This article was downloaded by: [Moskow State Univ Bibliote] On: 12 February 2014, At: 00:05 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Ergonomics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/terg20
Differential effects of head-mounted displays on visual performance a
a
a
b
Lutz Schega , Daniel Hamacher , Sandra Erfuth , Wolfgang Behrens-Baumann , Juliane b
bc
Reupsch & Michael B. Hoffmann a
Department of Training and Health, Institute of Sport Science, Otto-von-GuerickeUniversity Magdeburg, Brandenburger Str. 9, 39016Magdeburg, Germany b
Universitäts-Augenklinik, Leipziger Str. 44, 39120Magdeburg, Germany
c
Centre for Behavioral Brain Sciences, Magdeburg, Germany Published online: 12 Nov 2013.
To cite this article: Lutz Schega, Daniel Hamacher, Sandra Erfuth, Wolfgang Behrens-Baumann, Juliane Reupsch & Michael B. Hoffmann (2014) Differential effects of head-mounted displays on visual performance, Ergonomics, 57:1, 1-11, DOI: 10.1080/00140139.2013.853103 To link to this article: http://dx.doi.org/10.1080/00140139.2013.853103
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Ergonomics, 2014 Vol. 57, No. 1, 1–11, http://dx.doi.org/10.1080/00140139.2013.853103
Differential effects of head-mounted displays on visual performance Lutz Schegaa*, Daniel Hamachera, Sandra Erfutha, Wolfgang Behrens-Baumannb, Juliane Reupschb and Michael B. Hoffmannb,c a
Department of Training and Health, Institute of Sport Science, Otto-von-Guericke-University Magdeburg, Brandenburger Str. 9, 39016 Magdeburg, Germany; bUniversita¨ts-Augenklinik, Leipziger Str. 44, 39120 Magdeburg, Germany; c Centre for Behavioral Brain Sciences, Magdeburg, Germany
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
(Received 17 January 2013; accepted 2 October 2013) Head-mounted displays (HMDs) virtually augment the visual world to aid visual task completion. Three types of HMDs were compared [look around (LA); optical see-through with organic light emitting diodes and virtual retinal display] to determine whether LA, leaving the observer functionally monocular, is inferior. Response times and error rates were determined for a combined visual search and Go-NoGo task. The costs of switching between displays were assessed separately. Finally, HMD effects on basic visual functions were quantified. Effects of HMDs on visual search and Go-NoGo task were small, but for LA display-switching costs for the Go-NoGo-task the effects were pronounced. Basic visual functions were most affected for LA (reduced visual acuity and visual field sensitivity, inaccurate vergence movements and absent stereo-vision). LA involved comparatively high switching costs for the Go-NoGo task, which might indicate reduced processing of external control cues. Reduced basic visual functions are a likely cause of this effect. Practitioner Summary: We assessed how basic visual functions are affected in different head-mounted displays (HMDs) and how this relates to their efficiency in reality augmentation. In conclusion, look-around HMDs are an economical alternative to optical-see-through HMDs, if unpredictable external events are kept to a minimum. Keywords: consumer ergonomics; user needs analysis; human factors integration; risk assessment and management; vision and lighting
1.
Introduction
Augmented realities aim to optimise task completion in many aspects of both every day life and specialist applications. This is accomplished by combining the perceived real world with additional information yielding a fused augmented percept. Such augmented realities can be achieved using a head-mounted display (HMD). An HMD encloses a screen-like display component that can be integrated into different systems, e.g. glasses or helmet. It thus allows the user to embed contextual information that is superimposed onto their visual field (Azuma 1997). The areas of applications of augmented reality and HMDs are multifaceted (Azuma et al. 2001). They can be used for intraoperative monitoring by anesthesiologists (Liu et al. 2009), for endoscopy in microsurgery (Hiramatsu et al. 2005), for rehabilitation treatment in patients with Parkinson’s disease (Espay et al. 2010), for sensorimotor training in neurorehabilitation (Adamovich et al. 2009), gait retraining of patients with hip replacement (Schega, Hamacher, and Wagenaar 2011) and as mobile augmented reality in industrial applications (Tumler et al. 2008). There are different ways in which the reality can be augmented by the additional information provided on the HMD and in which reality augmentation can interfere with visual perception. First, simply wearing an HMD without any presentation of augmenting information can interfere directly with the perception of the environment. Second, the augmenting information can be provided without the need of a three-dimensional (3D) alignment of virtual and real objects, since the environmental target is not primarily relevant [e.g. situation dependent or independent textual or symbolic information (operating instructions, picking and packing lists, biofeedback in rehabilitation and sports) and, since the information is not available to others, protected information presentation (privacy data)]. Third, the augmenting information can be 3D aligned with real world objects (e.g. navigation in picking and packing, operating instruction aligned to the real target and navigation in medical applications). Sutherland’s initial pioneering work in the 1960s led to ‘ceiling mounted’ displays, i.e. ceiling mounted miniature cathode ray tubes (Sutherland 1968), which was followed by a continuous progress in the development of HMDs and manifold variants are at present available (Zhou, Dun, and Billinghurst 2008). In an ‘optical-see-through-HMD’ (OSTHMD), real and augmenting information are presented to the same eye. In a ‘look-around-HMD’ (LA-HMD), real and
*Corresponding author. Email: [email protected] q 2013 Taylor & Francis
2
L. Schega et al.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
augmenting information are presented to separate eyes. While this leaves the user effectively monocular for either input, LA-HMDs are a more economical variant and might therefore support the establishment of HMD-mediated augmented realities for a wider range of applications. This prompts the question whether performance in demanding visual tasks is reduced when LA-HMDs are used. Since the picture on the LA-HMD is presented monocularly, vergence becomes open loop while the accommodation stimulus is defined by the HMD scene. This may cause eye strain, discomfort, blurred image, altered perception of distance, size, depth and speed (Patterson, Winterbottom, and Pierce 2006). Another challenge with HMDs is that a mismatch of accommodative demands to vergence demands increases stereoscopic fusion time (So et al. 2011). This study aimed to compare the effects of LA-HMDs and OST-HMDs on visual performance. Thus it can be derived for which applications LA-HMDs might be preferable and for which the use of OST-HMDs might be critical. We focused on the relevance for visual perception of the environment and for tasks which did not require 3D alignment of real and virtual displays. Particular attention was paid to the performance during a demanding visual search task and to the assessment of basic aspects of visual function.
2. Methods 2.1. Subjects A total of 15 subjects without ophthalmological or neurological history (median age: 26 years; age range: 21– 30 years) were investigated. The subjects were emmetropic and had normal vision [decimal visual acuity $ 1.0, normal colour vision as tested with Velhagen plates, and normal stereo vision as tested with the TNO-test (Toegepast Natuurwetenschappelijk Onderzoek, NL)]. All subjects reported to be right-handed to exclude effects of handedness on the results (Peters and Ivanoff 1999). All subjects gave their informed written consent prior to the study. The procedures followed the tenets of the declaration of Helsinki (World Medical Association 2000) and were approved by the ethics committee of the University of Magdeburg, Germany. Since the experiments were carried out on different days, only 11 out of the 15 subjects included participated in all tests of E1, E2 and E4, and only 10 participated in E3.
2.2.
HMDs and experiments conducted
The influence of different HMDs on visual perception and function was tested: two different OST-HMDs, i.e. LitEye with organic light emitting diodes (LE750, termed here ‘oledOST’) and Nomad with a virtual retinal display (ND 2100, Microvision, termed here ‘vrdOST’), and one LA-HMD, Nikon Media Port (UP300x, termed here ‘LA’). Testing was performed with binocular viewing and HMDs were placed in front of the right eye, except for the reference condition in Experiment 1 with stimulus presentation on the monitor. The left eye always viewed the monitor. To minimise the disturbance of the visual perception (particularly of the left eye), the experimental setups were placed in front of an empty white wall in a room with moderate lighting. An overview about all test conditions is given in Table 1. Two sets of experiments were conducted: (i) assessment of the influence of HMDs (a) on the performance during a visual search task combined with a Go-NoGo task (Experiments 1 and 2) and (b) on vergence movements (Experiment 3) and (ii) quantification of the influence of above three HMDs on basic aspects of vision (stereo vision, visual acuity and visual field sensitivity; Experiment 4).
Table 1.
Viewing conditions in Experiments 1 and 2.
Experimental condition
Viewing condition
Left-eye view
Right-eye view
Experiment 1 – visual search & Go-NoGo (without switching)
Monitor HMD: OST
Search & Go-NoGo Black monitor
HMD: LA Monitor (task) HMD: OST (task)
Black monitor Search & Go-NoGo Black monitor
HMD: LA (task)
Black monitor
Search & Go-NoGo Search & Go-NoGo (background: black monitor) Search & Go-NoGo Search & Go-NoGo Search & Go-NoGo (background: black monitor) Search & Go-NoGo
Experiment 2 – visual search & Go-NoGo (with switching)
Ergonomics
3
(i) Influence of HMDs on the performance during visual search combined with different tasks (Experiments 1 and 2) and on vergence (Experiment 3).
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Supported by a chin rest, the volunteers were required to perform a visual search task which was combined with a Go-NoGo task. The search array, a 6 £ 6 matrix of red circles, was displayed either on a computer monitor at 1 m viewing distance or on the different HMDs in a randomised order. The matrix size was equal for the HMD and monitor, i.e. 9.88 £ 6.68 of visual angle (Figure 1(A)). Three experiments were conducted. 2.2.1. Experiment 1: influence of HMDs on performance in visual search and Go-NoGo task This experiment aimed to determine the influence of HMDs on the performance in a visual search and Go-NoGo task. The subjects were instructed to report via mouse click the presence of a target stimulus, a ‘0’, in the 6 £ 6 matrix with red ‘O’s (see Figure 1(A)), which was present in 50% of the trials. The target positions were randomly chosen on the inner 4 £ 4 matrix. Upon detection of the target ‘0’, the subjects had to respond by button press of the left mouse button with the left thumb. If no such target was detected, the right mouse button should be pressed with the right thumb. If the response time (RT) exceeded 1800 ms, the trial was aborted and auditory feedback given. In this case, the participants did not have to pass a make-up trial. The time delay of two successive matrices was 1000 ms. In addition to the visual search task, the subjects had to perform a Go-NoGo task, which had highest priority. In the centre of the matrix, a target was presented (see Figure 1(A)), either a target that should be responded to, ‘go-target’ (letter ‘P’) or a target that should be ignored, ‘nogotarget’ (letter ‘R’). As for the matrix task, the presentation probablilty of ‘P’ and ‘R’ was equal. Upon detection of the gotarget, the spacebar had to be pressed with the index finger, and no response should be given upon detection of the nogotarget. It should be noted that the Go-NoGo task required high visual acuity due to the target size (see Figure 1). As in Huckauf et al. (2010), error rates (ERs) and RT were analysed to assess processing of visual information presented either on a monitor, a LA or an oledOST. For each subject, each of these three conditions was tested in a single separate block of 10 trials. The four blocks were tested in randomised order to reduce sequential effects. At the beginning of each block, the subject passed two training trails (128 stimuli), as training can have a pronounced effect on visual search (Schuster et al. 2013). Thereafter, no significant RT changes were observed. The experiment lasted around 1.5 h for each subject. For each condition, there were at least 587 successful trials. In the analysis, four possible response types were differentiated for the search task: hits – correct response to present target; correct rejection – correct response absence to absent target; missed (‘m’) – incorrect response absence to present target; false alarm (‘fa’) – incorrect response to absent target. RTs were calculated for hits and correct rejections only. To assess the influence of the Go-NoGo task on the visual search task, this analysis was performed separately for the three different Go-NoGo task conditions (‘P’: P shown; ‘R’: R shown; ‘noPR’: neither ‘P’ nor ‘R’ shown). Similarly, the RTs were assessed. Finally, the ERs and RTs for the Go-NoGo task itself were assessed.
2.2.2.
Experiment 2: costs for randomly switching the display medium
This experiment aimed to determine the temporal costs for randomly switching the display medium, i.e. monitor or HMD. The 6 £ 6 matrix was presented on the monitor (Mon) or on the HMDs (Mon&LA vs Mon&oledOST and vs Mon&vrdOST) in random alternation. In each subject, each of these three conditions was tested in a single separate block of 10 trials. The four blocks were tested in randomised order to reduce sequential effects. The visual search/Go-NoGo task was the same as for Experiment 1. Care was taken to adjust the HMD such that the field of view was congruent with the monitor
Figure 1. Stimulus display and setup for eye-movement measurements. (A) Screenshot of 6 £ 6 matrix. Errors indicate target for ‘gonogo-task’ and for the search task in the matrix. (B) Subject with a HMD and Eyetracker. (C) Detection of pupil on the control monitor.
4
L. Schega et al.
in the background to prevent an inaccurate overlap. For this purpose, three crosses were simultaneously presented on the HMD and the monitor. Slight changes in head position were made until the subject reported a concurrent position of the crosses. As for Experiment 1, a RT of 1800 ms for each trial was allowed. The experiment lasted around 1.5 h for each participant. In order to avoid direct sequential effects, Experiments 1 and 2 were conducted in separate sessions on different days. The parameters were calculated as for Experiment 1.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
2.2.3.
Experiment 3: vergence movements
As vergence eye movements can be affected by HMDs (Kawara, Ohmi, and Yoshizawa 1996), this experiment aimed to determine vergence accuracy. For this purpose, both the eye tracker and the HMD were attached to the head (see Figure 1(B); EyeLink II, SR Research (Kanata, ON, Canada); sampling frequency 250 Hz). The subjects were instructed to fixate a cross for 1 min prior to the actual measurements, either on the monitor or on one of the three HMDs. Vergence movements were measured for 10 alternations between fixations of 1-s duration on the monitor and the HMD for each condition. The crosses did not overlap. Vergence was determined 750 ms after alternation. In each subject, each of these three conditions was tested in a single separate block of 10 trials. The vergence for fixation on the HMD was determined relative to that for fixation on the monitor. The blocks were tested in randomised order to reduce sequential effects. The visual axes may intersect at a vergence distance that deviates from the monitor plane. The deviation was calculated using the measured pupillary distance, the distance of the point of interception of the visual axis of each eye with the monitor (provided by the software), the distance of the subjects to monitor and the theorem of intersecting lines. The experiment lasted around 30– 45 min for each participant. 2.2.4. Experiment 4: influence of HMDs on stereo vision, visual acuity and visual field sensitivity Stereo vision was quantified by determining the minimal disparity necessary for a stereoscopic percept using the TNO-test, a stereo test based on random-dot stereograms viewed through red-green goggles. This test is based on red-green randomdot anaglyph images. The measurements were conducted for each of the conditions and then repeated with the inverted TNO-test cards to minimise learning effects for the second test. For the calculation of the descriptive statistics and the determination of the significance levels, disparity values were logarithmised and subsequently delogarithmised for the illustration of the descriptive statistics. Decimal visual acuities were determined binocularly at a 5-m viewing distance using an adaptive procedure (Freiburg Visual Acuity and Contrast Test, Bach 1996). Visual acuities were determined with and without HMD for each HMD-type in a separate session, following a counter-balanced design (following an A – B –B – A scheme for one half and an B – A –A – B for the other half of subjects) and yielding a total of four acuity measurements per HMD-type. These precautions were taken to exclude potential learning effects known for visual acuity measurements (Heinrich, Kruger, and Bach 2011). For the calculation of the descriptive statistics and the determination of the significance levels, decimal visual acuity values were logarithmised and subsequently delogarithmised for the depiction of the descriptive statistics. To assess visual field sensitivities, light spot detection thresholds were determined binocularly with automated static white on white perimetry (Octopus 101 Perimeter; Haag-Streit, Ko¨niz, Switzerland) using a standard paradigm, which was customised to include the position of the blind spot [background luminance: 4 asb; target size: Goldmann size III (0.48 diameter); target exposition: 100 ms]. At 97 visual field positions, evenly distributed in the visual field (^ 22.58) with 58 spacing, we determined the detection thresholds with the dynamic strategy (Weber and Klimaschka 1995) incorporated in Octopus 101. Threshold sensitivity is given as S[dB] ¼ 10 £ log(Lmax/Lstimulus), where Lmax equals 1000 asb. Subjects looked straight ahead during testing. Sensitivities were determined with and without HMD, for each HMD-type in a separate session, following a counter-balanced design across subjects (A– B sequence for one half and B – A sequence for the other half of the subjects) and yielding a total of four acuity measurements. These precautions were taken to exclude potential learning effects. The reliability factor of the perimetry measurements, i.e. the percentage of false positives and false negatives in the catch trials in the reference condition, without HMD, never exceeded 8.3%, and the median across subjects was 0%. 2.3.
Statistics
One-way repeated measures ANOVAs were conducted separately in Experiment 1 (factor: display) and in Experiment 2 for the condition switch to monitor and switch to HMD (factor display). Post hoc analyses were done with Scheffe-tests sequential Bonferroni corrected for multiple testing. To examine the ERs, Friedman tests [post hoc: Wilcoxon-test; sequentially Bonferroni corrected for multiple testing; (Holm 1979)] were applied. For Experiments 3 and 4, significance levels were determined with sequentially Bonferroni-corrected paired t-tests (Holm 1979).
Ergonomics
5
3. Results 3.1. Experiment 1: influence of HMDs on performance in visual search and Go-NoGo task
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Out of the variety of comparisons, only two significant effects of HMDs on RTs were evident (Figure 2), i.e. RT increases for oledOST (for response to matrix RT oledOST and LA: 1147 and 1004 ms, respectively; p ¼ 0.048) and LA (for response to Go-NoGo RT LA and monitor: 1073 and 950 ms, respectively; p ¼ 0.007). Furthermore, only two significant effects on ERs were evident, i.e. increased ER for oldedOST and more misses of the go-target for oledOST (Table 2). In summary, only few differences between HMDs were evident for visual search and Go-NoGo tasks, mainly slight deteriorations of performance for oledOST and one RT increase for LA. Presenting tasks monocularly in HMDs does generally not lead to a performance decrement compared to a conventional binocular monitor viewing condition, when no distracting stimuli are present in the environment. It should be noted that vrdOST was not tested in this experiment.
3.2. Experiment 2: costs for randomly switching the display medium The RTs and ERs were assessed for ‘response to matrix’ and ‘response to Go-NoGo’ while the mediums monitor and HMDs alternated in randomised order.
3.2.1. Response to matrix There were no RT differences between the HMDs. Effects on ER, i.e. false alarms, were observed for tasks both on HMDs and on the monitor, but only when the no-go-target (‘R’) was present. On the HMDs, ER were significantly greater for ‘Mon&vrdOST’ than for ‘Mon&LA’ ( p ¼ 0.006) and ‘Mon&oledOST’ ( p ¼ 0.006). On the monitor, ER were significantly greater for ‘Mon&LA’ than for ‘Mon&oledOST’ ( p ¼ 0.012).
3.2.2.
Response to Go-NoGo
Effects on both RT and ER, for both misses and false alarams, were observed (see Figures 3 and 4, respectively). RT were greater for LA, i.e. for tasks on the monitor RT were greater for switches from LA than from vrdOST ( p ¼ 0.016) and for tasks on the HMDs RT were greater for switches to LA than to oledOST ( p ¼ 0.009). ER, i.e. misses of the go-target, were higher for ‘Mon&LA’ than for ‘Mon&oledOST’ for both switches to the monitor (8.5 vs 2.0; p ¼ 0.005) and to the HMD (7.5 vs 1.5; p ¼ 0.028). False alarms for the no-go-target on monitor and HMDs were more frequent for ‘Mon&LA’ than for both ‘Mon&vrdOST’ (15.5 vs 2.7, p ¼ 0.024 and 16.00 vs 2.7, p ¼ 0.004, respectively) and ‘Mon&oledOST’ (15.5 vs 1.5; p ¼ 0.041 and 16.0 vs 1.5; p ¼ 0.008, respectively). In summary, only few and unsystematic dependences of media switching on HMD type were evident for response to matrix. In contrast, for the Go-NoGo task effects were more frequent and concentrated on LA. For LA, ER increased when switching to both monitor and HMD, and RT increased when switching to the monitor.
Figure 2. Experiment 1 – RT for monitor vs oledOST vs LA (n ¼ 11; mean ^ SEM). The RT to the matrix was significantly higher for oledOST than for LA for go-target (P) present ( p ¼ 0.048). Futhermore, the performance of the Go-NoGo task was faster for monitor than for LA ( p ¼ 0.007).
4 (9) 0 (2)
2.9 (11.9) 13.1 (22.8) 4.7 (21.5) 19.4 (19.7) 2.7 (7.4) 19.3 (21.4)
Monitor Median (interq. range) (%)
6 (15.5) 0 (4)
7.3 (35.5) 24.9 (11.9) 7.7 (34.2) 24.7 (22.3) 5.7 (13.8) 33.1 (21.4)
oledOST Median (interq. range) (%)
Experiment 1 – ERs for monitor vs oledOST vs LA.
Matrix fa NoPR m NoPR fa P m P fa R m R Go-NoGo P R
Task
Table 2.
6 (13) 1 (2)
3.6 (21.6) 9.0 (15.2) 1.6 (20.1) 12.0 (8.0) 3.5 (17.9) 11.4 (15.6)
LA Median (interq. range) (%)
7.52 (2) 0.08 (2)
3.250 (2) 10.750 (2) 5.871 (2) 4.750 (2) 4.710 (2) 5.250 (2)
X 2 (df)
0.023 0.961
0.197 0.005 0.053 0.093 0.095 0.072
p
0.150 0.072 0.126 0.484 0.108 0.100 0.036 1.00
2 2.524 2 0.962
P
2 1.960 2 2.100 2 1.859 2 .700 2 2.100 2 1.960
Z
Monitor vs oledOST
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
21.973 20.322
21.400 21.540 20.420 21.540 20.840 20.840
Z
0.098 0.748
0.322 0.123 0.674 0.246 0.802 0.401
P
Monitor vs LA
2 1.614 2 0.597
2 1.120 2 2.521 2 2.100 2 2.100 2 0.338 2 2.380
Z
0.106 1.00
0.263 0.036 0.108 0.108 0.735 0.051
P
oledOST vs LA
6 L. Schega et al.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Ergonomics
7
Figure 3. Experiment 2 – RT to the task ‘response to Go-NoGo’ (n ¼ 11; mean ^ SEM). The Go-NoGo task was presented on HMD (vrdOST or LA or oledOST) or monitor (Mon.). Presenting the go-target (‘P’) on the monitor yielded faster responses for vrdOST and monitor than for LA and monitor ( p ¼ 0.016). If the go-target was presented on the HMD, the RT was faster for vrdOST and monitor than for oledOST and monitor ( p ¼ 0.009).
Figure 4. Experiment 2 – ERs to the ‘response to Go-NoGo’ signal (n ¼ 11; mean ^ SEM). Misses of the go-target on the monitor were more frequent for ‘Mon&LA’ than for ‘Mon&oledOST’ (8.5 vs 2.0; p ¼ 0.005). On the HMD, they were more frequent for ‘Mon&LA’ vs ‘Mon&oledOST’ (7.5 vs 1.5; p ¼ 0.028). False alarms for the no-go-target on monitor and HMDs were more frequent for ‘Mon&LA’ than for ‘Mon&vrdOST’ (15.5 vs 2.7, p ¼ 0.024 and 16.00 vs 2.7, p ¼ 0.004, respectively) and ‘Mon&oledOST’ (15.5 vs 1.5; p ¼ 0.041 and 16.0 vs 1.5; p ¼ 0.008, respectively).
3.3. Experiment 3: vergence movements Although the vergence system is open loop for LA-HMDs and therefore instable, we observed effects that exceeded the increased scatter in the vergence data and consequently reached statistical significance. Only for LA, the eyes converged in front of the monitor plane ( p ¼ 0.038). For LA, the distance of the convergence point from the convergence on the monitor was significantly greater than for oledOST and vrdOST, p ¼ 0.038 and p ¼ 0.049, respectively (Figure 5).
3.4. 3.4.1.
Experiment 4: influence of HMDs on stereo vision, visual acuity and visual field sensitivity Visual acuity
The mean decimal visual acuity exceeded 1.6 in the reference measurements, i.e. without HMD. A significant effect of the HMD compared to reference was evident for LA, namely a reduction to 1.3 ( p , 0.001), while no significant effects were evident for the other two HMDs tested (see Figure 6(A)).
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
8
L. Schega et al.
Figure 5. Experiment 3 – vergence with HMDs (n ¼ 11; mean ^ SEM). The visual axes intersected significantly in front of the monitor only for LA ( p ¼ 0.038). Furthermore, for LA the distance of the convergence point to the monitor was significantly greater than for oledOST and vrdOST, p ¼ 0.038 and p ¼ 0.049, respectively (Figure 6).
3.4.2. Stereo vision The LA leaves the subject effectively monocular by completely covering one eye. As a matter of course none of the stereocharts was passed for LA. In contrast, stereo vision was retained for the other two HMDs. However, compared to unrestrained binocular viewing, significantly greater disparities were required for the recognition of the stereo targets ( p , 0.005; see Figure 6(B)). 3.4.3. Visual field sensitivity In analogy to Hoffmann, Seufert, and Schmidtborn (2007), visual field sensitivities were assessed at 11 visual field positions along the horizontal meridian (average of þ 2.58 and 2 2.58 elevation; see Figure 7(C)). Significant sensitivity reductions, compared to a reference condition without HMD, were evident at six positions for LA and at two positions for oledOST, but none were evident for vrdOST. It should be noted that non-significant sensitivity reductions were also induced in the left visual hemifield at around 158 of eccentricity and were particularly pronounced for LA and oledOST. These were presumably due to the blind spot in the left hemifield corresponding to the papilla of the left eye, as the right eye was covered by the HMD. The effect did not reach significance, probably due to the inter-individual variability of the size and location of the blind spot. To exclude this region from a more formal evaluation, the effects of the HMDs were assessed quantitatively for the average of the sensitivities within ^ 7.58 along the horizontal meridian. There was a significant sensitivity reduction for LA by 2.1 ^ 0.3 dB ( p , 0.001) and for oledOST by 0.8 ^ 0.2 dB ( p , 0.01). For vrdOST, there was a trend of a reduction by 0.6 ^ 0.4 dB, which failed to reach significance ( p ¼ 0.142). 4.
Discussion
A strong consistent effect of HMDs on visual search was absent and only minor effects were observed. In contrast, the integration of sporadically available additional visual information, i.e. the Go-NoGo task, was particularly affected by media switching to and from LA. This might be associated with the finding that basic aspects of visual function were most severely affected for LA. 4.1.
HMD effects on basic aspects of visual function
Basic aspects of visual function were affected for all HMDs, albeit to a variable degree. The effects were most pronounced for LA and least for vrdOST. For LA, visual function was reduced in all tests conducted, i.e. visual acuity, stereo vision (absence) and visual field sensitivities. This is due to the functional monocularity of observers with LA, which reduces visual acuity (Wood, Collins, and Carkeet 1992; Cagenello, Arditi, and Halpern 1993) and visual field sensitivities (Wood, Collins, and Carkeet 1992; Cagenello, Arditi, and Halpern 1993) and blocks binocular vision. Visual acuity and visual field sensitivities normally benefit from binocular summation, i.e. an increase of acuity in the order of 1.3 (Wood, Collins, and
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Ergonomics
9
Figure 6. Experiment 4 – influence of HMDs on stereo vision, visual acuity and visual field sensitivity. (A) Differential effect of wearing HMDs on visual acuity (n ¼ 11; mean ^ SEM). Only for LA the visual acuity was significantly reduced ( p , 0.001). (B) Differential effect of wearing HMDs on disparity thresholds for stereo vision (n ¼ 10; mean ^ SEM). While stereo targets were not detected for LA, they were detected both oledOST and vrdOST. However, in comparison to the condition ‘no HMD’ the disparity thresholds were significantly elevated ( p , 0.005). (C) Differential effect of wearing HMDs on visual field sensitivities (n ¼ 10; mean ^ SEM) along the horizontal meridian. Significant reductions were evident for LA and oledOST, but not for vrdOST. Significance levels: *p , 0.05; **p , 0.01; ***p , 0.001.
Carkeet 1992; Cagenello, Arditi, and Halpern 1993; Pardhan 1993; Vedamurthy et al. 2007) and visual field sensitivity of up to 1.4 (Wood, Collins, and Carkeet 1992). Consequently, LA visual acuities and visual field sensitivites are reduced. Further, blocking binocular vision results in an absence of stereo vision, which is also a plausible cause of the inaccurate vergence movements observed for LA. For oledOST, both stereo vision and visual field sensitivities were reduced. The latter is expected as the neutral density filter properties of oledOST reduce stimulus luminance by 30% (Azuma 1997). In contrast, the reduction of stimulus luminance for vrdOST is negligible (Patterson, Winterbottom, and Pierce 2006) and no reduction in visual field sensitivities was observed. For vrdOST, however, reductions of stereo sensitivity were evident. Consequently, stereovision was reduced for all HMDs, which indicates that monocular HMDs bear upon the integrity of high level binocular visual function that is essential for stereovision. 4.2.
Effects of HMD on demanding visual tasks
During the Go-NoGo task, the performance for the HMD LA was particularly inferior to that for the other HMDs. This effect was most pronounced during media switching. As basic aspects of visual function are most severely reduced for LA, these reductions might be a cause of the impaired processing of sporadically available information with highest priority and with high visual acuity demand. In particular, the reduced sensitivity for brief luminance changes, as assessed with static visual field perimetry, combined with inferior, since monocular, visual acuity for LA is a likely cause of the reduced
10
L. Schega et al.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
performance during the Go-NoGo task. An inaccurate vergence indicates a mismatch of the perceived spatial locations of the two media, i.e. monitor and HMD, which is likely associated with the absence of stereo vision. Furthermore, inaccurate vergence might, due to the coupling of vergence and accommodation, cause inaccurate accommodation (Drascic and Milgram 1996), further reducing visual acuity and hence target identification. This disadvantage of LA for the Go-NoGo task appears to be particularly relevant when switching between media. Consequently, LA has higher switching costs than the other HMDs tested. In conclusion, the economic HMD variant, i.e. LA, is an alternative to OST HMDs, if moderate reductions in visual function can be tolerated and if sporadic external events on the unattended medium are kept to a minimum. In the absence of distracting stimuli, monocularly presented tasks in LA-HMDs do not lead to a performance decrement. Consequently, ideal applications of LA HMDs are those that require the augmentation of visual input with embedded contextual information at predictable intervals in harzard-free situations without environmental stimuli. The use of LA-HMDs in hazard-free conditions should be particularly acceptable in applications without the need to fuse virtual objects and the real scene. However, if a 3D alignment of real and virtual objects is required, OST-HMDs would be applicable. Therefore, typical examples for LA HMDs applications would be the presentation of operating instruction (symbolic or textual) without 3D alignment to real world objects, biofeedback in rehabilitation and sports-training, and presentation of private data.
Acknowledgements The support by the Federal Ministry of Education and Research (01 IM 08 001 Z) is gratefully acknowledged. Furthermore, the authors thank Dr Huckauf and Mr Urbina from the Bauhaus-University of Weimar for their support.
References Adamovich, S. V., G. G. Fluet, E. Tunik, and A. S. Merians. 2009. “Sensorimotor Training in Virtual Reality: A Review.” NeuroRehabilitation 25 (1): 29 – 44. Azuma, R., Y. Baillot, R. Behringer, S. Feiner, S. Julier, and B. Macintyre. 2001. “Recent Advances in Augmented Reality.” IEEE Computer Graphics and Applications 21 (6): 34 – 47. Azuma, R. T. 1997. “A Survey of Augmented Reality.” Presence: Teleoperators and Virtual Environments 6 (4): 355– 385. Bach, M. 1996. “The Freiburg Visual Acuity Test – Automatic Measurement of Visual Acuity.” Optometry & Vision Science 73: 49– 53. Cagenello, R., A. Arditi, and D. L. Halpern. 1993. “Binocular Enhancement of Visual Acuity.” Journal of the Optical Society of America A: Optics, Image Science, and Vision 10 (8): 1841– 1848. Drascic, D., and P. Milgram. 1996. Perceptual Issues in Augmented Reality, 123–134. San Jose, CA: The International Society for Optical Engineering. Espay, A. J., Y. Baram, A. K. Dwivedi, R. Shukla, M. Gartner, L. Gaines, A. P. Duker, and F. J. Revilla. 2010. “At-Home Training with Closed-Loop Augmented-Reality Cueing Device for Improving Gait in Patients with Parkinson Disease.” Journal of Rehabilitation Research and Development 47 (6): 573– 581. Heinrich, S. P., K. Kruger, and M. Bach. 2011. “The Dynamics of Practice Effects in an Optotype Acuity Task.” Graefes Archive for Clinical and Experimental Ophthalmology 249 (9): 1319– 1326. Hiramatsu, K., T. Inui, M. Okada, T. Takeshima, H. Mishima, T. Sakaki, and S. Shiono. 2005. “New Device for Endoscopic Image Display During Microsurgical Clipping of Cerebral Aneurysms - Technical Note.” Neurologia Medico-Chirurgica (Tokyo) 45 (9): 487– 490. Hoffmann, M. B., P. S. Seufert, and L. C. Schmidtborn. 2007. “Perceptual Relevance of Abnormal Visual Field Representations – Static Visual Field Perimetry in Human Albinism.” The British Journal of Ophthalmology 91: 509– 513. Holm, S. 1979. “A Simple Sequentially Rejective Multiple Test Procedure.” Scandinavian Journal of Statistics 6: 65 – 70. Huckauf, A., M. Urbina, I. Bo¨ckelmann, L. Schega, R. Mecke, J. Grubert, F. Doil, and J. Tumler. 2010. “Perceptual Issues in Optical-SeeThrough Displays.” In Proceedings – APGV 2010: Symposium on Applied Perception in Graphics and Visualization, 41 – 48. New York: Association for Computing Machinery. Kawara, T., M. Ohmi, and T. Yoshizawa. 1996. “Effects on Visual Functions During Tasks of Object Handling in Virtual Environment with a Head Mounted Display.” Ergonomics 39 (11): 1370 –1380. Liu, D., S. A. Jenkins, P. M. Sanderson, M. O. Watson, T. Leane, A. Kruys, and W. J. Russell. 2009. “Monitoring with Head-Mounted Displays: Performance and Safety in a Full-Scale Simulator and Part-Task Trainer.” Anesthesia and Analgesia 109 (4): 1135– 1146. Pardhan, S. 1993. “Binocular Performance in Patients with Unilateral Cataract Using the Regan Test: Binocular Summation and Inhibition with Low-Contrast Charts.” Eye (London) 7 (Pt 1): 59 – 62. Patterson, R., M. D. Winterbottom, and B. J. Pierce. 2006. “Perceptual Issues in the Use of Head-Mounted Visual Displays.” Human Factors 48 (3): 555– 573. Peters, M., and J. Ivanoff. 1999. “Performance Asymmetries in Computer Mouse Control of Right-Handers, and Left-Handers with Leftand Right-Handed Mouse Experience.” Journal of Motor Behavior 31 (1): 86 – 94. Schega, L., D. Hamacher, and R. C. Wagenaar. 2011. “A Comparison of Effects of Augmented Reality and Verbal Information Based Interventions in Elderly Women After Hip Replacement.” Archives of Physical Medicine and Rehabilitation 92 (10): 1734– 1735. Schuster, D., J. Rivera, B. C. Sellers, S. M. Fiore, and F. Jentsch. 2013. “Perceptual Training for Visual Search.” Ergonomics. http://www. ncbi.nlm.nih.gov/entrez/query.fcgi?cmd¼Retrieve&db¼PubMed&dopt¼Citation&list_uids¼23650877
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Ergonomics
11
So, R. H. Y., W. S. Wong, R. Yip, A. K. C. Lam, and P. Ting. 2011. “Benefits of Matching Accommodative Demands to Vergence Demands in a Binocular Head-Mounted Display: A Study on Stereo Fusion Times.” Presence-Teleoperators and Virtual Environments 20 (6): 545– 558. Sutherland, I. E. 1968. “A Head-Mounted Three Dimensional Displayed.” In American Federation of Information Processing Societies Fall Joint Computer Conference, 757– 764, Part I. Washington, DC: Thompson Books. Tumler, J., F. Doil, R. Mecke, G. Paul, M. Schenk, E. A. Pfister, A. Huckauf, I. Bo¨ckelmann, and A. Roggentin. 2008. “Mobile Augmented Reality in Industrial Applications: Approaches for Solution of User-Related Issues.” In 7th IEEE International Symposium on Mixed and Augmented Reality Cambridge, 87 – 90. Washington, DC: IEEE Computer Society. Vedamurthy, I., C. M. Suttle, J. Alexander, and L. J. Asper. 2007. “Interocular Interactions During Acuity Measurement in Children and Adults, and in Adults with Amblyopia.” Vision Research 47 (2): 179– 188. Weber, J., and T. Klimaschka. 1995. “Test Time and Efficiency of the Dynamic Strategy in Glaucoma Perimetry.” German Journal of Ophthalmology 4 (1): 25 –31. Wood, J. M., M. J. Collins, and A. Carkeet. 1992. “Regional Variations in Binocular Summation Across the Visual Field.” Ophthalmic and Physiological Optics 12 (1): 46 – 51. World Medical Association. 2000. “Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects.” The Journal of the American Medical Association 284 (23): 3043–3045. Zhou, F., H. B.-L. Dun, and M. Billinghurst. 2008. “Trends in Augmented Reality Tracking, Interaction and Display: A Review of Ten Years of ISMAR.” In 7th IEEE International Symposium on Mixed and Augmented Reality, (ISMAR 2008), 193–202. Washington, DC: IEEE Computer Society.
Ergonomics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/terg20
Differential effects of head-mounted displays on visual performance a
a
a
b
Lutz Schega , Daniel Hamacher , Sandra Erfuth , Wolfgang Behrens-Baumann , Juliane b
bc
Reupsch & Michael B. Hoffmann a
Department of Training and Health, Institute of Sport Science, Otto-von-GuerickeUniversity Magdeburg, Brandenburger Str. 9, 39016Magdeburg, Germany b
Universitäts-Augenklinik, Leipziger Str. 44, 39120Magdeburg, Germany
c
Centre for Behavioral Brain Sciences, Magdeburg, Germany Published online: 12 Nov 2013.
To cite this article: Lutz Schega, Daniel Hamacher, Sandra Erfuth, Wolfgang Behrens-Baumann, Juliane Reupsch & Michael B. Hoffmann (2014) Differential effects of head-mounted displays on visual performance, Ergonomics, 57:1, 1-11, DOI: 10.1080/00140139.2013.853103 To link to this article: http://dx.doi.org/10.1080/00140139.2013.853103
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Ergonomics, 2014 Vol. 57, No. 1, 1–11, http://dx.doi.org/10.1080/00140139.2013.853103
Differential effects of head-mounted displays on visual performance Lutz Schegaa*, Daniel Hamachera, Sandra Erfutha, Wolfgang Behrens-Baumannb, Juliane Reupschb and Michael B. Hoffmannb,c a
Department of Training and Health, Institute of Sport Science, Otto-von-Guericke-University Magdeburg, Brandenburger Str. 9, 39016 Magdeburg, Germany; bUniversita¨ts-Augenklinik, Leipziger Str. 44, 39120 Magdeburg, Germany; c Centre for Behavioral Brain Sciences, Magdeburg, Germany
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
(Received 17 January 2013; accepted 2 October 2013) Head-mounted displays (HMDs) virtually augment the visual world to aid visual task completion. Three types of HMDs were compared [look around (LA); optical see-through with organic light emitting diodes and virtual retinal display] to determine whether LA, leaving the observer functionally monocular, is inferior. Response times and error rates were determined for a combined visual search and Go-NoGo task. The costs of switching between displays were assessed separately. Finally, HMD effects on basic visual functions were quantified. Effects of HMDs on visual search and Go-NoGo task were small, but for LA display-switching costs for the Go-NoGo-task the effects were pronounced. Basic visual functions were most affected for LA (reduced visual acuity and visual field sensitivity, inaccurate vergence movements and absent stereo-vision). LA involved comparatively high switching costs for the Go-NoGo task, which might indicate reduced processing of external control cues. Reduced basic visual functions are a likely cause of this effect. Practitioner Summary: We assessed how basic visual functions are affected in different head-mounted displays (HMDs) and how this relates to their efficiency in reality augmentation. In conclusion, look-around HMDs are an economical alternative to optical-see-through HMDs, if unpredictable external events are kept to a minimum. Keywords: consumer ergonomics; user needs analysis; human factors integration; risk assessment and management; vision and lighting
1.
Introduction
Augmented realities aim to optimise task completion in many aspects of both every day life and specialist applications. This is accomplished by combining the perceived real world with additional information yielding a fused augmented percept. Such augmented realities can be achieved using a head-mounted display (HMD). An HMD encloses a screen-like display component that can be integrated into different systems, e.g. glasses or helmet. It thus allows the user to embed contextual information that is superimposed onto their visual field (Azuma 1997). The areas of applications of augmented reality and HMDs are multifaceted (Azuma et al. 2001). They can be used for intraoperative monitoring by anesthesiologists (Liu et al. 2009), for endoscopy in microsurgery (Hiramatsu et al. 2005), for rehabilitation treatment in patients with Parkinson’s disease (Espay et al. 2010), for sensorimotor training in neurorehabilitation (Adamovich et al. 2009), gait retraining of patients with hip replacement (Schega, Hamacher, and Wagenaar 2011) and as mobile augmented reality in industrial applications (Tumler et al. 2008). There are different ways in which the reality can be augmented by the additional information provided on the HMD and in which reality augmentation can interfere with visual perception. First, simply wearing an HMD without any presentation of augmenting information can interfere directly with the perception of the environment. Second, the augmenting information can be provided without the need of a three-dimensional (3D) alignment of virtual and real objects, since the environmental target is not primarily relevant [e.g. situation dependent or independent textual or symbolic information (operating instructions, picking and packing lists, biofeedback in rehabilitation and sports) and, since the information is not available to others, protected information presentation (privacy data)]. Third, the augmenting information can be 3D aligned with real world objects (e.g. navigation in picking and packing, operating instruction aligned to the real target and navigation in medical applications). Sutherland’s initial pioneering work in the 1960s led to ‘ceiling mounted’ displays, i.e. ceiling mounted miniature cathode ray tubes (Sutherland 1968), which was followed by a continuous progress in the development of HMDs and manifold variants are at present available (Zhou, Dun, and Billinghurst 2008). In an ‘optical-see-through-HMD’ (OSTHMD), real and augmenting information are presented to the same eye. In a ‘look-around-HMD’ (LA-HMD), real and
*Corresponding author. Email: [email protected] q 2013 Taylor & Francis
2
L. Schega et al.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
augmenting information are presented to separate eyes. While this leaves the user effectively monocular for either input, LA-HMDs are a more economical variant and might therefore support the establishment of HMD-mediated augmented realities for a wider range of applications. This prompts the question whether performance in demanding visual tasks is reduced when LA-HMDs are used. Since the picture on the LA-HMD is presented monocularly, vergence becomes open loop while the accommodation stimulus is defined by the HMD scene. This may cause eye strain, discomfort, blurred image, altered perception of distance, size, depth and speed (Patterson, Winterbottom, and Pierce 2006). Another challenge with HMDs is that a mismatch of accommodative demands to vergence demands increases stereoscopic fusion time (So et al. 2011). This study aimed to compare the effects of LA-HMDs and OST-HMDs on visual performance. Thus it can be derived for which applications LA-HMDs might be preferable and for which the use of OST-HMDs might be critical. We focused on the relevance for visual perception of the environment and for tasks which did not require 3D alignment of real and virtual displays. Particular attention was paid to the performance during a demanding visual search task and to the assessment of basic aspects of visual function.
2. Methods 2.1. Subjects A total of 15 subjects without ophthalmological or neurological history (median age: 26 years; age range: 21– 30 years) were investigated. The subjects were emmetropic and had normal vision [decimal visual acuity $ 1.0, normal colour vision as tested with Velhagen plates, and normal stereo vision as tested with the TNO-test (Toegepast Natuurwetenschappelijk Onderzoek, NL)]. All subjects reported to be right-handed to exclude effects of handedness on the results (Peters and Ivanoff 1999). All subjects gave their informed written consent prior to the study. The procedures followed the tenets of the declaration of Helsinki (World Medical Association 2000) and were approved by the ethics committee of the University of Magdeburg, Germany. Since the experiments were carried out on different days, only 11 out of the 15 subjects included participated in all tests of E1, E2 and E4, and only 10 participated in E3.
2.2.
HMDs and experiments conducted
The influence of different HMDs on visual perception and function was tested: two different OST-HMDs, i.e. LitEye with organic light emitting diodes (LE750, termed here ‘oledOST’) and Nomad with a virtual retinal display (ND 2100, Microvision, termed here ‘vrdOST’), and one LA-HMD, Nikon Media Port (UP300x, termed here ‘LA’). Testing was performed with binocular viewing and HMDs were placed in front of the right eye, except for the reference condition in Experiment 1 with stimulus presentation on the monitor. The left eye always viewed the monitor. To minimise the disturbance of the visual perception (particularly of the left eye), the experimental setups were placed in front of an empty white wall in a room with moderate lighting. An overview about all test conditions is given in Table 1. Two sets of experiments were conducted: (i) assessment of the influence of HMDs (a) on the performance during a visual search task combined with a Go-NoGo task (Experiments 1 and 2) and (b) on vergence movements (Experiment 3) and (ii) quantification of the influence of above three HMDs on basic aspects of vision (stereo vision, visual acuity and visual field sensitivity; Experiment 4).
Table 1.
Viewing conditions in Experiments 1 and 2.
Experimental condition
Viewing condition
Left-eye view
Right-eye view
Experiment 1 – visual search & Go-NoGo (without switching)
Monitor HMD: OST
Search & Go-NoGo Black monitor
HMD: LA Monitor (task) HMD: OST (task)
Black monitor Search & Go-NoGo Black monitor
HMD: LA (task)
Black monitor
Search & Go-NoGo Search & Go-NoGo (background: black monitor) Search & Go-NoGo Search & Go-NoGo Search & Go-NoGo (background: black monitor) Search & Go-NoGo
Experiment 2 – visual search & Go-NoGo (with switching)
Ergonomics
3
(i) Influence of HMDs on the performance during visual search combined with different tasks (Experiments 1 and 2) and on vergence (Experiment 3).
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Supported by a chin rest, the volunteers were required to perform a visual search task which was combined with a Go-NoGo task. The search array, a 6 £ 6 matrix of red circles, was displayed either on a computer monitor at 1 m viewing distance or on the different HMDs in a randomised order. The matrix size was equal for the HMD and monitor, i.e. 9.88 £ 6.68 of visual angle (Figure 1(A)). Three experiments were conducted. 2.2.1. Experiment 1: influence of HMDs on performance in visual search and Go-NoGo task This experiment aimed to determine the influence of HMDs on the performance in a visual search and Go-NoGo task. The subjects were instructed to report via mouse click the presence of a target stimulus, a ‘0’, in the 6 £ 6 matrix with red ‘O’s (see Figure 1(A)), which was present in 50% of the trials. The target positions were randomly chosen on the inner 4 £ 4 matrix. Upon detection of the target ‘0’, the subjects had to respond by button press of the left mouse button with the left thumb. If no such target was detected, the right mouse button should be pressed with the right thumb. If the response time (RT) exceeded 1800 ms, the trial was aborted and auditory feedback given. In this case, the participants did not have to pass a make-up trial. The time delay of two successive matrices was 1000 ms. In addition to the visual search task, the subjects had to perform a Go-NoGo task, which had highest priority. In the centre of the matrix, a target was presented (see Figure 1(A)), either a target that should be responded to, ‘go-target’ (letter ‘P’) or a target that should be ignored, ‘nogotarget’ (letter ‘R’). As for the matrix task, the presentation probablilty of ‘P’ and ‘R’ was equal. Upon detection of the gotarget, the spacebar had to be pressed with the index finger, and no response should be given upon detection of the nogotarget. It should be noted that the Go-NoGo task required high visual acuity due to the target size (see Figure 1). As in Huckauf et al. (2010), error rates (ERs) and RT were analysed to assess processing of visual information presented either on a monitor, a LA or an oledOST. For each subject, each of these three conditions was tested in a single separate block of 10 trials. The four blocks were tested in randomised order to reduce sequential effects. At the beginning of each block, the subject passed two training trails (128 stimuli), as training can have a pronounced effect on visual search (Schuster et al. 2013). Thereafter, no significant RT changes were observed. The experiment lasted around 1.5 h for each subject. For each condition, there were at least 587 successful trials. In the analysis, four possible response types were differentiated for the search task: hits – correct response to present target; correct rejection – correct response absence to absent target; missed (‘m’) – incorrect response absence to present target; false alarm (‘fa’) – incorrect response to absent target. RTs were calculated for hits and correct rejections only. To assess the influence of the Go-NoGo task on the visual search task, this analysis was performed separately for the three different Go-NoGo task conditions (‘P’: P shown; ‘R’: R shown; ‘noPR’: neither ‘P’ nor ‘R’ shown). Similarly, the RTs were assessed. Finally, the ERs and RTs for the Go-NoGo task itself were assessed.
2.2.2.
Experiment 2: costs for randomly switching the display medium
This experiment aimed to determine the temporal costs for randomly switching the display medium, i.e. monitor or HMD. The 6 £ 6 matrix was presented on the monitor (Mon) or on the HMDs (Mon&LA vs Mon&oledOST and vs Mon&vrdOST) in random alternation. In each subject, each of these three conditions was tested in a single separate block of 10 trials. The four blocks were tested in randomised order to reduce sequential effects. The visual search/Go-NoGo task was the same as for Experiment 1. Care was taken to adjust the HMD such that the field of view was congruent with the monitor
Figure 1. Stimulus display and setup for eye-movement measurements. (A) Screenshot of 6 £ 6 matrix. Errors indicate target for ‘gonogo-task’ and for the search task in the matrix. (B) Subject with a HMD and Eyetracker. (C) Detection of pupil on the control monitor.
4
L. Schega et al.
in the background to prevent an inaccurate overlap. For this purpose, three crosses were simultaneously presented on the HMD and the monitor. Slight changes in head position were made until the subject reported a concurrent position of the crosses. As for Experiment 1, a RT of 1800 ms for each trial was allowed. The experiment lasted around 1.5 h for each participant. In order to avoid direct sequential effects, Experiments 1 and 2 were conducted in separate sessions on different days. The parameters were calculated as for Experiment 1.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
2.2.3.
Experiment 3: vergence movements
As vergence eye movements can be affected by HMDs (Kawara, Ohmi, and Yoshizawa 1996), this experiment aimed to determine vergence accuracy. For this purpose, both the eye tracker and the HMD were attached to the head (see Figure 1(B); EyeLink II, SR Research (Kanata, ON, Canada); sampling frequency 250 Hz). The subjects were instructed to fixate a cross for 1 min prior to the actual measurements, either on the monitor or on one of the three HMDs. Vergence movements were measured for 10 alternations between fixations of 1-s duration on the monitor and the HMD for each condition. The crosses did not overlap. Vergence was determined 750 ms after alternation. In each subject, each of these three conditions was tested in a single separate block of 10 trials. The vergence for fixation on the HMD was determined relative to that for fixation on the monitor. The blocks were tested in randomised order to reduce sequential effects. The visual axes may intersect at a vergence distance that deviates from the monitor plane. The deviation was calculated using the measured pupillary distance, the distance of the point of interception of the visual axis of each eye with the monitor (provided by the software), the distance of the subjects to monitor and the theorem of intersecting lines. The experiment lasted around 30– 45 min for each participant. 2.2.4. Experiment 4: influence of HMDs on stereo vision, visual acuity and visual field sensitivity Stereo vision was quantified by determining the minimal disparity necessary for a stereoscopic percept using the TNO-test, a stereo test based on random-dot stereograms viewed through red-green goggles. This test is based on red-green randomdot anaglyph images. The measurements were conducted for each of the conditions and then repeated with the inverted TNO-test cards to minimise learning effects for the second test. For the calculation of the descriptive statistics and the determination of the significance levels, disparity values were logarithmised and subsequently delogarithmised for the illustration of the descriptive statistics. Decimal visual acuities were determined binocularly at a 5-m viewing distance using an adaptive procedure (Freiburg Visual Acuity and Contrast Test, Bach 1996). Visual acuities were determined with and without HMD for each HMD-type in a separate session, following a counter-balanced design (following an A – B –B – A scheme for one half and an B – A –A – B for the other half of subjects) and yielding a total of four acuity measurements per HMD-type. These precautions were taken to exclude potential learning effects known for visual acuity measurements (Heinrich, Kruger, and Bach 2011). For the calculation of the descriptive statistics and the determination of the significance levels, decimal visual acuity values were logarithmised and subsequently delogarithmised for the depiction of the descriptive statistics. To assess visual field sensitivities, light spot detection thresholds were determined binocularly with automated static white on white perimetry (Octopus 101 Perimeter; Haag-Streit, Ko¨niz, Switzerland) using a standard paradigm, which was customised to include the position of the blind spot [background luminance: 4 asb; target size: Goldmann size III (0.48 diameter); target exposition: 100 ms]. At 97 visual field positions, evenly distributed in the visual field (^ 22.58) with 58 spacing, we determined the detection thresholds with the dynamic strategy (Weber and Klimaschka 1995) incorporated in Octopus 101. Threshold sensitivity is given as S[dB] ¼ 10 £ log(Lmax/Lstimulus), where Lmax equals 1000 asb. Subjects looked straight ahead during testing. Sensitivities were determined with and without HMD, for each HMD-type in a separate session, following a counter-balanced design across subjects (A– B sequence for one half and B – A sequence for the other half of the subjects) and yielding a total of four acuity measurements. These precautions were taken to exclude potential learning effects. The reliability factor of the perimetry measurements, i.e. the percentage of false positives and false negatives in the catch trials in the reference condition, without HMD, never exceeded 8.3%, and the median across subjects was 0%. 2.3.
Statistics
One-way repeated measures ANOVAs were conducted separately in Experiment 1 (factor: display) and in Experiment 2 for the condition switch to monitor and switch to HMD (factor display). Post hoc analyses were done with Scheffe-tests sequential Bonferroni corrected for multiple testing. To examine the ERs, Friedman tests [post hoc: Wilcoxon-test; sequentially Bonferroni corrected for multiple testing; (Holm 1979)] were applied. For Experiments 3 and 4, significance levels were determined with sequentially Bonferroni-corrected paired t-tests (Holm 1979).
Ergonomics
5
3. Results 3.1. Experiment 1: influence of HMDs on performance in visual search and Go-NoGo task
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Out of the variety of comparisons, only two significant effects of HMDs on RTs were evident (Figure 2), i.e. RT increases for oledOST (for response to matrix RT oledOST and LA: 1147 and 1004 ms, respectively; p ¼ 0.048) and LA (for response to Go-NoGo RT LA and monitor: 1073 and 950 ms, respectively; p ¼ 0.007). Furthermore, only two significant effects on ERs were evident, i.e. increased ER for oldedOST and more misses of the go-target for oledOST (Table 2). In summary, only few differences between HMDs were evident for visual search and Go-NoGo tasks, mainly slight deteriorations of performance for oledOST and one RT increase for LA. Presenting tasks monocularly in HMDs does generally not lead to a performance decrement compared to a conventional binocular monitor viewing condition, when no distracting stimuli are present in the environment. It should be noted that vrdOST was not tested in this experiment.
3.2. Experiment 2: costs for randomly switching the display medium The RTs and ERs were assessed for ‘response to matrix’ and ‘response to Go-NoGo’ while the mediums monitor and HMDs alternated in randomised order.
3.2.1. Response to matrix There were no RT differences between the HMDs. Effects on ER, i.e. false alarms, were observed for tasks both on HMDs and on the monitor, but only when the no-go-target (‘R’) was present. On the HMDs, ER were significantly greater for ‘Mon&vrdOST’ than for ‘Mon&LA’ ( p ¼ 0.006) and ‘Mon&oledOST’ ( p ¼ 0.006). On the monitor, ER were significantly greater for ‘Mon&LA’ than for ‘Mon&oledOST’ ( p ¼ 0.012).
3.2.2.
Response to Go-NoGo
Effects on both RT and ER, for both misses and false alarams, were observed (see Figures 3 and 4, respectively). RT were greater for LA, i.e. for tasks on the monitor RT were greater for switches from LA than from vrdOST ( p ¼ 0.016) and for tasks on the HMDs RT were greater for switches to LA than to oledOST ( p ¼ 0.009). ER, i.e. misses of the go-target, were higher for ‘Mon&LA’ than for ‘Mon&oledOST’ for both switches to the monitor (8.5 vs 2.0; p ¼ 0.005) and to the HMD (7.5 vs 1.5; p ¼ 0.028). False alarms for the no-go-target on monitor and HMDs were more frequent for ‘Mon&LA’ than for both ‘Mon&vrdOST’ (15.5 vs 2.7, p ¼ 0.024 and 16.00 vs 2.7, p ¼ 0.004, respectively) and ‘Mon&oledOST’ (15.5 vs 1.5; p ¼ 0.041 and 16.0 vs 1.5; p ¼ 0.008, respectively). In summary, only few and unsystematic dependences of media switching on HMD type were evident for response to matrix. In contrast, for the Go-NoGo task effects were more frequent and concentrated on LA. For LA, ER increased when switching to both monitor and HMD, and RT increased when switching to the monitor.
Figure 2. Experiment 1 – RT for monitor vs oledOST vs LA (n ¼ 11; mean ^ SEM). The RT to the matrix was significantly higher for oledOST than for LA for go-target (P) present ( p ¼ 0.048). Futhermore, the performance of the Go-NoGo task was faster for monitor than for LA ( p ¼ 0.007).
4 (9) 0 (2)
2.9 (11.9) 13.1 (22.8) 4.7 (21.5) 19.4 (19.7) 2.7 (7.4) 19.3 (21.4)
Monitor Median (interq. range) (%)
6 (15.5) 0 (4)
7.3 (35.5) 24.9 (11.9) 7.7 (34.2) 24.7 (22.3) 5.7 (13.8) 33.1 (21.4)
oledOST Median (interq. range) (%)
Experiment 1 – ERs for monitor vs oledOST vs LA.
Matrix fa NoPR m NoPR fa P m P fa R m R Go-NoGo P R
Task
Table 2.
6 (13) 1 (2)
3.6 (21.6) 9.0 (15.2) 1.6 (20.1) 12.0 (8.0) 3.5 (17.9) 11.4 (15.6)
LA Median (interq. range) (%)
7.52 (2) 0.08 (2)
3.250 (2) 10.750 (2) 5.871 (2) 4.750 (2) 4.710 (2) 5.250 (2)
X 2 (df)
0.023 0.961
0.197 0.005 0.053 0.093 0.095 0.072
p
0.150 0.072 0.126 0.484 0.108 0.100 0.036 1.00
2 2.524 2 0.962
P
2 1.960 2 2.100 2 1.859 2 .700 2 2.100 2 1.960
Z
Monitor vs oledOST
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
21.973 20.322
21.400 21.540 20.420 21.540 20.840 20.840
Z
0.098 0.748
0.322 0.123 0.674 0.246 0.802 0.401
P
Monitor vs LA
2 1.614 2 0.597
2 1.120 2 2.521 2 2.100 2 2.100 2 0.338 2 2.380
Z
0.106 1.00
0.263 0.036 0.108 0.108 0.735 0.051
P
oledOST vs LA
6 L. Schega et al.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Ergonomics
7
Figure 3. Experiment 2 – RT to the task ‘response to Go-NoGo’ (n ¼ 11; mean ^ SEM). The Go-NoGo task was presented on HMD (vrdOST or LA or oledOST) or monitor (Mon.). Presenting the go-target (‘P’) on the monitor yielded faster responses for vrdOST and monitor than for LA and monitor ( p ¼ 0.016). If the go-target was presented on the HMD, the RT was faster for vrdOST and monitor than for oledOST and monitor ( p ¼ 0.009).
Figure 4. Experiment 2 – ERs to the ‘response to Go-NoGo’ signal (n ¼ 11; mean ^ SEM). Misses of the go-target on the monitor were more frequent for ‘Mon&LA’ than for ‘Mon&oledOST’ (8.5 vs 2.0; p ¼ 0.005). On the HMD, they were more frequent for ‘Mon&LA’ vs ‘Mon&oledOST’ (7.5 vs 1.5; p ¼ 0.028). False alarms for the no-go-target on monitor and HMDs were more frequent for ‘Mon&LA’ than for ‘Mon&vrdOST’ (15.5 vs 2.7, p ¼ 0.024 and 16.00 vs 2.7, p ¼ 0.004, respectively) and ‘Mon&oledOST’ (15.5 vs 1.5; p ¼ 0.041 and 16.0 vs 1.5; p ¼ 0.008, respectively).
3.3. Experiment 3: vergence movements Although the vergence system is open loop for LA-HMDs and therefore instable, we observed effects that exceeded the increased scatter in the vergence data and consequently reached statistical significance. Only for LA, the eyes converged in front of the monitor plane ( p ¼ 0.038). For LA, the distance of the convergence point from the convergence on the monitor was significantly greater than for oledOST and vrdOST, p ¼ 0.038 and p ¼ 0.049, respectively (Figure 5).
3.4. 3.4.1.
Experiment 4: influence of HMDs on stereo vision, visual acuity and visual field sensitivity Visual acuity
The mean decimal visual acuity exceeded 1.6 in the reference measurements, i.e. without HMD. A significant effect of the HMD compared to reference was evident for LA, namely a reduction to 1.3 ( p , 0.001), while no significant effects were evident for the other two HMDs tested (see Figure 6(A)).
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
8
L. Schega et al.
Figure 5. Experiment 3 – vergence with HMDs (n ¼ 11; mean ^ SEM). The visual axes intersected significantly in front of the monitor only for LA ( p ¼ 0.038). Furthermore, for LA the distance of the convergence point to the monitor was significantly greater than for oledOST and vrdOST, p ¼ 0.038 and p ¼ 0.049, respectively (Figure 6).
3.4.2. Stereo vision The LA leaves the subject effectively monocular by completely covering one eye. As a matter of course none of the stereocharts was passed for LA. In contrast, stereo vision was retained for the other two HMDs. However, compared to unrestrained binocular viewing, significantly greater disparities were required for the recognition of the stereo targets ( p , 0.005; see Figure 6(B)). 3.4.3. Visual field sensitivity In analogy to Hoffmann, Seufert, and Schmidtborn (2007), visual field sensitivities were assessed at 11 visual field positions along the horizontal meridian (average of þ 2.58 and 2 2.58 elevation; see Figure 7(C)). Significant sensitivity reductions, compared to a reference condition without HMD, were evident at six positions for LA and at two positions for oledOST, but none were evident for vrdOST. It should be noted that non-significant sensitivity reductions were also induced in the left visual hemifield at around 158 of eccentricity and were particularly pronounced for LA and oledOST. These were presumably due to the blind spot in the left hemifield corresponding to the papilla of the left eye, as the right eye was covered by the HMD. The effect did not reach significance, probably due to the inter-individual variability of the size and location of the blind spot. To exclude this region from a more formal evaluation, the effects of the HMDs were assessed quantitatively for the average of the sensitivities within ^ 7.58 along the horizontal meridian. There was a significant sensitivity reduction for LA by 2.1 ^ 0.3 dB ( p , 0.001) and for oledOST by 0.8 ^ 0.2 dB ( p , 0.01). For vrdOST, there was a trend of a reduction by 0.6 ^ 0.4 dB, which failed to reach significance ( p ¼ 0.142). 4.
Discussion
A strong consistent effect of HMDs on visual search was absent and only minor effects were observed. In contrast, the integration of sporadically available additional visual information, i.e. the Go-NoGo task, was particularly affected by media switching to and from LA. This might be associated with the finding that basic aspects of visual function were most severely affected for LA. 4.1.
HMD effects on basic aspects of visual function
Basic aspects of visual function were affected for all HMDs, albeit to a variable degree. The effects were most pronounced for LA and least for vrdOST. For LA, visual function was reduced in all tests conducted, i.e. visual acuity, stereo vision (absence) and visual field sensitivities. This is due to the functional monocularity of observers with LA, which reduces visual acuity (Wood, Collins, and Carkeet 1992; Cagenello, Arditi, and Halpern 1993) and visual field sensitivities (Wood, Collins, and Carkeet 1992; Cagenello, Arditi, and Halpern 1993) and blocks binocular vision. Visual acuity and visual field sensitivities normally benefit from binocular summation, i.e. an increase of acuity in the order of 1.3 (Wood, Collins, and
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Ergonomics
9
Figure 6. Experiment 4 – influence of HMDs on stereo vision, visual acuity and visual field sensitivity. (A) Differential effect of wearing HMDs on visual acuity (n ¼ 11; mean ^ SEM). Only for LA the visual acuity was significantly reduced ( p , 0.001). (B) Differential effect of wearing HMDs on disparity thresholds for stereo vision (n ¼ 10; mean ^ SEM). While stereo targets were not detected for LA, they were detected both oledOST and vrdOST. However, in comparison to the condition ‘no HMD’ the disparity thresholds were significantly elevated ( p , 0.005). (C) Differential effect of wearing HMDs on visual field sensitivities (n ¼ 10; mean ^ SEM) along the horizontal meridian. Significant reductions were evident for LA and oledOST, but not for vrdOST. Significance levels: *p , 0.05; **p , 0.01; ***p , 0.001.
Carkeet 1992; Cagenello, Arditi, and Halpern 1993; Pardhan 1993; Vedamurthy et al. 2007) and visual field sensitivity of up to 1.4 (Wood, Collins, and Carkeet 1992). Consequently, LA visual acuities and visual field sensitivites are reduced. Further, blocking binocular vision results in an absence of stereo vision, which is also a plausible cause of the inaccurate vergence movements observed for LA. For oledOST, both stereo vision and visual field sensitivities were reduced. The latter is expected as the neutral density filter properties of oledOST reduce stimulus luminance by 30% (Azuma 1997). In contrast, the reduction of stimulus luminance for vrdOST is negligible (Patterson, Winterbottom, and Pierce 2006) and no reduction in visual field sensitivities was observed. For vrdOST, however, reductions of stereo sensitivity were evident. Consequently, stereovision was reduced for all HMDs, which indicates that monocular HMDs bear upon the integrity of high level binocular visual function that is essential for stereovision. 4.2.
Effects of HMD on demanding visual tasks
During the Go-NoGo task, the performance for the HMD LA was particularly inferior to that for the other HMDs. This effect was most pronounced during media switching. As basic aspects of visual function are most severely reduced for LA, these reductions might be a cause of the impaired processing of sporadically available information with highest priority and with high visual acuity demand. In particular, the reduced sensitivity for brief luminance changes, as assessed with static visual field perimetry, combined with inferior, since monocular, visual acuity for LA is a likely cause of the reduced
10
L. Schega et al.
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
performance during the Go-NoGo task. An inaccurate vergence indicates a mismatch of the perceived spatial locations of the two media, i.e. monitor and HMD, which is likely associated with the absence of stereo vision. Furthermore, inaccurate vergence might, due to the coupling of vergence and accommodation, cause inaccurate accommodation (Drascic and Milgram 1996), further reducing visual acuity and hence target identification. This disadvantage of LA for the Go-NoGo task appears to be particularly relevant when switching between media. Consequently, LA has higher switching costs than the other HMDs tested. In conclusion, the economic HMD variant, i.e. LA, is an alternative to OST HMDs, if moderate reductions in visual function can be tolerated and if sporadic external events on the unattended medium are kept to a minimum. In the absence of distracting stimuli, monocularly presented tasks in LA-HMDs do not lead to a performance decrement. Consequently, ideal applications of LA HMDs are those that require the augmentation of visual input with embedded contextual information at predictable intervals in harzard-free situations without environmental stimuli. The use of LA-HMDs in hazard-free conditions should be particularly acceptable in applications without the need to fuse virtual objects and the real scene. However, if a 3D alignment of real and virtual objects is required, OST-HMDs would be applicable. Therefore, typical examples for LA HMDs applications would be the presentation of operating instruction (symbolic or textual) without 3D alignment to real world objects, biofeedback in rehabilitation and sports-training, and presentation of private data.
Acknowledgements The support by the Federal Ministry of Education and Research (01 IM 08 001 Z) is gratefully acknowledged. Furthermore, the authors thank Dr Huckauf and Mr Urbina from the Bauhaus-University of Weimar for their support.
References Adamovich, S. V., G. G. Fluet, E. Tunik, and A. S. Merians. 2009. “Sensorimotor Training in Virtual Reality: A Review.” NeuroRehabilitation 25 (1): 29 – 44. Azuma, R., Y. Baillot, R. Behringer, S. Feiner, S. Julier, and B. Macintyre. 2001. “Recent Advances in Augmented Reality.” IEEE Computer Graphics and Applications 21 (6): 34 – 47. Azuma, R. T. 1997. “A Survey of Augmented Reality.” Presence: Teleoperators and Virtual Environments 6 (4): 355– 385. Bach, M. 1996. “The Freiburg Visual Acuity Test – Automatic Measurement of Visual Acuity.” Optometry & Vision Science 73: 49– 53. Cagenello, R., A. Arditi, and D. L. Halpern. 1993. “Binocular Enhancement of Visual Acuity.” Journal of the Optical Society of America A: Optics, Image Science, and Vision 10 (8): 1841– 1848. Drascic, D., and P. Milgram. 1996. Perceptual Issues in Augmented Reality, 123–134. San Jose, CA: The International Society for Optical Engineering. Espay, A. J., Y. Baram, A. K. Dwivedi, R. Shukla, M. Gartner, L. Gaines, A. P. Duker, and F. J. Revilla. 2010. “At-Home Training with Closed-Loop Augmented-Reality Cueing Device for Improving Gait in Patients with Parkinson Disease.” Journal of Rehabilitation Research and Development 47 (6): 573– 581. Heinrich, S. P., K. Kruger, and M. Bach. 2011. “The Dynamics of Practice Effects in an Optotype Acuity Task.” Graefes Archive for Clinical and Experimental Ophthalmology 249 (9): 1319– 1326. Hiramatsu, K., T. Inui, M. Okada, T. Takeshima, H. Mishima, T. Sakaki, and S. Shiono. 2005. “New Device for Endoscopic Image Display During Microsurgical Clipping of Cerebral Aneurysms - Technical Note.” Neurologia Medico-Chirurgica (Tokyo) 45 (9): 487– 490. Hoffmann, M. B., P. S. Seufert, and L. C. Schmidtborn. 2007. “Perceptual Relevance of Abnormal Visual Field Representations – Static Visual Field Perimetry in Human Albinism.” The British Journal of Ophthalmology 91: 509– 513. Holm, S. 1979. “A Simple Sequentially Rejective Multiple Test Procedure.” Scandinavian Journal of Statistics 6: 65 – 70. Huckauf, A., M. Urbina, I. Bo¨ckelmann, L. Schega, R. Mecke, J. Grubert, F. Doil, and J. Tumler. 2010. “Perceptual Issues in Optical-SeeThrough Displays.” In Proceedings – APGV 2010: Symposium on Applied Perception in Graphics and Visualization, 41 – 48. New York: Association for Computing Machinery. Kawara, T., M. Ohmi, and T. Yoshizawa. 1996. “Effects on Visual Functions During Tasks of Object Handling in Virtual Environment with a Head Mounted Display.” Ergonomics 39 (11): 1370 –1380. Liu, D., S. A. Jenkins, P. M. Sanderson, M. O. Watson, T. Leane, A. Kruys, and W. J. Russell. 2009. “Monitoring with Head-Mounted Displays: Performance and Safety in a Full-Scale Simulator and Part-Task Trainer.” Anesthesia and Analgesia 109 (4): 1135– 1146. Pardhan, S. 1993. “Binocular Performance in Patients with Unilateral Cataract Using the Regan Test: Binocular Summation and Inhibition with Low-Contrast Charts.” Eye (London) 7 (Pt 1): 59 – 62. Patterson, R., M. D. Winterbottom, and B. J. Pierce. 2006. “Perceptual Issues in the Use of Head-Mounted Visual Displays.” Human Factors 48 (3): 555– 573. Peters, M., and J. Ivanoff. 1999. “Performance Asymmetries in Computer Mouse Control of Right-Handers, and Left-Handers with Leftand Right-Handed Mouse Experience.” Journal of Motor Behavior 31 (1): 86 – 94. Schega, L., D. Hamacher, and R. C. Wagenaar. 2011. “A Comparison of Effects of Augmented Reality and Verbal Information Based Interventions in Elderly Women After Hip Replacement.” Archives of Physical Medicine and Rehabilitation 92 (10): 1734– 1735. Schuster, D., J. Rivera, B. C. Sellers, S. M. Fiore, and F. Jentsch. 2013. “Perceptual Training for Visual Search.” Ergonomics. http://www. ncbi.nlm.nih.gov/entrez/query.fcgi?cmd¼Retrieve&db¼PubMed&dopt¼Citation&list_uids¼23650877
Downloaded by [Moskow State Univ Bibliote] at 00:05 12 February 2014
Ergonomics
11
So, R. H. Y., W. S. Wong, R. Yip, A. K. C. Lam, and P. Ting. 2011. “Benefits of Matching Accommodative Demands to Vergence Demands in a Binocular Head-Mounted Display: A Study on Stereo Fusion Times.” Presence-Teleoperators and Virtual Environments 20 (6): 545– 558. Sutherland, I. E. 1968. “A Head-Mounted Three Dimensional Displayed.” In American Federation of Information Processing Societies Fall Joint Computer Conference, 757– 764, Part I. Washington, DC: Thompson Books. Tumler, J., F. Doil, R. Mecke, G. Paul, M. Schenk, E. A. Pfister, A. Huckauf, I. Bo¨ckelmann, and A. Roggentin. 2008. “Mobile Augmented Reality in Industrial Applications: Approaches for Solution of User-Related Issues.” In 7th IEEE International Symposium on Mixed and Augmented Reality Cambridge, 87 – 90. Washington, DC: IEEE Computer Society. Vedamurthy, I., C. M. Suttle, J. Alexander, and L. J. Asper. 2007. “Interocular Interactions During Acuity Measurement in Children and Adults, and in Adults with Amblyopia.” Vision Research 47 (2): 179– 188. Weber, J., and T. Klimaschka. 1995. “Test Time and Efficiency of the Dynamic Strategy in Glaucoma Perimetry.” German Journal of Ophthalmology 4 (1): 25 –31. Wood, J. M., M. J. Collins, and A. Carkeet. 1992. “Regional Variations in Binocular Summation Across the Visual Field.” Ophthalmic and Physiological Optics 12 (1): 46 – 51. World Medical Association. 2000. “Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects.” The Journal of the American Medical Association 284 (23): 3043–3045. Zhou, F., H. B.-L. Dun, and M. Billinghurst. 2008. “Trends in Augmented Reality Tracking, Interaction and Display: A Review of Ten Years of ISMAR.” In 7th IEEE International Symposium on Mixed and Augmented Reality, (ISMAR 2008), 193–202. Washington, DC: IEEE Computer Society.
