Switching-related costs are redistributed across processing stages and behavioral measures, and age-related modulation of semantic task-set reconfiguration emerges more clearly during preparation than task execution, particularly under continuous switching demands.
Key Findings
Results
Reaction time and error rate showed reliable age effects but no significant interactions between age and switching condition across both experiments.
Experiment 1 included 36 young and 30 older adults in an externally paced paradigm; Experiment 2 included 28 young and 28 older adults in a self-paced paradigm.
Age effects on RT and ER were consistent across experiments, indicating older adults were generally slower and/or less accurate.
The absence of age-by-condition interactions in RT and ER suggests switching-related costs in execution measures were not differentially larger for older adults.
Results
Local switch costs were expressed in error rate in Experiment 1 (externally paced) but shifted to reaction time in Experiment 2 (self-paced).
This cross-experiment difference indicates that the behavioral measure capturing local switch costs depends on the task structure and pacing conditions.
Local switch costs reflect the comparison between task-switch trials and task-repeat trials within a mixed block.
The shift from ER-based to RT-based local switch costs when participants control stimulus onset suggests temporal redistribution of processing demands.
Results
Global switch costs were observed in execution measures (RT and/or ER) only in Experiment 1 but not in Experiment 2.
Global switch costs were measured by comparing all-switch blocks to all-repeat blocks.
In Experiment 1 (externally paced), global switch costs appeared in execution-stage measures.
In Experiment 2 (self-paced), global switch costs were not reliably observed in RT or ER, suggesting they were absorbed into preparation time when participants controlled timing.
Results
In Experiment 2, preparation time showed reliable mixing, local, and global switching effects, with age-related amplification emerging specifically for global switching.
Preparation time was measured as the interval between participants' initiation of stimulus onset and the stimulus appearance, available only in the self-paced Experiment 2.
Mixing effects (mixed blocks vs. single-task blocks), local switch effects (switch vs. repeat trials within mixed blocks), and global switch effects (all-switch vs. all-repeat blocks) were all significant in preparation time.
Age-related amplification was found specifically for global switching in preparation time, indicating older adults took disproportionately longer to prepare under continuous switching demands.
No age-by-condition interactions were found in RT or ER in Experiment 2, underscoring that preparation time captured age differences that execution measures missed.
Discussion
Age-related differences in semantic task-set reconfiguration are underestimated when only execution measures (RT and ER) are examined, due to confounding of preparatory and execution processes.
Traditional task-switching paradigms that do not allow participant-controlled preparation conflate preparatory and execution-stage processes.
When participants could self-initiate stimulus onset (Experiment 2), age-related differences emerged in preparation time but not in RT or ER.
The authors conclude that 'age-related modulation of semantic task-set reconfiguration may emerge more clearly during preparation than task execution, particularly under continuous switching demands.'
Discussion
Switching-related costs are redistributed across processing stages (preparation vs. execution) and behavioral measures depending on task structure and pacing.
The pattern of results across two experiments demonstrates that the same underlying switching demands can manifest differently in RT, ER, or preparation time depending on paradigm design.
The authors interpret preparation time 'cautiously as reflecting participant-regulated preparatory processes rather than a pure measure of preparation efficiency.'
This redistribution phenomenon may explain why prior literature often reports minimal age-related differences in switch costs: execution-focused paradigms may obscure age effects that reside in preparatory processing.
Discussion
The conclusion that switching-related control processes are relatively preserved in aging may be an artifact of paradigms that do not separately measure preparation.
Prior task-switching studies relying on externally paced designs without preparation time measurement have generally found minimal age-related switch cost differences.
The current findings suggest this conclusion is based on paradigms that 'confound preparatory and execution processes.'
Separating preparation time from execution measures revealed age-related amplification specifically for global switching, a finding invisible in execution measures alone.
What This Means
This research suggests that the common scientific conclusion that older adults switch between mental tasks just as well as younger adults may be misleading. Most previous studies measured how quickly and accurately people responded after being shown a task cue, but these measures blend together two separate processes: getting ready for a task (preparation) and actually doing it (execution). When the researchers gave participants control over when they saw the task stimulus — allowing them to take as much preparation time as they wanted — a different picture emerged. Older adults took significantly longer to prepare specifically when they had to keep switching tasks continuously, even though their response times and accuracy during actual task execution looked similar to younger adults.
The study also found that the way switching difficulty shows up in behavior depends heavily on the design of the experiment. In a standard externally paced version, switching costs appeared as more errors; in the self-paced version, they appeared as slower responses. Global switching demands (always switching versus never switching) showed up in response measures in the standard version but disappeared from those measures and reappeared in preparation time in the self-paced version. This pattern suggests that mental effort related to task switching gets 'redistributed' across different stages of processing depending on how much control people have over their own timing.
These findings matter because they suggest that aging does affect the cognitive processes involved in switching between tasks, but these effects are hidden when researchers only look at how fast and accurately people respond. The age-related difficulty appears most strongly during the preparation phase, especially when people must sustain continuous task-switching demands. This has implications for how researchers design studies of cognitive aging and how we interpret claims about which mental abilities are or are not preserved as people get older.
Cha C, Hsieh S. (2026). Temporal redistribution of control reveals age-related differences in task switching at the level of preparation.. Psychological research. https://doi.org/10.1007/s00426-026-02374-7