Demographics
Two older adults were excluded for not completing the entire virtual environment task (one male and one female older adult). The final sample comprised of thirty-one young adults (15 males and 16 females; mean age: 22.68 ± 3.07) and thirty older adults (13 males and 17 females; mean age: 67.30 ± 5.10) (Table 1). Young and older adults were well matched in gender ratio, years of education, and MMSE. Moreover, older adults showed less exposure to computers and virtual environment than young adults. Young adults showed significantly better performance than older adults in WAIS-block design and digit span (both forward and backward versions) with medium-to-large effect size. No significant differences were found between young and older adults in all three subtests of MMQ. Twenty-one young adults (67.7%) needed extra 1 to 4 practice trials (Mean = 1.10, SD = 1.06), and twenty-six older adults (86.7%) needed additional 1 to 7 practice trials (Mean = 2.57, SD = 1.93), the percentage did not show significantly differences (χ2 = 3.088, p = .079). However, older adults required significantly more practice trials than young adults (t = -3.306, p = .002). Therefore, the number of practice trials was analyzed as the covariate in the following ANOVA.
Table 1. Demographic information and cognitive measures of young and older adults
Young adults
(n = 31)
|
Older adults
(n = 30)
|
t
|
Effect size (d)
|
M
|
SD
|
M
|
SD
|
Age (years)
|
22.68
|
3.07
|
67.30
|
5.10
|
-41.24***
|
-0.98
|
Education (years)
|
15.61
|
2.09
|
15.27
|
1.39
|
0.76
|
0.10
|
Computer exposure
|
1.23
|
0.43
|
1.73
|
0.78
|
-3.13**
|
-0.81
|
Virtual reality exposure
|
2.48
|
0.72
|
2.97
|
0.18
|
-3.60**
|
-0.94
|
MMSE
|
28.68
|
1.14
|
28.70
|
1.32
|
-0.07
|
-0.01
|
MMQ-contentment
|
50.55
|
12.19
|
46.50
|
11.04
|
1.36
|
0.17
|
MMQ-ability
|
54.45
|
9.60
|
50.18
|
10.83
|
1.63
|
0.20
|
MMQ-strategy
|
33.94
|
15.08
|
30.42
|
14.57
|
0.94
|
0.12
|
WAIS block design
|
42.87
|
4.62
|
33.33
|
6.71
|
6.42***
|
0.64
|
Forward digit span
|
8.74
|
0.58
|
7.73
|
0.83
|
5.51***
|
0.576
|
Backward digit span
|
6.74
|
1.53
|
5.20
|
1.42
|
4.081**
|
0.463
|
M: mean; SD: standard deviation; MMSE: Mini-Mental State Examination; MMQ: Multifactorial Memory Questionnaire; WAIS: Wechsler Adult Intelligence Scale-IV. ** p < .01, *** p < .001
Strategy defined by the probe trial
According to the strategy discrimination criterion in probe trial, the current study comprised four groups: egocentric older adults, non-egocentric older adults, egocentric young adults and non-egocentric young adults. The final sample comprised of seven egocentric older adults (23.33%), twenty-three non-egocentric older adults (76.67%), seventeen egocentric young adults (54.84%) and fourteen non-egocentric young adults (45.16%), χ2 = 6.34, p = .012. Two strategy groups did not differ in age. Specifically, egocentric older adults age from 61 to 75 (mean age: 66.86 ± 5.640, 5 females and 2 males), non-egocentric older adults age from 57 to 78 (mean age: 67.43 ± 5.053, 12 females and 11 males), egocentric young adults age from 18 to 29 (mean age: 23.59 ± 3.144, 8 females and 9 males), non-egocentric young adults age from 18 to 28 (mean age: 21.57 ± 2.681, 7 females and 4 males) (Table 2).
Table 2. Demographic information of egocentric and non-egocentric strategy navigators in young and older adults.
Young adults (n = 28)
|
Older adults (n = 30)
|
Egocentric
(n = 17)
|
Non-egocentric (n = 11)
|
Egocentric
(n = 7)
|
Non-egocentric (n = 23)
|
Age (year)
|
21.57 ± 2.68
|
23.59 ± 3.14
|
67.43 ± 5.05
|
66.86 ± 5.64
|
Female/Male
|
8/9
|
7/4
|
12/11
|
5/2
|
3 young adults (2 males and 1 female) exceeded three-standard deviations (SDs) in the 6th and 7th learning trials and were excluded from the analyses.
Theoretically, the absence of distal landmarks in probe trial should only impair strategies relied on landmarks (i.e. allocentric strategy or guidance strategy), leaving the egocentric strategy unimpaired. As shown in Figure 2, the performance of egocentric strategy users (including both young and older adults) did not differ between the 9th learning trial and the probe trial, whereas non-egocentric strategy users performed significantly worse in the probe trial than in the 9th learning trial, including slower navigation speed (t = 2.273, p = .029), more rotations (t = -10.382, p < .001), higher distance error (t = -9.113, p < .001) and lower percentage of successful trials (t = 6.508, p < .001). This specific impairment of non-egocentric navigation from the 9th learning trial to the probe trial highly supported the strategy discrimination of the probe trial.
To be noted, we carefully checked all the trajectories of the participants. All participants classified into non-egocentric group entered more irrelevant alleys in probe trial compared with the 9th learning trial, except for 2 older adults, who entered the same number of irrelevant alleys during the probe trial and the 9th trial, indicating that the two older adults used “serial strategy” [52] by keeping turning left in all the intersections (Supplementary Figure 1). This strategy needs less cognitive load and more like a trial-and-error procedure (e.g., I will certainly reach the destination if I keep turning left in each intersection), while egocentric strategy requires the navigators to combine a series of stimulus-response association (e.g., turn left in the first intersection and turn right when I see a forest). Therefore, these two older adults did not fully form an egocentric strategy according to their route and were classified as non-egocentric strategy users.
Navigation performance in the learning phase
According to the results of repeated-measure ANOVA (Table 3), the data did not fulfil the sphericity assumption (all ps < .001), we chose the Huynh-Feldt epsilon coefficient to adjust the degrees of freedom (ε = .654 for speed, ε = .655 for rotation, ε = .533 for distance error, ε = .619 for percentage of successful trials, separately). We found that the main effect of learning trials were significant in speed [F(5, 282) = 10.860, p < .001, η2=.167], rotation [F(4, 195) = 2.535, p = .047, η2=.045], and distance error [F(4, 230) = 2.796, p = .024, η2=.049] (Figure 3a, b, c). The main effect of learning trials did not reach significance in the percentage of successful trials [F(5, 268) = .627, p = .678, η2=.011] (Figure 3d). Given that the percentages of successful trials were close to 100% during the entire learning phase, the absence of learning effect may be the consequence of the ceiling effect.
Table 3. Repeated-measure ANOVA with age and strategy as between factors and trial as within factor (n = 58)
Speed (vm/ second)
|
Rotation (°)
|
Distance error (%)
|
Successful trials (%)
|
Age
|
F
|
25.906
|
2.711
|
2.235
|
3.584
|
p
|
< 0 .001***
|
0.105
|
0.141
|
0.064
|
Effect size (η2)
|
0.32
|
0.048
|
0.04
|
0.062
|
Strategy
|
F
|
0.385
|
6.969
|
6.326
|
1.96
|
p
|
0.538
|
0.011*
|
0.015*
|
0.167
|
Effect size (η2)
|
0.007
|
0.114
|
0.105
|
0.035
|
Trial
|
F
|
10.86
|
2.535
|
2.796
|
0.627
|
p
|
< 0 .001***
|
0.047*
|
0.024*
|
0.678
|
Effect size (η2)
|
0.167
|
0.045
|
0.049
|
0.011
|
Age × Strategy
|
F
|
0.524
|
7.385
|
7.182
|
4.037
|
p
|
0.472
|
0.009**
|
0.010**
|
0.050*
|
Effect size (η2)
|
0.010
|
0.120
|
0.117
|
0.070
|
Age × Trial
|
F
|
1.284
|
0.123
|
0.279
|
0.094
|
p
|
0.269
|
0.966
|
0.902
|
0.993
|
Effect size (η2)
|
0.023
|
0.002
|
0.005
|
0.002
|
Strategy × Trial
|
F
|
0.525
|
0.150
|
0.304
|
0.600
|
p
|
0.765
|
0.952
|
0.885
|
0.698
|
Effect size (η2)
|
0.010
|
0.003
|
0.006
|
0.011
|
Age × Strategy × Trial
|
F
|
1.136
|
1.024
|
1.160
|
1.321
|
p
|
0.342
|
0.392
|
0.330
|
0.256
|
Effect size (η2)
|
0.021
|
0.019
|
0.021
|
0.024
|
3 young adults (2 males and 1 female) were excluded from the repeated-measure ANOVA because of exceeding three-standard deviations in rotation. * p < .05, ** p < .01, *** p < .001.
The main effect of age was significant in speed [F(1, 54) = 25.906, p < .001, η2=.324]. Young adults navigated significantly faster than older adults (Figure 4a). None of the other main effects of age reached significance, including rotation [F(1, 54) = 2.711, p = .105, η2=.048], distance error [F(1, 54) = 2.235, p = .141, η2=.040] and percentage of successful trials [F(1, 54) = 3.584, p = .064, η2=.062].
The main effect of strategy was significant in rotation [F(1, 54) = 6.969, p = .011, η2=.114] and distance error [F(1, 54) = 6.326, p = .015, η2=.105]. Non-egocentric strategy users completed the virtual star maze task with more rotations and higher distance error. No significant effects were found in speed [F(1, 54) = .385, p = .538, η2=.007] and percentage of successful trials [F(1, 54) = 1.960, p = .167, η2=.035].
The interactions between age and strategy were significant in rotation [F(1, 54) = 7.385, p = .009, η2=.120], distance error [F(1, 54) = 7.182, p = .010, η2=.117] and percentage of successful trials [F(1, 54) = 4.037, p = .050, η2=.070]. As shown in Figure 4, all simple effect tests revealed similar results: the performances of non-egocentric older adults were significantly worse than those of egocentric older adults and non-egocentric young adults. Compared with egocentric older adults, non-egocentric older adults completed the task with more rotations [F(1, 57) = 20.81, p < .0011, η2=1.568], higher distance error [F(1, 57) = 19.35, p < .001, η2=1.574] and lower percentage of successful trials [F(1, 57) = 10.49, p = .002, η2=.910], suggesting that the non-egocentric strategy was less accurate than the egocentric strategy in older adults. Compared with non-egocentric young adults, non-egocentric older adults completed the task with more rotations [F(1, 57) = 15.49, p < .001, η2=1.241], higher distance error [F(1, 57) = 14.31, p < .001] , η2=1.226 and lower percentage of successful trials [F(1, 57) = 11.94, p = .001, η2=1.075], showing that older adults were less accurate than young adults in using an non-egocentric strategy. Other interactions were not significant (all ps > .05).
Given that young and older adults showed differences in WAIS block design, digit span, virtual environment exposure, computer exposure, and practice trials, another repeated-measure ANOVA was performed with all of them controlled as covariates. We found the interactions were nonsignificant for all of these covariates, indicating these covariates had no significant impacts on age-related navigation performance. The main effect of age showed that older adults navigated significantly slower than young adults [F(1, 49) = 6.168, p = .017, η2=.114]. The significant main effects of strategy revealed that non-egocentric strategy users completed the task with more rotations [F(1, 49) = 4.698, p = .035, η2=.086] and more distance error [F(1, 49) = 6.643, p = .013, η2=.108]. Moreover, the significant interactions between age and strategy revealed that non-egocentric older adults completed the star maze with more rotations [F(1, 49) = 6.072, p = .017, η2=.107] and distance error [F(1, 49) = 7.372, p = .009, η2=.116].
Visuo-spatial ability in older adults
All the Pearson correlations in young adults were not significant (Table 4). For older adults, the WAIS block-design score was positively correlated with navigation speed (r = .370, p = .044) and the percentage of successful learning trails (r = .417, p = .022) (Table 4, Figure 5), none of the other correlations were significant. These significant correlations suggest that better visuo-spatial ability in older adults was correlated with faster navigation speed and higher possibility to successfully complete the star maze.
Table 4. Correlations between navigation performance and cognitive tests in young and older adults
WAIS block design
|
Forward digit span
|
Backward digit span
|
r
|
p
|
r
|
p
|
r
|
p
|
Young adults
|
Speed (vm/ second)
|
-0.10
|
0.618
|
-0.064
|
0.747
|
-0.054
|
0.787
|
Rotation (°)
|
0.0
|
0.703
|
0.003
|
0.988
|
0.018
|
0.928
|
Distance error (%)
|
0.132
|
0.502
|
-0.167
|
0.397
|
0.015
|
0.938
|
Successful trials (%)
|
-0.144
|
0.465
|
0.021
|
0.915
|
-0.068
|
0.732
|
Older adults
|
Speed (vm/ second)
|
0.370
|
0.044*
|
-0.123
|
0.519
|
0.207
|
0.272
|
Rotation (°)
|
-0.355
|
0.054
|
-0.03
|
0.876
|
-0.222
|
0.238
|
Distance error (%)
|
-0.267
|
0.154
|
-0.053
|
0.779
|
-0.19
|
0.313
|
Successful trials (%)
|
0.417
|
0.022*
|
-0.193
|
0.307
|
0.151
|
0.426
|
WAIS: Wechsler Adult Intelligence Scale-IV; * p < .05.
Furthermore, the independent sample t tests revealed that egocentric older adults scored higher on the WAIS-block design test than non-egocentric older adults, suggesting better visuo-spatial ability in egocentric older adults (Table 5). However, two groups did not differ in forward digit span (t = -.785, p = .457) and backward digit span (t = .041, p = .967). These results suggest that non-egocentric older adults have a specific deficit in visuo-spatial ability. In addition, the non-egocentric older adults showed more satisfaction and confidence about their memory (MMQ1) and reported less trouble about memory during the last fortnight (MMQ2), whereas the egocentric older adults preferred to adopt different memory strategies during daily life (MMQ3, thought this result does not reach significant, the effect size is large), which may indicate a preserved ability to adopt or switch to an efficient strategy.
Table 5. Independent sample t tests between egocentric older adults and non-egocentric older adults
Egocentric older adults (n = 7)
|
Non-egocentric older adults (n = 23)
|
t
|
Effect size (d)
|
M
|
SD
|
M
|
SD
|
WAIS block design
|
37.57
|
3.31
|
31.74
|
7.1
|
3.008**
|
1.05
|
Forward digit span
|
7.43
|
1.13
|
7.78
|
0.67
|
-0.785
|
-0.38
|
Backward digit span
|
5.29
|
1.38
|
5.26
|
1.39
|
0.041
|
0.02
|
MMQ-contentment
|
38.57
|
10.45
|
48.26
|
9.83
|
-2.252*
|
-0.96
|
MMQ-ability
|
41.79
|
9.21
|
52.78
|
10.12
|
-2.565*
|
-1.14
|
MMQ-strategy
|
39.29
|
14.73
|
27.41
|
13.48
|
1.999
|
0.84
|
M: mean; SD: standard deviation; WAIS: Wechsler Adult Intelligence Scale-IV; MMQ: Multifactorial Memory Questionnaire. * p < .05, ** p < .01.