112 / 218
110
Chapter 5
Figure 5.3
The effect of Reward on Task switching diminishes with age in a subgroup
matched for maximum available reward
The results in figure 2 were supported by an additional analysis, in which we only included the
participants who all received the exact same version of the task (study A, C, D in table 1).
A:
This analysis confirmed that the effect of Reward on Task switching is reduced with age. The outline
of the data points indicates the study in which each participant participated (table 1).The red and blue
data points refer to the younger and older group in a subsequent median split analysis (B), * The black
solid line indicates the regression slope (Pearson) for the subgroup (N = 68), the Spearman correlation
is reported.
B:
Analysis with age as a between subject factor (median split) revealed significant effects of Reward on
Task switching in the younger group, but not the older group.
also reflect these age-related differences in the maximum time participants had to respond
on each trial type. However, this age-related effect in response deadlines was not different for
repeat and switch trials (Age x Task switching in terms of response deadline: r (ρ) = 0.050, p
= 0.588).
Next, we assessed whether the difference between the average response times during
the actual test and the average response times during the practice (i.e. average response
deadlines) changed with age. Indeed, this difference increased with age (Age x Response
time difference: r (ρ) = 0.467, p < 0.001), suggesting that the average response deadline was
less close to their maximum speed for older than younger participants. To test whether this
can account for the higher earnings, we assessed whether the response time difference was
associated with the total earnings. Indeed, a larger response deadline difference (i.e. increased
room for improvement on test compared with practice) was associated with higher earnings
(Total earnings x Response time difference: r (ρ) = 0.480, p < 0.001). Crucially however, the
difference in response times between practice and test was not associated with the task effects
(Rewarded Task switching effects in terms of SAT x Response time difference (r (ρ) = -0.037,
A C D
< 26 > 25
study
Age (years)
age group
-1
0
1
2
10 20 30 40 50 60 70
r = - 0.360,
p
= 0.003 *
more speed on
high reward trials
higher accuracy on
high reward trials
-0.2
0
0.2
A
B
(SAT)
high (switch - repeat)
- low (switch - repeat)
Reward e ect on task switching
high reward - low reward
(SAT)
Younger Older
Age group
N = 35
N = 33
p
< 0.001
p
> 0.05
p
< 0.001
Reward e ect on task switching