No Directed Forgetting of Stimulus-Response Associations
Note: This manuscript is a pre-print, and may not exactly replicate the final publication of this
Hard to forget: No Directed Forgetting of Stimulus-Response Associations
Hannah Dames¹,², Marco Ragni3, Andrea Kiesel² & Christina U. Pfeuffer4
University of Freiburg
¹ University of Zurich, Department of Psychology, Zurich, Switzerland
² Albert-Ludwigs-Universität Freiburg, Department of Psychology, Freiburg, Germany
3 Technische Universität Chemnitz, Predictive Analytics, Chemnitz, Germany
4 Catholic University of Eichstätt-Ingolstadt, Department of Psychology, Eichstätt, Germany
This research was supported by funding from the German Research Foundation, DFG
[grant number RA1934/5-1 and RA1934/8-1] to M. Ragni, from the Swiss Confederation as part
of the Swiss Government Excellence Scholarships for Foreign Scholars and Artists to H. Dames,
and by the German Academic Exchange Service (DAAD) as part of an Exchange Scholarship to
All experiments were preregistered prior to data collection. The data as well as materials
are publicly available on OSF (Dames et al., 2021): https://osf.io/ta6qv/
Department of Psychology
University of Zurich
8050 Zürich, Switzerland
No Directed Forgetting of Stimulus-Response Associations 2
Humans can intentionally forget previously-learned declarative information such as words:
Memory for to-be-remembered (TBR) words is typically better than for to-be-forgotten (TBF)
ones (directed forgetting effect). The role of intention in the learning and retrieval of procedural
bindings is, however, less clear. Here, we combined item-method directed forgetting and item-
specific S-R priming to investigate whether people have intentional control over the formation
and/or retrieval of stimulus-response (S-R) associations: By categorizing stimuli in a learning
phase participants formed S-R associations. A memory cue following participants’ responses
indicated to remember or forget the stimulus for a later memory test. In the following test phase,
participants responded to the same stimuli, but the required response item-specifically repeated
or switched. In four experiments, reaction time differences between item-specific repetitions and
switches (item-specific S-R effect) did not differ between TBR and TBF stimuli. This was the
case when S-R associations already existed before the memory instruction was given (Exp. 1),
when stimuli, responses, and memory cues were paired multiple times (Exp. 2), when forming
new S-R associations (Exp. 3), and when categorizing word stimuli rather than images (Exp. 4).
Thus, intentionally up- and/or downregulating the memory strength of a stimulus representation
does not necessarily go hand in hand with the strengthening of the associated response. These
results can be explained by two accounts: either top-down control can selectively operate on
different levels of representations (and does not incidentally affect an entire event file) or that
separate memory systems store procedural and declarative content.
Keywords: stimulus-response associations, directed forgetting, memory, learning and retrieval
No Directed Forgetting of Stimulus-Response Associations 3
Internalized routines and habits make up a great proportion of our everyday life’s actions.
When preparing a cup of coffee, for instance, we do not put much thought into the individual
steps required to reach our goal: to have a hot cup of coffee. We seem to be able to exert such
actions, like preparing a cup of coffee or driving a car, more or less automatically – that is,
without controlled processing and deliberate intention (e.g., see Moors & De Houwer, 2006, for
theories of automaticity and corresponding debates). Evidence for the automatic retrieval of
actions comes from studies investigating stimulus–response (S-R) associations (e.g., Henson et
al., 2014, for a review). When stimuli (e.g., a button on the coffee machine lighting up signals
that the machine heated up properly) and responses (e.g., pressing the button to start the coffee
making process) repeatedly co-occur, they bind together, forming S-R associations. In the lab,
the formation and retrieval of such S-R associations is reflected in the observation that
participants’ responses are typically faster for stimuli that consistently require the same response
as opposed to a response different from the previously-executed one (referred to as repetition
priming effect, see Henson et al., 2014; Logan, 1988, 1990). Whereas the automatic retrieval of
S-R associations enables us to complete many of our everyday tasks without much cognitive
effort (e.g., when preparing a cup of coffee), there are numerous situations in which people try to
overcome such existing mental shortcuts (e.g., to break the old habit of checking the phone while
working). In addition, not every pairing of a stimulus and a response is equally relevant for our
goals (e.g., when learning new rules some may be more important than others). In the present
study, we investigated whether, and if so, to what extent, people have intentional control over the
formation of new and the retrieval of already-existing S-R associations. To address this question,
we combined the item-method of directed forgetting (for a review, see R. A. Bjork, 1998;
No Directed Forgetting of Stimulus-Response Associations 4
MacLeod, 1998) and an item-specific priming paradigm (see Hsu & Waszak, 2012;
Moutsopoulou et al., 2015). In the following, we will first discuss findings of list-method
directed forgetting on declarative and procedural memory content before explaining the item-
method of directed forgetting and our specific paradigm in detail.
Directed Forgetting of Declarative Content. Decades of research has already established
that humans are able to intentionally forget declarative memory content such as lists of words, as
often assessed using the directed forgetting paradigm (e.g., R. A. Bjork, 1970, 1998; Geiselman,
1974; Geiselman et al., 1983; MacLeod, 1998). For instance, in the list-method of the directed
forgetting paradigm, participants learn two lists of items, such as words. After studying the first
list (List 1), half of the participants are instructed to forget List 1 (referred to as the forget group),
whereas the other half is told to continue remembering that list (referred to as the remember
group). Afterwards, both groups study a second list of items (List 2). Typically, in a subsequent
memory test for both lists, participants in the forget group recall fewer List 1 items than the
remember group (List 1 forgetting costs). Moreover, participants in the forget group often recall
more List 2 items than the remember group (List 2 forgetting benefits: e.g., E. L. Bjork et al.,
1998; Geiselman et al., 1983; Sahakyan et al., 2013).
The observation that List 1 forgetting costs and List 2 forgetting benefits can occur
independent from one another (e.g., Pastötter & Bäuml, 2010; Sahakyan & Delaney, 2003, 2005)
suggests that processes contributing to the two effects are dissociable. Often, List 1 forgetting
costs are associated with retrieval processes (i.e., the memory strength of an item representation
itself is not affected but that memory representation cannot be accessed/retrieved; e.g., due to
inhibitory mechanisms, missing retrieval cues, etc.). For instance, the prominent context-change
hypothesis (Sahakyan & Kelley, 2002) postulates that upon the forget instruction, participants
No Directed Forgetting of Stimulus-Response Associations 5
deliberately change their mental context to intentionally forget List 1 items, thereby introducing
a larger-than-normal context change between acquisition of List 1 and learning of List 2 in the
forget than in the remember group. At test, the current context mismatches the List 1 context
impeding the recall of List1 items in the forget group. Alternatively, inhibitory accounts propose
that upon the forget instruction, List 1 items are inhibited, impairing retrieval of List1 items at
test (e.g., R. A. Bjork, 1989; Geiselman et al., 1983). In contrast, List 2 forgetting benefits have
been attributed to differences in encoding processes. For instance, it has been argued that after a
forget instruction, a reset of encoding processes enhances encoding of early List 2 items (e.g.,
due to reduced working memory load and reduced inattention; see Pastötter et al., 2017; Pastötter
& Bäuml, 2010).
Directed Forgetting of Procedural Memory Content. In contrast to this long tradition of
directed forgetting for declarative memory content, studies exploring the question whether we
can likewise intentionally forget procedural memory content (such as habits or routines) are rare.
The few studies (Dreisbach & Bäuml, 2014; Schmidt et al., 2021; Tempel & Frings, 2016)
investigating the impact of directed forgetting on procedural memory conducted so far used only
the list-method of directed forgetting. Results with this method revealed first evidence for
directed forgetting of procedural memory. Tempel and Frings (2016) observed that the
instruction to forget a whole learning episode enhances the recall of subsequently learned motor
sequences. Yet, List 1 forgetting costs for motor sequences were absent in their first experiment
and only emerged in a second experiment when introducing a distinct, prolonged break
separating List 1 and List 2. This finding is in line with previous studies on list-method directed
forgetting for verbal material (i.e., declarative memory content). There, List 1 forgetting costs
could be abolished under certain circumstances, such as when reinstating the learning context of
No Directed Forgetting of Stimulus-Response Associations 6
List 1, suggesting that retrieval of (i.e., the access to) to-be-forgotten (TBF) information is
impaired because of a contextual switch between List 1 and List 2 (Sahakyan & Kelley, 2002), as
So far, only a single study by Dreisbach and Bäuml (2014) has provided some initial
evidence that incidentally-learned habits in terms of S-R associations (Tempel & Frings, 2016
and Schmidt et al., 2022 investigated the impact of directed forgetting on motor sequences on not
S-R associations) may be reduced by means of list-method directed forgetting (see Dames et al.,
2022, for a failed attempt to conceptually replicate this finding). That is, participants received the
remember/forget instruction after several pairings of stimulus and response (and before encoding
a new set of S-R associations; thus, the memory cue concerned an entire learning episode, called
List 1). The authors observed List 1 forgetting costs because item-specific S-R effects in a testing
phase were significant in the remember but not (Experiment 1; or reduced effects, Experiment 2)
in the forget group. Dreisbach and Bäuml (2014) argued that participants were able to reduce
retrieval of inappropriate habits via a form of retroactive control.
We consider the three studies of Tempel and Frings (2016), Schmidt et al. (2021) and
Dreisbach and Bäuml (2014) as initial evidence that humans may have some intentional control
over the retrieval of a whole learning episode of procedural memory content (i.e., multiple motor
sequences learned in List 1). Importantly, however, the described prior studies used the list-
method directed forgetting. Using this method, participants in the forget and remember group
encode List 1 information – irrespective of the memory instruction – in the same way, as they do
not yet know whether they will have to remember or forget that information later on.
Consequently, we can assume that (up until the forget cue) TBF information is stored in long-
term memory the same way as to-be-remembered (TBR) information of List 1, but that retrieval
No Directed Forgetting of Stimulus-Response Associations 7
of TBF information is impaired at test (e.g., either because it is stored under a different context;
e.g., Sahakyan & Delaney, 2003; Sahakyan & Kelley, 2002 or it is inhibited; e.g., R. A. Bjork,
1998; Geiselman et al., 1983). That is, in the list-method, forgetting costs are usually not
attributed to differential encoding of the TBF in memory but rather to lower accessibility due to
on mechanisms operating at retrieval. Therefore, it is still unknown whether intention matters for
the encoding of S-R associations.
Parallel to this work, a recent, completely different line of research found that we cannot
remove outdated task-rules from procedural working memory (Abrahamse et al., 2022). This
contrasts with the findings of the described prior studies investigating the effects of list-method
directed forgetting on procedural memory content. Using the inducer-diagnostic paradigm
(introduced by Liefooghe et al., 2012), Abrahamse et al. observed that the Instruction-Based
Congruency (IBC) effect in the diagnostic task remained when the S-R rules of the inducer task
became irrelevant or were cancelled – although a general speed-up was observed indicating that
the cancellation cues were processed. This finding suggests that, even after a task-rule became
irrelevant, the instructed S-R mappings of the inducer task remained in working memory and
impacted performance. Different from the long-term list-method directed-forgetting studies
discussed above, Abrahamse et al.’s (2022) study thus implies that at least newly-formed
procedural representations cannot be removed from working memory by directed-forgetting -like
instructions. Assuming that S-R associations are also maintained in working memory before (or
while) entering long-term memory, it remains to be assessed whether we have intentional control
over their encoding into procedural memory.
The conflicting results of prior studies suggest that participants may intentionally impair
retrieval of irrelevant, procedural learning episodes (lists of motor sequences; Tempel & Frings,
No Directed Forgetting of Stimulus-Response Associations 8
2016; Schmidt et al., 2022; or lists of S-R associations; Dreisbach & Bäuml, 2014), but that S-R
associations themselves may still be intact. That is, directed forgetting may not affect the
encoding or maintenance of procedural associations, but only their retrieval when using a list-
wise directed-forgetting procedure. This assumption stands in contrast with findings regarding
declarative working memory content. Here, participants seem to be remarkably efficient in
intentionally forgetting outdated information in working memory (i.e., removing that information
from working memory; Dames & Oberauer, 2022).
The lack of studies on the intentional forgetting of procedural representations in contrast
to declarative representation (see Table 1) is especially surprising as procedural memory forms
the very basis of habits – which, in everyday life, we often try to intentionally overcome (e.g.,
not to snack while watching TV). Therefore, further assessments are essential to close this gap of
knowledge and determine whether S-R associations can intentionally be forgotten and by what
mechanisms. Doing so will also increase our understanding of similarities and differences
between declarative and procedural memory content. In the present study, we combined the item-
method of directed forgetting (e.g., Bjork, 1970) and item-specific stimulus response priming
(e.g., Moutsopoulou et al., 2015). Doing so will also increase our understanding of similarities
and differences between declarative and procedural memory content.
Item-Method of Directed Forgetting. In studies using the item-method of directed
forgetting (see MacLeod, 1998 for a review), participants are instructed to memorize sequentially
presented stimuli. After the offset of each stimulus, a memory cue informs participants to either
remember (TBR) or forget (TBF) the stimulus for a later memory test. In a subsequent memory
test for all items, participants typically recall fewer TBF than TBR items. Whereas the specific
mechanisms underlying the item-method directed-forgetting effect are still debated
No Directed Forgetting of Stimulus-Response Associations 9
controversially (e.g., Fawcett & Taylor, 2008; Tan et al., 2020), recent studies suggest that for
declarative memory content, both upregulation of TBR and downregulation of TBF information
play a role in item-method directed-forgetting (both in long-term memory; Fellner et al., 2020 as
well as in working memory, Dames & Oberauer, 2022).
The item-method of directed forgetting has some desirable properties when it comes to
investigating the intentional control over encoding and retrieving procedural memory content.
First, in the item-method, worse memory for TBF than TBR words has been attributed to
differential encoding mechanism between TBR and TBF stimuli (e.g., upon the forget cue,
participants selectively rehearse TBR but not TBF words – that is, upon the forget cue,
participants stop committing the TBF stimuli to memory; or participants downregulate/inhibit the
TBF and boost the TBR information by some other mechanism Basden et al., 1993; Basden &
Basden, 1996; MacLeod, 1975; Woodward et al., 1973). That is, different from the list-method
discussed above, which is likely driven by processes operating at retrieval (because the memory
instruction is given after an entire list has been encoded, List 1 stimuli in the remember and
forget group are encoded the same way up until the remember/forget instruction), we can also
observe effects of intentional forgetting at encoding.
Second, the item-method of directed forgetting allows to specifically target which
memory representation is cued as TBR or TBF (e.g., is the item or the binding/association
TBR/TBF). A great majority of the item-method directed-forgetting literature investigated how a
remember or forget instruction affected memory for single items (e.g., words). When
investigating the effects of directed forgetting on S-R associations, however, we are particularly
interested in how the directed-forgetting instruction impacts the encoding or retrieval of the
No Directed Forgetting of Stimulus-Response Associations 10
association between stimuli and responses. That is, we are interested in directed forgetting of
associative information (i.e., here the binding/associations between a stimulus and a response).
Figure 1 and Table 1 contrast the few studies on directed forgetting of procedural memory
to the existing evidence on directed forgetting of declarative memory while distinguishing the
effects of directed forgetting on items and/or bindings (associations). First, in Figure 1, we
distinguish whether the information of interest assessed in a study is assumed to be stored in
declarative (e.g., words, pictures) or procedural (e.g., task-rules, S-R associations) memory.
Second, we distinguish whether a study investigated the effect of directed forgetting on item
(e.g., the directed-forgetting instruction tackled a single stimulus or a list of stimuli) or binding
(e.g., the directed-forgetting instruction concerned the binding between two information, e.g.,
stimulus-stimulus or stimulus-response bindings) memory. Third, we differentiate whether the
list-method or item-method of directed forgetting was used (see above). Fourth, we distinguished
whether the learning of the information was explicit (e.g., for a word-location binding:
“remember/forget at which location a word was presented”) or implicit (e.g., “remember/forget
the word”, no instruction concerning the location was given). Please note, that here the learning
of the information could be explicit/implicit and/or the directed-forgetting instructions could
specifically concern the item and/or the binding. Naturally, when no associative information was
assessed in a study, in Figure 1, there is no path for “bindings”. Finally, in Table 1, we illustrate
some of the work and their results for each path. In Figure 1, supporting evidence for an effect of
directed forgetting is highlighted with a green checkmark. Evidence against an effect of directed
forgetting in Table 1 is highlighted with a red “X”. A question mark indicates that there is no
work on this topic.
No Directed Forgetting of Stimulus-Response Associations 11
As illustrated, besides the robust directed-forgetting effects observed on item memory for
both list and item-method, there is already some (yet occasionally mixed) evidence that
participants can likewise intentionally forget associative, declarative information. For instance,
Bancroft et al. (2013) found that a directed-forgetting instruction affected not only item
recognition but also associative recognition suggesting that people have intentional control over
strengthening associative memory (see also Hockley et al., 2016). Importantly, this was the case
even when memory for the associative information was not explicitly instructed and thus formed
without any deliberate attempt (note that in Figure 1, we also distinguish which representations
the directed-forgetting instruction tackles, e.g., “forget the item” vs. “forget the
binding/association”; or whether the stimuli/bindings were intentionally or incidentally learned).
Adding to this initial research, a recent study demonstrated that the intention to remember creates
stronger memory for the bindings between stimuli and their context even when using a deep
semantic processing task (as compared to a process-only condition; Popov & Dames, 2022;
please note that there was however no forget instruction). Therefore, for verbal material, there is
some evidence that at least the intention to remember strengthens declarative bindings.
Here, we systematically investigated whether the intention to remember or forget a single
declarative stimulus affects the encoding and/or retrieval of a response bound to that stimulus.
That is, we assessed whether forgetting stimuli also led to the forgetting of corresponding S-R
associations. Please note, that in the present study, we instructed participants to remember or
forget the stimulus and not the whole S-R associations for two reasons. First, when instructing
participants to remember the link between stimuli and their responses, they may likely also store
an explicit memory of the S-R mapping in their declarative memory. Thereby, besides making
the encoding of S-R associations explicit (and not implicit as in the present study), it would be
No Directed Forgetting of Stimulus-Response Associations 12
unclear whether any memory instruction affected the retrieval of procedural (implicit) S-R copies
or declarative (explicit) S-R representations. That is, in the present study, we aimed for the
incidental encoding of S-R associations. Second, we base our theoretical reasoning on The
Binding and Retrieval in Action Control (BRAC) framework (Frings et al., 2020). Here, it is still
unknown whether up- or downregulating the memory strength of one single component within
an event-file (see below) affects information bound to that stimulus. Specifically, targeting the
memory strength of the stimulus information only allowed us to investigate to which extent top-
down control in form of a declarative memory instruction affects procedural components bound
to that information.
No Directed Forgetting of Stimulus-Response Associations 13
Illustration of Different Directed Forgetting Manipulations on Different Types of Memory
Representations and the Respective (Mixed) Results.
Note. S = Stimulus, R = Response, C = Context, DF = Directed Forgetting. The numbers in
brackets refer to studies listed in Table 1. In the present work, we only consider the DF of
procedural memory content in the form of associative representations. Hence, we did not
further illustrate the “item” representation route (the representation of a single response, R) for
the procedural memory system.
No Directed Forgetting of Stimulus-Response Associations 14
Studies Illustrating Directed Forgetting (DF) Effects as Listed in Figure 1
Exemplary Reviews or Studies
Robust list-wise DF of stimuli (costs and benefits) using words (or pictures, e.g., Basden & Basden,
1996) as stimuli (S) (reviews: E. L. Bjork et al., 1998; MacLeod, 1998; Sahakyan et al., 2013);
observed when stimuli are not intentionally learned but only judged (e.g., Abel & Bäuml, 2019;
Geiselman et al., 1983); often absent in tests of item recognition.
Robust item-wise DF of stimuli using words (or pictures/scenes, e.g., Ahmad et al., 2019; Hauswald
& Kissler, 2008) as stimuli (e.g., MacLeod, 1999; Burgess et al., 2017; review: MacLeod, 1998)
Mixed: list-wise DF when learning of bindings (e.g., S-S/S-Context(C)) is explicitly instructed
- Gottlob & Golding (2007): DF of S-C (case/color) bindings
- Hanczakowski et al. (2012): No DF effect on associative recognition of word pairs (both when
encoding of associations were promoted or not) but forgetting costs for list discrimination
- Whitlock et al., (2020; Exp. 4): No DF effect on a behavioral level for stimulus-scene bindings
Unclear/positive: When stimulus-list bindings are incidentally learned, list-wise DF effects on list-
membership have been observed (e.g., Dames, Brand, et al., 2022; Lehman & Malmberg, 2009)
Positive: item-wise DF of bindings when learning of bindings (e.g., S-S/S-C) was explicitly
instructed or encouraged via a practice associative memory test
- Bancroft et al. (2013, Exp. 4): DF of associative recognition of word pairs
- Hockley et al. (2016, Exp. 2): DF of associative recognition of word pairs
- Whitlock. et al. (2022, Exp. 1): DF effect for object-scene bindings
- Popov & Dames (2022): intention to remember strengthens S-C bindings even when using a
deep semantic processing task (as compared to a process-only condition)
Mixed: item-wise DF of bindings when learning of bindings (e.g., S-S/S-C) was incidental
- Hourihan et al. (2007), DF: Recognition of TBF but not TBR words was significantly better
when they were tested in the same screen location as study compared to a different location
- Burgess et al. (2017), no DF: providing the original background scenes for the words at test
benefited memory for TBR and TBF words equally
- Whitlock. et al. (2022, Exp. 2): no DF effect on associative memory for object-scenes
- Hockley et al. (2016, Exp 1A & 1B): DF on associative recognition of word pairs
Positive: list-wise DF of procedural bindings (e.g., S-R or R-R-R-R) that were explicitly instructed
- Dreisbach & Bäuml (2014): forgetting costs but no benefits for S-R associations (unclear, how
explicit S-R instruction were; see main text); Measure: item-specific S-R effects
- Tempel & Frings (2016): forgetting costs for motor sequences (R-R-R-R) observed in Exp.2
(prolonged break between List 1 and 2) but not Exp. 1; Measure: Recall (cued by list)
- Schmidt et al. (2021, three-list version): forgetting costs for motor sequences when using the
same but not a different effector (hand) in List 3 in List 2, Measure: Recall (cued by list)
- Both Tempel & Frings (2016) and Schmidt et al. (2021) observed L2 forgetting benefits
No studies (Dreisbach & Bäuml, 2014: level of intentionality of S-R learning was unclear)
No Directed Forgetting of Stimulus-Response Associations 15
The Present Study. Our novel experimental procedure consisted of three major phases:
(1) A learning, (2) a distraction, and (3) a test phase. During the learning phase, participants
classified images of objects (e.g., as containing a mechanism or not). By categorizing the stimuli,
participants encoded S-R associations (stimulus-action, S-A, associations between stimuli and
motor outputs according to the terminology provided by Moutsopoulou et al., 2015). Importantly,
whether a left or right response is required for the classification of a specific stimulus differed
between stimuli and was indicated by a task cue detailing the classification-response mapping
prior to the stimulus’ presentation. In some blocks (depended on the experiment) participants
were also instructed to memorize some of the object images for a later memory test and to forget
others. For this, directly after the response to a stimulus, a memory cue informed participants to
either forget or remember the stimulus. In a subsequent distractor task intended to spurge
working memory, participants solved a visual working memory task for 1.5 minutes. In the
subsequent test phase, participants were presented once more with the stimuli from the learning
phase (both TBF and TBR objects). Importantly, for half of the stimuli per memory condition,
the same response was required to correctly classify the stimulus (the S-R mapping was the same
as in the learning phase; item-specific response repetition between learning and test). For the
other half of the stimuli, a different response was required as compared to the learning phase (the
S-R mapping was the opposite; item-specific response switch between learning and test).
How we Measure Directed Forgetting of S-R Associations. In typical item-specific S-
R priming experiments (e.g., Moutsopoulou et al., 2015; Pfeuffer et al., 2017), RTs and error
rates for item-specific response repetitions between learning and test phase trials are lower as
compared to item-specific response switches, indicating that S-R associations have been formed
during the learning phase and were retrieved in the test phase. We refer to these effects as item-
No Directed Forgetting of Stimulus-Response Associations 16
specific S-R effects. We refer to this effect as item-specific S-R effects (repetition priming). In
the present study, we used the difference in RTs in the test phase between trials that require an
item-specific response switch and an item-specific response repetition to determine a directed-
forgetting effect based on the remember/forget cues. That is, we measured the strength of item-
specific S-R effects for TBF as compared to TBR stimuli.
Intentional Control over Binding and Retrieval in Action Control?
Current theories on item-method directed forgetting of declarative memory content
revolve around the notion that either (or both; see Fellner et al., 2020) the selective rehearsal of
TBR information (e.g., MacLeod, 1975; Woodward et al., 1973) or/and the
inhibition/downregulation of TBF declarative memory (e.g., Fawcett & Taylor, 2008, 2012)
contribute to directed-forgetting effects. Both accounts share the assumption that the intention to
remember or forget can regulate the memory strength of the target stimulus representation (and
does not “only” impair the retrieval of that stimulus representations via e.g., missing/impaired
Similar to the ways by which directed forgetting could affect declarative memory content,
directed-forgetting effects on procedural memory content may stem from several possible
underlying mechanisms. Regarding potential mechanisms, we base our theoretical reasoning on
The Binding and Retrieval in Action Control (BRAC) framework (Frings et al., 2020). BRAC is
based on the notion that various features of an event are bound together and later retrieved when
one or more of the features re-occur. Features (e.g., stimulus, response, and effect) integrated in
these bindings are presumed to be coded in a common representational format (e.g., Hommel,
2004; Hommel et al., 2001). This allows for one feature to retrieve the respective other features
bound to it. Crucially, it is assumed that binding and retrieval of event-files can be separately
No Directed Forgetting of Stimulus-Response Associations 17
modulated by not only bottom-up but also top-down influences. Particularly, the BRAC
framework postulates several processes allowing the top-down control of event-files (e.g.,
different levels of action representation, attentional weighting of stimulus features, and other
forms of meta-cognitive control; Frings et al., 2020). Yet, there is scarce knowledge about the
mechanisms by which intentional control over event-files can specifically impact processes of
feature binding and retrieval (see Hommel, 2022). At present, although the binding of
information into an even-file seems to be unselective (i.e., to happen automatically; see work on
distractor-response binding; see Hommel, 2022, for a review), one study showed that the
saliency of a stimulus might affect corresponding S-R bindings (Schmalbrock et al., 2021).
Importantly, it is still unclear if and how top-down intentional control (e.g., intention to
remember or forget) – as opposed to the one study one study on bottom-up processes such as
saliency – affects the strength of information integration into an event file or how it affects the
retrieval of that information. In theory, intentional control might also selectively affect
representations or bindings already integrated into an event file. However, to this day, there is no
evidence that components of an event file can be intentionally removed from the event-file once
they have been integrated. Thus, the present study represents an important step towards
understanding how top-down intentional control can manage what has been already bound or
what is retrieved in an event-file.
No Directed Forgetting of Stimulus-Response Associations 18
Illustration of A) Theoretical Accounts and B) Derived Predictions
Note. Red colors and hatched areas indicate that the memory strength of the corresponding
representation/weight (specific impact + separate memory system account) or entire event file
(global impact account) are downregulated/inhibited. Conversely, green colors indicate an
upregulation of memory strength.
No Directed Forgetting of Stimulus-Response Associations 19
We considered three different theoretical accounts how a memory instruction could affect
S-R associations as illustrated in Figure 2. Let us assume that participants have responded to a
stimulus and, by doing so, formed a S-R association (i.e., an event file including the
corresponding association). Next, participants receive the instruction to remember or forget the
stimulus they just categorized.
Global Impact. First, the instruction to remember or forget the stimulus could have a
global impact on the entire event-file (i.e., both the stimulus and response representation as well
as their binding were affected). If so, any control process would not operate on specific
representations within an event-file but always impact all integrated information. That is, the
whole event-file would become downregulated/inhibited upon intending to forget one feature
(e.g., impairing retrieval of that information in the test phase; as illustrated by red and hatched
areas in Figure 2) and/or all information bound in an event-file would be boosted upon intending
to remember one feature (strengthening the whole event-file, thereby enhancing later item-
specific S-R effects; as illustrated by green areas in Figure 2). Regardless of the exact underlying
mechanisms (i.e., whether boosting the memory strength of TBR information, e.g., due to
selective rehearsal and/or inhibition/downregulation of TBF information play a role), according
to this account, a global impact of the memory instruction should result in larger item-specific S-
R effects for TBR as compared to TBF stimuli at test.
Specific Impact. Another possibility is that the memory strength of representations bound
within an event-file can be specifically up- (e.g., as illustrated by the green colored stimulus
representation in Figure 2) or downregulated (red and hatched areas in Figure 2) without
impacting associated information. If so, the instruction to remember or forget the stimulus would
influence selectively the memory strength of the stimulus itself but would not affect the binding
No Directed Forgetting of Stimulus-Response Associations 20
or the associated response. As a consequence, item-specific S-R effects should not be modulated
by the memory instruction in this case (because the item is presented as a retrieval cue and does
not need to be retrieved from memory itself).
Separate Memory Systems. The third account assumes separate memory systems for
declarative and procedural memory content. That is, S-R associations are stored in procedural
memory, but the representation of the stimulus information is additionally kept in a separate,
declarative memory system. If so, one might additionally assume that the instruction to
remember or to forget a stimulus (that is itself declarative information) affects only declarative
memory contents. Then, any operation carried out on the stimulus representation (e.g.,
instruction to remember that stimulus for a later memory test) in declarative memory would have
any impact on the S-R associations stored in procedural memory.
In sum, while the global impact account predicts a modulation of the item-specific S-R
effects by the memory instruction, both the specific impact and separate memory systems
One may reason that reduced memory strength (or inhibition) of the stimulus representation
itself could also impair the retrieval of the associated response (despite having an unaffected
associated weight between the stimulus and the response). If that were the case, one may also
expect reduced S-R effects for TBF as compared to TBF stimuli. However, we deem that
scenario unlikely because during test phase, the stimuli are presented again and do not have to be
freely retrieved from memory. If anything, better LTM for the items may enhance item
No Directed Forgetting of Stimulus-Response Associations 21
account do not (see Figure 2). To foreshadow, our results speak against the global impact
General Structure of Experiments 1-3
All Experiments were pre-registered (see individual experiment description for the links).
Our new task design combined two paradigms used to investigate item-method long-term
memory directed forgetting (e.g., Bjork, 1970) and item-specific stimulus response priming (e.g.,
Moutsopoulou et al., 2015). The four experiments reported here followed a similar experimental
procedure that is described below. The individual experiments varied with respect to whether
participants had already learned S-R associations (Experiment 1), how often a memory cue was
paired with a specific response (Experiment 2 and Experiment 3), as well as the task and
materials (Experiment 4). Because Experiment 4’s task deviated from the previous three
experiments, we will introduce the structure of Experiment 4 later in the manuscript.
Material and Participants
The experiments were conducted in German (Experiment 1) and in English (Experiment
1-4). Stimuli were drawn from a set of 128 images (256 x 256 pixels) of everyday objects that
contained 64 mechanical and 64 non-mechanical objects that we have used in previous studies
(Dames et al., 2022; stimulus set originally from Brady et al., 2008; Moutsopoulou et al., 2015;
Pfeuffer et al., 2017). Participants’ task was to classify these images according to whether they
contained a mechanism or not (e.g., wheels, levers, electronic parts; “mechanisch”, Engl.
mechanic vs. not, “nicht-mechanisch”, Engl. non-mechanic) using the ‘S’ (left key) and ‘L’ (right
key) keys on their keyboard. The experiments were programmed using jsPsych (de Leeuw, 2015)
and were administered online through Prolific. As Experiment 1 was the first study to investigate
the impact of item-method directed forgetting on the retrieval of S-R associations, we had no
No Directed Forgetting of Stimulus-Response Associations 22
clear indicator for our sample size estimation. For list-method directed forgetting, Dreisbach and
Bäuml (2014) observed a partial eta squared of ηp2 = .14 and ηp2 = .06 for the critical
interaction between memory instruction and compatibility (similar to item-specific response
repetitions vs. response switches as in the present study) in their RT analyses. Power analyses
using G*Power (Faul et al., 2007) on the mean ηp2 = .10 suggested a sample size of N = 76
(with α = .05 and 1-β = .80) to detect a significant interaction assuming that item-method and
list-method directed forgetting would have similar impact on item-specific S-R effects.
Participants gave informed consent form prior to the study, and they were debriefed at the end.
The present study was conducted in adherence to the Declaration of Helsinki and the guidelines
set by the local ethics committee.
The following pre-registered exclusion criteria were defined prior to running the
experiments. We excluded participants who 1) switched tabs/windows away from the experiment
more than twice, 2) indicated that they did not follow the instructions, 3) stated that they did not
participate seriously in the experiment or did not agree that we could use their data, 3) failed to
answer correctly to two attention checks that were presented within the instructions (only
introduced after Experiment 1), 4) committed errors or response omissions on more than 30% of
the trials in the learning or test phase, 5) had any suspicion about the purpose of the experiment
or who reported having suspected that TBF stimuli would be tested, or 6) could not correctly
state what they were supposed to do when the remember/forget cue was presented. Furthermore,
we excluded participants who had less than six observations per cell remaining for any analysis
(due to the applied trial exclusion criteria; see Data Analysis section below). All participants
excluded according to these exclusion criteria were replaced with new ones.
No Directed Forgetting of Stimulus-Response Associations 23
We extended the item-method of directed-forgetting paradigm (for a review, see R. A.
Bjork, 1998; MacLeod, 1998) to S-R mappings using a classification task (drawing on an item-
specific priming paradigm; Hsu & Waszak, 2012; Moutsopoulou et al., 2015; Pfeuffer et al.,
2017). Our general experimental design consisted of three major phases (see Figure 3): (1) A
learning, (2) a distraction, and (3) a test phase. Instructions were given on screen and were
summarized prior to the beginning of each block. Participants first received instructions
regarding the classification task. In a practice block, participants were familiarized with the
categorization task. After the practice block, participants had the possibility to reread the
instructions and redo the practice block or proceed with the experiment. After the practice phase,
the main experimental blocks followed where participants categorized objects and also tried to
memorize some of the object images for a later memory test.
No Directed Forgetting of Stimulus-Response Associations 24
General Structure of Experimental Phases and Trials of the Learning and Test Phase
Note. The example depicts the general structure of the experiments.
No Directed Forgetting of Stimulus-Response Associations 25
Learning phase. During the learning phase, participants’ task was to classify images of objects
as mechanic versus non-mechanic (Experiment 1-3) or words as describing animate versus
inanimate living beings/objects (Experiment 4). By categorizing the stimuli, participants encoded
S-R associations (stimulus-action, S-A, associations between stimuli and motor outputs
according to the terminology provided by Moutsopoulou et al., 2015). The trial structure of the
learning phase for Experiment 1-3 is illustrated in Figure 3: First, a task cue was presented for
700 ms that indicated the classification-response mapping. For example, the task cue “M + N”
indicated that a left response was required for mechanic objects (M) and a right response for non-
mechanic (N) ones. Conversely, the task cue “N + M” indicated that a right response was
required for mechanic objects and a left response for non-mechanic ones. Then, the stimulus (the
image of an object or a word describing an object/living being) followed, and participants
classified the image via a left/ right key press according to the preceding task cue. Stimuli were
displayed until a response (classification via a left/right key press) was given or for a maximum
duration of 2000 ms. Next, only for incorrect trials, a feedback screen was presented stating
“error!” for incorrect responses or “too slow!” in case of response omissions (red, 500 ms).
Immediately after the feedback offset for incorrect trials and after the response for correct trials,
each stimulus was followed by either a TBR or a TBF instruction for 2000 ms (note the
differences between experiments described in the sections of the respective experiments). The
instruction (memory cue) to remember a stimulus was symbolized by an open blue circle (o), and
the instruction to forget was symbolized by an orange slashed-circle (ø). Participants were
instructed to remember only the TBR stimuli for a later memory test and to forget the TBF
stimuli, because they would not be tested in the memory test afterwards. Crucially, participants
No Directed Forgetting of Stimulus-Response Associations 26
were not instructed to memorize any mappings/associations between the stimuli and their
responses. They were only instructed to memorize the stimuli.
(2) Distractor phase. In the following distractor phase intended to purge short-term
memory, participants solved a visual working memory task for 1.5 minutes (an adapted,
computerized version of the Corsi blocktapping Task; CORSI; forward span; Corsi, 1973).
(3) Test phase. In the test phase, participants were presented once more with the stimuli
from the learning phase (both TBF and TBR stimuli). Trial structure and task instructions were
equivalent to the learning phase (except that, starting in Experiment 2, no memory cues were
presented in the test phase anymore). Importantly, for the old stimuli (previously-encountered in
the learning phase), for half of the stimuli per memory condition, the required response (S-R
mapping) was the same as in the learning phase (item-specific response repetition between
learning and test, e.g., the stimulus requires a right key press in both the learning and test phase).
For the other half of the stimuli, the required response (S-R mapping) was the opposite (item-
specific response switch between learning and test, e.g., the stimulus requires a right key press in
the learning phase, but a left key press in the test phase). In all experiments, stimuli were
presented only once in the test phase In Experiment 3, novel stimuli were additionally presented
in the test phase (intermixed with old stimuli).
Using post-experimental questions, we checked whether participants followed the
instructions and had any suspicion about the purpose of the experiment. Followings these
questions, participants were properly debriefed about the reason behind the directed-forgetting
No Directed Forgetting of Stimulus-Response Associations 27
On a trial level (trials that were excluded, but the participants remained for the analysis),
the following data were discarded: First, trials with response omissions were removed. Second,
for the analyses of trials in the test phase only trials with correct responses for the corresponding
stimuli in the learning phase were used. Third, for all RT analyses, error trials were discarded.
Fourth, trials from the practice blocks were removed. Fifth, within each RT analysis trials with
exceptionally high or low RTs that is, trials with RTs above/below 3SD from the individual cell
mean were discarded.
We analyzed the data of all experiments using mixed models (for an overview of mixed
models see Baayen et al., 2008; Judd et al., 2012). The correctness of responses was analyzed
using generalized linear mixed models (GLMMs) assuming a Bernoulli data distribution
predicted by a linear model through a logistic link function. Please note that because we could
not find statistically significant evidence for item-specific S-R effects in participants' error rates,
we do not report these findings in this manuscript. Error rate patterns, however, did not
contradict RT patterns. RTs were log-transformed and analyzed using linear mixed models
(LMMs) assuming a Gaussian data distribution. Experiment 1 and 2 used a frequentist mixed
model approach, whereas Experiment 3 and 4 used a Bayesian approach. All categorical
predictors were coded as sum-to-zero contrasts.
For the main analyses, our models included the fixed factors response (repetition vs.
switch) and instruction (remember vs. forget) as well as their interaction. The random effects
structure for our design was given by random intercepts for participants and stimuli and by-
participant random slopes for response and instruction including their interaction. In addition, we
added a by-stimulus random slope for instruction. In case the models did not converge, as pre-
No Directed Forgetting of Stimulus-Response Associations 28
registered, we first removed the correlation among random slopes and then stepwise removed
random slopes (beginning with the by-stimulus random slope and then the by-participant random
slopes starting with the removal of the interaction term and then the response term) until the
models converged. In Experiment 1, the final model that converged included the full random
effect structure excluding all correlations among random effects. In Experiment 2, the final
model that converged included random intercepts for participants and images and a by-
participant random slope for the response term, excluding all correlations among random effects.
In Experiment 3 and 4, using Bayesian LMMs, the models converged using their full random
The frequentist LMMs were analyzed using mainly the R packages lme4 (Bates et al.,
2015) and afex (Singmann et al., 2015). We obtained model predictions and graphics with the
support of the ggeffects package (Lüdecke, 2018), and emmeans (Lenth et al., 2018). For the RT
analyses and the frequentist LMMs, we evaluated the significance of the fixed effect using the
Kenward-Roger approximation to estimate the denominator degrees of freedom with the
standard p < .05 criterion.
We analyzed the data of Experiment 3 and 4 using Bayesian LMMs using the
BayesFactor package with default settings (50,000 iterations; Morey et al., 2015; note that we
initially pre-registered to run this analysis using the brms package, but switched to the
BayesFactor package for reasons of time efficiency). We fitted two competing, nested models for
each hypothesis: One including the factor of interest and one that did not. Using the same
package, we then calculated Bayes Factors (BFs) to estimate the strength of evidence for the null
and the alternative hypothesis. We considered BF > 3 as substantial evidence for one hypothesis
over the other (see Bayes factor conventions of Jarosz & Wiley, 2014).
No Directed Forgetting of Stimulus-Response Associations 29
In Experiment 1, we tested whether the intention to remember or forget a stimulus
affected already-learned S-R associations. To do so, we had participants respond to a stimulus
with the same response three times in three learning blocks before receiving a remember or
forget instruction after a fourth S-R pairing (4 learning instances before the remember/forget
instruction). We assessed their item-specific S-R effects in a later one-instance test phase.
Methods, hypotheses, and data analysis procedures were preregistered and can be found
online (https://osf.io/y5bxd). There, we also report detailed information on the randomization
In total, 135 native English speakers completed the whole experiment via the web-based
participation platform Prolific (+ additional 16 participants from the local participant pool at the
University of Freiburg). After applying the preregistered exclusion criteria, the final sample
consisted of 74 participants (Mage = 27.0, SDage = 6.6, range 18–40 years; 51.4% female, 47.3%
male, 1.4% diverse).
Participants first practiced the categorization task (mechanic vs. non-mechanic) on eight
new stimuli (see task description). Next, the learning phase followed which consisted of three
learning blocks without memory instructions in order to build sufficiently strong S-R
associations as well as one fourth learning block including remember/forget instructions for each
stimulus. In total, participants responded to 64 objects. In each block the stimuli were presented
once (and thus four times in total; 256 trials; rerandomized order per block). After the third
No Directed Forgetting of Stimulus-Response Associations 30
learning block, participants were told that it would now also be their task to remember some of
the images of objects they classified for a later memory test and that they could forget others (see
general task description for the specific instructions of the memory task). In the fourth and last
block of the learning phase, instead of a feedback screen, each object was followed by either a
TBR or a TBF instruction. Participants then practiced this extended task (classification and
memorization) on the same eight stimuli as in the practice at the beginning of the experiment.
The practice was followed by a recognition test for the four TBR stimuli only (random
presentation of the four TBR practice stimuli in addition to four new stimuli). Besides
familiarization with the task, this practice recognition test was intended to substantiate
participants’ belief that the instruction to forget was genuine. Participants then performed the
fourth block of the learning phase (classification including remember/forget instruction). After
the distractor phase, the test phase followed. Again, participants classified all of the objects from
the learning phase as containing a mechanism or not (the same task as in the learning phase).
Importantly they were not informed that this phase differed from the fourth block of the learning
phase. They were told that they would see the same stimuli as before and were again asked to
only remember the TBR stimuli and to forget the TBF stimuli. All stimuli that were a TBR or
TBF stimulus in the fourth learning block were again presented with a corresponding TBR or
TBF memory cue, respectively. Crucially, in this test phase, the required response per stimulus
could item-specifically either repeat or switch. Participants were told that after this fifth block,
the memory test would follow after performing another short visual working memory task.
A second distractor phase was followed by a recognition test where participants were
again presented with all objects from the learning/test phase (both TBF and TBR stimuli) and 64
new objects (total of 128 trials; randomly split into two blocks of 64 trials). Participants were
No Directed Forgetting of Stimulus-Response Associations 31
told that they would be presented with all the images from the learning phase (note, that they
were not informed that there was a preceding test phase) as well as new images one at a time.
They were instructed to decide for each image whether it had been presented during the
experiment so far or not – regardless of its associated memory instruction – by pressing either
left (“S”-key) or right (“L”-key) response key. The recognition task was introduced as another
“classification task”. On a given trial, an old or new stimulus image was presented in the center
of the screen, and participants were instructed to categorize the stimulus as quickly and
accurately as possible as either “old” or “new” according to the category labels that were
displayed in the right or left upper screen corner. Whether a right or left response was required
for an “old” or “new” categorization was randomly assigned for each participant, but always
remained the same throughout the recognition test for a participant (e.g., “old” always requires a
“S” key press). There was no time limit for the recognition test.
Results and Discussion
Mean log-RTs are displayed in Figure 4A. Untransformed RTs were faster for response
repetitions as compared to response switches for both TBR (Mrepetition = 750ms, Mswitch = 763ms)
and TBF stimuli (Mrepetition = 744ms, Mswitch = 758ms).
No Directed Forgetting of Stimulus-Response Associations 32
Estimated (model-based) mean log reaction times (RTs) across conditions of (A) Experiment 1
and (B) Experiment 2
Table 1 shows the results and model statistics of the log-RT model. To estimate the item-
specific S-R effects, we compared RTs in test trials with response repetitions and response
switches. As predicted, RTs were significantly faster for response repetitions as compared to
response switches (p = .039, t = 2.06). However, there were no significant differences in RTs
when responding to a TBR or TBF image. Likewise, the interaction of response and instruction
Note. Error bars represent the standard errors (SE) of the model.
No Directed Forgetting of Stimulus-Response Associations 33
was not significant. That is, we observed no evidence that the intention to remember or forget the
stimulus affected the observed item-specific S-R effects.
Results of the Linear Mixed Model for the Log Reaction Times (RTs) in Experiment 1
6.54 – 6.62
-0.02 – -0.00
-0.01 – 0.01
Response x Instruction
-0.01 – 0.01
Note. CI = 95% confidence intervals.
Last, we also analyzed memory performance in the subsequent recognition test, to test
whether the memory instruction affected participants’ declarative memory for the images. Note,
that we did not necessarily expect to find a difference in recognition performance between TBR
and TBF stimuli, because participants had already responded to each stimulus five times
throughout the experiment (3 times without any memory instruction, 2 times with a memory
instruction – once each in the learning and test phase). That is, the stimuli should have been
deeply processed regardless of the memory instruction resulting in high familiarity for these
stimuli and thus potential ceiling effects. Nevertheless, we found a significant difference in
memory performance between TBR and TBF words (GLMM on participants’ responses’
correctness; correct recognition) with instruction, despite this difference being small (3.1%;
GLMM was fitted with a binomial errors distribution and a logistic linking function; with a
random intercepts for participants and stimuli and by-participant random slopes for instruction,
p = .038). Together with previous studies supporting the generality and robustness of directed
No Directed Forgetting of Stimulus-Response Associations 34
forgetting for pictures or scenes (e.g., Ahmad et al., 2019; Hauswald & Kissler, 2008), this
finding suggests that participants clearly formed the intention to remember or forget the stimuli
and that this intention resulted in a directed-forgetting effect for the declarative stimulus
representation despite prior in-depth processing of the stimuli during five learning instances.
To summarize, in Experiment 1, despite observing a directed-forgetting effect on
declarative memory, we observed item-specific S-R effects but no effect of directed forgetting on
the strength of these item-specific S-R effects.
In Experiment 1, we did not observe an effect of directed forgetting on already-learned S-
R associations. The aim of Experiment 2 was to test whether the intention to remember or forget
the stimulus would impact the encoding or retrieval of S-R associations when the stimulus was
paired multiple times with a response and the intention to remember or forget it. Furthermore, we
wanted to test if directed forgetting affected item-specific S-R effects (i.e., underlying S-R
associations) if memory cues were paired with the stimuli right from the first stimulus encounter.
Thus, in Experiment 2, during the learning phase, participants responded to each stimulus
four times (in four blocks) in the exact same way (i.e., same S-R mapping). Right from the start,
participants were also instructed to memorize some of the object images for a later memory test
and forget others. That is, already during the first learning instance, a memory cue informed
participants to either forget or remember the stimuli participants categorized. As a consequence,
stimuli, responses, and intention were paired four times. Like in Experiment 1, in the test phase,
participants were once more presented with the stimuli from the learning phase (both TBF and
TBR objects) and required responses that item-specifically repeated versus switched. Different
from Experiment 1, no memory cues were presented anymore in the test phase. We chose to
No Directed Forgetting of Stimulus-Response Associations 35
change this, because in Experiment 1, RTs were generally slow in the test phase, implying that
participants may have additionally tried to memorize the stimuli in the test phase for the later
memory test. Potentially, such additional processing and intentional slowing of responses may
have overshadowed a tentative influence of the memory instruction on the retrieval of item-
specific S-R associations.
Methods, hypotheses, and the data analysis procedures were preregistered and can be
found online: https://osf.io/xm583.
The same preregistered inclusion and exclusion criteria as in Experiment 1 were applied
and we only recruited participants that had not participated in the previous experiment. From the
initial sample (N = 152, see the General Methods section for our preregistered exclusion criteria),
n = 73 participants remained in the final sample after applying all exclusion criteria (Mage = 27.3,
SD = 5.6, range: 18 – 40 years; 56.8% female, 40.5% male, 2.7% diverse).
Material and Procedure
Stimuli and procedure were identical to Experiment 1 with the following exceptions:
After familiarizing themselves with the classification task in a first practice block as in
Experiment 1, participants were told that it was also their task to remember some of the images
of objects they classify for a later memory test. The memory instructions were the same as the
ones given after the third learning block in Experiment 1. Participants then practiced this
extended task (classification and memorization) on the same eight stimuli as in the first practice.
The practice task was followed by a recognition test for TBR stimuli only (random presentation
of the four TBR practice stimuli in addition to four new ones).
No Directed Forgetting of Stimulus-Response Associations 36
The learning phase consisted of four blocks and each item was presented four times in the
learning phase (once per block) – stimuli, response, and memory cues were thus paired four
times. In the test phase, participants were presented with all stimuli of the learning phase (both
TBF and TBR stimuli). Again, participants were instructed to classify the objects as containing a
mechanism or not by pressing a left/right key (the same task as in the learning phase). However,
this time, they were not presented with a memory cue. Participants were instructed that they
would only need to respond to stimuli fast and accurately and that there would be no memory
cues for this phase. Participants were also informed that there would be no later memory test.
They were told that they will respond to both TBR and TBF stimuli and that they should simply
respond as fast and accurately as possible regardless of the previous memory instruction.
Table 2 shows the results and model statistics of the log-RT model. Untransformed RTs
were faster for response repetitions as compared to response switches for both TBR
(Mrepetition = 611ms, Mswitch = 629ms) and TBF stimuli (Mrepetition = 618ms, Mswitch = 624ms).
Results of the Linear Mixed Model for the Log Reaction Times (RTs) in Experiment 2
6.36 – 6.43
-0.01 – -0.00
-0.00 – 0.01
Response x Instruction
-0.00 – 0.01
Note. CI = 95% confidence intervals.
No Directed Forgetting of Stimulus-Response Associations 37
As illustrated by the mean log-RTs in Figure 4B, removing the memory cues from the
test phase reduced RTs substantially. Still, whereas RTs for response repetitions were
significantly faster than for response switches (p = .007), this effect was not modulated by the
intention to remember or forget the stimulus (see Table 2).
In Experiments 1 and 2, directed forgetting did not impact the encoding or retrieval of S-
R associations (i.e., corresponding item-specific S-R effects). In Experiment 3, we aimed to
replicate this finding while ruling out potential confounds that may have hindered the
observation of an influence of directed forgetting.
First, to strengthen the directed-forgetting manipulation, we changed the stimulus set
slightly, selecting images that seemed particularly distinct from each other and memorable.
Second, we ruled out that participants did not have enough time to process the memory
cues before the next trial started. Furthermore, we worried that some participants may have
wrongfully understood that the memory cue concerned the subsequent trial and not the object
image they had just categorized (despite being clear on this in the instruction, one participant
mentioned such an understanding of the task instructions in the previous experiment). To rule out
both potential confounds, we introduced a “Next round!” window (500ms) between the offset of
the memory cue and the onset of the next trial. Furthermore, we added a picture to the task
instructions clarifying with arrows that the memory cue indicated whether the just-categorized
object image should be remembered or forgotten. Additionally, we added a corresponding
question at the end of the experiment. We showed participants the memory cues and asked them
whether the respective cue meant they should remember/forget the object image of the current or
the next trial (we excluded participants who did not respond correctly to this question).
No Directed Forgetting of Stimulus-Response Associations 38
Third, we added new images to the test phase which had not been categorized before.
Doing so allowed us to compare RTs of TBR and TBF images to that baseline condition,
estimating to what extent participants formed item-specific S-R associations at all. If they did,
RTs to both TBR and TBF stimuli should be faster than to new ones.
Last, we tested whether the instruction to forget or to remember the stimulus (hence, an
additional memory task) potentially even weakened the encoding of item-specific S-R
associations (as compared to a condition where no such additional memory task was present).
When participants aim to remember the stimulus itself but not the S-R associations, they may
rehearse the respective stimulus thereby re-activating its representation. We assume that upon
responding to a stimulus, a S-R association is formed (e.g., via Hebbian learning; for an
overview see Abrahamse et al., 2016; Verguts, 2017). If the stimulus is now activated again (e.g.,
being rehearsed to conform with the memory instruction) but no response is given, the
distributed associative memory (Rumelhart et al., 1986) may send signals through the existing
connection weights and predict a response (the network’s output based on the previously learned
S-R association). The predicted response would then be compared to the actual responses which
in this case is absent. Hence, an error is produced that may, in turn, result in weight changes
between the stimuli and the associated responses (the weight matrix). As a result, intending to
memorize the stimulus without also reactivating the associated response, may indeed result in the
loss of the previously-learned S-R association (in particular, when this association is weak to
To test this possibility, participants were divided into two groups: One group (no-memory
group) worked only on the item-specific priming task without memory instructions (e.g., Hsu &
Waszak, 2012; Moutsopoulou et al., 2015; Pfeuffer, Moutsopoulou et al., 2017). The other group
No Directed Forgetting of Stimulus-Response Associations 39
(memory group) did the same but in addition received the instruction to remember half of the
stimuli and forget the other half. That is, in the learning phase, half of the participants were
instructed to memorize some of the images for a later memory test and forget others. Like in
Experiment 1 and 2, directly after the response to a stimulus, a memory cue informed
participants to either forget or remember the stimulus (memory-group only). For the other half of
the participants, the same cues were presented after participants’ responses, but they were told
that these stimuli signaled the end of a trial and did not receive a memory instruction that was
associated with the cues (no-memory group). Thereby, stimulus timing and visual input were
equivalent between the two groups.
If the memory instruction negatively influenced the retrieval and/or encoding of S-R
associations, performance differences between item-specific response repetitions and response
switches (S-R effect) in the test phase should be smaller for the memory group (also when only
considering the TBR stimuli) as compared to no-memory group.
Methods, hypotheses, and the data analysis procedures were preregistered and can be
found online (https://osf.io/v7zk6).
Inclusion and exclusion criteria were the same as before. Of the initial sample (N = 244),
n = 180 participants remained in the final sample after applying all exclusion criteria (Mage =
28.8, SD = 6.2, range: 18-40 years; 71.6% female, 27.2% male, 1.2% diverse). Participants were
randomly assigned to either the group receiving a memory instruction concerning the cues
No Directed Forgetting of Stimulus-Response Associations 40
(memory group, n = 93) or the group receiving no additional memory instruction concerning the
cues (no-memory group, n = 87).
Material and Procedure
Material and procedure were the same as in Experiment 2 with the following exceptions:
After the first practice block (categorization task), only participants in the memory group were
told that it was also their task to remember some of the images they classify for a later memory
test. For these participants, instructions concerning the memory cues were the same as in the
previous experiments. Participants in the no-memory group were presented with the same cues
but were simply instructed that they mark the end of a trial and that the color/shape does not have
any meaning. We also told them that the cues were included for another group of participants,
where the cues had a meaning. For both groups, for incorrect trials or responses outside the
response frame, the memory cue was presented after a feedback screen (feedback in red:
“error!”/“too slow!”). Participants then practiced this (extended) task (classification and
memorization for the memory-group only) on 16 new stimuli.
For the memory group, the practice was followed by a recognition test for TBR stimuli
only (random presentation of the four TBR practice stimuli in addition to four new ones). Like in
Experiment 2, besides familiarization with the task, this practice recognition test for the memory
group was intended to increase participants’ belief that the instruction to forget was genuine. To
equate completion times, for the no-memory group, the practice was followed by a distractor task
where participants solved a visual working memory task for 1.5 minutes instead (CORSI;
forward span; Corsi, 1973).
The trial structure was the same as in Experiment 2, but in order to increase the time
participants had to process the memory-cues, we presented an additional filler screen (500ms)
No Directed Forgetting of Stimulus-Response Associations 41
that informed participants that a new object would be presented (“Next Round!”). In contrast to
Experiment 2, stimuli, responses, and memory instructions were paired only once during the
learning phase. Furthermore, we split the 64 object images over two blocks to facilitate the
memorization task. That is, in the learning phase, participants responded in total to 64 object
images over two blocks (32 stimuli each). The test phase, following the distractor task, was the
same as in Experiment 1 and 2 with the exception that participants were presented not only with
all 64 stimuli of the learning phase but also with 32 new stimuli (all intermixed in two blocks).
Furthermore, like in Experiment 2, participants did not receive any memory cues during the test
In addition to the post-experimental questions already used in Experiment 1 and 2, we
further checked whether participants correctly understood the task by asking them about their
comprehension of the instructions and the task. Specifically, we asked participants in the
memory group what they were supposed to do and what they actually did when they saw the
Results of Bayesian Analysis
Any comparison concerning effects for forget cues in the no-memory group were not
meaningful as many participants reported interpreting the forget cues as “error” feedback
resulting in no measurable item-specific S-R effects (see Figure 5). Hence, we refrained from
analyzing trials where forget cues were presented in the no-memory group (but we nevertheless
plotted the data in Figure 5 and analyzed trials with a remember cue in the no-memory group). In
the memory group, untransformed RTs for TBR stimuli were Mrepetition = 641ms and
Mswitch = 654ms and for TBF stimuli, Mrepetition = 642ms and Mswitch = 646ms. The mean RT for
new stimuli was Mnew = 665ms.
No Directed Forgetting of Stimulus-Response Associations 42
First, we fitted a Bayesian LMMs on participants’ log-RTs on TBR trials only, including
the fixed factors response (switch vs. repetition) and group (memory vs. no-memory), as well as
their interaction. The model further included random intercepts for participants and stimuli as
well as a by-participants slope for the factor response. To analyze whether item-specific S-R
effects for items followed by a remember cue differed between the memory vs. no-memory
group, we tested whether a model including the interaction between the factors response and
group was superior to a model without this interaction, which was not the case (BF01 = 14.6).
In a second model, we analyzed the effect of instruction (remember vs. forget) on the
data from the memory group only (that is, we excluded all data from the non-memory group).
There, we could replicate the findings of Experiment 1 and 2: The instruction to remember or
forget the stimulus did not affect item-specific S-R effects (BF01 = 11.1). This time, however,
there was generally no noteworthy difference between response repetitions and response
No Directed Forgetting of Stimulus-Response Associations 43
switches (BF01 = 8.3). Again, there was no difference in RTs between TBR and TBF stimuli
(BF01 = 17.0).
Estimated (model-based) Mean Log Reaction Times (RTs) Across Groups and Conditions in
In Experiment 3, we again found evidence against an influence of memory instructions
on item-specific S-R effects. Furthermore, we could show that the memory instruction itself
(dual task) did not obscure our findings, as RT differences between response repetitions and
response switches after a remember and forget cue did not differ between these participants who
Note. Error bars represent the within subject 95% confidence intervals.
No Directed Forgetting of Stimulus-Response Associations 44
received a memory instruction and those who did not. The goal of Experiment 4 was to replicate
the absence of an impact of directed forgetting on S-R associations using a different stimulus set
and task. Potentially, particularly the selective, verbal rehearsal of TBR information may impact
the encoding strength of S-R associations. Hence, in Experiment 4, we asked participants to
classify words (instead of images) as animate versus inanimate (instead of mechanic vs. non-
mechanic), thereby encoding S-R associations. At the same time, we instructed participants to
remember some of the words for a later memory test. Like in Experiment 1 and 2, we had only
one memory-group and no “control” group without memory instructions.
Methods, hypotheses, and the data analysis procedures were preregistered and can be
found online (https://osf.io/3nad2).
The same inclusion and exclusion criteria as in Experiment 3 were applied. From the
initial sample (N = 106), n = 69 remained in the final sample (Mage = 28.6, SD = 6, range: 18-39
years; 50.7% female, 46.4% male, 2.9% diverse).
Material and Procedure
Frequent English nouns referring to animate or inanimate concrete words with a length of
4 to 5 letters served as stimuli. Participants’ task was to classify the stimuli as animate/inanimate
and to remember/forget them according to the corresponding cue.
The experimental procedure was the same as for the memory group in Experiment 3 with
the following differences: The second practice block (the practice block with remember/forget
cues) was followed by a recall test for TBR words only and not by a recognition test.
No Directed Forgetting of Stimulus-Response Associations 45
As before, in each trial a task cue was presented for 700 ms that indicated the
classification-response mapping. For example, the task cue “A + I” indicated that a left response
was required for words referring to something animate and a right response for words referring
to something inanimate. Conversely, the task cue “I + A” indicated that a right response as
required for words referring to an animate entity and a left response for inanimate ones. Then,
the stimulus (the word) followed, and participants classified the stimulus via a left/ right key
press according to the preceding task cue. In total, participants responded to 64 words in one
block. Everything else (e.g., presentation times) was the same as for the memory-group in
Experiment 3 and there was only one pairing of stimuli, responses, and cues in the learning
Results of Bayesian Analyses
Mean log-RTs across conditions are shown in Figure 6. Untransformed RTs for TBR
stimuli were Mrepetition = 690ms and Mswitch = 691ms and for TBF stimuli, Mrepetition = 689ms and
Mswitch = 682ms. The mean RT for new stimuli was Mnew = 709ms. The instruction to remember
or forget the stimulus again did not affect item-specific S-R effects (BF01 = 18.4). Other than in
No Directed Forgetting of Stimulus-Response Associations 46
the previous experiments, however, there was no RT difference between response repetitions and
response switches (BF01 = 22.3; that is, we found no S-R effect).
Estimated (model-based) Mean Log Reaction Times (RTs) across Conditions in Experiment 4
To summarize, the results of Experiment 4 replicated the absence of an influence of the
intention to remember or to forget the stimulus on corresponding item-specific S-R effects (i.e.,
the underlying S-R associations) when using a different task (animacy task) as well as words
instead of images as stimuli. It should be noted, however, that item-specific S-R effects were
Note. Error bars represent the within subject 95% confidence intervals.
No Directed Forgetting of Stimulus-Response Associations 47
generally absent in this experiment. Nevertheless, RTs were still faster when responding to old
stimuli, indicating that item-specific memory representations were encoded and retrieved
(though apparently no response-related representations). Our finding that RTs generally did not
differ for TBR and TBF stimuli suggests that the intention to remember or forget had no impact
on these representations.
Decades of research on directed forgetting has well established that humans are able to
intentionally forget declarative memory content (e.g., R. A. Bjork, 1970, 1998; Geiselman, 1974;
Geiselman et al., 1983; MacLeod, 1998). The present study investigated to what extent the
intention to remember or to forget also affects procedural memory representations. To this end,
we studied whether the intentional remembering or forgetting of a stimulus affected responses
associated with that stimulus. To do so, we used a new experimental directed-forgetting design
for procedural S-R associations combining item-specific priming and the item-method of
In four experiments, the intention to remember or to forget a stimulus did not affect the
formation and/or retrieval of corresponding S-R associations. This was the case when S-R
associations already existed before the memory instruction was given (Experiment 1), when a
stimulus was paired multiple times with a response and the intention to remember or forget it
(Experiment 2), when learning new S-R associations with only one pairing (Experiment 3), as
well as when using a different categorization task with word stimuli (Experiment 4). This finding
has strong implications for our perspective on whether humans have intentional control over
what is stored in and retrieved from procedural memory – a fundamental question not only for
researchers investigating binding and retrieval processes in action control (e.g., the BRAC
No Directed Forgetting of Stimulus-Response Associations 48
framework, Frings et al., 2020), but also for our everyday life where daily trying to overcome old
habits is a challenge we often face.
Directed Forgetting of Stimulus-Response Associations
In contrast to the few previous studies that used the list-method of directed forgetting to
investigate intentional forgetting of motor memory (Dreisbach & Bäuml, 2014; Schmidt et al.,
2021; Tempel & Frings, 2016), we found no effect of directed forgetting on procedural memory
representations (as measured by item-specific S-R effects) when using the item-method of
directed-forgetting procedure. These findings are not necessarily in conflict: Whereas forgetting
costs in the list-method directed-forgetting procedure have been associated with impaired
retrieval of the TBF learning episodes (e.g., Sahakyan & Kelley, 2002), item-method directed-
forgetting effects have been attributed to differences in encoding of TBF and TBR information
(e.g., Fellner et al., 2020; Tan et al., 2020). The present study assessed how the up- or
downregulation of memory strength for the stimulus representation affected the corresponding S-
R associations. To investigate this question, item-method directed forgetting was the more
suitable method as it allowed us to assess the effect of remembering and forgetting shortly after
responding to a stimulus. That is, we could investigate the impact of intention on S-R
associations immediately after they had been formed (and were likely still in working memory)
and/or when they were consolidated into long-term memory. Our findings suggest that, whereas
retrieval of a whole learning episode of procedural information such as a S-R association could
be intentionally impaired (e.g., Dreisbach & Bäuml, 2014; but see Dames et al., 2022 for a
critical discussion of the applied method and a failure to replicate the effect), the encoding
strength of individual S-R associations does not seem to differ depending on the intention to
remember or forget the stimulus.
No Directed Forgetting of Stimulus-Response Associations 49
Implications for Research on Action Control and Memory
In the introduction, we contrasted three different accounts regarding how the instruction
to remember or forget the stimulus could affect S-R associations/bindings (e.g., in a respective
event-file). Applying the observations previously made in directed-forgetting paradigms
assessing declarative memory content (see MacLeod, 1998)
to the context of S-R associations,
one may have assumed that performance differences between item-specific response repetitions
and response switches (item-specific S-R effects indicating S-R associations) in the test phase
would be smaller for TBF as compared to TBR stimuli. This was not the case. Thus, our results
speak against the notion of a global impact of intention on the encoding or retrieval of an event-
file. That is, once features of an event become bound together (presumably in a common
representational format e.g., Hommel, 2004; Hommel et al., 2001), up- or downregulating the
memory strength of one component of the event-file does not influence the associative weights
for other information bound to that stimulus. Instead, we consider two alternative accounts:
According to the specific impact account (Figure 2), representations integrated in an
event-file could be selectively up- or downregulated without affecting the associative weights
that bind stimulus features to response features in the same event-file. If so, this finding has
profound implications for the recently raised question which processes allow for the top-down
Note, that a recent study could also show that the intention to remember strengthens item-
context bindings (as compared to a process-only condition) even when using a deep semantic
processing task similar to the one used in the present study (Popov & Dames, 2022; please note
that there was however no forget instruction).
No Directed Forgetting of Stimulus-Response Associations 50
control of event-files (e.g., Frings et al., 2020). Our experiments suggest that information in an
event-file could be managed/controlled specifically. However, this conclusion would also imply
that it is still unclear how information that receives more attention/upregulated (e.g., via a
directed-forgetting instruction) could become more integrated into an event-file (or can be
intentionally removed from an event-file once it has been formed) in a top-down fashion. At
least, when intending to forget the stimulus, this does not seem to be possible.
It could be argued that the memory strength of information stored in an event-file can be
controlled when participants intend to remember or forget the S-R association itself and not just
the stimulus as in the current experiments. However, when instructing participants to remember
the link between stimuli and their responses, they may likely also store an explicit memory of the
S-R mapping in their declarative memory. Thus, it would be unclear whether any memory
instruction affected the encoding or retrieval of procedural (implicit) S-R associations or
declarative (explicit) S-R representations. It would be difficult to interpret the results of such an
experimental endeavor. Hence, the question remains whether we can control the memory
strength of features after they have been bound in an event-files at all. It does not seem unlikely
that such kind of implicit, procedural memory is stored “automatically” (e.g., like one-shot
learning of context information, Malmberg & Shiffrin, 2005) without the possibility of intention
interfering with this process (Burgess et al., 2017). This line of reasoning raises the question of
“where” and “how” S-R associations are represented in general, as we further discuss in the
According to alternative separate memory systems account, a declarative memory
instruction (e.g., such as asking participants to remember or forget stimuli for a later memory
test) operates on a separate copy of the stimulus representation and not on the representation of
No Directed Forgetting of Stimulus-Response Associations 51
the stimulus that is associated to the response and stored in procedural memory. That is, our
results would be in line with the assumptions that S-R associations (as a form of procedural
memory) and stimulus representations (as a form of declarative memory) are stored separately
without interfering with each other. Consequently, up- or downregulating the memory strength of
the declarative stimulus representation (e.g., as proposed by studies on directed forgetting of
declarative content; Fellner et al., 2020; and as proposed by the results of the declarative memory
test of Exp. 1) would not affect the memory strength of S-R associations which are kept
Whereas the scenario of a selective influence of information within an event-file is still
unexplored, the notion that memory can be separated into a declarative and a procedural part has
been previously proposed (e.g., in working memory; Oberauer, 2009). For instance, according to
Oberauer (2009), working memory can be separated into a declarative part which maintains the
objects of our thoughts (e.g., bindings between object and context features) and a procedural part
which is responsible for the mental and motor operations performed on this information. The
stimulus information for the later memory test could be maintained in declarative working
memory and S-R bindings could be held in procedural working memory. However, it has also
been argued that procedural and declarative working memory operate in analogous ways
(Oberauer et al., 2013). Our finding may hint towards a dissociation between the two memory
systems: while we may have intentional control over what is kept in declarative working memory
(Dames & Oberauer, 2022), this may not be the case for procedural working memory. Future
studies are needed to further investigate the impact of directed forgetting on procedural working
memory to test this hypothesis. First evidence for this notion comes from a recent study by
Abrahamse et al. (2022) who observed that task-rules still linger in working memory even after
No Directed Forgetting of Stimulus-Response Associations 52
becoming irrelevant. Their finding suggests that, at least in working memory, our cognitive
system has little control over the impact of newly-encoded S-R rules. Extending Abrahamse et al.
(2022)’s finding in working memory to long-term memory, both their study and the present study
do not provide support for the existence of intentional control at the level of procedural working
memory or long-term memory.
Taken together, our study calls for a more explicit description of (1) how and (2) where S-R
bindings/associations and (corresponding) event-files are stored in the short- and in the long-
term. As long as researchers cannot come up with or agree on such a representational format, it is
difficult to speculate on how top-down processes could affect the memory strength of
associations incorporating such procedural representations. Clearly, more work is needed to
clarify previously-formulated claims regarding how top-down processes may affect action
control (see Frings et al., 2020).
In the present study, we showed that specifically up- or downregulating memory strength for
stimulus features does not impact the encoding or retrieval of the associated response (as
measured via item-specific S-R effects). This finding suggests that we either can specifically
control the memory strength of components integrated in an event-file without this affecting the
associative weight of information bound to that component (provoking numerous further
questions and assumptions on how S-R associations are represented, as discussed above).
Alternatively, our study may be taken as evidence that procedural S-R associations are stored in a
memory system separate from declarative stimulus representations, so that processes operating
on one or the other representation do not affect each other. If so, our findings question the
assumption that we have any intentional control at all over the encoding and retrieval of
No Directed Forgetting of Stimulus-Response Associations 53
procedural S-R associations. The present study thus demonstrates a failure of top-down
intentional control to impact on procedural representations incorporated into
bindings/associations and thereby provides novel insights into how top-down control can manage
what is bound into and retrieved from an event-file.
Abel, M., & Bäuml, K.-H. T. (2019). List-method directed forgetting after prolonged retention
interval: Further challenges to contemporary accounts. Journal of Memory and
Language, 106, 18–28. https://doi.org/10.1016/j.jml.2019.02.002
Abrahamse, E., Braem, S., De Houwer, J., & Liefooghe, B. (2022). Tenacious instructions: How
to dismantle newly instructed task rules? Journal of Experimental Psychology: General.
Abrahamse, E., Braem, S., Notebaert, W., & Verguts, T. (2016). Grounding cognitive control in
associative learning. Psychological Bulletin, 142(7), 693–728.
Ahmad, F. N., Tan, P., & Hockley, W. E. (2019). Directed forgetting for categorised pictures:
Recognition memory for perceptual details versus gist. Memory, 27(7), 894–903.
Baayen, R. H., Davidson, D. J., & Bates, D. M. (2008). Mixed-effects modeling with crossed
random effects for subjects and items. Journal of Memory and Language, 59(4), 390–
Bancroft, T. D., Hockley, W. E., & Farquhar, R. (2013). The longer we have to forget the more
we remember: The ironic effect of postcue duration in item-based directed forgetting.
No Directed Forgetting of Stimulus-Response Associations 54
Basden, B. H., & Basden, D. R. (1996). Directed Forgetting: Further Comparisons of the Item
and List Methods. Memory, 4(6), 633–653. https://doi.org/10.1080/741941000
Basden, B. H., Basden, D. R., & Gargano, G. J. (1993). Directed forgetting in implicit and
explicit memory tests: A comparison of methods. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 19(3), 603–616. https://doi.org/10.1037/0278-
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting Linear Mixed-Effects Models
Using lme4. Journal of Statistical Software, 67(1), 1–48. https://doi.org/10.18637/jss.v067.i01
Bjork, E. L., Bjork, R. A., & Anderson, M. C. (1998). Varieties of goal-directed forgetting.
Intentional Forgetting: Interdisciplinary Approaches., 103–137.
Bjork, R. A. (1970). Positive forgetting: The noninterference of Items intentionally forgotten.
Journal of Verbal Learning and Verbal Behavior, 9(3), 255–268.
Bjork, R. A. (1989). Retrieval inhibition as an adaptive mechanism in human memory. Varieties
of Memory and Consciousness: Essays in Honour of Endel Tulving., 309–330.
Bjork, R. A. (1998). Intentional forgetting in perspective: Comments, conjectures, and some
directed remembering. Intentional Forgetting: Interdisciplinary Approaches, 453–481.
Brady, T. F., Konkle, T., Alvarez, G. A., & Oliva, A. (2008). Visual long-term memory has a
massive storage capacity for object details. Proceedings of the National Academy of
Sciences, 105(38), 14325–14329. https://doi.org/10.1073/pnas.0803390105
Burgess, N., Hockley, W. E., & Hourihan, K. L. (2017). The effects of context in item-based
directed forgetting: Evidence for “one-shot” context storage. Memory & Cognition, 45(5),
No Directed Forgetting of Stimulus-Response Associations 55
Corsi, P. M. (1973). Human memory and the medial temporal region of the brain (Vol. 34, Issues
2-B, p. 891). ProQuest Information & Learning.
Dames, H., Brand, D., & Ragni, M. (2022). Evidence for Multiple Mechanisms Underlying List-
Method Directed Forgetting [Preprint]. PsyArXiv. https://doi.org/10.31234/osf.io/5gpmv
Dames, H., Kiesel, A., Pfeuffer, C. U., & Ragni, M. (2022). Intentional Forgetting of Habits?
Combining List-Method Directed Forgetting and Item-Specific Stimulus-Response
Priming. In J. Culbertson, A. Perfors, H. Rabagliati & V. Ramenzoni (Eds.), Proceedings
of the 44th Annual Conference of the Cognitive Science Society.
Dames, H., & Oberauer, K. (2022). Directed forgetting in working memory. Journal of
Experimental Psychology: General. Advance online publication.
Web browser. Behavior Research Methods, 47(1), 1–12. https://doi.org/10.3758/s13428-
Dreisbach, G., & Bäuml, K.-H. T. (2014). Don’t Do It Again! Directed Forgetting of Habits.
Psychological Science, 25(6), 1242–1248. https://doi.org/10.1177/0956797614526063
Fawcett, J. M., & Taylor, T. L. (2008). Forgetting is effortful: Evidence from reaction time
probes in an item-method directed forgetting task. Memory & Cognition, 36(6), 1168–
Fawcett, J. M., & Taylor, T. L. (2012). The control of working memory resources in intentional
forgetting: Evidence from incidental probe word recognition. Acta Psychologica, 139(1),
No Directed Forgetting of Stimulus-Response Associations 56
Fellner, M.-C., Waldhauser, G. T., & Axmacher, N. (2020). Tracking selective rehearsal and
active inhibition of memory traces in directed forgetting. Current Biology, 30(13), 2638-
Frings, C., Hommel, B., Koch, I., Rothermund, K., Dignath, D., Giesen, C., Kiesel, A., Kunde,
W., Mayr, S., Moeller, B., Möller, M., Pfister, R., & Philipp, A. (2020). Binding and
Retrieval in Action Control (BRAC). Trends in Cognitive Sciences, 24(5), 375–387.
Geiselman, R. E. (1974). Positive forgetting of sentence material. Memory & Cognition, 2(4),
Geiselman, R. E., Bjork, R. A., & Fishman, D. L. (1983). Disrupted retrieval in directed
forgetting: A link with posthypnotic amnesia. Journal of Experimental Psychology:
General, 112(1), 58–72. https://doi.org/10.1037/0096-34184.108.40.206
Gottlob, L. R., & Golding, J. M. (2007). Directed forgetting in the list method affects recognition
memory for source. Quarterly Journal of Experimental Psychology, 60(11), 1524–1539.
Hanczakowski, M., Pasek, T., & Zawadzka, K. (2012). Context-dependent impairment of
recollection in list-method directed forgetting. Memory, 20(7), 758–770.
Hauswald, A., & Kissler, J. (2008). Directed forgetting of complex pictures in an item method
paradigm. Memory, 16(8), 797–809. https://doi.org/10.1080/09658210802169087
Henson, R. N., Eckstein, D., Waszak, F., Frings, C., & Horner, A. J. (2014). Stimulus–response
bindings in priming. Trends in Cognitive Sciences, 18(7), 376–384.
No Directed Forgetting of Stimulus-Response Associations 57
Hockley, W. E., Ahmad, F. N., & Nicholson, R. (2016). Intentional and incidental encoding of
item and associative information in the directed forgetting procedure. Memory &
Cognition, 44(2), 220–228. https://doi.org/10.3758/s13421-015-0557-8
Hommel, B. (2004). Event files: Feature binding in and across perception and action. Trends in
Cognitive Sciences, 8(11), 494–500. https://doi.org/10.1016/j.tics.2004.08.007
Hommel, B. (2022). The Control of Event-File Management. Journal of Cognition, 5(1), 1.
Hommel, B., Müsseler, J., Aschersleben, G., & Prinz, W. (2001). The Theory of Event Coding
(TEC): A framework for perception and action planning. Behavioral and Brain Sciences,
24(5), 849–878. Cambridge Core. https://doi.org/10.1017/S0140525X01000103
Hourihan, K. L., Goldberg, S., & Taylor, T. L. (2007). The role of spatial location in
remembering and forgetting peripheral words. Canadian Journal of Experimental
Psychology/Revue Canadienne de Psychologie Expérimentale, 61(2), 91–101.
Hsu, Y.-F., & Waszak, F. (2012). Stimulus-classification traces are dominant in response
learning. International Journal of Psychophysiology, 86(3), 262–268.
Jarosz, A. F., & Wiley, J. (2014). What Are the Odds? A Practical Guide to Computing and
Reporting Bayes Factors. The Journal of Problem Solving, 7(1).
Judd, C. M., Westfall, J., & Kenny, D. A. (2012). Treating stimuli as a random factor in social
psychology: A new and comprehensive solution to a pervasive but largely ignored
No Directed Forgetting of Stimulus-Response Associations 58
problem. Journal of Personality and Social Psychology, 103(1), 54–69.
Lehman, M., & Malmberg, K. J. (2009). A global theory of remembering and forgetting from
multiple lists. Journal of Experimental Psychology: Learning, Memory, and Cognition,
35(4), 970–988. https://doi.org/10.1037/a0015728
Lenth, R., Singmann, H, Love, J., Buerkner, P., & Herve, M. (2018). Emmeans: Estimated
Marginal Means, AKA Least-squares Means. R Package Version 1.1. 3; 2018.
Liefooghe, B., Wenke, D., & De Houwer, J. (2012). Instruction-based task-rule congruency
effects. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38(5),
Logan, G. D. (1988). Toward an instance theory of automatization. Psychological Review, 95(4),
Logan, G. D. (1990). Repetition priming and automaticity: Common underlying mechanisms?
Cognitive Psychology, 22(1), 1–35. https://doi.org/10.1016/0010-0285(90)90002-L
Lüdecke, D. (2018). sjPlot: Data visualization for statistics in social science. R Package Version,
MacLeod, C. M. (1975). Long-term recognition and recall following directed forgetting. Journal
of Experimental Psychology: Human Learning and Memory, 1(3), 271–279.
MacLeod, C. M. (1998). Directed forgetting. Intentional Forgetting: Interdisciplinary
No Directed Forgetting of Stimulus-Response Associations 59
Malmberg, K. J., & Shiffrin, R. M. (2005). The “One-Shot” Hypothesis for Context Storage.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 31(2), 322–336.
Moors, A., & De Houwer, J. (2006). Automaticity: A Theoretical and Conceptual Analysis.
Psychological Bulletin, 132(2), 297–326. https://doi.org/10.1037/0033-2909.132.2.297
Morey, R. D., Rouder, J. N., Jamil, T., & Morey, M. R. D. (2015). Package ‘bayesfactor.’ URLh
Http://Cran/r-Projectorg/Web/Packages/BayesFactor/BayesFactor Pdf i (Accessed 1006
Moutsopoulou, K., Yang, Q., Desantis, A., & Waszak, F. (2015). Stimulus–classification and
stimulus–action associations: Effects of repetition learning and durability. Quarterly
Journal of Experimental Psychology, 68(9), 1744–1757.
Oberauer, K. (2009). Chapter 2 Design for a Working Memory. In Psychology of Learning and
Motivation (Vol. 51, pp. 45–100). Academic Press. https://doi.org/10.1016/S0079-
Oberauer, K., Souza, A. S., Druey, M. D., & Gade, M. (2013). Analogous mechanisms of
selection and updating in declarative and procedural working memory: Experiments and
a computational model. Cognitive Psychology, 66(2), 157–211.
Pastötter, B., & Bäuml, K.-H. (2010). Amount of postcue encoding predicts amount of directed
forgetting. Journal of Experimental Psychology: Learning, Memory, and Cognition,
36(1), 54–65. https://doi.org/10.1037/a0017406
No Directed Forgetting of Stimulus-Response Associations 60
Pastötter, B., Tempel, T., & Bäuml, K.-H. T. (2017). Long-Term Memory Updating: The Reset-
of-Encoding Hypothesis in List-Method Directed Forgetting. Frontiers in Psychology, 8,
Pfeuffer, C. U., Moutsopoulou, K., Pfister, R., Waszak, F., & Kiesel, A. (2017). The power of
words: On item-specific stimulus–response associations formed in the absence of action.
Journal of Experimental Psychology: Human Perception and Performance, 43(2), 328–
Popov, V., & Dames, H. (2022). Intent matters: Resolving the intentional versus incidental
learning paradox in episodic long-term memory. Journal of Experimental Psychology:
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning representations by back-
propagating errors. Nature, 323(6088), 533–536. https://doi.org/10.1038/323533a0
Sahakyan, L., & Delaney, P. F. (2003). Can encoding differences explain the benefits of directed
forgetting in the list method paradigm? Journal of Memory and Language, 48(1), 195–
Sahakyan, L., & Delaney, P. F. (2005). Directed Forgetting in Incidental Learning and
Recognition Testing: Support for a Two-Factor Account. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 31(4), 789–801.
Sahakyan, L., Delaney, P. F., Foster, N. L., & Abushanab, B. (2013). Chapter Four - List-Method
Directed Forgetting in Cognitive and Clinical Research: A Theoretical and
Methodological Review. In B. H. Ross (Ed.), Psychology of Learning and Motivation
No Directed Forgetting of Stimulus-Response Associations 61
(Vol. 59, pp. 131–189). Academic Press. https://doi.org/10.1016/B978-0-12-407187-
Sahakyan, L., & Kelley, C. M. (2002). A contextual change account of the directed forgetting
effect. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28(6),
Schmalbrock, P., Laub, R., & Frings, C. (2021). Integrating salience and action – Increased
integration strength through salience. Visual Cognition, 29(2), 91–104.
Schmidt, M., Frings, C., & Tempel, T. (2021). Selective directed forgetting of motor sequences.
Acta Psychologica, 218, 103352. https://doi.org/10.1016/j.actpsy.2021.103352
Tan, P., Ensor, T. M., Hockley, W. E., Harrison, G. W., & Wilson, D. E. (2020). In support of
selective rehearsal: Double-item presentation in item-method directed forgetting.
Psychonomic Bulletin & Review, 27(3), 529–535. https://doi.org/10.3758/s13423-020-
Tempel, T., & Frings, C. (2016). Directed forgetting benefits motor sequence encoding. Memory
& Cognition, 44(3), 413–419. https://doi.org/10.3758/s13421-015-0565-8
Verguts, T. (2017). Computational Models of Cognitive Control. In T. Egner (Ed.), The Wiley
Handbook of Cognitive Control (pp. 125–142). John Wiley & Sons, Ltd.
Whitlock, J., Chiu, J. Y.-C., & Sahakyan, L. (2022). Directed forgetting in associative memory:
Dissociating item and associative impairment. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 48(1), 29–42. https://doi.org/10.1037/xlm0001027
No Directed Forgetting of Stimulus-Response Associations 62
Whitlock, J., Lo, Y.-P., Chiu, Y.-C., & Sahakyan, L. (2020). Eye movement analyses of strong
and weak memories and goal-driven forgetting. Cognition, 204, 104391.
Woodward, A. E., Bjork, R. A., & Jongeward, R. H. (1973). Recall and recognition as a function
of primary rehearsal. Journal of Verbal Learning and Verbal Behavior, 12(6), 608–617.