maximum (HM + 30). The data shown on the left are for the ‘young’ and ‘middle aged’ drivers (i are for the ‘older’ drivers and ‘oldest’ drivers (60-69 and 70+ years). Closed symbols restricted right horizontal peripheral vision of less than 30”.
many of the young drivers were able to turn their head for more than 90” and would have been able to focus on the traffic without additional eye movements at the largest distance on the scale of the figure (> 100 m). With an additional 15” of eye movement to the right (one natural saccade), many older drivers (N=20 or 50%) still would not be able to focus centrally on approaching traffic which is further away than 40 m, while only six of the younger drivers would not be able to do so. Even after large eye movements of 30” (two natural saccades), eight of the older drivers would not be able to fixate on such oncoming traffic (40 m away), while only three drivers (all three were in the 40-59 years group) of the younger group would have been unable to do so. Very often, in addition to their restricted head movements, the oldest drivers also had restricted right horizontal peripheral vision of less than 30” (see closed symbols in Fig. 3). Four of the oldest drivers who had such restricted head movements that they would not have been able to judge the speed and distance of oncoming car at a distance of 50 m, even with more than two natural eye movements, also had severely restricted peripheral vision of less than 30”.
DISCUSSION The findings of the present study indicate that there is a significant age-related decrement in maximum achieved head movement and in several visual functions relating to UFOV in drivers aged 60 years or more. Analysing the effects of head movement loss on the UFOV of these participants as if they were arriving at the decision point of a T-intersection
and 70+ Years
Movement
60
90
100
110
(deg)
means of only maximum head head movement and additional 30 years and 40-59 years) and represent participants who had
revealed that many older drivers would have difficulty bringing vehicles beyond 50 m into central focus. As part of the assessment process, participants also had responded to a brief questionnaire focussing on their self-reported driving history, number of accidents during the last five years, and self-judged driving ability. The questionnaire response pattern that stood out was that of the young male drivers (< 30 years). This particular group reported the highest number of accidents and, more than for drivers in any of the other groups, considered themselves to be above average drivers. None of the drivers in any age group rated themselves as below average drivers. Such response patterns match well with earlier research findings and recent accident statistics (e.g. Guerin, 1994; Page et al., 1992). The response patterns of the oldest age group did not differ substantially from the middle-aged (40-59 years) and older (60-69 years) groups. However, caution may be needed when interpreting the self-reported data of older drivers as they may be reluctant to reveal any information which could question their driving competence. For example, Sloane et al. (1990), cited in Owsley et al. ( 1991) found that most of the older drivers with the highest number of recorded accidents were men who tended to under-report accident involvement in response to driving history questionnaires. This may be a matter of pride or of fear that disclosure will lead to licence revocation. The results of the vision tests were instructive. While all of the drivers in the sample had acceptable binocular visual acuity (better than 20/40), the monocular vision tests revealed that 12 had severely impaired visual acuity in one of their eyes. Four of
Age related effects of restricted head movements these were in the older (60-69 years) group and seven were in the oldest (70 or more years) group, showing increasing deficit with increasing age. Four of the latter group had such significant levels of impairment (20/200 or worse) that they could have been classified as one-eyed. The stereovision data also revealed an age-related pattern of increasing deficit. Of the 16 (20%) whose scores were less than 75%, 10 came from the oldest group. A statistically significant age effect was obtained for horizontal peripheral vision, with the 22 of the 29 participants having less than 30” falling in the older (9) and oldest (13) age groups. Lack of stereovision seemed to be a predominant feature of half of the drivers aged over 70 years, and loss of visual acuity in one eye could be a primary factor. Judgement of depth, in the absence of stereovision, requires use of monocular depth cues, such as texture, gradient, and relative movement. Gradual loss of vision in one eye may not be noticed, and the driver may fail consciously to use monocular cues to judge depth and distance as an adaptive response to the loss. Loss of horizontal peripheral vision, also a significant factor in our oldest drivers, Further narrows the range of visual perception in functioning eyes. In all, these results reveal considerable deterioration in monocular visual acuity, stereovision, and horizontal peripheral vision in drivers aged 60 years and over, with the number of those showing deficits increasing as age increases. The seriousness of the deficit also increases with age. The present findings support those of Klein ( 1991) and Shinar and Schieber (1991), who reported in some detail on the causes of specific vision deficits in older drivers and called for better assessment of older drivers’ vision and more and better research into their effects on driving. The various elements of reduced visual compentence (e.g. loss of monocular, stereo-, and horizontal peripheral vision) may compound to create conditions which make effective and safe judgements at intersections and in manouvering difficult for many of the oldest drivers. Kosnik et al. (1990) suggest that older drivers recognise loss of vision and decide to give up driving, but there are issues of gradual, undetected onset and decline and of disincentives in the form of loss of mobility and independence which may delay self-revocation of a driver’s licence to a time beyond the point at which the deficit first became problematic or unsafe for the driver and other road users. This lack of awareness or use of denial seems evident in the present study’s 60-years-and-over participants when comparing their measured deficits against their ratings of their own driving ability as average or above average. This tendency to either be unaware
799
of, or to deny, declining ability is supported by the findings of Barr (1991), Garber and Srinivasan ( 199 1), and Owsley et al. ( 199 1) in respect of accident rates of older drivers, especially at intersections. Johnson and Keltner (1983), cited in Shinar and Schieber ( 199 1) examined 10,000 volunteer California licence applicants and found significant deterioration among drivers aged over 60 years and that drivers with visual field loss in both eyes had accident and conviction rates more than twice as high as age- and sex-matched drivers with normal fields. Use of more extensive and thorough vision and driving skill testing procedures may more accurately reveal both the extent of visual dysfunction among older drivers and the point at which such vision defects begin to contribute to high accident risk. The results of the head movement tests revealed significant age and gender related differences between drivers in their ability to rotate their heads, with those aged 60 years or more having the greater restriction and with males usually performing less well than their female counterparts. The average decrement in the oldest participants was a loss ofabout one-third of head movement in the right lateral plane compared with the group aged under 30 years. Arthritic changes in cervical vertebrae, changes in muscle tone, and neurological conditions affecting head turning may well account for the reduction. Decrements in head rotation of drivers have not previously been studied, but it is a common observation that older drivers have difficulty in turning their head at intersections, often relying on body movement to compensate. The importance of head movements to safe driving were studied by Allan et al. (1974), who found that, in comparing performance of head-movement-restricted drivers with head-unrestricted drivers in a driving simulator, restricted drivers made more driving errors, fewer driving adjustment responses to critical events in the left field (under US driving conditions), and shifted their fixations less often. While average values are useful for detecting overall age related trends in visual and physical impairments, it also is important to determine the extent to which individuals have combinations of deficits. The nature of the vision testing equipment available for the present study meant that it was only possible to produce a scattergram comparing degrees of head rotation and degrees of horizontal peripheral vision. Those data (Fig. 2) clearly show age-related impairment of neck articulation and peripheral vision, such that, in general, the youngest participants had both more head movement and a wider peripheral visual field, and that as age increased, the range of both declined. The data also show great variability between participants, suggesting that broad
800
R. B. ISLER et al.
generalizations cannot be made and that specific testing for deficits is essential. Ball and colleagues (Ball et al., 1988, 1990; Ball and Owsley, 1991) noted that standard perimetry may significantly underestimate the restriction in vision of drivers aged over 60 years, indicating that assessments based on tests of functional vision, such as UFOV, reveal a three-fold reduction compared with young (18-25 years) adults. As a result, it is important to consider how the deficits detected in the present sample may have affected UFOV. If, as part of an overall aging process, a number of the sensory and physical systems contributing to effective search for, identification and perception of, and response to driving hazards become increasingly inoperative, the driver’s UFOV is likely to be considerably reduced. Driving safely through an intersection involves attending, perceiving, and responding to crucial road and traffic information which usually is distributed on a much larger area than drivers’ useful field of view covers. For instance, using the concept of the functional visual field (Sanders, 1993), oncoming traffic appears first in the far peripheral head field so that head movements are necessary in order to bring the oncoming vehicle first into peripheral vision (eye-field) so that it can be detected. Eye movements and, possibly, more head movements are then necessary in order to shift the detected vehicle from peripheral vision (eye field) to central binocular vision (stationary field), where visual acuity and depth perception is best. Restricted neck articulation and vision deficits are likely to interfere with this process and might result, first, in delaying the detection of oncoming vehicles and, second, in delaying the fixation of the vehicle with central vision until it is possibly too close for an accurate judgement of its distance and speed to be made to permit safe clearance of the intersection. In a study of driver head and eye movement in the intersection environment, Kito et al. (1989) found that a driver of a small vehicle preparing to make a right turn (in Europe and the US the equivalent manoeuvre would be a left turn) increased both head and eye movements and gazes (head-eye movement of greater than 20”), especially toward the right, during the approach to, and initial turning at, the intersection. Presumably, these behaviours are intended to enhance detection and response to approaching vehicles. Although saccades provide quicker shifts than head movements (Sanders, 1993) and frequent, rapid eye movements are considered to be more effective in scanning intersections than head movements (Rahini et al., 1990), one saccade has a natural magnitude of only 15” or less (Bahill et al., 1975). Consequently, in the typical intersection
environment in which roads intersect at 90”, saccadic movements alone are not likely to provide a sufficient scan for safe detection of approaching vehicles and gazes (head and eye movement) are likely to be required. The data from the present analysis, set out in Fig. 3, indicate that when the head is rotated to the maximum degree possible, without additional eye movements, many of the drivers aged less than 60 years would be able to bring an approaching vehicle into central vision when it was at a distance greater than 20 m, but that only one of the drivers aged 60 years and over could do so, and then only just. Adding the equivalent of one saccade would have allowed almost all of the under 60-year-olds to bring a vehicle at a distance of over 50 m into central vision, while only a few of the 60-years-and-over drivers would have been able to have done so. Moving the eye the equivalent of two saccades (30”) would still suggest that 25% of the older drivers would be unable to bring a target vehicle approaching at a distance beyond 50 m into central vision. This means that attention to, and perception of, approaching vehicles would be delayed, with the delay becoming more problematic and dangerous in a rural environment, where vehicle approach speeds are higher. It is important for drivers at an intersection to be able to bring oncoming traffic into their (stationary field) central vision in order to perform the speed and distance estimations necessary for them to decide whether a gap is large enough to safely enter the traffic flow. A driver whose visual field and head movements are restricted would, in the intersection setting, have to try to detect and bring approaching traffic into central vision by turning the upper torso and by turning the eyes as far as possible. However, the car seat and seat belt may constrain the degree of torso turning and, in drivers requiring corrective lenses, the eye may turn beyond the limit of the lens and/or have vision obstructed by the spectacle frame. As a consequence, these alternatives may not effectively compensate for loss of head movement and visual functions. Another possible compensatory response, making more rapid ocular movements, may not be available to some older drivers. For example, Carter et al. (1983) found an age-related increase in the latency of ocular movements, which would place older persons at a disadvantage in everyday situations involving visual search. Age-related restriction of head movement must compound visual deficits in stereovision and peripheral vision, increasing the difficulty, for older drivers, of bringing the image of an approaching vehicle into central (foveal) vision, where visual acuity is highest. This combination of factors may,
801
Age related effects of restricted head movements
in part, explain why older drivers have higher reported rates of intersection accidents that resulted in injury or death in a number of highly motorized countries (Bernhoft, 1990; Evans, 199 1b; Faulkner, 1975 cited in Hills, 1980; Frith, 1991; Mackie, 1972 cited in Hills, 1980; Garber and Srinivasan, 1991). It may also be a factor in the older drivers’ higher rates of lane-change (Garber and Srinivasan, 1991) and U-turn (Frith, 1991) accidents, since head movements and good vision are essential for checking mirrors and minimizing blind-angles. Given that there are increasing numbers of older drivers on the roads in many developed countries (Barr, 1991; Waller, 1991), and that the probability of accidents increases with age, as do defects in vision and restrictions in head movement, it is important that better provisions be made for the assessment of these various functions as part of the regular driver’s licence examinations of older drivers. Acknowledgements~The authors wish to thank Rob Bakker for technical assistance. This research was commissioned by Transit New Zealand.
REFERENCES Allan, J. A., Schroeder, R. S. and Ball, P. G. (1974) Effects of head restriction on drivers eye movements and errors in simulated dangerous situations. Journal of Applied Psychology
59, 5, 643-648.
Ball, K., Beard, B., Roenker, D., Miller, R. and Griggs, D. (1988) Age and visual search: expanding the useful field of view. Journal of the Optical Society of America 5, 2210-2219.
Ball, K., Owsley, C. and Beard, B. (1990) Clinical visual perimetry underestimates peripheral field problems in older drivers. Clinical Vision Science 5, 2, 113-125. Ball, K. and Owsley, C. (1991) Identifying correlates of accident involvement for the older driver. Human Factors 33, 5, 5833595.
Bahill, A. T., Adler, D. and Stark, L. (1975) Most naturally occurring human saccades have magnitudes of 15 degrees or less. Investigative Ophthalmology 14, 6, 468-469.
Barr, R. A. (1991) Recent changes in driving among older adults. Human Factors 33 and 533 (5), 597-600. Bernhoft, I. M. (1990) Elderly drivers: results from a Nordic in-depth study on elderly car drivers. In Third European Workshop on Recent Developments in Road Safety Research. Linkoping: VTI Report 3664 Road Safety Research, pp. 11 l-130. Linkoping, VT1 Report 366A.
Carter, J. E., Obler, L., Woodward, S. and Albert, M. (1983) Effect of increasing age on the latency of saccadic eye movements. Journal of Gerontology 38, 318-320.
Cohen, A. S. (1996) Psychisch bedingte Strassenbreite. Tuef Rheinland Publishers, Koeln. Drummond, A. E. and Yeo, E. (1992) The risk of driver crash involvement as a function of driver age. Monash University Accident Research Centre Research Report, Vol. 49, Melbourne, Australia. Evans, L. (1991a) Older-driver risks to themselves and to
other road users. Transportation Research Record 1325 34-41, TRB, National Research Council, Washington, D.C. Evans, L. ( 1991b) TrafJic Safety and the Driver. Van Nostrand Reinhold Publishers, New York. Frith, W. J. (1991) The safety of the elderly road user in New Zealand relative to road users in other age groups. In Proceedings of the International Conference on Road Safety, pp. 548-551. New Delhi, India. Garber, N. J. and Srinivasan, R. (1991) Characteristics of accidents involving elderly drivers at intersections. Transportation Research Record, 1325, 8-16, TRB, National Research Council, Washington, D.C. Guerin, B. (1994) What do people think about the risks of driving? Implications for road safety interventions. Journal of Applied Social Psychology 24, 99441021. Hills, B. L. ( 1980) Vision, visibility, and perception in driving. Perception 9, 1833216. Ikedo, M. and Takeuchi, T. (1975) Influence of fovea1 load on the functional visual field. Perception and Psychophysics 18, 2555260.
Kito, T., Haraguchi, M., Funatsu, T., Sato, M. and Kondo, M. (1989) Measurements of gaze movements while driving. Perceptual and Motor Skills 68, 19-25. Klein, R. ( 199 1) Age related eye disease, visual impairment, and driving in the elderly. Human Factors 33, 5, 521-525.
Kosnik, W. D., Sekular, R. and Kline, D. W. (1990) Selfreported visual problems of older drivers. Human Factors 32, 5, 597-608.
Leibowitz, H. and Appelle, S. (1969) The effect of a central task on the luminance thresholds of peripherally presented stimuli. Human Factors 11, 387-392. Mackworth, N. H. (1965) Stimulus density limits the useful field of vision. Psychonomic Sciences 3, 67-68. Mourant, R. R. and Rockwell, T. H. (1972) Strategies of visual search by novice and experienced drivers. Human Factors 14, 325-335.
Owsley, C., Ball, K., Sloane, M. E., Roenker, D. L. and Bruni, J. R. (1991) Visual perceptual/cognitive correlates of vehicle accidents in older drivers. Psychology and Aging 6, 4033415.
Page, D. G., McDonald, C. W. and Ryba, K. A. (1992) Road safety and young drivers. A report to the Task group on road safety for young drivers. Creative Challenge Services Ltd, Palmerston North, New Zealand. Rahini, M., Briggs, R. P. and Thorn, D. R. (1990) A field evaluation of driver eye and head movement strategies toward environment targets and distracters. Applied Ergonomics
21, 4, 2677274.
Sanders, A. F. (1993) Processing information in the functional visual field. In Perception and Cognition, eds G. d’ydewalle and J. Van Rensbergen. Elsevier, Amsterdam. Scialfa, C. T., Kline, D. W. and Lyman, B. (1987) Age differences in target identification as a function of retinal location and noise level: an examination of the useful field of view. Psychology and Aging 2, 14-19. Shinar, D. and Schieber, F. (1991) Visual requirements for safety and mobility of older drivers. Human Factors 33, 5, 5077519.
Waller, P. (1991) The older driver. Human Factors 33, 5, 499-505.
Wood, J. M. and Troutbeck, R. (1992) Effect of restriction of the binocular visual field on driving performance. Ophthalamic
and Physiological
Optics 12, 291-298.
Williams, L. J. (1982) Cognitive load and the functional field of view. Human Factors 24, 6, 6833692.
Pergamon PII: s0001-4575(97)00048-1
AGE RELATED EFFECTS OF RESTRICTED HEAD MOVEMENTS ON THE USEFUL FIELD OF VIEW OF DRIVERS ROBERT B. ISLER, BARRY S. PARSONSON and GLENN J. HANSSON Universityof Waikato,PsychologyDepartment,Trafficand Road Safety Research Group, Private Bag 3 105, Hamilton, New Zealand
Abstract-Eighty drivers in a sample of four groups of 20 (10 males, 10 females), aged under 30 years (‘young’), 40-59 (‘middle aged’), 60-69 (‘older’), and 70 years and over (‘oldest’), participated in tests of head rotation and of several visual functions relevant to safe driving. Head rotation data showed that the oldest drivers had lost about l/3 of movement and that the loss tended to be more evident in males. Second attempts almost always produced slightly better results. All participants had at least 20/40 binocular vision, however, tests of monocular visual acuity, stereovision, and horizontal peripheral vision revealed that the poorest performers were aged 60 years and over and that the degree of decrement increased with age. Many of the older and oldest drivers in the sample were both severely restricted in their ability to turn their head and evidenced a loss of visual functioning. Analysis of the effect of reduced head movement on the useful field of view indicated that, for the drivers aged 60 years and over, there was an evident restriction on the distances at which approaching traffic could be brought into the central, stationary field, so that even at maximum head rotation plus one saccade (lY), approaching vehicles would not be clearly perceived beyond a distance of 50 m. The findings are discussed in relation to older drivers’ involvement in intersection accidents. 0 1997 Elsevier Science Ltd. Keywords-Traffic
accident,
Older driver, Head movement,
INTRODUCTION
Useful field of view, Intersection
Driving safely through intersections requires gathering distal and proximal visual information from many different angles, normally received through eye and/or head movements and, for visual areas behind the drivers, through the use of a vehicle’s mirrors. Blind angles cover areas in the traffic scene from which the drivers cannot extract information. Such angles are determined to some extent by external physical factors such as the vehicle’s design (especially A, B, and C pillar location and thickness, and rear window area, shape, and slope) and the quality, size, and location of the mirrors or barriers such as buildings, fences, trees, or land contour features. However, to a much larger degree, blind angles vary according to the physical and perceptual capacities of the drivers. They are minimized if drivers have good head turning and eye movement capabilities and are readily able to move their head and eyes towards peripherally located road scenes. In order to perceive important information in detail, such as traffic signs and relevant traffic, it needs to be first brought into central vision, that small area of the visual field (less than three degrees) in which visual acuity is highest. Eye movement
The number of drivers on the road in developed countries who are aged 60 years and older is increasing, and there also is evidence of an increasing involvement in road crashes among such drivers (Barr, 1991; Drummond and Yeo, 1992; Waller, 1991). International traffic accident statistics indicate that a major contributing driving error is failure to yield right of way at intersections (Evans, 1991a), with older drivers over represented in such accidents. Similarly, Garber and Srinivasan ( 1991) found that the probability of older drivers committing violations at an intersection or during change of lane manoeuvres was much higher than for younger drivers. Intersections are particularly demanding traffic environments for drivers, requiring a series of complex visual perception and decision processes. When drivers approach intersections, visual scanning using eye and head movements occurs (Kito et al., 1989), providing data for directional and ‘Go/No Go’ decisions. The directional decisions follow detection and interpretation of road signs and lane markings while the ‘Go/No Go’ decisions are based on judgements of the speed and distance of approaching traffic. 793
194
R. B. ISLER et al.
studies in real traffic situations show that drivers continually scan the road scene for useful information (e.g. Mourant and Rockwell, 1972). When peripheral information enters the visual field, attentional processes determine which information needs to be further inspected in detail using central vision. For some roading information, its appearance in the visual field does not necessarily mean that it will be perceived. For example, in some ‘looked but failed to see’ incidents, the drivers cannot perceive information in their visual fields because of attentional deficits, or because the information was presented to a particular area of the retina which is functionally blind. A number of visual deficits occur naturally with aging, including, cataracts, macular degeneration, open-angle glaucoma, and diabetic retinopathy, and are likely to affect driving performance (Klein, 199 1; Wood and Troutbeck, 1992). Visual and behavioural deficits (e.g. hemiagnosias) that affect driving performance can result from strokes. Drivers in the older age groups are the most likely candidates for these types of impairment and it seems that they often are not aware of such visual loss because of gradual onset or may not reveal it for fear of loss of their licence. Keeping the high accident rate of older drivers at intersections and during change of lane manoeuvres in mind, along with the changes in visual acuity and head movement ability, it seems reasonable to assume that the range, in degrees, of blind angles increases with age. Many intersection and change of lane accidents involving older drivers may, therefore, be caused by their not perceiving some of the important information essential for making safe decisions in those traffic environments. Recent research by Owsley et al. (1991) found that, in a sample of 53 older drivers (57-83 years), the size of useful field of view (UFOV) correlated well with accident frequency. The size of the full UFOV is smaller than the visual field as determined by clinical perimetry and does not cover the whole area of visual sensitivity (Ball et al., 1990). It consists of a small binocular stationary field and a larger eye field (binocular in more central areas) and refers to the total visual field area within which particular information can be acquired without eye and head movements, i.e. within one eye fixation (Sanders, 1993). According to Sanders, two stimuli within a small angular separation can be processed as a single one in the stationary field, while in the eye field subjects can obtain only an hypothesis from a peripheral stimulus before centrally fixating. Clearly defined eye movements can then be initiated in order to inspect peripheral stimuli in detail. Such eye movements are found to be much faster than for a new percept without an hypothesis.
Important boundaries lie between the eye field, which is part of the UFOV, and the two laterally extended monocular eye fields and head fields, from which no information can be obtained while attending to a central task. Appropriate eye movements may give access to stimuli in the extended eye fields, while head movements and, possibly, more eye movements would be necessary in order to detect peripheral stimuli in the head fields (Sanders, 1993). The boundaries between the eye fields and the extended eye fields are not clearly defined because they vary substantially across individual drivers and situations. Research has shown that increase of central cognitive load, stress, noise, and aging are factors, which may result in a generalized shrinkage of the (peripheral) eye field (see, for example, Williams, 1982; Ikedo and Takeuchi, 1975; Leibowitz and Appelle, 1969; Mackworth, 1965; Scialfa et al., 1987). In such situations, effective eye and head movements are particularly important as a ‘narrower beam of vision’ needs to be moved around faster to get the same information from the driving environment. It is a common observation that many older drivers appear to find it increasingly difficult to turn their head beyond certain angles owing to physical neck conditions (e.g., arthritis, lower muscle tone, neurological disorders). In severe cases, where head rotation is strictly limited, very little information from the head field is accessible, as body movements, which normally would serve to compensate for impaired head movements, are likely to be restricted by the car seat and seat belt. Another option would be to compensate head movement restrictions with larger eye movements. However, as Bahill et al. (1975) report, most of the naturally occurring eye movements (saccades) have an amplitude of only 15” or less and larger eye movements would put considerable strain on the saccadic system. Additionally, Carter et al. (1983) found that older people have longer latencies of saccadic eye movements. Surprisingly, while much attention has been given to research on older drivers’ visual deficits, there is no similar body of research available which examines the effects of restricted head movements on older drivers’ driving performance. As head movement restrictions seem predominantly to be a problem of older drivers and as these drivers are over-represented in intersection and manoeuvring accidents, where the accessibility to the extended eye fields and head fields of the field of view is particularly important, research in this area seems pertinent. This study aimed to assess the degree of head movement restriction in drivers in four different age
Age related
effects of restricted
groups and to test several visual functions relevant to UFOV in order to examine the relation between age, head movement, and UFOV.
METHOD Participants
The sample consisted of 80 drivers, comprising 10 male and 10 female participants in each of four age groups, ‘young’ (under 30 years, mean age M= 23.4), ‘middle aged’ (40 to 59 years, M=49.9); ‘older’ (60 to 69 years, M=65.3), and ‘oldest’ (70 years and over, M=76.1). In addition to equal numbers of male and female participants, an effort was made to include participants in the broader community, from various ethnic backgrounds and occupations. Participants were recruited as volunteers (with travel costs reimbursed) through an advertisement in a local newspaper, through contacts with Age Concern (a self help organization of older people), and by word of mouth. Apparatus
A brief driver behaviour questionnaire sought information on participants’ self-reported number of accidents over the last five years. Additionally, there was a question regarding participants’ self-judged driving ability with three response categories; below average, average, and above average. A consent form was used which outlined what was required of the participants, assured confidentiality, and acknowledged other ethical obligations. Most vision testing (binocular and monocular, visual acuity and stereovision) was conducted with the aid of a vision screener (Keystone VS-II, model 1135A from Keystone View, Division of Mast/Keyestone, Inc., NV, USA). A standard perimeter (Lafayette Instrument Company, Inc., IN, USA) was used to measure horizontal peripheral vision. Different coloured target disks (red, yellow, green, blue and white, 5 mm diameter) served as targets, however, only the data obtained with the white disk were analysed for this study. In order to measure the degree of maximum head turning, a head turning measurement device was developed. It consisted of a cycle helmet with a stylus pointer protruding from the centre front, parallel with the centre of the wearer’s line of vision when the eyes were looking directly ahead. A plumb bob suspended on the cord from the stylus allowed a reading to be taken of head rotation (in degrees of arc). The 210” arc, marked in five-degree segments from zero at the front centre to 105” on either side, was made of transparent plastic and mounted on a shoulder harness.
795
head movements
Procedure
After the participants had signed the consent form and completed the questionnaire, they performed the vision tests, followed by the test which measured their maximal degree of head movement. Visual acuity. Far point (6 metres) monocular and binocular visual acuity testing was performed. Seven ratings, Snellen values from 20/20 (highest visual acuity) to 20/200, were possible. Stereovision (depth perception). Stereovisual performance defines the extent to which objects may be discriminated in three dimensions. In this test, the participants identified which of five different symbols appeared to stand out from the others. Five levels of stereovision were tested, with the first being the easiest (10% stereovision), followed by three intermediary levels (30, 60, and 75%) and the final being the most difficult (85% stereovision). Seventy-five percent stereovision (level 4) or greater, was considered acceptable according to criteria set by the Keystone VS-II Vision Screener standards guide. Horizontal peripheral vision. The degree of peripheral vision was determined with five different coloured target stimuli (red, yellow, green, blue and white). Starting always from one or other peripheral side (90° from centre), each of the three target stimuli were moved along the perimeter towards the centre until the participant was able to perceive and name the colour of the test stimulus. Only the values for the white target stimulus were considered for the analysis. Maximum head movements. The head turning measurement device was used to determine the range (in degrees) of head movements, separately to the left and the right side, for each participant. Once the participant had turned their head to the maximum of articulation on either side, a record of the degree of turn was made. Data were obtained from two trials on each side.
RESULTS The results from the questionnaire on accident frequency during the last five years and self-judged driving ability revealed that the eighty participants reported a total of 23 accidents during that period (Table 1). Eight (35% of the sample) involved the youngest group (~30 years), of these, five of the 10 young male drivers reported having at least one accident during the last five years. Eight drivers (40%) from the 40 to 59 year-old age group each reported a single accident over that time span. The older and oldest participants (60-69 years and 70 + years of age) reported 7 (33%) accidents, however 25 drivers
196
R. B.
Table 1. Results from the Driver Behaviour Vision Testing
Questionnaire
and the
Age groups Total N=80 N
<30 n=20
40-59 n=20
60-69 n=20
70+ n=20
40
8
40
3
15
4
20
45 55
7 13
35 65
10 14
50 50
5 5
25 25
0 55 45
0 9 11
0 45 55
0 10 10
0 50 50
0 12 8
0 60 40
0
1
5
4
20
7
35
10
1
5
3
15
10
50
15
5
25
9
45
13
65
n%n%n%n%
Number of accidents last five years 23 8 Distance travelled per year 41 9 < 10,000 km 11 > 10,000 km 39 Self reported driving ability below average 0 0 average 42 11 9 above average 38 Monocular visual acuity less than 20/40 12 0 Stereovision 2 less than 75% 16 Horizontal peripheral vision 3 less than 30” 29
(63%) from these groups estimated their travel distance to be less than 10,000 km per year. The question regarding participants’ self-judged driving ability revealed that not one person in the sample of 80 drivers considered him or herself as a ‘below average’ driver. Eight of the 10 young male drivers rated themselves as above average (not shown on Table l), while almost all young female drivers identified themselves as average drivers. About equal numbers in the remaining age groups rated themselves either as average or above average drivers. There was no clear gender effect on the two eldest age groups. Table 1 also shows the frequency distribution of visual acuities tested for both eyes (binocular) and separately for the left and right eye (monocular) by age. All drivers had binocular visual acuity above the minimal standard of 20/40 required for driving in most countries. However, 12 drivers had severely impaired visual acuity in one of their eyes. Four of these were older drivers and seven were in the oldest age group. Of the latter, four had such a severe visual acuity impairment in one eye (20/200 or less) that they could be considered as functionally one-eyed drivers. Table 1 further shows the frequency distribution for stereovision levels (expressed in percentages of full stereovision) by age. Sixteen participants (20% of the whole sample) had less than 75% stereovision (unacceptable according to criteria set for the Keystone VS-II Screener). Ten ( 13%) of these drivers were from the oldest age group. The results of the horizontal peripheral vision task also revealed an age-related loss of such vision. As Table 1 shows, the percentage of drivers with less than 30” of peripheral
ISLER
et
al.
vision increased for the older (45%) and oldest (65%) age groups. Figure 1 shows mean degrees of maximum head movement, for the first and second attempt, by age and gender. A preliminary analysis (t-test) did not reveal a difference between the degrees of maximum left and right head movements pooled for all age groups, so the data for these two conditions were combined. The graph shows a clear age effect, with a gradual decline in mean degrees of maximum head movement across the age groups. The youngest participants had a mean maximum head movement of 86”, which is 14” more than the corresponding mean value of the 40 to 59 year-old participants, 19” more than the 60 to 69 year-olds, and 27” more than the 70 year-old and above participants. Except for those in the 60 to 69 year-old group, the female participants in each group were normally able to turn their head further than their male counterparts. Second attempts typically produced higher degrees of head movement (mean value of improvement taken over all groups: M= lY), with the middle-aged female group (40-59 years) as the only exception. A 4 x 2 x 2 (age groups x gende x trial, first versus second attempt) analysis of variance on the degrees of maximum head movement with repeated measures on the factor Trial confirmed main effects for Age, F(3, 72) =24.3, ~~0.01, Trial F(1, 72)=10.3, ~~0.01, but not for Gender F( 1, 72) =2.6, p>O.O5. There were no significant interactions. Scheffe’s post hoc tests revealed that the youngest age group had significantly greater maximum head movement compared with the middle, older and oldest age groups (all ps ~0.05). There were no significant differences between the maximum head movement when comparing either the middle age group with the older age group or the older age group with the oldest age group. However, the middle aged group produced significantly greater head movements than the oldest age group (~~0.05). Figure 2 shows a scatterplot for the whole sample with their degrees of maximum head movement in the right plane on the x-axis and the individual size of their right horizontal peripheral visual field as measured with a standard perimeter, on the y-axis. The graph reveals that in most cases the older people with severely restricted head movements (e.g. 60” or less) also had small degrees of horizontal peripheral vision. A significant Pearson’s product-moment correlation between these two variables was obtained, r=0.34, p ~0.05. Most young drivers (< 30 years) had a maximum head movement of more than 70” while the majority of the older and oldest drivers had maximum head movement of less than 70”. Three in the sample of oldest drivers had less than 50” of head
Age related
m -30
f <30
effects of restricted
m40-59
Gender Fig. 1. Mean degrees
40
50
60
70
of maximum
80
90
f 40-59
(m=male;
head movement,
100
110
Maximum Head Movement (deg) Fig. 2. Scatterplot and regression line with 95% confidence interval of individual degrees of maximum head movement plotted against sizes of right horizontal peripheral visual field.
movement and two of these drivers had also less than 20” horizontal peripheral vision. A further analysis was attempted in order to evaluate the potential effects of restricted head movements on the visual field of the sample of drivers assuming that they were arriving at the decision point of a T-intersection with an assumed road width of 8 m. Cohen (1996) showed that such a road width would be most adequate for rural roads. At the decision point drivers would need to judge the distance and speed of the oncoming traffic from the right in order to find a safe gap to cross the intersection and merge with the traffic on the left lane (drivers in New Zealand keep to the left side of the road).
797
head movements
m60-69
f=female)
f 60-69
m70+
f 70+
J
and Age (years)
for the first and second attempt,
by age and gender.
First, when looking straight ahead, for all drivers the oncoming traffic would appear almost perpendicular to their line of sight, in their extended lateral monocular eye or head field and no distance and speed judgements would be possible. Normally, much head movement, possibly supported by some eye movements, would then be initiated in order to bring the traffic into central vision to allow speed and distance judgements. In cases of restricted head movement, however, larger performed horizontal eye movement angles would be needed in order to compensate for the head movement loss. Figure 3 shows for each participant in each age group their maximum achieved head movement angle and their calculated critical detection distance, Dcrit, by which oncoming traffic would be brought into central vision assuming that no additional eye movements were performed (round symbols). Dcrit was calculated as
Dcrit = Ld/tan ~1 where Ld=lateral distance (about 4 m) from participant to oncoming car if road width was 8, c1=maximum achieved head movement angle. Figure 3 further shows Dcrit for each driver assuming that they would support their maximum head movement with a 15” horizontal eye movement (squared symbols). Such a movement corresponds to one natural saccade (Bahill et al., 1975). Finally, Fig. 3 also shows Dcrit values (triangular symbols) assuming that drivers’ maximum head movement was supported by 30” of horizontal eye movement (two natural saccades). The data in Fig. 3 show that the maximum achieved head movement angles of the older drivers would not be sufficient to bring oncoming traffic in central vision at distances exceeding 20 m without additional eye movements to the right. In contrast,
798
R. B. ISLER et al.
Drivers
50
<30 Years
60 Maximum
Fig. 3. Calculated movement (HM), 30” eye movement those on the right
and 40-59
70 Head
Movement
Years
60
Dnvers
90
100
110
(deg)
60-69
50
Years
60 Maximum
70 Head
critical distance by which oncoming traffic would be brought into central vision by maximum head movement and additional 15” eye movement (HM+ 1.5), and maximum (HM + 30). The data shown on the left are for the ‘young’ and ‘middle aged’ drivers (i are for the ‘older’ drivers and ‘oldest’ drivers (60-69 and 70+ years). Closed symbols restricted right horizontal peripheral vision of less than 30”.
many of the young drivers were able to turn their head for more than 90” and would have been able to focus on the traffic without additional eye movements at the largest distance on the scale of the figure (> 100 m). With an additional 15” of eye movement to the right (one natural saccade), many older drivers (N=20 or 50%) still would not be able to focus centrally on approaching traffic which is further away than 40 m, while only six of the younger drivers would not be able to do so. Even after large eye movements of 30” (two natural saccades), eight of the older drivers would not be able to fixate on such oncoming traffic (40 m away), while only three drivers (all three were in the 40-59 years group) of the younger group would have been unable to do so. Very often, in addition to their restricted head movements, the oldest drivers also had restricted right horizontal peripheral vision of less than 30” (see closed symbols in Fig. 3). Four of the oldest drivers who had such restricted head movements that they would not have been able to judge the speed and distance of oncoming car at a distance of 50 m, even with more than two natural eye movements, also had severely restricted peripheral vision of less than 30”.
DISCUSSION The findings of the present study indicate that there is a significant age-related decrement in maximum achieved head movement and in several visual functions relating to UFOV in drivers aged 60 years or more. Analysing the effects of head movement loss on the UFOV of these participants as if they were arriving at the decision point of a T-intersection
and 70+ Years
Movement
60
90
100
110
(deg)
means of only maximum head head movement and additional 30 years and 40-59 years) and represent participants who had
revealed that many older drivers would have difficulty bringing vehicles beyond 50 m into central focus. As part of the assessment process, participants also had responded to a brief questionnaire focussing on their self-reported driving history, number of accidents during the last five years, and self-judged driving ability. The questionnaire response pattern that stood out was that of the young male drivers (< 30 years). This particular group reported the highest number of accidents and, more than for drivers in any of the other groups, considered themselves to be above average drivers. None of the drivers in any age group rated themselves as below average drivers. Such response patterns match well with earlier research findings and recent accident statistics (e.g. Guerin, 1994; Page et al., 1992). The response patterns of the oldest age group did not differ substantially from the middle-aged (40-59 years) and older (60-69 years) groups. However, caution may be needed when interpreting the self-reported data of older drivers as they may be reluctant to reveal any information which could question their driving competence. For example, Sloane et al. (1990), cited in Owsley et al. ( 1991) found that most of the older drivers with the highest number of recorded accidents were men who tended to under-report accident involvement in response to driving history questionnaires. This may be a matter of pride or of fear that disclosure will lead to licence revocation. The results of the vision tests were instructive. While all of the drivers in the sample had acceptable binocular visual acuity (better than 20/40), the monocular vision tests revealed that 12 had severely impaired visual acuity in one of their eyes. Four of
Age related effects of restricted head movements these were in the older (60-69 years) group and seven were in the oldest (70 or more years) group, showing increasing deficit with increasing age. Four of the latter group had such significant levels of impairment (20/200 or worse) that they could have been classified as one-eyed. The stereovision data also revealed an age-related pattern of increasing deficit. Of the 16 (20%) whose scores were less than 75%, 10 came from the oldest group. A statistically significant age effect was obtained for horizontal peripheral vision, with the 22 of the 29 participants having less than 30” falling in the older (9) and oldest (13) age groups. Lack of stereovision seemed to be a predominant feature of half of the drivers aged over 70 years, and loss of visual acuity in one eye could be a primary factor. Judgement of depth, in the absence of stereovision, requires use of monocular depth cues, such as texture, gradient, and relative movement. Gradual loss of vision in one eye may not be noticed, and the driver may fail consciously to use monocular cues to judge depth and distance as an adaptive response to the loss. Loss of horizontal peripheral vision, also a significant factor in our oldest drivers, Further narrows the range of visual perception in functioning eyes. In all, these results reveal considerable deterioration in monocular visual acuity, stereovision, and horizontal peripheral vision in drivers aged 60 years and over, with the number of those showing deficits increasing as age increases. The seriousness of the deficit also increases with age. The present findings support those of Klein ( 1991) and Shinar and Schieber (1991), who reported in some detail on the causes of specific vision deficits in older drivers and called for better assessment of older drivers’ vision and more and better research into their effects on driving. The various elements of reduced visual compentence (e.g. loss of monocular, stereo-, and horizontal peripheral vision) may compound to create conditions which make effective and safe judgements at intersections and in manouvering difficult for many of the oldest drivers. Kosnik et al. (1990) suggest that older drivers recognise loss of vision and decide to give up driving, but there are issues of gradual, undetected onset and decline and of disincentives in the form of loss of mobility and independence which may delay self-revocation of a driver’s licence to a time beyond the point at which the deficit first became problematic or unsafe for the driver and other road users. This lack of awareness or use of denial seems evident in the present study’s 60-years-and-over participants when comparing their measured deficits against their ratings of their own driving ability as average or above average. This tendency to either be unaware
799
of, or to deny, declining ability is supported by the findings of Barr (1991), Garber and Srinivasan ( 199 1), and Owsley et al. ( 199 1) in respect of accident rates of older drivers, especially at intersections. Johnson and Keltner (1983), cited in Shinar and Schieber ( 199 1) examined 10,000 volunteer California licence applicants and found significant deterioration among drivers aged over 60 years and that drivers with visual field loss in both eyes had accident and conviction rates more than twice as high as age- and sex-matched drivers with normal fields. Use of more extensive and thorough vision and driving skill testing procedures may more accurately reveal both the extent of visual dysfunction among older drivers and the point at which such vision defects begin to contribute to high accident risk. The results of the head movement tests revealed significant age and gender related differences between drivers in their ability to rotate their heads, with those aged 60 years or more having the greater restriction and with males usually performing less well than their female counterparts. The average decrement in the oldest participants was a loss ofabout one-third of head movement in the right lateral plane compared with the group aged under 30 years. Arthritic changes in cervical vertebrae, changes in muscle tone, and neurological conditions affecting head turning may well account for the reduction. Decrements in head rotation of drivers have not previously been studied, but it is a common observation that older drivers have difficulty in turning their head at intersections, often relying on body movement to compensate. The importance of head movements to safe driving were studied by Allan et al. (1974), who found that, in comparing performance of head-movement-restricted drivers with head-unrestricted drivers in a driving simulator, restricted drivers made more driving errors, fewer driving adjustment responses to critical events in the left field (under US driving conditions), and shifted their fixations less often. While average values are useful for detecting overall age related trends in visual and physical impairments, it also is important to determine the extent to which individuals have combinations of deficits. The nature of the vision testing equipment available for the present study meant that it was only possible to produce a scattergram comparing degrees of head rotation and degrees of horizontal peripheral vision. Those data (Fig. 2) clearly show age-related impairment of neck articulation and peripheral vision, such that, in general, the youngest participants had both more head movement and a wider peripheral visual field, and that as age increased, the range of both declined. The data also show great variability between participants, suggesting that broad
800
R. B. ISLER et al.
generalizations cannot be made and that specific testing for deficits is essential. Ball and colleagues (Ball et al., 1988, 1990; Ball and Owsley, 1991) noted that standard perimetry may significantly underestimate the restriction in vision of drivers aged over 60 years, indicating that assessments based on tests of functional vision, such as UFOV, reveal a three-fold reduction compared with young (18-25 years) adults. As a result, it is important to consider how the deficits detected in the present sample may have affected UFOV. If, as part of an overall aging process, a number of the sensory and physical systems contributing to effective search for, identification and perception of, and response to driving hazards become increasingly inoperative, the driver’s UFOV is likely to be considerably reduced. Driving safely through an intersection involves attending, perceiving, and responding to crucial road and traffic information which usually is distributed on a much larger area than drivers’ useful field of view covers. For instance, using the concept of the functional visual field (Sanders, 1993), oncoming traffic appears first in the far peripheral head field so that head movements are necessary in order to bring the oncoming vehicle first into peripheral vision (eye-field) so that it can be detected. Eye movements and, possibly, more head movements are then necessary in order to shift the detected vehicle from peripheral vision (eye field) to central binocular vision (stationary field), where visual acuity and depth perception is best. Restricted neck articulation and vision deficits are likely to interfere with this process and might result, first, in delaying the detection of oncoming vehicles and, second, in delaying the fixation of the vehicle with central vision until it is possibly too close for an accurate judgement of its distance and speed to be made to permit safe clearance of the intersection. In a study of driver head and eye movement in the intersection environment, Kito et al. (1989) found that a driver of a small vehicle preparing to make a right turn (in Europe and the US the equivalent manoeuvre would be a left turn) increased both head and eye movements and gazes (head-eye movement of greater than 20”), especially toward the right, during the approach to, and initial turning at, the intersection. Presumably, these behaviours are intended to enhance detection and response to approaching vehicles. Although saccades provide quicker shifts than head movements (Sanders, 1993) and frequent, rapid eye movements are considered to be more effective in scanning intersections than head movements (Rahini et al., 1990), one saccade has a natural magnitude of only 15” or less (Bahill et al., 1975). Consequently, in the typical intersection
environment in which roads intersect at 90”, saccadic movements alone are not likely to provide a sufficient scan for safe detection of approaching vehicles and gazes (head and eye movement) are likely to be required. The data from the present analysis, set out in Fig. 3, indicate that when the head is rotated to the maximum degree possible, without additional eye movements, many of the drivers aged less than 60 years would be able to bring an approaching vehicle into central vision when it was at a distance greater than 20 m, but that only one of the drivers aged 60 years and over could do so, and then only just. Adding the equivalent of one saccade would have allowed almost all of the under 60-year-olds to bring a vehicle at a distance of over 50 m into central vision, while only a few of the 60-years-and-over drivers would have been able to have done so. Moving the eye the equivalent of two saccades (30”) would still suggest that 25% of the older drivers would be unable to bring a target vehicle approaching at a distance beyond 50 m into central vision. This means that attention to, and perception of, approaching vehicles would be delayed, with the delay becoming more problematic and dangerous in a rural environment, where vehicle approach speeds are higher. It is important for drivers at an intersection to be able to bring oncoming traffic into their (stationary field) central vision in order to perform the speed and distance estimations necessary for them to decide whether a gap is large enough to safely enter the traffic flow. A driver whose visual field and head movements are restricted would, in the intersection setting, have to try to detect and bring approaching traffic into central vision by turning the upper torso and by turning the eyes as far as possible. However, the car seat and seat belt may constrain the degree of torso turning and, in drivers requiring corrective lenses, the eye may turn beyond the limit of the lens and/or have vision obstructed by the spectacle frame. As a consequence, these alternatives may not effectively compensate for loss of head movement and visual functions. Another possible compensatory response, making more rapid ocular movements, may not be available to some older drivers. For example, Carter et al. (1983) found an age-related increase in the latency of ocular movements, which would place older persons at a disadvantage in everyday situations involving visual search. Age-related restriction of head movement must compound visual deficits in stereovision and peripheral vision, increasing the difficulty, for older drivers, of bringing the image of an approaching vehicle into central (foveal) vision, where visual acuity is highest. This combination of factors may,
801
Age related effects of restricted head movements
in part, explain why older drivers have higher reported rates of intersection accidents that resulted in injury or death in a number of highly motorized countries (Bernhoft, 1990; Evans, 199 1b; Faulkner, 1975 cited in Hills, 1980; Frith, 1991; Mackie, 1972 cited in Hills, 1980; Garber and Srinivasan, 1991). It may also be a factor in the older drivers’ higher rates of lane-change (Garber and Srinivasan, 1991) and U-turn (Frith, 1991) accidents, since head movements and good vision are essential for checking mirrors and minimizing blind-angles. Given that there are increasing numbers of older drivers on the roads in many developed countries (Barr, 1991; Waller, 1991), and that the probability of accidents increases with age, as do defects in vision and restrictions in head movement, it is important that better provisions be made for the assessment of these various functions as part of the regular driver’s licence examinations of older drivers. Acknowledgements~The authors wish to thank Rob Bakker for technical assistance. This research was commissioned by Transit New Zealand.
REFERENCES Allan, J. A., Schroeder, R. S. and Ball, P. G. (1974) Effects of head restriction on drivers eye movements and errors in simulated dangerous situations. Journal of Applied Psychology
59, 5, 643-648.
Ball, K., Beard, B., Roenker, D., Miller, R. and Griggs, D. (1988) Age and visual search: expanding the useful field of view. Journal of the Optical Society of America 5, 2210-2219.
Ball, K., Owsley, C. and Beard, B. (1990) Clinical visual perimetry underestimates peripheral field problems in older drivers. Clinical Vision Science 5, 2, 113-125. Ball, K. and Owsley, C. (1991) Identifying correlates of accident involvement for the older driver. Human Factors 33, 5, 5833595.
Bahill, A. T., Adler, D. and Stark, L. (1975) Most naturally occurring human saccades have magnitudes of 15 degrees or less. Investigative Ophthalmology 14, 6, 468-469.
Barr, R. A. (1991) Recent changes in driving among older adults. Human Factors 33 and 533 (5), 597-600. Bernhoft, I. M. (1990) Elderly drivers: results from a Nordic in-depth study on elderly car drivers. In Third European Workshop on Recent Developments in Road Safety Research. Linkoping: VTI Report 3664 Road Safety Research, pp. 11 l-130. Linkoping, VT1 Report 366A.
Carter, J. E., Obler, L., Woodward, S. and Albert, M. (1983) Effect of increasing age on the latency of saccadic eye movements. Journal of Gerontology 38, 318-320.
Cohen, A. S. (1996) Psychisch bedingte Strassenbreite. Tuef Rheinland Publishers, Koeln. Drummond, A. E. and Yeo, E. (1992) The risk of driver crash involvement as a function of driver age. Monash University Accident Research Centre Research Report, Vol. 49, Melbourne, Australia. Evans, L. (1991a) Older-driver risks to themselves and to
other road users. Transportation Research Record 1325 34-41, TRB, National Research Council, Washington, D.C. Evans, L. ( 1991b) TrafJic Safety and the Driver. Van Nostrand Reinhold Publishers, New York. Frith, W. J. (1991) The safety of the elderly road user in New Zealand relative to road users in other age groups. In Proceedings of the International Conference on Road Safety, pp. 548-551. New Delhi, India. Garber, N. J. and Srinivasan, R. (1991) Characteristics of accidents involving elderly drivers at intersections. Transportation Research Record, 1325, 8-16, TRB, National Research Council, Washington, D.C. Guerin, B. (1994) What do people think about the risks of driving? Implications for road safety interventions. Journal of Applied Social Psychology 24, 99441021. Hills, B. L. ( 1980) Vision, visibility, and perception in driving. Perception 9, 1833216. Ikedo, M. and Takeuchi, T. (1975) Influence of fovea1 load on the functional visual field. Perception and Psychophysics 18, 2555260.
Kito, T., Haraguchi, M., Funatsu, T., Sato, M. and Kondo, M. (1989) Measurements of gaze movements while driving. Perceptual and Motor Skills 68, 19-25. Klein, R. ( 199 1) Age related eye disease, visual impairment, and driving in the elderly. Human Factors 33, 5, 521-525.
Kosnik, W. D., Sekular, R. and Kline, D. W. (1990) Selfreported visual problems of older drivers. Human Factors 32, 5, 597-608.
Leibowitz, H. and Appelle, S. (1969) The effect of a central task on the luminance thresholds of peripherally presented stimuli. Human Factors 11, 387-392. Mackworth, N. H. (1965) Stimulus density limits the useful field of vision. Psychonomic Sciences 3, 67-68. Mourant, R. R. and Rockwell, T. H. (1972) Strategies of visual search by novice and experienced drivers. Human Factors 14, 325-335.
Owsley, C., Ball, K., Sloane, M. E., Roenker, D. L. and Bruni, J. R. (1991) Visual perceptual/cognitive correlates of vehicle accidents in older drivers. Psychology and Aging 6, 4033415.
Page, D. G., McDonald, C. W. and Ryba, K. A. (1992) Road safety and young drivers. A report to the Task group on road safety for young drivers. Creative Challenge Services Ltd, Palmerston North, New Zealand. Rahini, M., Briggs, R. P. and Thorn, D. R. (1990) A field evaluation of driver eye and head movement strategies toward environment targets and distracters. Applied Ergonomics
21, 4, 2677274.
Sanders, A. F. (1993) Processing information in the functional visual field. In Perception and Cognition, eds G. d’ydewalle and J. Van Rensbergen. Elsevier, Amsterdam. Scialfa, C. T., Kline, D. W. and Lyman, B. (1987) Age differences in target identification as a function of retinal location and noise level: an examination of the useful field of view. Psychology and Aging 2, 14-19. Shinar, D. and Schieber, F. (1991) Visual requirements for safety and mobility of older drivers. Human Factors 33, 5, 5077519.
Waller, P. (1991) The older driver. Human Factors 33, 5, 499-505.
Wood, J. M. and Troutbeck, R. (1992) Effect of restriction of the binocular visual field on driving performance. Ophthalamic
and Physiological
Optics 12, 291-298.
Williams, L. J. (1982) Cognitive load and the functional field of view. Human Factors 24, 6, 6833692.