115 / 218
113
Reward modulation of cognitive function: aging
methods exist which can be used to assess speed accuracy tradeoffs (Salthouse and Hedden,
2002; Forstmann et al., 2011; Heitz, 2014). The method used in this study is an elementary
formalization of a speed accuracy tradeoff. Nevertheless, the results clearly show that taking
into account both speed and accuracy can reveal differences that would not be revealed when
response times and accuracy are assessed separately. However, future work should extend these
findings using more sophisticated approaches (e.g. mathematical decision making models)
and by experimentally manipulating speed accuracy strategies (Heitz, 2014). Such model-
based approaches might be more sensitive, as they take into account trial-by-trial changes
in speed and accuracy. The consequence of the simple composite score approach used in the
current study is that changes in speed and accuracy are assumed to contribute equally to the
decision process. It is therefore hard to interpret the SAT measure used in the current study
without taking note of the separate response and accuracy measures. Finally, we cannot rule
out completely that older participants show reduced motivational effects because they value
money less than young participants. However, we argue that this is unlikely for two reasons.
First, age was associated with contrasting effects of reward on switch and repeat trials. Second,
older participants actually earned more money than did the younger participants.
In the current study we controlled for general age-related differences in processing speed
(Salthouse, 1996) by determining each individual’s response deadline, by using a within-
subject comparison of conditions, and by taking into account the speed-accuracy tradeoff.
In line with previous work (Bijleveld et al., 2010; Forstmann et al., 2011), older participants
were more cautious (i.e. slower and more accurate than younger participants). This cautious
strategy was already evident during practice and allowed older participants more time to
respond accurately during test. It is unlikely that allowing slower (older) participants more
time has induced differences in reward-related speed-accuracy strategies, given the equal
response deadline for high and low reward (i.e. only arrow/word x switch/repeat trials were
adapted). In addition, the response deadline adaptations did not result in an age-related
difference in switch and repeat trials, suggesting it is unlikely that this adaptation has changed
the accuracy-over-speed strategy during the integration of reward and task switching
differently in older and younger participants. The cautiousness of the older participants was
further corroborated by the observation that the overall response deadline of the younger
participants was closer to their maximum performance in terms of speed (on test). Crucially
however, this difference was not related to the age-related adaptation to the task conditions.
Here, we show for the first time that age-related changes in response strategies can be observed
when participants need to flexibly adapt to changing task demands. Specifically, we observed
an age-related decrease in the degree to which older participants use information about
rewards to change their speed-accuracy strategy in changing cognitive control conditions;
i.e., older participants no longer used rewards to adapt cognitive control processes. As such
the present study goes beyond prior work focusing commonly merely on cognitive deficits
(Salthouse, 1996; West, 1996; Kray et al., 2002; Park et al., 2002). An obvious next step would
be to unravel the neural mechanisms underlying these changes.