The role of temporal delay and repeated prospective memory cue

advertisement
Psychological Research (2014) 78:584–596
DOI 10.1007/s00426-013-0510-z
ORIGINAL ARTICLE
The role of temporal delay and repeated prospective memory cue
exposure on the deactivation of completed intentions
Moritz Walser • Franziska Plessow
Thomas Goschke • Rico Fischer
•
Received: 28 February 2013 / Accepted: 24 July 2013 / Published online: 7 August 2013
Ó Springer-Verlag Berlin Heidelberg 2013
Abstract Previous studies have shown that completed
prospective memory (PM) intentions entail aftereffects in
terms of ongoing-task-performance decrements in trials
containing repeated PM cues which previously served as
PM cues triggering the intended action. Previous research
reported that PM aftereffects decrease over time, thus
revealing a specific time course of PM aftereffects. In the
present study, we tested two accounts for this pattern,
assuming either that the decline of aftereffects is related to
the temporal distance to PM task completion or may be a
result of the repeated exposure of repeated PM cues in the
ongoing task. In three experiments, we manipulated both
the temporal distance to PM task completion and the frequency of repeated PM cues and demonstrated that aftereffects of completed intentions declined with repeated
exposure of formerly relevant PM cues. In addition, effects
of repeated exposure were not only limited to the repetition
of specific PM-cue exemplars but also generalized to other
semantically related PM cues within the PM-cue category.
Together, findings show that decreased aftereffects of
completed intentions are not related to the temporal duration of the subsequent test block, but crucially depend on
the repeated exposure of the previously relevant PM cues.
Introduction
The ability to form, maintain, and retrieve an intended
action in a specific situation in the future, such as mailing a
letter when coming across a mailbox, is known as eventM. Walser (&) F. Plessow T. Goschke R. Fischer
Department of Psychology, Technische Universität Dresden,
01062 Dresden, Germany
e-mail: walser@psychologie.tu-dresden.de
123
based prospective memory (PM) and is an essential ability
for every day functioning. While many prospective memory studies investigated the involvement of maintenance
and monitoring processes in performing a delayed intended
action (e.g., Smith, 2003), previous studies also highlighted
the role of a specific PM cue (e.g., mailbox) as a trigger
signal for the retrieval of the intended action (Einstein et al.
2005; McDaniel & Einstein, 2000). The attractiveness of
such a view lies in the contextual trigger condition that
does not necessarily require an active and demanding
monitoring mechanism. The influential multiprocess
framework (McDaniel & Einstein, 2000), e.g., proposes
that PM retrieval is supported by specific features of the
PM cue. Salient and focal PM cues support a rather
spontaneous retrieval of the intended action and increase
the probability for successful PM performance. Non-salient
and non-focal cues, on the other hand, require rather
resource-demanding monitoring processes. Such a PM cuebased focus of successful delayed intention retrieval and
PM performance, however, requires a strong association
between the PM cue (e.g., mailbox) and the to-be-performed action (e.g., mail a letter) that is maintained during
the retention interval until the required PM-cue event
occurs and the intention will have been completed successfully (Einstein et al., 2005). Such a view is also in line
with findings showing that not only the content (i.e., action
plan) of the intention (Goschke & Kuhl, 1993) but also PM
cues are stored in an increased sub threshold activation in
long-term memory compared to other memory contents
(Marsh, Hicks, & Watson, 2002).
A strong reliance on environmental cues as trigger for
action implementation, however, raises the question of the
susceptibility to reoccurring PM cues after the intended
action has been completed. The successful intention
deactivation (e.g., deactivating the link between intention
Psychological Research (2014) 78:584–596
and retrieval cue) plays an important role for every day
functioning because a failure to deactivate completed
intentions would be dysfunctional as it might provoke
erroneous retrieval of the already completed intended
action (i.e., commission errors) or interfere with subsequent tasks.
Not surprisingly, researchers have started to investigate
aftereffects of completed intentions and the particular role
of repeated occurrence of PM cues when the link between
the PM cue and the intended action has become irrelevant
(e.g., Pink & Dodson, 2013; Scullin, Bugg, & McDaniel,
2012; Walser, Fischer, & Goschke, 2012; Walser, Goschke, & Fischer, 2013).
Studies on PM deactivation
Typical studies investigating the deactivation of completed
intentions use classical laboratory PM paradigms (e.g.,
Einstein & McDaniel, 1990) in which participants perform
a choice reaction time task, such as lexical-decision or
number categorization tasks (ongoing task), and an additional PM task. The PM task requires participants to suspend the ongoing task and to press a different key in
response to a PM cue such as a rarely appearing prespecified word or symbol (e.g., Scullin & Bugg, 2013;
Walser et al., 2012). At the end of the PM task participants
are instructed that the task has been completed and is of no
more relevance. Subsequently, no-more-relevant PM cues
from the previous PM task are embedded in an ongoing
task as PMREPEATED trials. Aftereffects of completed
intentions are measured as ongoing task performance differences on PMREPEATED trials compared to control trials
(i.e., oddball trials).
On the basis of such paradigms, many researchers found
persisting effects of completed intentions, such as
increased RTs and/or (commission) errors on PMREPEATED
trials suggesting that performing and completing the
intended action does not lead immediately to a complete
deactivation of the link between PM cue and intended
action (Beck, Ruge, Walser, & Goschke, 2013; Bugg,
Scullin, & McDaniel, 2013; Pink & Dodson, 2013; Scullin
et al., 2012; Scullin, Bugg, McDaniel, & Einstein, 2011;
Scullin & Bugg, 2013; Walser et al., 2012, 2013). The idea
here is that the intention representation (i.e., PM cue,
intended action and/or the link between the PM cue and the
intended action) remains in a heightened state of residual
activation (Walser et al., 2012) and therefore, still triggers
the old PM response on PMREPEATED trials resulting in
response conflict with the ongoing task, and thus in
increased response times (RTs) or error rates (PM aftereffects). Of note, given that intention representations
comprise several distinct components, including representations of the intended goal, the to-be-performed action,
585
and its execution conditions or trigger cues, persisting postenactment activation could in principle refer to each of
these components. While it is an interesting question,
which particular components of completed intentions may
persist in a state of increased activation, in the present
paper, we focus on a different unresolved question, namely
whether aftereffects of completed intentions gradually
decline with increasing temporal distance to the completion
of a PM task, or whether deactivation of completed
intentions depends on the repeated exposure to repeated
(no longer relevant) PM cues.
Because the repeated activation of a no longer relevant
action is often dysfunctional, the identification of conditions and factors that modulate or minimize PM aftereffects
is of major interest to researchers in the prospective
memory field. For example, several factors have been
recognized to result in increased aftereffects of completed
intentions: Strong PM-cue salience (Scullin et al., 2012),
similarity between the PM task and the subsequent task in
which aftereffects are measured (Scullin et al., 2012),
impaired cognitive control ability (Scullin et al., 2011,
2012) as well as personality factors such as a tendency
toward state orientation as compared to action orientation
(Walser et al., 2013). Furthermore, it has also been shown
that the extent of aftereffects of completed intentions can
be modulated by either reflecting upon no-more-relevant
PM cues after PM task completion (i.e., increased aftereffects) or by performing a cognitively demanding working
memory task (i.e., reduced aftereffects) (Walser et al.,
2013).
Time course of PM aftereffects
Regarding the modulation of PM aftereffects, inconsistent
results have been reported with respect to the time course
of PM aftereffects. For example an intention representation
may gradually decay as a function of delay after intention
completion due to the mere passage of time and/or interference from other memoranda (for this discussion on
short-term memory see e.g., Berman, Jonides, & Lewis,
2009; Campoy, 2012). The idea of delay-dependent
intention deactivation is supported by findings of similar
forgetting curves for other memory contents (for a similar
decay of PM performance over time see McBride, Beckner,
& Abney, 2011; for an overview of retrospective memory
forgetting see Rubin & Wenzel, 1996). Indeed, a gradual
decrease of PM aftereffects over time was recently reported
by Walser et al. (2012). Aftereffects were assessed within
blocks containing six PMREPEATED trials (compared to six
oddball trials). In four experiments, RT and/or error
aftereffects were increased during the early first three
PMREPEATED encounters as compared to the later last three
PMREPEATED encounters (for a similar finding see Beck
123
586
et al., 2013). Alternatively, and in contrast to a temporal
dependency, the reduction of intention aftereffects over
time, however, might also result from PM-cue effectiveness washing out with increasing number of cue encounters. Since specific PMREPEATED trials are linked during
intention formation and PM performance with the associated intended action, this specific PM cue-action link could
be weakened when the specific PMREPEATED cue is
repeatedly bound to a new action (i.e., the ongoing task).
Evidence against a temporal dependency of PM aftereffects was provided in a study by Scullin and Bugg (2013)
(see also Scullin et al., 2011). Scullin and Bugg (2013)
used a between-subjects design and measured commission
errors on a single PMREPEATED trial to investigate the effect
of delay interval length after intention completion on
intention aftereffects. Importantly, the probability of
commission errors did not differ between short (40 trials)
versus long delay intervals (258 trials).
Our goal in the present study was to investigate why results
from previous studies were apparently contradictory to each
other by differentiating between the two competing explanations of delay versus repeated exposure on the degradation
of aftereffects of completed intentions over time.
Experiment 1
We adapted the paradigm of Walser et al. (2012) (see
Fig. 1). During an initial PM block, participants performed
a digit parity-judgment task as ongoing task. In addition,
they had to respond on PM trials, which were two different
PM cue exemplars of the same category (e.g., PM A
cue = square, PM B cue = rhombus, PM-cue category = quadrangles), by pressing the X key. Following an
instruction that the PM task had been completed, aftereffects of completed intentions were measured in two
Fig. 1 Example trials of the
prospective memory (PM) block
and Test blocks 1 and 2 for
Experiment 1. Participants were
required to perform parity
judgments on all trials except
for PM A and PM B trials in
which they had to press the
X key. In Experiment 2, PM
AREPEATED trials and oddball A
trials did not occur in Test
blocks 1 and 2. Note, framing of
trial types were not present in
the experiment but serve
exclusively to illustrate different
trial types in this figure
123
Psychological Research (2014) 78:584–596
subsequent test blocks (Test block 1 and Test block 2) in
which participants had to perform the ongoing task during
all trials. Whereas in Test block 1 only one exemplar
served as repeated PM cue (PM AREPEATED), in Test block
2 both exemplars (PM AREPEATED and PM BREPEATED)
were included. To calculate comparable aftereffects triggered by PM AREPEATED and PM BREPEATED cues, the
oddball trials were categorically matched. That is, from an
unrelated category (e.g., punctuation marks) one arbitrary
exemplar served as oddball A (e.g., two question marks)
and another as oddball B (e.g., two exclamation marks),
respectively. This allowed us to calculate PM aftereffects
for exemplars that served in both Test blocks (A-items; i.e.,
PM AREPEATED vs. oddball A) and PM aftereffects for
exemplars that served exclusively in Test block 2 (B-items;
i.e., PM BREPEATED vs. oddball B).
Comparing PM aftereffects for A-items and B-items in
Test block 2 enabled us to differentiate between the two
hypotheses. If the temporal delay is of crucial importance for
the decline of PM aftereffects, PM aftereffects in Test block 2
should be similarly reduced for A-items and B-items, because
both item sets share the same temporal distance to PM task
completion. If, on the other hand, PM aftereffects decline as a
function of repeated exposure, PM aftereffects should be
smaller for A-items that also occurred in Test block 1 compared to B-items that were only introduced in Test block 2.
Method
Participants
Twenty-eight students of the Technische Universität
Dresden (3 male, age M = 24.46 years, SD = 6.35)
attended a single experimental session lasting for about
50 min. Participants had normal or corrected-to-normal
sight and received 5 € or course credit.
Psychological Research (2014) 78:584–596
Apparatus and stimuli
The digits 2–9 served as stimuli (Arial font, visual angle
2.2° with a viewing distance of approximately 60 cm) and
were presented in black against a gray background on a
19-inch monitor. One of two exemplar deviant stimuli that
were members of one out of ten categories (i.e., two
punctuation marks, circles, quadrangles, two letters, triangle, cylinder, mathematical symbols, one single currency
symbol, in parenthesis, with an arrow) could appear around
or next to the ongoing-task digits as PM A trials, PM B
trials, or control trials (i.e., oddball A trials, oddball B
trials) (visual angle 4.5°), whereby for each participant five
categories were randomly drawn to serve as PM A/PM B
trials and five categories to serve as oddball A/oddball B
trials. Further 15 different deviant stimuli (e.g., @-symbols, a star) that accompanied ongoing-task digits were
used as further control and/or irrelevant filler trials (i.e.,
oddball trials). Participants responded with the ‘‘,’’ key
(right index finger), the ‘‘.’’ key (right middle finger) and
the ‘‘X’’ key (left index finger) on a standard German
(QWERTZ) keyboard.
Procedure and design
The experiment started with a practice block to familiarize
the participants with the ongoing task, which required them
to categorize digits according to parity with the right index
finger for odd digits and the right middle finger for even
digits. The practice block contained 16 standard and 8
symbolic deviant trials. After the initial practice block
participants completed five cycles with each a PM block,
Test block 1 and Test block 2 (Fig. 1).
PM block At the beginning of the PM block participants
received the PM task instruction to respond to members of
a deviant PM-cue category (e.g., quadrangles) with the left
index finger instead of performing the ongoing task. Of
each PM-cue category two members were presented each
four times as PM cues (e.g., PM A cue = square, PM B
cue = rhombus). Additionally on four trials an irrelevant
oddball appeared. The PM block contained 48 trials. Each
trial started with a fixation cross (500 ms) followed by the
imperative stimulus which remained until a response was
given. If an incorrect or no response was given within
3,000 ms, a high-pitch tone (750 Hz) was given through
headphones as feedback for 200 ms. At the end of the PM
block an instruction was shown that the PM task was
completed.
Test block 1 Test block 1 (144 trials) started after a short
break (20 s) with an instruction to perform the parityjudgment task on all trials. Only one of the two PM cues of
587
the previous PM block served as PMREPEATED trial in Test
block 1 (PM AREPEATED). From an unrelated category (e.g.,
punctuation marks) a single item served as oddball (oddball
A, e.g., two question marks). PM AREPEATED trials and
oddball A trials occurred each six times. In addition, Test
block 1 contained two further unrelated oddball trial types
(each 6 trials) to ensure a similar amount of deviant trials in
Test blocks 1 and 2. Test block 1 served to measure
aftereffects on PM AREPEATED (compared to oddball A
trials).
Test block 2 Test block 2 (144 trials) started after a short
break (10 s). Importantly, this block contained the same
PM AREPEATED trials and oddball A trials as in Test block
1. In addition, it contained six PM BREPEATED trials (e.g.,
rhombus) and six oddball B trials (e.g., two exclamation
marks) which served to measure aftereffects of the second
category member (i.e., B-items).
Each digit (2–9) was presented in random order 6 times
in each PM block and 18 times in each Test block 1 and
Test block 2, respectively. PM trials, PMREPEATED trials
and oddball trials were randomly interspersed within
blocks, the only constraint being that they could not appear
during the first four trials of the PM block and Test blocks,
respectively. The only variation of the experimental task
over the five cycles was that different deviant stimulus
categories were used in each cycle. Specifically, during
each cycle one out of the ten deviant stimulus categories
served as PM A/PM B trials and another as oddball
A/oddball B trials.
Results
Error trials (6.2 %) and trials with RTs 2.5 standard deviations (SDs) above or below a participant’s mean RT for a
given trial type (PM block: 2.7 %; Test block 1: 2.7 %;
Test block 2: 2.4 %) were excluded prior to RT analyses.
PM block
Participants performed equally well on PM A trials and PM
B trials, as shown by similar RTs, t(27) = -0.84,
p = .411, d = -0.11; and error rates, t(27) = 0.63,
p = .537, d = 0.10. Overall, responses were 303 ms
slower, t(27) = 13.99, p \ .001, d = 2.50; and 18.1 %
more erroneous, t(27) = 7.32, p \ .001, d = 1.68, on
oddball trials compared to standard trials (Fig. 2; Table 1).
Test block 1
We conducted repeated-measures ANOVAs with the factor
trial type (standard, oddball A, PM AREPEATED) on RTs and
error rates of the ongoing task. For RT analysis, trial type
123
588
Psychological Research (2014) 78:584–596
Fig. 2 Results for Experiments
1 and 2. Mean response times
(RT) and error rates for the PM
block in PM A and PM B trials,
for Test block 1 (oddball A, PM
AREPEATED) and Test block 2
(oddball A, PM AREPEATED,
oddball B, PM BREPEATED) as a
function of trial type. Error bars
represent standard errors
Table 1 Mean RTs and error rates for the PM block, Test block 1
and Test block 2 by trial type in Experiments 1 and 2 (standard
deviations in parentheses)
Experiment 1
Experiment 2
RT (ms)
Error (%)
RT (ms)
Error (%)
Standard
583 (71)
6.2 (3.9)
593 (79)
5.5 (3.5)
Oddball
886 (156)
24.3 (14.7)
921 (172)
15.3 (13.6)
PM A
597 (66)
7.0 (7.9)
647 (91)
7.8 (6.6)
PM B
605 (83)
6.3 (6.6)
629 (87)
9.1 (10.2)
546 (75)
5.3 (3.6)
PM block
Test block 1
Standard
534 (57)
5.5 (3.6)
Oddball A
PM AREPEATED
589 (88)
617 (92)
5.2 (4.0)
6.4 (6.2)
Most important, we found aftereffects of completed
intentions (M = 28 ms), as shown by slower RTs on PM
AREPEATED trials than on oddball A trials, F(1,
27) = 15.41, p = .001, g2 = .36. To test for a decrease of
aftereffects as a function of cue repetitions within Test
block 1, we compared aftereffects between the first three
encounters and last three encounters of PM AREPEATED
trials (for a similar analysis see also Walser et al., 2012).
We found aftereffects for early encounters (M = 60 ms,
t[27] = 6.00, p \ .001, d = 0.58) but not for late
encounters
(M = 0 ms,
t[27] = -0.03,
p = .973,
d = 0.00), as indicated by a Trial type 9 Block position
interaction, F(1, 27) = 15.59, p = .001, g2 = .37.
Test block 2
Test block 2
Standard
525 (56)
6.1 (3.9)
Oddball A
547 (72)
6.7 (5.9)
548 (71)
5.9 (3.4)
PM AREPEATED
549 (60)
6.1 (5.2)
Oddball B
564 (81)
PM BREPEATED
567 (85)
7.0 (5.9)
602 (90)
6.0 (6.2)
6.6 (5.9)
646 (120)
7.7 (6.2)
reached significance, F(2, 54) = 50.87, p \ .001,
g2 = .65. Repeated contrasts revealed an orientation
response (55 ms) on oddball A trials in terms of slower
RTs than on standard trials, F(1, 27) = 38.00, p \ .001,
g2 = .56. At the same time, it should be noted that RTs on
the two further oddball trial types, which served exclusively as filler trials to ensure the same ratio of standard
and deviant stimuli during Test block 1 and Test block 2,
did not differ from RTs on oddball A trials, Fs \ 1
(planned contrasts).
123
We compared aftereffects of completed intentions in Test
block 2 by computing 2 (trial type: PMREPEATED, oddball) 9 2 (item set: A-items, B-items) repeated-measures
ANOVAs on RTs and error rates. For RTs the factor item set
reached significance, F(1, 27) = 7.38, p = .011, g2 = .21,
indicating smaller overall RTs for A-items (M = 548 ms),
that also appeared in Test block 1, compared to B-items
(M = 565 ms) that were only presented in Test block 2. The
factor trial type and the Trial type 9 Item set interaction did
not reach significance, Fs \ 1, indicating no aftereffects and
hence complete intention deactivation for both, A-items
(M = 2 ms) and B-items (M = 3 ms), respectively.
Further analyses
To disregard explanations on the basis of the use of
repeated cycles of PM and Test blocks in the present
Psychological Research (2014) 78:584–596
design,1 in a subsequent step, we repeated the ANOVAs using
only the very first cycle of Test block 1 and Test block 2. Most
importantly and in line with the analysis including all cycles,
we found aftereffects for A-items in Test block 1 (M = 42 ms;
PM AREPEATED trials: M = 651 ms, SD = 122 ms; oddball A
trials: M = 609 ms, SD = 105 ms), F(1, 27) = 7.17,
p = .012, g2 = .21. In Test block 2, no PM aftereffects were
found for A-items (M = 0 ms; PM AREPEATED trials:
M = 553 ms, SD = 91 ms; oddball A trials: M = 553 ms,
SD = 91 ms) nor for B-items (M = -6 ms; PM BREPEATED
trials: M = 592 ms, SD = 130 ms; oddball B trials: M =
598 ms, SD = 159 ms), Fs \ 1. Further analyses revealed
that for A-items during Test blocks 1 and 2 and for B-items
during Test block 2 aftereffects did not vary over the course of
the experimental session, as no Trial type 9 Repeated cycles
interactions were found, Fs \ 1. An additional RT analysis on
PM trials (PM A and PM B together) as a function of repeated
cycle was significant, F(4, 108) = 8.65, p \ .001, g2 = .24.
Repeated contrast showed that RTs were increased in the first
cycle (M = 653 ms, SD = 93 ms) compared to second cycle
(M = 594 ms, SD = 75 ms), F(1, 27) = 14.90, p \ .001,
g2 = .36; but did not differ between subsequent cycles,
Fs \ 1. In sum, these analyses revealed that repeating cycles
led to faster responses on PM trials, whereas they did not affect
aftereffects of completed intentions.
To disregard that missing aftereffects in Test block 2
were due to a complete shielding of deviant stimulus
information, we computed repeated-measures ANOVAs
with the factor trial type (standard, oddball A, oddball B)
on RTs and error rates, which reached significance,
F(2, 54) = 12.42, p \ .001, g2 = .32. Planned contrasts
revealed shorter RTs on standard trials than on oddball A
trials, F(1, 27) = 12.54, p = .001, g2 = .32, indicating an
orientation reaction on deviant stimuli and thus that participants could not ignore deviant stimuli completely during Test block 2. There was a tendency—albeit not
significant—toward increased RTs on oddball B compared
to oddball A trials, F(1, 27) = 3.25, p = .083, g2 = .10.
Comparison of PM aftereffects in Test block 1 and Test
block 2 (A-items)
We conducted 2 (trial type: PM AREPEATED, oddball
A) 9 2 (Test block: 1, 2) repeated-measures ANOVAs on
589
RTs and error rates to analyze the fade of aftereffects for
A-items between Test blocks 1 and 2. Overall RTs on PM
AREPEATED trials (M = 583 ms) were increased compared
to oddball A trials (M = 568 ms), F(1, 27) = 12.80,
p = .001, g2 = .32. Overall RTs were slower in Test block
1 (M = 603 ms) than in Test block 2 (M = 548 ms), F(1,
27) = 33.31, p \ .001, g2 = .55. Most important, the Trial
type 9 Test block interaction was significant, F(1, 27) =
5.32, p = .029, g2 = .17, indicating that PM aftereffects
for A-items decreased from Test block 1 to Test block 2.
As hardly any commission errors (0.02 %) were made
we only computed an overall error analysis. For all corresponding analyses on error rates of Test blocks 1 and 2,
however, no significant effects were found, all ps [ .268.
Discussion
In Experiment 1, category members of PM cues that were
presented as PMREPEATED cues in Test block 1 and Test
block 2 (A-items) revealed PM aftereffects only in Test
block 1 but not in the subsequent Test block 2. Category
members of PM cues that were presented as PMREPEATED
cues exclusively in Test block 2 (B-items) did also not
show PM aftereffects in Test block 2. Although these
results seem to suggest that sufficient temporal delay
between intention completion and measurement of PM
aftereffects determines the decrease of PM aftereffects,
such an interpretation has to be handled with care, because
two alternative interpretations need to be considered:
First, whereas PM BREPEATED served as PM B cues in
the PM block, oddball B cues in the Test block 2 were
never presented before. Therefore, one could argue that the
novelty of a first-time presentation of oddball B trials in
Test block 2 resulted in a stronger orientation response
(i.e., slowed RTs to oddball B trials) and thus, eliminated
the PM aftereffect for B-items. To disregard this possibility, we conducted a control replication experiment in which
oddball A trials and oddball B trials were presented already
in the PM block. Results were virtually identical to
Experiment 1. Most importantly, in Test block 2 no PM
aftereffects were found for B-items.2
Second, effects of repeated exposure of PM AREPEATED
trials in Test block 1 may have transferred to PM
2
1
The repeated cycles of PM and Test blocks differ to other
approaches such as single-cycle paradigms (e.g., Scullin & Bugg,
2013). One could argue, e.g., that using repeated cycles, PM
aftereffects may be overestimated when participants are ambiguous
whether the intention has really finished and persist monitoring
(Walser et al., 2012). In contrast, aftereffects may also be underestimated, because over the course of the experimental session,
participants might learn to increasingly shield the ongoing task from
deviant PMREPEATED and oddball stimuli during the Test blocks.
Sixteen participants participated in the experiment (1 male, age
M = 22.56 years, SD = 4.08). In Test block 1, RTs on PM
AREPEATED trials (M = 609 ms, SD = 121 ms) were slower than
on oddball A trials (M = 550 ms, SD = 89 ms), F(1, 15) = 28.65,
p \ .001, g2 = .66, indicating aftereffects of completed intentions
(M = 59 ms). In Test block 2, no aftereffects were found for A-items
(6 ms, PM AREPEATED trials: M = 553 ms, SD = 88 ms; oddball A:
M = 547 ms, SD = 85 ms) nor for B-items (10 ms, PM BREPEATED
trials: M = 548 ms, SD = 79 ms; oddball B: M = 538 ms,
SD = 80 ms). The factor trial type was not significant. Also, Trial
type 9 Item set did not interact, both Fs \ 1.
123
590
BREPEATED trials in Test block 2. That is, because PM A
cues and PM B cues were exemplar items of the same
semantic category, it is conceivable that diminished PM
aftereffects for B-items in Test block 2 are a consequence
of generalized transfer effects from semantically related
A-items. More specifically, during the repeated exposure to
PM AREPEATED trials participants might have deactivated
not only the specific S-R link between PM A cues of the
categorical intention and the associated intended action.
Instead, participants might have formed a more abstract
semantic intention representation during the categorical
PM task instruction (see also Walser et al., 2012). Consequently, this might have enabled a transfer of repeated
exposure of PM AREPEATED trials to PM BREPEATED trials
resulting even in complete intention deactivation of PM
BREPEATED trials in Test block 2.
Experiment 2
We conducted Experiment 2 to test for the assumption of a
within-category transfer effect. To rule out this alternative
explanation, we adapted Experiment 1 by omitting all
A-items from Test block 1 and Test block 2. Therefore, Test
block 1 did not contain any PMREPEATED trials, eliminating
the possibility of repeated exposure and of within-category
transfer. Only in Test block 2, B-items (PM BREPEATED and
oddball B) were included. If intention deactivation in
Experiment 1 would have been due to a delay effect, no
aftereffects on PM BREPEATED trials should occur in
Experiment 2. If intention deactivation in Experiment 1
would have been due to a transfer from PM AREPEATED trial
repetitions to PM BREPEATED trials, in Experiment 2 aftereffects on PM BREPEATED trials should be observed.
Method
Participants
Sixteen new students of the Technische Universität Dresden (2 male; age M = 20.63 years, SD = 3.01) participated in Experiment 2.
Apparatus and stimuli
In Experiment 2, we used the same apparatus and stimuli as
in Experiment 1.
Psychological Research (2014) 78:584–596
trials were shown. Consequently, Test block 1 contained 132
standard trials and two different specific oddball trial types,
each during six trials. Test block 2 contained 132 standard
trials, 6 PM BREPEATED trials and 6 oddball B trials.
Results
Error trials (5.8 %) and trials with RTs 2.5 SDs above or
below a participant’s mean RT for a given trial type (PM
block: 2.8 %; Test block 2: 2.7 %) were excluded prior to
RT analyses.
PM block
RTs and error rates on PM A and PM B trials did not differ,
t(15) = 1.45, p = .166, d = 0.20; and t(15) = -0.84,
p = .411, d = -0.15, respectively. Participants responded
slower and made more errors on oddball trials than on
standard trials, t(15) = 11.10, p \ .001, d = 2.45, and
t(27) = 3.02, p = .009, d = 0.99, respectively (Fig. 2;
Table 1).
Test block 2
We conducted repeated-measures ANOVAs with the factor
trial type (standard, oddball B, PM BREPEATED) on RTs and
error rates. The ANOVA on RTs was significant, F(2,
30) = 21.46, p \ .001, g2 = .59. RTs on oddball B trials
were slower than on standard trials, denoting an orientation
response (M = 54 ms), F(1, 15) = 32.80, p \ .001,
g2 = .69 (repeated contrast). Most important, we found
aftereffects of completed intentions (M = 44 ms) in terms
of increased RTs on PM BREPEATED trials as compared to
oddball B trials, F(1, 15) = 6.01, p = .026, g2 = .29. In
the corresponding analysis on error data, trial type did not
reach significance, F \ 1.
Similar to Experiment 1, we again compared RT aftereffects between the first three encounters and the last three
encounters of PM BREPEATED trials. Presumably due to the
small sample size and thus lack of statistical power, the
Trial type 9 Block position interaction slightly missed
significance, F(1, 15) = 3.68, p = .074, g2 = .20. On a
descriptive level, however, aftereffects were larger for
early encounters (M = 95 ms, t[15] = 2.50, p = .025,
d = 0.64) than those of late encounters (M = 28 ms,
t[15] = 1.90, p = .076, d = 0.28).
Further analyses
Procedure and design
The procedure of Experiment 2 was similar to the one of
Experiment 1 except the following changes. During Test
block 1 and Test block 2 no PM AREPEATED and oddball A
123
In addition, we re-analyzed aftereffects of Experiment 2 by
exclusively using data from the first cycle of PM block,
Test block 1 and Test block 2, to rule out the possibility
that the repeated cycle of PM block and Test blocks
Psychological Research (2014) 78:584–596
affected intention aftereffects. Importantly, even for the
very first cycle, we found significant PM aftereffects in
Test block 2 (M = 52 ms; PM BREPEATED trials: M =
668 ms, SD = 132 ms; oddball B trials: M = 616 ms,
SD = 105 ms), F(1, 15) = 7.09, p = .013, g2 = .34. In
addition, aftereffects did not vary over the course of the
experimental session, as no Trial type 9 Repeated cycles
interaction was found in a subsequent analysis, F \ 1.
Between-experiment comparison
We conducted a 2 9 2 mixed ANOVA with trial type
(PMREPEATED, oddball) as within-subjects factor and
experiment (Experiment 1: PM AREPEATED, oddball A in
Test block 1; Experiment 2: PM BREPEATED, oddball B in
Test block 2) as between-subjects factor to test whether
aftereffects in Experiment 1 (M = 26 ms) and Experiment
2 (M = 44 ms) varied as a function of delay after intention
completion when ruling out the influence of repeated
exposure. The factor experiment, F \ 1; and the Experiment 9 Trial type interaction, F(1, 39) = 1.10, p = .299,
g2 = .03, were not significant, indicating no differences in
aftereffects after a short (Experiment 1) and long delay
(Experiment 2).
To test whether Item-B aftereffects in Test block 2
varied as a function of repeated Item-A exposure, we
computed a between-experiment comparison of Item-B
aftereffects in Test block 2. Importantly, the mixed
ANOVA with trial type (PM BREPEATED, oddball B) as
within-subjects factor and experiment (Experiment 1,
Experiment 2) as between-subjects factor revealed smaller
Item-B aftereffects in Experiment 1 (M = 2 ms) compared
to Experiment 2 (M = 44 ms), as indicated by a Trial
type 9 Experiment interaction, F(1, 42) = 5.66, p = .022,
g2 = .12.
Discussion
Surprisingly and in contrast to Experiment 1, we observed
aftereffects of completed intentions for PM BREPEATED
trials during Test block 2. These aftereffects did not differ
from those of PM AREPEATED trials during Test block 1 in
Experiment 1 with a much shorter temporal delay to
intention completion. Further, PM BREPEATED aftereffects
were substantially increased compared to those of Experiment 1, when ruling out the role of delay. Consequently,
the disappearance of aftereffects in Test block 2 found in
Experiment 1 cannot be accounted for by a delay effect.
Findings from Experiment 2 rather indicate that effects of
repeated exposure (i.e., response reconfiguration) affect not
only specific members of an intention but that response
reconfiguration may generalize from one PM cue to other
PM cues of the same semantic category (see ‘‘General
591
discussion’’ for further implications and alternative explanations of this finding).
Experiment 3
Experiment 3 served to provide further and more direct
evidence for the assumption that decreased aftereffects of
completed intentions are specifically related to repeated
exposure to PMREPEATED items and less so to temporal distance to intention completion. For this, we directly manipulated repeated exposure and temporal delay in a single
experiment using only PM-cue exemplars. In particular, for
this, participants had to perform the PM task in response to
specific PM cues (e.g., a square) instead of different members
of a PM-cue category as in Experiments 1 and 2. We used a
single Test block to measure aftereffects. We tested the
influence of delay versus repeated cue exposure on aftereffects using a 2 (block length: short, long) 9 2 (frequency: 4
PMREPEATED trials, 12 PMREPEATED trials) within-subjects
design. If aftereffects fade as a function of delay after PM
task completion, they should be smaller in the long- than the
short-block condition. If in contrast aftereffects fade as a
function of PMREPEATED trial repetitions, they should be
smaller in the 12- than the 4-PMREPEATED trials condition.
Method
Participants
Twenty-four new students of the Technische Universität
Dresden (8 male; age M = 23.79 years, SD = 1.05) participated for 12 € or course credit in two experimental
sessions lasting about 1 h each.
Apparatus and stimuli
In Experiment 3, we used the same apparatus as in
Experiment 1. However, instead of categorical PM cues,
exemplar PM cues were used. That is, 36 different deviants
(e.g., square, circle, @-symbols, stars, triangles) served as
PM cues, PMREPEATED cues and oddballs.
Procedure and design
Participants attended two experimental sessions each
comprising 12 cycles with each a PM block and a single
Test block. In contrast to Experiments 1 and 2 participants
received at the beginning of each cycle a PM instruction to
press the X key instead of performing the ongoing parityjudgment task in response to a specific PM cue (e.g., digits
surrounded by a square). The PM block (48 trials) contained
4 PM trials and 4 oddball trials. We used a 2 9 2 design to
manipulate block length (short: 48 trials, long: 144 trials)
123
592
and PMREPEATED trial frequency (4 trials, 12 trials) in the
subsequent Test block. To separate the orientation response
from intention aftereffects and thus realize a comparable
baseline condition a similar number of oddball trials (i.e., 4
trials vs. 12 trials) were presented in the Test block. During
each of the 12 cycles, one of the 36 deviant stimuli was
assigned to serve as PM/PMREPEATED trial, one as oddball
trial during the PM block and one as oddball trial during the
Test block. Two experimental sessions were conducted to
increase statistical power. Both sessions similarly contained
each of the four Block length 9 PMREPEATED trial frequency conditions three times. The only difference between
sessions was the order of conditions for each participant.
Two to five days passed between the sessions. Similar to
Experiments 1 and 2, each experimental session started with
a practice block, in which participants were made familiar
with the parity-judgment task.
Results
Error trials (4.9 %) and trials with RTs 2.5 SDs above or
below a participant’s mean RT for a given trial type (PM
block: 2.6 %; Test block: 2.7 %) were excluded prior to
RT analyses.
PM block
Performance on PM trials was comparable to previous
experiments. RTs on oddball trials were 233 ms slower
Fig. 3 Mean response time
(RT) and percent error as a
function of trial type
[prospective memory (PM),
standard] in the PM block and
as a function of PMREPEATED
trial frequency (4 trials, 12
trials), block length (short: 48
trials, long: 144 trials) and trial
type (standard, oddball,
PMREAPEATED) in the Test block
of Experiment 3. Error bars
represent standard errors
123
Psychological Research (2014) 78:584–596
than on standard trials, indicating an orientation response,
t(23) = 13.55, p \ .001, d = 1.92. Participants made
6.1 % more errors on oddball trials than on standard trials,
t(23) = 7.29, p \ .001, d = 1.17 (Fig. 3; Table 2).
Test block
We conducted repeated-measures ANOVAs with the factors trial type (PMREPEATED, oddball), block length (short,
long) and frequency (4 trials, 12 trials) on RTs and error
data of the ongoing task. The ANOVA on RTs revealed a
main effect of trial type, F(1, 23) = 43.58, p \ .001,
g2 = .66, indicating an overall aftereffect of completed
intentions (M = 27 ms). RTs were faster on short blocks
(M = 581 ms) than on long blocks (M = 602 ms), F(1,
23) = 9.28, p = .006, g2 = .29. Responses were slower
in the low frequency (M = 615 ms) compared to the highfrequency condition (M = 569 ms), F(1, 23) = 43.58,
p \ .001, g2 = .66. Most important, aftereffects did not
vary as a function of block length, as Trial type 9 Block
length did not interact, F(1, 23) = 1.20, p = .285, g2 =
.05. In contrast, aftereffects were substantially reduced in
the 12 PMREPEATED trials condition (M = 12 ms, t[23] =
2.66, p = .014, d = 0.11) compared to the 4 PMREPEATED
trials condition (M = 40 ms, t[23] = 4.89, p \ .001, d =
0.33), as indicated by a significant Trial type 9 Frequency
interaction, F(1, 23) = 18.04, p \ .001, g2 = .44. There
was no Trial type 9 Block length 9 Frequency interaction, F \ 1.
Psychological Research (2014) 78:584–596
593
Table 2 Mean RTs and error rates for the PM block by trial type, and
for the Test block by block length (short: 48 trials, long: 144 trials),
PMREPEATED trial number and trial type in Experiment 3 (standard
deviations in parentheses)
RT (ms)
Error (%)
PM block
Standard
541 (88)
4.9 (3.5)
Oddball
774 (147)
11.0 (6.5)
PM
617 (82)
10.3 (7.2)
Standard
513 (85)
4.1 (3.4)
Oddball
575 (110)
5.4 (7.0)
PMREPEATED
610 (139)
4.3 (6.1)
Standard
522 (88)
4.5 (4.1)
Oddball
566 (109)
6.1 (6.1)
575 (113)
6.6 (8.0)
517 (86)
4.9 (4.3)
Test block
Short blocks
4 PMREPEATED trials
12 PMREPEATED trials
PMREPEATED
Long blocks
4 PMREPEATED trials
Standard
Oddball
614 (113)
5.6 (6.2)
PMREPEATED
662 (153)
6.6 (8.0)
Standard
519 (89)
4.6 (3.4)
Oddball
559 (93)
5.8 (4.9)
PMREPEATED
576 (112)
6.1 (6.1)
12 PMREPEATED trials
Further analyses
Similar to Experiments 1 and 2, we re-analyzed aftereffects
of Experiment 3 only including the first cycle of PM block
and Test block. Visual inspection suggests that the main
findings can also be obtained in the first cycle. That is,
aftereffects were not affected by block length, F \ 1. In
contrast, aftereffects were at least numerically smaller in
the 12 PMREPEATED trials condition (M = 27 ms, t[11] =
1.42, p = .182, d = 0.18) than in the 4 PMREPEATED trials
condition (M = 88 ms, t[11] = 2.09, p = .060, d = 0.45).
Due to a lack of power, however, this difference failed
to reach statistical significance, F(1, 20) = 1.61, p = .219,
g2 = .07.3
Subsequently, it was tested whether aftereffects in the 4
PMREPEATED trials condition and 12 PMREPEATED trials
condition were comparable when the analysis was restricted to the first four PMREPEATED encounters. Importantly,
the Trial type 9 Frequency interaction was no more significant, F(1, 23) = 3.12, p = .090, g2 = .12, suggesting
3
We thank Julie Bugg for suggesting this analysis.
that aftereffects did not (at least statistically) differ anymore between the 12 PMREPEATED trials condition (M =
27 ms) and 4 PMREPEATED trials condition (M = 40 ms),
thereby fostering the role of repeated exposure on
aftereffects.4
To further test if the deactivation depends on repeated
exposure to PMREPEATED trials, we re-analyzed aftereffects
for early and late PMREPEATED trials (see also ‘‘Experiment
1’’ and ‘‘Experiment 2’’). Irrespective of block length,
aftereffects decreased from the early six encounters
(M = 21 ms, t[23] = 2.99, p = .007, d = 0.18) to the late
six encounters (M = 4 ms, t[23] = 1.17, p = .254,
d = 0.04) in the 12 PMREPEATED trials condition, F(1,
23) = 7.42, p = .012, g2 = .24. Similarly, for the four
PMREPEATED trials condition aftereffects decreased from
the early two encounters (M = 62 ms, t[23] = 5.21,
p \ .001, d = 0.42) to the late two encounters (M =
21 ms, t[23] = 2.42, p = .024, d = 0.19), as indicated by
a Trial type 9 Block position interaction, F(1, 23) =
12.09, p = .002, g2 = .35.
An additional ANOVA on standard trials only as a
function of block length and frequency revealed that RTs
did not vary as a function of block length, F \ 1. RTs on
standard trials were slightly increased in the high-frequency condition (M = 520 ms) compared to the lowfrequency condition (M = 515 ms), F(1, 23) = 5.76,
p = .025, g2 = .20.5
We only computed an overall error analysis, because
participants hardly made any commission errors (0.02 %).
For all corresponding analyses on error rates of the Test
block, no significant effects were found, all ps [ .092.
4
Although not significant, the at least numerically somewhat larger
aftereffects in the 4 compared to the 12 PMREPEATED trials condition
may be due to the task structure, as we cannot control for a potential
imbalance of encountered oddball trials prior to the first 4
PMREPEATED trials in both conditions. It is conceivable, e.g., that in
the 12 PMREPEATED trials condition participants have an increased
experience of task-irrelevant deviants prior to the first 4 PMREPEATED
trials, which should decrease aftereffects and RTs in trials including
deviants in general (oddball trials and PMREPEATED trials alike). In
fact, this is supported by a main effect of frequency, with faster
responses to deviant trials in the high-frequency (12 repeat) condition
compared to the low-frequency (4 repeat) condition.
5
This might have been caused by the increased number of postdeviant standard trials (i.e., 24) following the 24 deviant stimuli (12
PMREPEATED, 12 oddballs) in the high-frequency condition compared
to the 8 post-deviant standard trials following the 8 deviant stimuli (4
PMREPEATED, 4 oddballs). Switching attention back from deviant
stimuli to subsequent standard trials might have caused reorientation
costs (see also Meier & Rey-Mermet, 2012). Re-analyzing standard
trial RTs while excluding post-deviant standard trials strongly
diminished the RT difference between low- and high-frequent
conditions to a non-significant level, F(1, 23) = 2.71, p = .113,
g2 = .10.
123
594
Discussion
In Experiment 3, we measured aftereffects of completed
intentions in a single Test block after intention completion,
varying orthogonally the block length and the frequency of
PMREPEATED trials. Aftereffects did not vary as a function
of block length, arguing against the hypothesis that aftereffects might decrease as a function of delay after intention
completion. In contrast, aftereffects were reduced in conditions with 12 PMREPEATED trials compared to conditions
with 4 PMREPEATED trials supporting the assumption that
aftereffects might decrease as a function of repeated
exposure of the no-more-relevant PM cue during the
ongoing task.
General discussion
Our aim in the present study was to shed light on the
controversy of whether aftereffects of completed intentions
fade as a function of delay after intention completion or as
a function of repeated exposure of PMREPEATED trials. In
Experiment 1, aftereffects for both PM-cue category
members vanished in Test block 2 independently of whether they were repeatedly exposed (i.e., A-items) or never
shown (i.e., B-items) during Test block 1. At first sight, this
finding was consistent with the assumption that aftereffects
fade as a function of delay after intention completion.
However, in Experiment 2, in which we prevented repeated
exposure by omitting all PMREPEATED trials from Test
block 1, aftereffects re-occurred during Test block 2. This
finding indicates that aftereffects did not cease as a function of temporal distance to intention completion. Instead,
in Experiment 1 response reconfiguration processes due to
repeated exposure of A-items during Test block 1 transferred from one PM-cue category member to another. We
found conforming evidence for this assumption in Experiment 3, in which we orthogonally tested delay and repeated exposure in a single experiment. Importantly, delay in
terms of block length did not affect the size of aftereffects.
Instead, aftereffects of completed intentions were strongly
affected by repeated exposure as indicated by reduced
aftereffects in high-frequency conditions in which PMREPEATED trials were shown 12 times as compared to lowfrequency conditions in which PMREPEATED trials were
shown only four times.
Interpretations of the repeated exposure effect
Our findings enable explaining contradictory results of
previous studies. First and consistent with studies, in
which a between-subjects manipulation of delay interval
length between PM task completion and measurement of
123
Psychological Research (2014) 78:584–596
aftereffects was used (Scullin et al., 2011; Scullin & Bugg,
2013), we did not find evidence for a delay effect, neither
in a between-experiment comparison between Experiments 1 and 2, nor by manipulating block length in
Experiment 3. Secondly, we found decreasing aftereffects
that were associated with the repeated exposure of
PMREPEATED trials in both, Experiment 1 (A-items), and in
Experiment 3 with smaller aftereffects in the high frequency compared to the low-frequency condition, irrespective of the temporal distance to intention completion.
We assume that the first PMREPEATED encounters triggered
retrieval of the associated intended action, resulting in a
response conflict and thus increased ongoing task RTs
and/or commission errors. Over the course of PMREPEATED
trial repetitions the link between the PM cue and the nomore-relevant PM task (e.g., pressing the X key) was
destabilized resulting in decreasing reactivation of the old
PM response during PMREPEATED encounters. It is even
conceivable that during this response reconfiguration,
participants formed a new link between PMREPETEAD trials
and performing an ongoing task response. This interpretation of a response reconfiguration effect is consistent
with studies showing that new stimulus–response links can
be acquired within only a few stimulus repetitions (De
Baene, Kühn, & Brass, 2012; Ruge & Wolfensteller,
2010).
Interpretations of the transfer effect of repeated
exposure
The finding of a transfer effect of repeated exposure from
PM AREPEATED trials to PM BREPEATED trials is very
informative. First, it is in line with the findings from a
previous experiment (Walser et al., 2012, Experiment 4), in
which participants received a categorical PM instruction
and performed the PM task exclusively on one of two PMcue category members. Interestingly, aftereffects of completed intentions were also found for another PM-cue category member that never served as PM cue during the PM
block. Findings from Walser et al. (2012) and the transfer
effect observed in the present study indicate that PM
intentions might not only be stored as specific links
between the PM cue and its associated intended action.
Rather, PM intentions might be stored on a more abstract,
semantic level in episodic memory (Goschke & Kuhl,
1993) and/or specific links might generalize and transfer to
other related items.
The assumption of an abstract semantic intention representation is, however, not mandatory for explaining
transfer effects in Experiment 1. That is, it is also conceivable that participants deactivated their left-hand
responses during Test block 1, thereby leading not only to
no more aftereffects on PM AREPEATED trials, but also on
Psychological Research (2014) 78:584–596
PM BREPEATED trials in Test block 2.6 At the same time it
should be noted, though, that other studies reported aftereffects of completed intentions when a single hand was
used to complete both, the ongoing task and the PM
response, respectively (Scullin et al., 2012; Scullin &
Bugg, 2013). Still, this alternative explanation highlights
that it remains a theoretically extremely important question
what exactly is deactivated (e.g., the abstract semantic
intention representation; a specific stimulus–response link
between PM cue and motor action on a procedural level)
after intention completion, how intentions are represented
in memory, and which aspects of intentions are responsible
for aftereffects of completed intentions.
In addition, one might assume that the transfer effect of
repeated exposure might alternatively be explainable by
retrieval-induced forgetting,7 suggesting that the retrieval
of a practice-item impairs the activation of related items
(Anderson, Bjork, & Bjork, 1994). Although we cannot
entirely exclude this possibility, we render this explanation
as rather unlikely. That is, in previous work applying virtually the same experimental design, PMREPEATED trials,
which were members of a PM-cue category and never
presented during the PM block and thus ‘‘unpracticed’’,
yielded reliable aftereffects (Walser et al., 2012, Experiment 4).
Furthermore, given the possibility of a transfer of the
repeated exposure effect, it might play an important role in
interpreting findings of a seemingly delay effect in previous studies on aftereffects of completed intentions. For
instance, Förster, Liberman, and Higgins (2005) measured
aftereffects of a PM cue (e.g., the symbol of glasses) using
semantically related words (e.g., professor, read, sun). In a
first test block after intention completion, aftereffects were
found in terms of increased lexical-decision RTs on words
related to glasses as compared to control words (and
interpreted as an inhibition effect). In a subsequent second
test block, lexical-decision RTs were similar on intentionrelated words and control words. The authors interpreted
their finding as intention aftereffects disappearing as a
function of delay after intention completion. Against the
backdrop that repeated exposure might transfer to semantically related aspects of a completed intention, these
results have to be interpreted with caution and a transfer
effect of repeated exposure has to be taken into consideration as an alternative explanation. More specifically, the
construct glasses and its semantically related items might
have been bound to the ongoing lexical-decision task with
repeated exposure, leading to a disappearance of aftereffects. In the present study, we did not aim to directly test
this assumption. Nevertheless, our finding of transfer
6
7
We thank Michael Ziessler for raising this point.
We thank Julie Bugg for this suggestion.
595
effects clearly calls for future research on the implications
for previous findings and on its generalizability. For
instance, it remains an empirical question whether
observed transfer effects from encoding at a category level
(Experiment 2) to related PM exemplar cues are restricted
to the specific encoding at the category level or whether a
similar transfer would also be observable from encoding at
an exemplar level to other related PM exemplar cues.
The role of delay and interference on intention
deactivation
The present and previous experiments on deactivation of
completed intentions did not find unambiguous proof for
delay-dependent deactivation processes (Beck et al., 2013;
Förster et al., 2005; Scullin et al., 2011; Scullin & Bugg,
2013; Walser et al., 2012). Despite the clear evidence for
repeated exposure on intention deactivation, we do not
deny that decay might play a role for deactivation of
completed intentions. Given the relatively short time
intervals investigated so far, future research might investigate the role of delay by systematically varying the timeinterval length between intention completion and measurement of intention aftereffects.
At the same time, determining the specific time course
of PM aftereffect decline is not trivial, as other mechanisms have to be considered as well. For example, it is
conceivable that some aspects of the intention representation (i.e., the readiness of the PM cue, intended action, and/
or PM cue–intended action link) might lose their strength
as a mere function of delay after intention completion.
Furthermore, interference from other memoranda might
play a crucial role for intention deactivation to work. In
line with the latter idea, it might be crucial for intention
deactivation what a person does after intention completion.
Consistent with this idea Walser et al. (2013) recently
showed smaller intention aftereffects when participants
were required to perform a resource demanding working
memory task between intention completion and measurement of aftereffects. In addition, aftereffects could even be
increased when participants reflected upon the no-morerelevant PM cues. A similar assumption has been discussed
for forgetting in other memory fields such as short-term
memory (e.g., Berman et al., 2009; Campoy, 2012;
McKeown & Mercer, 2012). Consequently, future research
might more systematically investigate the role of interference and delay on intention deactivation.
Conclusions
The present study integrates heterogeneous findings from
previous studies that were apparently opposing each other.
123
596
That is, decreasing aftereffects in studies using multiple
PMREPEATED trials (Beck et al., 2013; Walser et al., 2012,
2013) might be explained with repeated exposure and are
thus consistent with studies using a between-subject comparison and/or a single PMREPEATED trial that did not find
differences in aftereffects (Scullin et al., 2011; Scullin &
Bugg, 2013). Further, the present findings indicate that
repeated exposure might also transfer between semantically related (but not identical) PMREPEATED trials (Förster
et al., 2005). The present study suggest that previous
findings of decreasing aftereffects in studies using multiple
PMREPEATED trials cannot be explained by a delay-dependent decay of intention aftereffects.
Acknowledgments We are grateful to Julia Kleindienst, Laura
Pepernick and Sarah Richter for assistance in data collection. We
thank Julie Bugg and Michael Ziessler for thoughtful comments on an
earlier version of this manuscript. This research was partly supported
by the German Research Foundation (DFG) (SFB 940/1-2013).
References
Anderson, M. C., Bjork, R. A., & Bjork, E. L. (1994). Remembering
can cause forgetting: Retrieval dynamics in long-term memory.
Journal of Experimental Psychology: Learning, Memory, and
Cognition, 20(5), 1063–1087. doi:10.1037/0278-7393.20.5.1063
Beck, S. M., Ruge, H., Walser, M., & Goschke, T. (2013). The
functional neuroanatomy of spontaneous retrieval and strategic
monitoring of delayed intentions (in revision).
Berman, M. G., Jonides, J., & Lewis, R. L. (2009). In search of decay
in verbal short-term memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(2), 317–333. doi:10.
1037/a0014873.
Bugg, J. M., Scullin, M. K., & McDaniel, M. A. (2013). Strengthening encoding via implementation intention formation increases
prospective memory commission errors. Psychonomic Bulletin
& Review, 20(3), 522–527. doi:10.3758/s13423-013-0378-3
Campoy, G. (2012). Evidence for decay in verbal short-term memory:
A commentary on Berman, Jonides, and Lewis (2009). Journal
of Experimental Psychology: Learning, Memory, and Cognition,
38(4), 1129–1136. doi:10.1037/a0026934.
De Baene, W., Kühn, S., & Brass, M. (2012). Challenging a decade of
brain research on task switching: Brain activation in the taskswitching paradigm reflects adaptation rather than reconfiguration of task sets. Human Brain Mapping, 33(3), 639–651. doi:10.
1002/hbm.21234.
Einstein, G. O., & McDaniel, M. A. (1990). Normal aging and
prospective memory. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 16(4), 717–726. doi:10.
1037/0278-7393.16.4.717
Einstein, G. O., McDaniel, M. A., Thomas, R., Mayfield, S., Shank,
H., Morrisette, N., et al. (2005). Multiple processes in prospective memory retrieval: Factors determining monitoring versus
spontaneous retrieval. Journal of Experimental Psychology:
General, 134(3), 327–342. doi:10.1037/0096-3445.134.3.327.
Förster, J., Liberman, N., & Higgins, E. T. (2005). Accessibility from
active and fulfilled goals. Journal of Experimental Social
Psychology, 41(3), 220–239. doi:10.1016/j.jesp.2004.06.009.
123
Psychological Research (2014) 78:584–596
Goschke, T., & Kuhl, J. (1993). Representation of intentions:
Persisting activation in memory. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 19(5),
1211–1226. doi:10.1037/0278-7393.19.5.1211.
Marsh, R. L., Hicks, J. L., & Watson, V. (2002). The dynamics of
intention retrieval and coordination of action in event-based
prospective memory. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 28(4), 652–659. doi:10.
1037/0278-7393.28.4.652.
McBride, D. M., Beckner, J. K., & Abney, D. H. (2011). Effects of
delay of prospective memory cues in an ongoing task on
prospective memory task performance. Memory & Cognition,
39(7), 1222–1231. doi:10.3758/s13421-011-0105-0.
McDaniel, M. A., & Einstein, G. O. (2000). Strategic and automatic
processes in prospective memory retrieval: A multiprocess
framework. Applied Cognitive Psychology, 14(7), S127–S144.
doi:10.1002/acp.775.
McKeown, D., & Mercer, T. (2012). Short-term forgetting without
interference. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 38(4), 1057–1068. doi:10.1037/
a0027749.
Meier, B., & Rey-Mermet, A. (2012). Beyond monitoring: Aftereffects of responding to prospective memory targets. Consciousness and Cognition, 21(4), 1644–1653. doi:10.1016/j.concog.
2012.09.003.
Pink, J. E., & Dodson, C. S. (2013). Negative prospective memory:
Remembering not to perform an action. Psychonomic Bulletin &
Review, 20(1), 184–190. doi:10.3758/s13423-012-0337-4.
Rubin, D. C., & Wenzel, A. E. (1996). One hundred years of
forgetting: A quantitative description of retention. Psychological
Review, 103(4), 734–760. doi:10.1037/0033-295X.103.4.734.
Ruge, H., & Wolfensteller, U. (2010). Rapid formation of pragmatic
rule representations in the human brain during instruction-based
learning. Cerebral Cortex, 20(7), 1656–1667. doi:10.1093/
cercor/bhp228.
Scullin, M. K., & Bugg, J. M. (2013). Failing to forget: Prospective
memory commission errors can result from spontaneous retrieval
and impaired executive control. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 39(3),
965–971. doi:10.1037/a0029198.
Scullin, M. K., Bugg, J. M., & McDaniel, M. A. (2012). Whoops, I
did it again: Commission errors in prospective memory.
Psychology and Aging, 27(1), 46–53. doi:10.1037/a0026112.
Scullin, M. K., Bugg, J. M., McDaniel, M. A., & Einstein, G. O.
(2011). Prospective memory and aging: Preserved spontaneous
retrieval, but impaired deactivation, in older adults. Memory &
Cognition, 39, 1232–1240. doi:10.3758/s13421-011-0106-z.
Smith, R. E. (2003). The cost of remembering to remember in eventbased prospective memory: Investigating the capacity demands
of delayed intention performance. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 29(3), 347–361.
doi:10.1037/0278-7393.29.3.347.
Walser, M., Fischer, R., & Goschke, T. (2012). The failure of
deactivating intentions: Aftereffects of completed intentions in
the repeated prospective memory cue paradigm. Journal of
Experimental Psychology: Learning, Memory, and Cognition,
38(4), 1030–1044. doi:10.1037/a0027000.
Walser, M., Goschke, T., & Fischer, R. (2013). The difficulty of
letting go: Moderators of the deactivation of completed intentions. Psychological Research. doi:10.1007/s00426-013-0509-5
Download