Embodied Grounding of Memory: Toward the Effects of Motor Execution on Memory consolidation

Abstract

Behavioural and neuroscientific research has provided evidence for a strong functional link between the neural motor system and lexical-semantic processing of action-related language. It remains unclear, however, whether the impact of motor actions is restricted to online language comprehension or whether sensorimotor codes are also important in the formation and consolidation of persisting memory representations of the word's referents. The current study now demonstrates that recognition performance for action words is modulated by motor actions performed during the retention interval. Specifically, participants were required to learn words denoting objects that were associated with either a pressing or a twisting action (e.g., piano, screwdriver) and words that were not associated to actions. During a 6-8-minute retention phase, participants performed an intervening task that required the execution of pressing or twisting responses. A subsequent recognition task revealed a better memory for words that denoted objects for which the functional use was congruent with the action performed during the retention interval (e.g., pepper mill-twisting action, doorbell-pressing action) than for words that denoted objects for which the functional use was incongruent. In further experiments, we were able to generalize this effect of selective memory enhancement of words by performing congruent motor actions to an implicit perceptual (Experiment 2) and implicit semantic memory test (Experiment 3). Our findings suggest that a reactivation of motor codes affects the process of memory consolidation and emphasizes therefore the important role of sensorimotor codes in establishing enduring semantic representations.

Full text

This article was downloaded by: [Universitaetsbibliothek Potsdam]
On: 07 January 2014, At: 08:31
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
The Quarterly Journal of Experimental
Psychology
Publication details, including instructions for authors and subscription
information:
http://www.tandfonline.com/loi/pqje20
Embodied grounding of memory: Toward
the effects of motor execution on memory
consolidation
Wessel O. van Dama, Shirley-Ann Rueschemeyera, Harold Bekkeringa &
Oliver Lindemanna
a Donders Institute for Brain, Cognition and Behavior, Radboud University
Nijmegen, Nijmegen, The Netherlands
Accepted author version posted online: 28 Mar 2013.Published online: 11
Apr 2013.
To cite this article: Wessel O. van Dam, Shirley-Ann Rueschemeyer, Harold Bekkering & Oliver Lindemann
(2013) Embodied grounding of memory: Toward the effects of motor execution on memory consolidation, The
Quarterly Journal of Experimental Psychology, 66:12, 2310-2328, DOI: 10.1080/17470218.2013.777084
To link to this article: http://dx.doi.org/10.1080/17470218.2013.777084
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”)
contained in the publications on our platform. However, Taylor & Francis, our agents, and our
licensors make no representations or warranties whatsoever as to the accuracy, completeness, or
suitability for any purpose of the Content. Any opinions and views expressed in this publication
are the opinions and views of the authors, and are not the views of or endorsed by Taylor &
Francis. The accuracy of the Content should not be relied upon and should be independently
verified with primary sources of information. Taylor and Francis shall not be liable for any
losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities
whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial
or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use
can be found at http://www.tandfonline.com/page/terms-and-conditions

Embodied grounding of memory: Toward the effects of
motor execution on memory consolidation
Wessel O. van Dam, Shirley-Ann Rueschemeyer, Harold Bekkering, and
Oliver Lindemann
Donders Institute for Brain, Cognition and Behavior, Radboud University Nijmegen, Nijmegen,
The Netherlands
Behavioural and neuroscientific research has provided evidence for a strong functional link between the
neural motor system and lexical–semantic processing of action-related language. It remains unclear,
however, whether the impact of motor actions is restricted to online language comprehension or
whether sensorimotor codes are also important in the formation and consolidation of persisting
memory representations of the word’s referents. The current study now demonstrates that recognition
performance for action words is modulated by motor actions performed during the retention interval.
Specifically, participants were required to learn words denoting objects that were associated with either a
pressing or a twisting action (e.g., piano, screwdriver) and words that were not associated to actions.
During a 6–8-minute retention phase, participants performed an intervening task that required the
execution of pressing or twisting responses. A subsequent recognition task revealed a better memory
for words that denoted objects for which the functional use was congruent with the action performed
during the retention interval (e.g., pepper mill–twisting action, doorbell–pressing action) than for words
that denoted objects for which the functional use was incongruent. In further experiments, we were able
to generalize this effect of selective memory enhancement of words by performing congruent motor
actions to an implicit perceptual (Experiment 2) and implicit semantic memory test (Experiment 3).
Our findings suggest that a reactivation of motor codes affects the process of memory consolidation
and emphasizes therefore the important role of sensorimotor codes in establishing enduring semantic
representations.
Keywords: Embodiment; Semantics; Action; Memory.
How linguistic meaning is processed by the brain
remains an open question. Much research in the
field focuses on the nature of lexical–semantic rep-
resentations, or on the process of accessing
meaning from the mental lexicon; however, very
little has been done towards understanding
whether the formation and consolidation of persist-
ing memory representations of a word’s referents
are affected by cross-modal interactions between
perception/action and the meaning of the word to
be encoded. This last point is particularly relevant
given the large body of evidence recently amassed
Correspondence should be addressed to Wessel O. van Dam, P.O. Box 9104, 6500 HE Nijmegen, The Netherlands.
E-mail: VANDAMW@mailbox.sc.edu
The authors thank Pascal de Water and Gerard van Oijen for technical support and Markus van Ackeren for help in creating stimu-
lus material. The authors would also like to thank Eelco van Dongen, Michael Masson, Art Glenberg, and Diane Pecher for
helpful comments on an earlier draft of this manuscript. The study was supported by the Dutch Organization for Scientific
Research NWO-VICI grant (453-05-001) to the third author and NWO-VENI grant (016-094-053) to the last author.
2310 #2013 The Experimental Psychology Society
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013
Vol. 66, No. 12, 2310–2328, http://dx.doi.org/10.1080/17470218.2013.777084
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

in favour of so-called embodied theories of
language.
Embodied theories of language propose that
sensorimotor processes are involved in representing
lexical–semantic content (Barsalou, 1999;
Glenberg, 1997; Pulvermueller, 1999). For
example, action words are suggested to be rep-
resented in part in neural motor areas (e.g., Hauk
& Pulvermueller, 2004), while language denoting
specific visual features are suggested to activate
neural visual pathways (e.g., Simmons et al.,
2007). This position is fundamentally different
from a purely symbolic view, which suggests that
meaning is amodal (i.e., not based on action or per-
ception), and that lexical–semantic meaning can be
derived purely from other words (Kintsch, 2008).
In the past decade a plethora of studies using
various experimental techniques have lent weight
to the embodied language framework.
For instance, Stanfield and Zwaan (2001) show
that participants automatically engage in visual simu-
lation while reading. Specifically, they demonstrate
that the comprehension of sentences describing
objects in a certain orientation (e.g., John put the
pencil in the cup: implying a vertical orientation) facili-
tates responses to pictures of objects with the same
orientation (see also Zwaan, Stanfield, & Yaxley,
2002). In a similar vein, Glenberg and Kaschak
(2002) show that reading sentences describing
actions (e.g. “Open the drawer”) facilitates the execution
of congruent motor responses (in this example, a move-
ment of the hand towards the body), suggesting that
language comprehension has a direct and implicit
effect on a comprehender’s motor performance.
While these two studies investigated language–action
interaction at the sentence level, numerous studies
have shown that single words also modulate action per-
formance (Rueschemeyer, Pfeiffer, & Bekkering,
2010; Van Dam, Rueschemeyer, Lindemann, &
Bekkering, 2010).
Behavioural studies thus demonstrate that
language can indeed modulate action performance;
however, they do not specify the nature of the inter-
action between action and language. Neuroimaging
studies have demonstrated that words denoting
actions activate some neural areas also activated
by action execution. More specifically, the
comprehension of action verbs (Kemmerer,
Castillo, Talavage, Patterson, & Wiley, 2008;
Van Dam, Rueschemeyer, & Bekkering, 2010),
action sentences (e.g., Desai, Binder, Conant, &
Seidenberg, 2010), and words denoting manipul-
able objects (e.g., Chao & Martin, 2000;
Rueschemeyer, Rooij, Lindemann, Willems, &
Bekkering, 2010; Saccuman et al., 2006) reliably
activate the inferior parietal and/or premotor cor-
tices—that is, parts of the cerebral motor system.
On the basis of neuroimaging data it has therefore
been argued that language and action share
common conceptual representations in the brain.
A great number of language studies have recently
focused on the impact of sensorimotor activations
on lexical–semantic processing. In a similar vein,
ample evidence has been provided for the role of sen-
sorimotor activations in enhancing memory encod-
ing processes. Surprisingly, the role of sensorimotor
processes in memory consolidation processes has
been largely ignored. In particular, while sensorimo-
tor regions do appear to be important for retrieving
conceptual knowledge, it is unspecified whether
they aid in the establishment and consolidation of
lasting conceptual representations.
Previous research by Engelkamp and colleagues
demonstrates, for instance, that action phrases or
nouns are better recalled when participants perform
the denoted motor action or carry out a prototypical
action corresponding to each noun (subject per-
formed tasks) than when executing an unrelated
action or a verbal task while encoding (Engelkamp,
1997; Engelkamp & Zimmer, 1983; Zimmer &
Engelkamp, 2003). Engelkamp (1997) furthermore
shows that the effect of self-performed tasks is not
merely due to an intention to act, but due to the
actual execution of an action during the study
phase. This enactment effect has been accounted
for by controversial explanations like the proposals
that memory for action should be regarded as a
different class of memory events from verbal
memory (e.g., Cohen, 1981), that memory for
action events is particularly rich due to its multimodal
nature (e.g., Bäckman, Nilsson, & Chalom, 1986),
and that a specialized action output system is
involved in subject-performed tasks but not in
verbal tasks (Engelkamp & Zimmer, 1983).
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2311
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

However, independent from this debate, the studies
reviewed above suggest that encoding of action-
related language can benefit from sensorimotor pro-
cesses specific to the action denoted by the sentence
or prototypical action corresponding to the noun.
Some further evidence for the role of sensorimo-
tor representations in memory processing has been
provided recently by Pecher, Van Dantzig, Zwaan,
& Zeelenberg (2009). The authors asked their par-
ticipants to make sensibility judgements on sen-
tences that implied either a particular shape or
orientation of an object (e.g., The handyman made
a hole in the wall with his drill, implying a drill in a
horizontal orientation) and found that the perform-
ance in a surprise picture recognition task after a 40–
50-minute delay was better for visually presented
objects that matched in shape or orientation with
the shape or orientation implied by the sentence
during the previous incidental encoding phase. In
addition, Pecher, Zanolie, and Zeelenberg (2007)
showed that recognition memory for pictures of
concepts was enhanced by earlier presentation of
visual properties of those concepts. The above-men-
tioned results suggest that the influence of sensory
simulations during language processing is not
restricted to online language comprehension and
continues to be evident after longer delays (but see
also Zwaan & Taylor, 2006).
Previous studies thus have provided evidence for
the role of enactment and sensory simulation on
memory encoding and the role of sensorimotor pro-
cesses in lexical–semantic comprehension (e.g.,
Rueschemeyer, Rooij, et al., 2010). A question
that remains is whether memory for action-related
semantic information (e.g., words denoting actions
or tools) crucially depends on direct associations
with sensory and motor codes (see Hommel,
Muesseler, Aschersleben, & Prinz, 2001, for a
similar notion of shared representations in the
context of visual perception and cognitive control).
In order to test the hypothesis that memory for
action-related information is based on sensorimotor
codes, we investigated the impact of motor behav-
iour on the processes of memory consolidation.
That is, the current study examined whether the
execution of different actions and the resulting acti-
vation of specific sensorimotor codes during a
retention interval affect recognition performance
for words denoting tools in a delayed recognition
task. To this end, we instructed our participants to
remember action-related object words (study
phase). The typical functional use of these objects
was associated either with a pressing (e.g., doorbell)
or with a twisting (e.g., pepper mill) action. As a
control condition, we presented object words that
were not related to a motor action (e.g., fence).
Before the memory performance was tested in a rec-
ognition task (test phase), participants were asked to
perform one of two motor actions (i.e., twisting or
pressing) as response to a go/no-go magnitude
classification task (retention phase).
In Experiment 1, we investigated whether recog-
nition performance for words of the two action cat-
egories interacts with the type of motor actions
(congruent versus incongruent movement) per-
formed during the retention interval. Specifically,
we expect that the sensitivity or memory strength
(measured by means of d′) will be greater for words
that denote objects for which the functional use is
congruent with the action performed during the
retention interval than for words that denote objects
for which the functional use is incongruent with the
action performed during the retention interval.
Experiments 2 and 3 elaborated on this effect in
further detail. In particular we determine whether
an effect of congruent actions during the retention
phase is the result of an enhancement in memory
consolidation or memory retrieval processes.
EXPERIMENT 1
Method
Participants
In the first experiment, 21 right-handed subjects
(average age of 20.6 years) participated. All subjects
were students at the Radboud University Nijmegen
and received 7.50 euros or course credits for their par-
ticipation. No subject was aware of the purpose of the
experiment. The data from one subject were excluded
from analysis because of an excessive amount of errors
(overall accuracy level of less than 55%).
2312 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

Stimuli
All word and number stimuli were presented
centrally in white Arial fonts on a black back-
ground. The average length of the words in the
study and test phase was 9 letters. From a
viewing distance of 100 cm, the word stimuli
subtended on average a visual angle of 0.57 ×
2.86 degrees. The experimental stimuli were
comprised of 75 real Dutch words denoting fam-
iliar objects (stimuli with English translations can
be seen in Table 1). Stimuli belonged to one of
three experimental conditions: (a) words denoting
Table 1. Dutch words associated with a pressing movement, a turning movement, or no movement
Pressing words Schroevendraaier Screwdriver
Piano Piano Puntenslijper Pencil sharpener
Orgel Organ Opwindklok Wind-up clock
Flipperkast Pinball machine Contactslot Ignition switch
Toetsenbord Keyboard Bankschroef Vice
Rekenmachine Calculator Moer Nut
Deurbel Doorbell Waterkraan Water tap
Muisknop Mouse button Waterpomptang Water pump pliers
Spuitbus Spray can Slagroomklopper Cream whipper
Pinapparaat Electronic payment terminal Tafelvoetbalbak Table football
Mobieltje Mobile phone Bout Bolt
Stopwatch Stopwatch Kauwgomballenautomaat Chewing gum machine
Perforator Perforator
Stemkastje Voting machine Nonmanipulable words
Typemachine Typewriter Huis House
Fototoestel Camera Heide Heath/moor
Laptop Laptop Pilaar Pillar
Nietmachine Stapler Plein Square
Claxon Horn Hotel Hotel
Pepperspray Pepperspray Terras Outdoor cafe
Punaise Drawing pin Vijver Pond
Touchscreen Touchscreen Blokhut Log cabin
Klavecimbel Harpsichord Piramide Pyramid
Aanknop On-button Rotsblok Boulder
Cassetterecorder Recorder Zendmast Transmitter mast
Afstandsbediening Remote control Schutting Fence
Standbeeld Statue
Twisting words Stoeprand Kerbstone
Tol Top Kerktoren Church tower
Stuur Steering wheel Regenpijp Drainpipe
Garde Rod/birch Strandtent Beach kiosk
Ventiel Air-valve Koeienstal Cowshed
Pepermolen Pepper mill Uithangbord Sign board
Huissleutel Key Kasteelmuur Castel wall
Kurkentrekker Bottle opener Bloementuin Flower garden
Passer Compasses Lantaarnpaal Lamppost
Jampot Jar Treinstation Railway station
Volumeknop Volume button Rioleringsbuis Sewage pipe
Blikopener Tin opener Schotelantenne Satellite dish
Cijferslot Combination lock
Eierwekker Egg timer
Note: Pressing movement: pressing words. Turning movement: twisting words. No movement: nonmanipulable words. English
translations are printed in italics.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2313
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

objects for which the functional use is associated
with a pressing movement (e.g., typewriter); (b)
words denoting objects for which the functional
use is associated with a turning movement (e.g.,
pepper mill); and (c) words denoting objects that
were not associated with any movement (e.g.,
fence).
To test that stimuli were truly perceived as
being associated with a particular action, an
online questionnaire was administered to 36
native Dutch speakers who did not participate in
the subsequent experiment (see Table 2 for
results). In this questionnaire, participants were
asked (a) whether they thought the word denoted
an object that was associated with an action (1 =
yes, 2 =no), and (b) if yes, with which type of
action (1 =pressing action, 2 =twisting action,
3=other type of action). Participants confirmed
that pressing and twisting words denoted objects
that are highly associated with an action, whereas
nonmanipulable words were not (percentage of
yes-responses: pressing, M=92.3; twisting, M=
87.9; nonmanipulable, M=4.4; all means signifi-
cantly differed from 50% as indicated by one-
sample ttests; all ps ,.001). Additionally, partici-
pants confirmed that pressing words denoted
objects that are highly associated with a pressing
action (pressing action: M=83.8; twisting
action: M=1.0; other type of action: M=15.2),
and twisting words denoted objects highly associ-
ated with a twisting action (pressing action: M=
8.5; twisting action: M=71.7; other type of
action: M=19.9). Paired-sample ttests indicated
that pressing words had a stronger association
with a pressing than with a twisting, t(35) =
23.08, or other type of action, t(35) =10.69,
whereas twisting words had a stronger association
with a twisting than with a pressing, t(35) =
13.62, or other type of action, t(35) =7.42.
Words were also matched across conditions for
length (pressing: M=9.6 letters; twisting: M=
10.0; nonmanipulable: M=8.8), F(2, 23) =1.76,
p=.194.
In order to obtain a measure of word frequency,
we calculated the mean logarithmic lemma fre-
quency per million for each condition using the
CELEX lexical database (Baayen, Piepenbrock,
& van Rijn, 1993). This value was 0.55 for pressing
words, 0.61 for twisting words, and 3.08 for non-
manipulable object words. A repeated measures
analysis of variance (ANOVA) indicated that
there were no reliable differences in the mean log-
arithmic lemma frequency per million between any
of the experimental conditions, F(2, 23) =2.84,
p=.079.
As number stimuli for the go/no-go magnitude
classification task in the retention phase, we used
the Arabic digits 1 to 9 (except the 5). Each digit
was presented 24 times, resulting in a total
number of 192 trials.
Design
The experiment comprised 25 object words from
each condition (pressing, twisting, and neutral),
which were all presented during the test phase of
the experiment. A subset of this word list (15 in
each condition) was presented during the study
phase. Therefore, during the test phase there were
15 “old”items (also presented during study phase)
and 10 “new”items (not presented during the
study phase) for each condition.
Procedure
An experimental session consisted of three con-
secutive phases. No breaks were made between
the phases. The experiment lasted on average
about 35 min.
Study phase. Participants were instructed to memor-
ize a list of 45 object words (i.e., 15 words from
each condition). Each word was presented for
Table 2. Mean ratings of the pretests
Measure Pressing Twisting Nonmanipulable
Length (no. letters) 9.6 10.0 8.8
Log lemma frequency 0.55 0.61 3.08
Associated with an
action (%)
92.3 87.9 4.4
Pressing action (%) 83.8 1.0 15.2
Twisting action (%) 8.5 71.7 19.9
Other type of action (%) 0 0 100.0
2314 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

7,000 ms preceded by a fixation cross (500 ms).
Words were presented in random order.
Retention phase. After studying the word list, par-
ticipants performed a go/no-go numerical magni-
tude classification task. Each trial started with a
fixation cross (500 ms). Afterwards, a target
number was presented for a maximum of 2,000
ms, and participants were required to indicate
whether the number was smaller or larger than
5. Importantly, number judgements had to be
made by means of one of two actions. That is,
half of the participants responded by means of a
twisting action, and the other half of the partici-
pants responded by means of a pressing action.
The twisting action was performed on a device
that had to be rotated for 90 degrees in a clockwise
fashion. The pressing action was performed on a
response button. After an intertrial interval of 500
ms, the next trial started. The trail order was ran-
domized. The retention phase lasted on average
about 6–8 minutes.
Test phase. At the end of the experiment, partici-
pants performed a delayed recognition memory
task on all 75 object words (see Table 1 for the
full list of stimuli). Each trial started with the pres-
entation of a fixation cross for 500 ms. Then the
object word was presented and remained visible
until the participant responded or a maximum of
3,000 ms. Participants indicated whether the
object word was already presented in the test
phase by pressing the left arrow key (“old”) on the
computer keyboard or whether it was a newly pre-
sented word (“new”) by pressing the right arrow
key. The order in which the object words were
presented was randomized. The experiment com-
prised 25 object words from each condition (press-
ing, twisting, and neutral) from which 15 object
words had been presented during the study phase
—that is, 15 “old”items and 10 “new”items from
each condition.
Results
The mean hit rates (i.e., responding “old”to a pre-
viously presented object word) and mean
false-alarm rates (i.e., responding “old”to a newly
presented object word) in the recognition task are
presented in Table 3. In order to obtain measures
of sensitivity or memory strength for a particular
word category, we calculated the d′scores (see
MacMillan & Creelman, 1991) for each condition
and each participant (see Figure 1 for means). The
d′scores were submitted to a two-way mixed-
model ANOVA with the between-subject factor
action (twisting vs. pressing) and within-subject
factor congruency (congruent, neutral, and incon-
gruent). The significance criterion for all analyses
reported here was set to α=.05.
In line with our hypothesis, the ANOVA
revealed a significant main effect of congruency,
F(2, 18) =8.31, p=.003, MSE =0.260,
η
2
p
=.480. To further explore this main effect, we
calculated post hoc paired-sample ttests. Words
for which the functional use was congruent with
the action performed during the retention interval
showed a higher d′than words that denoted
objects for which the functional use was incongru-
ent, t(20) =4.15, p=.0005. No difference was
observed between congruent and nonmanipulable
object words, t(20) =1.72, p=.102, or between
Table 3. Average hit rates, false-alarm rates, and dprimes for the items of Experiment 1 for the factors congruency and action
Pressing Twisting
Congruency Hit rates False-alarm rates d′Hit rates False-alarm rates d′
Congruent .83 (.03) .06 (.02) 2.57 (0.19) .82 (.04) .06 (.02) 2.58 (0.20)
Neutral .79 (.06) .13 (.04) 2.43 (0.26) .81 (.04) .06 (.02) 2.04 (0.35)
Incongruent .76 (.04) .04 (.02) 2.30 (0.12) .77 (.04) .04 (.02) 2.14 (0.15)
Note: Congruency: congruent, neutral, and incongruent words. Action: pressing versus twisting action. Standard errors in parentheses.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2315
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

incongruent and nonmanipulable object words,
t(20) =0.11, p=.917. There was no main effect
for the factor action, p=.487, and no interaction
between the factors, p=.519. Post hoc simple
effect analyses revealed that d′was higher for press-
ing words than for twisting words in the pressing
condition, t(10) =2.64, p=.025, and that d′was
higher for twisting words than for pressing words
in the twisting condition, t(9) =3.19, p=.011.
Discussion
The results of Experiment 1 showed that sensi-
tivity or memory strength (measured by means of
d′)washigherforwordsforwhichthefunctional
use was congruent with the action performed
during the retention phase than for words for
which the functional use was incongruent with
the action performed during the retention phase.
The d′scores for the nonmanipulable object
words strongly resembled the scores of the incon-
gruent items; however, probably due to a high var-
iance within the nonmanipulable object condition,
the difference between the nonmanipulable object
condition and the congruent condition did not
reach significance. These findings suggest that a
reactivation of motor information during
retention interval affects memory consolidation
and the strength of a specificmemorytrace.It
could, however, be argued that the performance
of a specific action during the retention phase
merely leads to a heightened familiarity of all
items associated with a congruent action, irrespec-
tive of an existing memory trace for a specificitem.
Thefactthatweobtainedaneffectinourd′(i.e., a
bias free measure) renders this possibility highly
unlikely.
EXPERIMENT 2
The results of the first experiment showed that
memory performance in an explicit episodic recog-
nition test is enhanced for words that denote
objects for which the functional use is congruent
with the action performed during the retention
interval. That is, we show that performing an
action during the retention phase of an explicit epi-
sodic recognition task enhances memory for words
with an action feature congruent to the action per-
formed. For example, the word pepper mill is
remembered better after a retention phase in
which the participant performed a twisting
motion with his/her hand than after a retention
Figure 1. Average d′s for congruent, neutral, and incongruent words in both the pressing and twisting condition of Experiment 1 (pressing
versus twisting action during retention phase).
2316 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

phase in which the participant performed an action
not usually executed in conjunction with pepper
mills.
There are two possible explanations for this
effect: (a) Sensorimotor processes enhance
memory consolidation by strengthening and
enriching associations of the memory items to sen-
sorimotor information during the retention phase.
This would be in line with reconsolidation theory
suggesting that memory reactivation provides the
possibility to modify and update old memories, by
means of a transient destabilization of an existing
memory trace. (b) Alternatively, the action con-
gruency effect might emerge during memory retrie-
val in the recognition test, because participants were
semantically primed and biased in their recollection
processes due to salient action treatment during the
retention phase.
The aim of the second experiment was to dis-
tinguish between these two alternatives. If the
action congruency effect is driven by differences
in encoding associations between semantic and
sensorimotor codes, then action congruency
effects should not be restricted to tests measuring
explicit episodic memory performance and general-
ize to implicit tests as well as to tests that are sensi-
tive to perceptual priming. Consequently, we varied
the task requirements of the test phase and
employed an implicit perceptual memory test in
Experiment 2. To do so, we used the picture frag-
ment completion test (cf. Snodgrass & Feenan,
1990), in which participants are required to name
pictures that are initially masked and become
increasingly more recognizable. In this task, partici-
pants typically show a perceptual priming effect—
that is, their performance is higher for items that
were previously perceived (i.e., old items) than for
new items. Importantly, however, picture fragmen-
tation tests are insensitive to semantic priming
(Jacoby, Baker, & Brooks, 1989; Roediger &
McDermott, 1993). The size of the perceptual
priming effect is considered to be a measure of
implicit memory given that participants may con-
sciously retrieve the study experience but, hence,
are not required to in order to show priming
(Snodgrass & Feenan, 1990).
Method
Participants
In the second experiment, 40 right-handed subjects
(average age of 19.5 years) participated. All subjects
were students at the Radboud University Nijmegen
and received 7.50 euros or course credits for their
participation. No subject was aware of the
purpose of the experiment.
Stimuli
Fifty-four pictures corresponding to a subset of
the 75 object words (pictures from the three
different conditions were matched with regard
to recognition rates, leaving 18 stimuli for each
condition) were created. The drawings were pre-
sented in black and white or greyscale. For a
viewing distance of 80 cm, the stimuli subtended
a visual angle of 7.15 ×5.11°. To test that the
picture stimuli were matched with regard to rec-
ognition difficulty, an online questionnaire was
administered to 31 native Dutch speakers who
did not participate in the subsequent experiment.
In this pretest we tested how easily participants
recognized each picture, by letting them specify
the correct name corresponding to each picture.
Subsequently, we calculated the mean percentage
correct responses for each condition. A repeated
measures ANOVA indicated that there were no
differences in the recognition difficulty of pic-
tures (pressing action: M=88.5; twisting
action: M=85.5; other type of action: M=
86.0) of the experimental conditions, F(2,
29) =2.27, p=.121.
For the picture fragment completion test
(Snodgrass & Feenan, 1990), stimuli were
drawn from a set of 54 pictures of objects (18
stimuli for each condition). We applied a frag-
mentation algorithm, which plotted on each
picture (340 ×260 pixel in size) subsequently
221 black squares (20 ×20 pixels in size). Since
newly added squares never overlapped with pre-
viously plotted ones, the picture was at the end
fully masked. The order of the square positions
was randomized. Using this method, we con-
structedaseriesofimagesat20different
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2317
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

fragmentation levels (0% to 100% masked). At
the moment that the first picture appeared it
was fully masked; every 150 ms an additional
5% of pixel blocks became visible.
Design
The design was identical to that of Experiment 1
except that the experiment comprised 18 objects
from each condition (pressing, twisting, and
neutral). That is, a subset of 18 from the 25
object words were presented as pictures (the
subset of 18 pictures from each condition were
matched with regard to recognition difficulty).
Half of the pictures from each condition (i.e., 9
“old”items) had already been presented as words
during the study phase. During the test phase
there were 9 “old”items and 9 “new”items for
each condition.
Procedure
An experimental session consisted of three con-
secutive phases. No breaks were made between
the phases. The experiment lasted on average
about 35 minutes.
Study phase. Participants were instructed to memor-
ize a list of 45 object words (i.e., 9 words of each
condition) and an additional 18 filler items. Each
word was presented for 7,000 ms proceeded by a
fixation cross (500 ms). Words were presented in
random order.
Retention phase. The retention phase of Experiment 2
was identical to the retention phase in Experiment 1.
Test phase. Instead of a delayed recognition
memory task, participants performed an implicit
memory test—namely, a picture fragment
Figure 2. Trial overview for Experiment 2.
2318 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

completion test (Snodgrass & Feenan, 1990).
Participants started a trial by keeping the middle
button of a button box depressed; after 2,000 ms
a picture emerged at the centre of the screen (see
Figure 2 for a trial overview). At the moment
the picture first appeared it was fully masked;
every 150 ms an additional 5% of the picture
was unmasked. That is, the participants had the
impression of a gradually emerging picture out of
a black background. Participants were instructed
to react as fast as possible if they identified the
object that was depicted. Participants responded
by releasing the middle button of the button
box. Participants were able to respond up to 2 s
after the picture was fully visible. The order in
which the object pictures were presented during
the picture identification task was randomized.
The experiment comprised 18 object pictures
from each condition (pressing, twisting, and
neutral) from which 9 object words had been pre-
sented during the study phase—that is, 9 “old”
items and 9 “new”items from each condition.
Crucially, the lists of “old”and “new”pictures
within a category were counterbalanced over
participants.
Results
We calculated the mean picture identification accu-
racy and clarification level per condition for each
participant (see Table 4). The picture identification
accuracy and clarification level scores were averaged
for each participant in each condition (see Figure 3
for means) and were submitted to a two-way
repeated measures ANOVA with the factors
action (twisting vs. pressing) and congruency (con-
gruent, neutral, and incongruent).
In line with our hypothesis, the ANOVA on
picture identification accuracy revealed a signifi-
cant main effect of congruency, F(2, 37) =5.50,
p=.008, MSE =87.25, η
2
p
=.229. To further
explore this main effect, we calculated post hoc
paired-sample ttests. Words for which the func-
tional use was congruent with the action per-
formed during the retention interval showed a
higher accuracy than words that denoted objects
for which the functional use was incongruent,
t(39) =2.39, p=.022. No difference was observed
between congruent and nonmanipulable object
words, t(39) =1.09, p=.284. Nonmanipulable
object words showed a higher accuracy than
words for which the functional use was incongru-
ent with the action performed during the retention
interval, t(39) =2.92, p=.006. Post hoc simple
effect analyses revealed that accuracy was higher
for pressing words than for twisting words in the
pressing condition, t(19) =2.60, p=.017,
whereas simple effect analyses did not reveal a
difference in the accuracy of pressing and twisting
words in the twisting condition, t(19) =0.96,
p=.349.
A repeated measures ANOVA on the clarifica-
tion level scores revealed a significant effect of con-
dition, F(2, 37) =15.67, p,.001, MSE =21.27,
η
2
p
=.459. Post hoc paired-sample ttests revealed
no difference in the clarification level scores
between congruent and incongruent object words,
t(39) =0.69, p=.495. Nonmanipulable object
words had a lower clarification level score than con-
gruent object words, t(39) =3.57, p=.001, and
nonmanipulable object words had a lower
Table 4. Average recognition accuracy and percentage clarification level for the items of Experiment 2 for the factors congruency and action
Congruency
Pressing Twisting
Accuracy (%) Clarification (%) Accuracy (%) Clarification (%)
Congruent 87.0 (2.57) 73.23 (2.26) 89.5 (2.85) 72.66 (2.09)
Neutral 93.0 (1.70) 68.84 (2.57) 87.5 (2.43) 69.05 (1.90)
Incongruent 80.0 (2.67) 75.35 (2.29) 87.5 (2.53) 72.04 (1.89)
Note: Congruency: congruent, neutral and incongruent words. Action: pressing versus twisting action. Standard errors in parentheses.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2319
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

clarification level score than incongruent object
words, t(39) =5.32, p,.001.
It could, however, be argued that participants are
merely better at detecting anything that is associ-
ated with a congruent action, irrespective of an
existing memory trace for a specific item. A
repeated measures ANOVA on the picture identi-
fication accuracy to “new”(i.e., pictures not pre-
sented as word during the learning phase) object
pictures revealed a main effect of congruency,
F(2, 37) =5.72, p=.007, MSE =124.06,
η
2
p
=.236. To further explore this main effect, we
calculated post hoc paired-sample ttests. No differ-
ence was observed between congruent and incon-
gruent object words, t(39) =0.74, p=.466.
Nonmanipulable objects words showed a higher
picture identification accuracy than words for
which the functional use was congruent with
the action performed during the retention interval,
t(39) =2.24, p=.031, and nonmanipulable object
words showed a higher picture identification accu-
racy than words for which the functional use was
incongruent with the action performed during the
retention interval, t(39) =3.21, p=.003.
Discussion
The results of Experiment 2 showed that picture
identification accuracy was higher for pictures cor-
responding to congruent object words than for pic-
tures corresponding to incongruent object words.
Picture fragmentation tests are known to be insen-
sitive to semantic priming. Therefore, we render it
unlikely that the observed pattern of enhanced
Figure 3. Average recognition rates (% correct) and percentage clarification level for congruent, neutral, and incongruent words for the pressing
and twisting condition for both the old and the new items.
2320 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

detections for pictures, congruent to the action per-
formed during the retention phase, merely reflects a
semantic bias in retrieval processes. Pictures corre-
sponding to nonmanipulable object words,
however, also were recognized more accurately
than pictures corresponding to incongruent object
words. A similar effect for the nonmanipulable
object words was observed in the clarification level
scores. This pattern of results seems to suggest
that a reactivation of motor information during
the retention interval does not have a facilitatory
effect on memory consolidation of items highly
associated with congruent action information but
rather suggests that incongruent motor information
has an interfering effect on memory consolidation
of items highly associated with the action infor-
mation. In order to better understand whether the
differences in implicit memory performance are
inhibitory or facilitatory, we performed an analysis
of the picture identification accuracy of the “new”
items. This analysis showed that picture identifi-
cation accuracy of nonmanipulable object words
was higher than picture identification accuracy of
congruent and incongruent object words. These
findings suggest that the effect observed for the
“old”nonmanipulable object condition is driven
by the general difference in picture identification
accuracy between nonmanipulable pictures and
congruent/incongruent object pictures.
EXPERIMENT 3
Experiments 1 and 2 demonstrated an interaction
between object word and the action performed
during the retention interval. Specifically, the
memory performance in an explicit and implicit epi-
sodic recognition test was enhanced for words that
denoted objects for whichthe functional use was con-
gruent with the action performed during the reten-
tion interval. Experiment 2 showed, moreover, that
a similar congruency effect can be observed in a per-
ceptual memory test (i.e., a picture fragment com-
pletion test). This excludes the possibility that
observed effects are driven by semantic priming of
recollection processesin the test phase and rather sup-
ports our assumption of a specific impact of the motor
actions on memory consolidation processes. If the
effects indeed reflect differences in memory consoli-
dation, it is expected that effects generalize to all
types of explicit and implicit tests of memory. The
aim of the third experiment was therefore to
examine whether the action congruency effect is
also present in an implicit memory test (i.e., a word
fragment completion test), in which participants are
required to give the names corresponding to degraded
word forms. In this task, participants typically show a
higher performance for previously experienced (i.e.,
old) than for not previously experienced (i.e., new)
items. This is considered to be a truly implicit
memory test (Roediger, Weldon, Stadler, &
Riegler, 1992). The word fragment completion test
is known to be sensitive to semantic priming and in
fact is often used to show semantic effects (e.g.,
Cies´licka, 2004; Gellatly, Parker, Blurton, &
Woods, 1994; Schütz, Schendzielarz, Zwitserlood,
& Vorberg, 2006; Smith, 1991).
Method
Participants
In the third experiment, 24 new right-handed sub-
jects (average age of 20.7 years) participated. All
subjects were students at the Radboud University
Nijmegen and received 7.50 euros or course
credits for their participation. No subject was
aware of the purpose of the experiment.
Stimuli and design
The stimuli and design were identical to those of
Experiment 1. Word fragments were constructed
by omitting one or more (1–7) letters of the original
word stimuli (see Table 5 for word fragments).
Procedure
Also the study and retention phase were unchanged.
Instead of a delayed recognition task in the test
phase, participants were required to perform an
implicit memory test—namely, a word fragment
completion test (Roediger et al., 1992; Tulving,
Schacter, & Stark, 1982). Participants were told
that they would see word fragments with missing
letters and that they should fill in missing letters
and form an existing Dutch word. Subjects were
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2321
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

not told that some of the test items had been pre-
sented as part of the studied list (see Table 5 for
word fragments). Participants were instructed to be
as fast as possible on the word fragment completion
task. They had to respond by means of a left mouse
button click. Hereafter, a box would appear on the
computer screen in which they had to type the com-
pleted word fragment.
Results
Data analysis
We measured the word fragment completion
latencies, defined as the time difference between
word fragment onset and pressing of the left
mouse button. Trials with response times that
deviated more than two standard deviations from
the mean reaction time (RT) were treated as outliers
and were excluded from further analysis.
Subsequently, we subtracted the mean reaction
times to nonstudied items from all conditions to
derive a measure of the semantic priming effect in
our implicit memory test (see Table 6). Due to the
long response latencies in the word fragment com-
pletion task, the RT data were clearly not normally
distributed. To correct for the non-normal distri-
bution and to increase the robustness of the F-test,
we log-transformed our RT data. In addition, we
measured the mean percentage of word fragments
that were completed. Reaction times and errors
were averaged for each participant in each condition
(see Figure 4 for means) and were submitted to a
two-way repeated measures ANOVA with the
factors action (twisting vs. pressing) and congruency
(congruent, neutral, and incongruent).
The ANOVA performed on the accuracy data
did not produce a reliable main effect of congruency,
F(2, 21) =1.24, p=.311. In line with our
Table 5. Dutch words associated with a pressing movement, a turning movement, or no movement and their corresponding word fragments
Pressing words Twisting words Nonmanipulable words
Piano _i_no Tol to_ Huis hui_
Orgel o__el Stuur st__r Heide he_de
Flipperkast f_ip_erka_t Garde g_r_e Pilaar pi_a_r
Toetsenbord to__sen_o_d Ventiel ve__iel Plein p_e_n
Rekenmachine re_e__achi_e Pepermolen pe_e_mo_en Hotel ho_e_
Deurbel de_r_el Huissleutel h_is_leu_el Terras t__ras
Muisknop mu_sk_op Kurkentrekker k_r_entr__ker Vijver vi_ve_
Spuitbus sp_itbu_ Passer p_s_er Blokhut blo__ut
Pinapparaat pi_ap_ar_at Jampot ja_po_ Piramide pi_a_i_e
Mobieltje mo_ielt_e Volumeknop v_lu_ekno_ Rotsblok r_ts_lok
Stopwatch s_opwa__h Blikopener b_iko_ene_ Zendmast _endma_t
Perforator per_ora_or Cijferslot c_j_ersl_t Schutting sc_ut_ing
Stemkastje s_em_ast_e Eierwekker e_erwe_ker Standbeeld s_and_e_ld
Typemachine t_pema_hin_ Schroevendraaier sc_roe_endr_aier Stoeprand s_oep_and
Fototoestel f_tot_este_ Puntenslijper _unt_nsli_per Kerktoren k__k_oren
Laptop _a_top Opwindklok opwi_dk__k Regenpijp _e_en_ijp
Nietmachine _ie_mac_ine Contactslot _onta_ts_ot Strandtent s_ran_te_t
Claxon c_ax_n Bankschroef b_nks_hro_f Koeienstal ko_iens_al
Pepperspray p__persp_ay Moer m_er Uithangbord u_t_ang_ord
Punaise _unai_e Waterkraan wa_er_r_an Kasteelmuur k_s_ee_mu_r
Touchscreen to__hsc_een Waterpomptang _aterp_m__ang Bloementuin _loe_ent_in
Klavecimbel _lave_im_el Slagroomklopper sla_roo__lop_er Lantaarnpaal lan_a_rnpa_l
Aanknop a_nk_op Tafelvoetbalbak ta_elvoe__alba_ Treinstation t_ein_tati_n
Cassetterecorder c_sse_tereco_der Bout _out Rioleringsbuis ri_lerin_sb_is
Afstandsbediening af__andsbe_ieni_g Kauwgomballenautomaat ka__go_b_lle__uto_aat Schotelantenne _chotel_nte_ne
Note: Pressing movement: pressing words. Turning movement: twisting words. No movement: Nonmanipulable words.
2322 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

hypothesis, the analysis of reaction times revealed a
significant main effect of congruency, F(2, 21) =
3.59, p=.046, MSE =0.012, η
2
p
=.229. Post hoc
paired-sample ttests revealed that priming effects
were larger for words denoting objects for which
the functional use was congruent with the action
performed than for words that denoted objects for
which the functional use was incongruent, t(23) =
2.68, p=.013. No difference in the semantic
priming was observed between congruent and non-
manipulable object words, t(23) ,1, and between
incongruent and nonmanipulable object words,
t(23) =1.35, p=.189.
A repeated measures ANOVA on the word
fragment completion latencies to “new”word frag-
ments revealed a main effect of congruency, F(2,
21) =4.44, p=.025, MSE =0.015, η
2
p
=.297. To
further explore this main effect, we calculated post
hoc paired-sample ttests. No difference was observed
between congruent and incongruent word fragments,
t(23) =0.416, p=.681, and between nonmanipul-
able word fragments and incongruent word frag-
ments, t(23) =1.57, p=.129. Nonmanipulable
word fragments showed higher word fragment
latencies than congruent word fragments, t(23) =
2.51, p=.020.
Table 6. Average accuracy and priming effects for the items of Experiment 3 for the factors congruency and action
Congruency
Pressing Twisting
Accuracy (%) Priming effect Accuracy (%) Priming effect
Congruent 88.89 (2.88) 1,507 (251) 79.44 (3.71) 933 (228)
Neutral 80.00 (3.48) 1,500 (339) 82.78 (2.24) 525 (333)
Incongruent 78.89 (3.37) 926 (305) 85.56 (1.61) 726 (383)
Note: Congruency: congruent, neutral, and incongruent words. Action: pressing versus twisting action. Standard errors in parentheses.
Figure 4. Average priming effects (compared to nonstudied base rates) for congruent, neutral, and incongruent words in both the pressing and
twisting conditions of Experiment 3 (pressing versus twisting action during retention phase).
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2323
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

Discussion
The results of Experiment 3 show a facilitation
effect in the mean reaction times for congruent
object word fragments as compared with incongru-
ent object word fragments. Given that we were able
to extend the action congruency effect observed in
Experiment 1 to a classical implicit memory test
(i.e., a word fragment completion test), it is unlikely
that the observed congruency effect merely reflects
biased memory retrieval. The results rather suggest
that the semantic action congruency effect is driven
by differences in encoding associations between
semantic and sensorimotor codes.
Effects in word stem completion tests are often
explained as being due to the fact that prime and
target share the same orthographic features.
However, word stem completion tests have also
been shown to be sensitive to semantic priming
and are often used to show semantic effects (Cies´
licka, 2004; Schütz et al., 2006; Smith, 1991).
For example, Schütz et al. (2006) showed that
both form primes and semantic primes exert an
influence on word stem completion performance.
It was shown that form primes exerted stronger
effects than semantic primes, which, however,
showed statistically stronger effects than a neutral
word condition. In the case of the present study,
we think it is not plausible to assume that a modu-
lation of the observed priming effects by the differ-
ent motor responses (pressing vs. twisting) reflects a
difference in the representation or processing of
orthographic features. Alternatively, we think that
the observed modulation of priming effects by the
motor execution task occurred because of the
semantic content that is shared by the prime and
target.
GENERAL DISCUSSION
The present study demonstrates an effect of exe-
cuted motor actions during a retention interval on
performance on subsequent explicit and implicit
memory tasks. The observed action congruency
effect in Experiment 1 shows that memory strength
was enhanced for words that denoted objects for
which the functional use was congruent with the
action performed during the retention interval
(e.g., pepper mill–twisting action, doorbell–
pressing action) compared with words that
denoted objects for which the functional use was
incongruent (e.g., pepper mill–pressing action,
doorbell–twisting action). Interestingly, the
results of Experiment 2 show that the action con-
gruency effect is not restricted to explicit episodic
recognition tasks and demonstrates a similar
impact of motor information during retention on
an implicit perceptual memory task. That is, we
found enhanced perceptual priming of the memor-
ized words for line drawings in a picture fragmen-
tation test (Snodgrass & Feenan, 1990), if the
functional use of depicted object was associated
with an action that was similar to the action per-
formed during the retention interval. Similarly,
the results of Experiment 3 showed a facilitation
effect in the mean reaction times for congruent
object word fragments as compared with incongru-
ent object word fragments. That is, in Experiment
3 we were able to extend the action congruency
effect to a classical implicit memory test (i.e., a
word fragment completion test). These findings
suggest that a reactivation of motor information
during the retention interval does not merely
prime subsequent retrieval processes, but affects
memory consolidation and the strength of a
memory trace and emphasizes therefore the impor-
tant role of sensorimotor information in establish-
ing enduring representations.
Previous research has shown that both action
phrases and nouns are better recalled/recognized
when participants perform the denoted action or
carry out the prototypical action corresponding to
each noun, whereas performing an unrelated
action does not (Dijkstra & Kashak, 2006;
Engelkamp & Zimmer, 1983; Zimmer &
Engelkamp, 2003). The authors argued for the
involvement of a specialized action output system
in subject-performed tasks but not in verbal tasks
(Engelkamp & Zimmer, 1983). The main idea is
that the overt performance of a movement (which
has preexperimentally been shown to be associated
with a particular item) during the encoding phase
can lead to an integration of action-specific
2324 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

information into the memory record of an item.
This additional action-specific information in
turn is sufficient for enhancing subsequent recog-
nition and recollection memory (Zimmer &
Engelkamp, 2003). These results seem to suggest
that the modality-specific (in this case motor)
information can play an important role in enhan-
cing memory encoding processes.
In a similar vein, numerous studies have argued
for the involvement of sensorimotor brain regions
in language comprehension (e.g., Desai et al.,
2010; Tettamanti et al., 2005). Results from
Pecher et al. (2009) suggest that the involvement
of sensorimotor processes does not seem to be
restricted to online language comprehension.
That is, recognition performance after a 40–
50-min delay, in a surprise recognition task, was
better for pictures of objects that matched in
shape or orientation with the shape or orientation
implied by the sentence during the study phase.
These findings suggest that memory performance
is affected by the overlap between the visual simu-
lation during encoding of the linguistic information
and the physical appearance of the picture infor-
mation at test. Such a cross-modality memory
effect is in line with the transfer-appropriate pro-
cessing principle (Durgunog˘lu & Roediger, 1987;
Morris, Bransford, & Franks, 1977). This principle
entails that memory performance is modulated by
the relationship between how information is
initially encoded and later retrieved. That is,
memory performance will be enhanced if the pro-
cesses engaged during encoding are similar to
those engaged in during retrieval.
Importantly, the action congruency effect in the
present study cannot be explained by the transfer-
appropriate processing principle, because the
encoding and retrieval phase were kept constant.
Our finding goes therefore beyond previous
reports of sensorimotor effects in memory (Pecher
et al., 2009). The crucial manipulation in the
current study was the modality-specific information
activated by the motor task during the retention
interval. Moreover, the results of Experiment 2
showed that a congruency effect can be observed
in a perceptual memory test (i.e., a picture fragment
completion task), which excludes the possibility
that observed effects are driven by a semantic
priming of recollection processes in the test
phase. Our results suggest that memory perform-
ance can be enhanced by the activation of motor
information that is specific to the to-be-remem-
bered item, even if the enactment of the prototypi-
cal action corresponding to each noun was
separated in time from the presentation of the
studied words. These findings are in line with the
idea that reactivation supports the consolidation
of memories (Kuhl, Shah, DuBrow, & Wagner,
2010; Rasch & Born, 2007). In the context of
reconsolidation theory it has been suggested that
memory reactivation provides the possibility to
modify and update old memories, by means of a
transient destabilization of an existing memory
trace. Support for this notion has been provided
by studies showing that reactivated memories can
be strengthened if newly encoded information
matches the existing memory trace (Moscovitch,
Nadel, Winocur, Gilboa, & Rosenbaum, 2006)
and can become generalized if reactivation occurs
in a different environment (Gordon, 1981). On
the other hand, reactivated memories can be wea-
kened if newly encoded information is conflicting
with the existing memory trace (Diekelmann,
Büchel, Born, & Rasch, 2011; Walker,
Brakefield, Hobson, & Stickgold, 2003) and in
some cases even lead to memory blockade
(Schiller et al., 2010).
In line with this research, we argue that the
observed action congruency effect demonstrates
that a reactivation of motor information during
the retention interval modulates subsequent per-
formance on a delayed recognition task. This
finding suggests that motor simulation plays an
important role in establishing enduring represen-
tations of a word’s referents. In our study, the
action performed during the retention interval
might have acted as a specific cue, reactivating
memory representations of the items associated
with the same motor information. Consequently,
these memory representations became destabilized
and susceptible to interferences. A specific
strengthening of memory representations, as
observed in the present study, might have taken
place given that no interfering information was
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2325
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

presented after reactivation. Secondly, the cue that
presumably exerted a reactivating influence
matched the existing memory trace and therefore
might have led to a specific strengthening of
memory representations. However, it could be
argued that instead of having an effect on memory
consolidation, the intervening motor action influ-
enced performance during the test phase, making
items with congruent motor actions now seem
more familiar. We render this possibility unlikely
given that we did not observe a difference in the
false-alarm rates of congruent and incongruent
object words in Experiment 1. That is, if the inter-
vening motor action would purely heighten the fam-
iliarity of all items that are associated with a
congruent motor action, then it would be expected
that the false-alarm rate of congruent object words
would be higher than the false-alarm rate of incon-
gruent object words. In a similar vein, an analysis
of “new”items of Experiment 2 and Experiment 3
did not reveal a difference between congruent and
incongruent object words.
The present study, however, cannot exclude the
possibility that participants used a motor imagery
strategy during the encoding of the list of words
and would therefore be aware of the twisting and
pressing categories. That is, one might argue that
the reason for observing a congruency effect
between the action performed during the retention
interval and the prototypical action carried out on
the word’s referent is because the person voluntarily
images the functional use of the referent object
during the encoding of the study list. However, a
post hoc verbal questionnaire also casts serious
doubt on the explanation that the effects are driven
by the participants’awareness of the stimulus cat-
egories and therefore intentional memory strategies.
That is, no participant indicated to be aware of the
relevant dimension according to which words
could be categorized (i.e., functional object use).
We also cannot exclude the possibility that the
motor action performed during the retention
phase was used as an additional cue to retrieve
words from memory. However, given that no par-
ticipant indicated to be aware of the relevant dimen-
sion according to which words could be categorized,
we also render this possibility to be unlikely.
CONCLUSION
The current study demonstrates that the sensitivity
or memory strength for object words was modu-
lated by the activation of modality-specific infor-
mation during the retention interval. This effect
generalized to different types of memory tasks
and could in particular be shown for explicit and
implicit memory tests as well as for semantic and
perceptual tasks. The data suggest that activation
of modality-specific (in this case motor) infor-
mation during word memorization can play an
important role in establishing more enduring rep-
resentations of a word’s referents, and that a reacti-
vation of this motor information during retention
interval can modulate subsequent performance on
a delayed recognition task.
Original manuscript received 25 July 2012
Accepted revision received 27 November 2012
First published online 11 April 2013
REFERENCES
Baayen, R., Piepenbrock, R., & van Rijn, H. (1993). The
CELEX lexical database [CD-ROM]. Philadelphia,
PA: University of Pennsylvania, Linguistic Data
Consortium.
Bäckman, L., Nilsson, L. G., & Chalom, D. (1986).
New evidence on the nature of the encoding of
action events. Memory & Cognition,14, 339–346.
Barsalou, L. W. (1999). Perceptual symbol systems.
Behavioral Brain Sciences,22, 577–660.
Chao, L. L., & Martin, A. (2000). Representation of
manipulable man-made objects in the dorsal stream.
NeuroImage,12, 478–484.
Cies´licka, A. (2004). Semantic priming effects in bilin-
gual word fragment completion. Poznan Studies in
Contemporary Linguistics,40, 111–133.
Cohen, R. L. (1981). On the generality of some memory
laws. Scandinavian Journal of Psychology,22, 267–281.
Desai, R. H., Binder, J. R., Conant, L. L., & Seidenberg,
M. S. (2010). Activation of sensory-motor areas in
sentence comprehension. Cerebral Cortex,20,
468–478.
Diekelmann, S., Büchel, C., Born, J., & Rasch, B.
(2011). Labile or stable: Opposing consequences for
2326 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

memory when reactivated during waking and sleep.
Nature Neuroscience,14, 381–386.
Dijkstra, K., & Kashak, M. P. (2006). Encoding in
verbal, enacted and autobiographical tasks in young
and older adults. The Quarterly Journal of
Experimental Psychology,59(8), 1338–1345.
Durgunog˘lu, A. Y., & Roediger, H. L. (1987). Test
differences in accessing bilingual memory. Journal of
Memory and Language,26, 377–391.
Engelkamp, J. (1997). Memory for to-be-performed
tasks versus memory for performed tasks. Memory &
Cognition,25(1), 117–124.
Engelkamp, J., & Zimmer, H. D. (1983). Zum Einfluß
von Wahrnehmen und Tun auf das Behalten von
Verb–Objekt-Phrasen [The influence of perception
and enactment on the retention of verb–object
phrases]. Sprache & Kognition,2, 117–129.
Gellatly, A., Parker, A., Blurton, A., & Woods, C.
(1994). Word stem and word fragment completion
following semantic activation and elaboration.
Journal of Experimental Psychology: Learning,
Memory, and Cognition,20(5), 1099–1107.
Glenberg, A. M. (1997). What memory is for. Behavioral
Brain Sciences,20, 1–55.
Glenberg, A. M., & Kaschak, M. P. (2002). Grounding
language in action. Psychonomic Bulletin & Review,9,
558–565.
Gordon, W. C. (1981). Mechanisms of cue-induced
retention enhancements. In N. E. Spear & J. A.
Kleim (Eds.), Information processing in animals:
Memory mechanisms (pp. 319–339). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Hauk, O., & Pulvermueller, F. (2004). Neurophysiological
distinction of action words in the fronto-central cortex.
Human Brain Mapping,21, 191–201.
Hommel, B., Muesseler, J., Aschersleben, G., & Prinz,
W. (2001). The theory of event coding (TEC): A fra-
mework for perception and action. Behavioral and
Brain Sciences,24, 849–937.
Jacoby, L. L., Baker, J. G., & Brooks, L. R. (1989).
Episodic effects on picture identification: Implications
for theories of concept learning and theories of
memory. Journal of Experimental Psychology: Learning,
Memory, and Cognition,15(2), 275–281.
Kemmerer, D., Castillo, J. G., Talavage, T., Patterson,
S., & Wiley, C. (2008). Neuroanatomical distribution
of five semantic components of verbs: Evidence from
fMRI. Brain and Language,107, 16–43.
Kintsch, W. (2008). Symbol systems and perceptual rep-
resentations. In M. De Vega, A. Glenberg, &
A. Graesser (Eds.), Symbols and embodiment
(pp. 145–164). Oxford: Oxford University Press.
Kuhl, B. A., Shah, A. T., DuBrow, S., & Wagner, A. D.
(2010). Resistance to forgetting associated with hip-
pocampus-mediated reactivation during new learn-
ing. Nature Neuroscience,13, 501–506.
Macmillan, N. A., & Creelman, C. D. (1991). Detection
theory: A user’s guide. Cambridge, UK: Cambridge
University Press.
Morris, D. C., Bransford, J. D., & Franks, J. J. (1977).
Levels of processing versus transfer appropriate pro-
cessing. Journal of Verbal Learning and Verbal
Behavior,16, 519–533.
Moscovitch, M., Nadel, L., Winocur, G., Gilboa, A., &
Rosenbaum, S. (2006). The cognitive neuroscience of
remote episodic, semantic and spatial memory.
Current Opinion in Neurobiology,16, 179–190.
Pecher, D., Van Dantzig, S., Zwaan, R. A., &
Zeelenberg, R. (2009). Language comprehenders
retain implied shape and orientation of objects. The
Quarterly Journal of Experimental Psychology,62(6),
1108–1114.
Pecher, D., Zanolie, K., & Zeelenberg, R. (2007).
Verifying visual properties in sentence verification
facilitates picture recognition memory. Experimental
Psychology,54(3), 173–179.
Pulvermueller, F. (1999). Words in the brain’s language.
Behavioral and Brain Sciences,22, 253–336.
Rasch, B., & Born, J. (2007). Maintaining memories by
reactivation. Current Opinion in Neurobiology,17,
698–703.
Roediger, H. L., & McDermott, K. B. (1993). Implicit
memory in normal human subjects. In F. Boller &
J. Grafman (Eds.), Handbook of neuropsychology
(Vol. 8, pp. 63–131). Amsterdam: Elsevier.
Roediger,H.L.,Weldon,M.S.,Stadler,M.L.,&
Riegler, G. L. (1992). Direct comparison of two
implicit memory tests: Word fragment and word
stem completion. Journal of Experimental
Psychology: Learning, Memory, and Cognition,18(6),
1251–1269.
Rueschemeyer, S.-A., Pfeiffer, C., & Bekkering, H.
(2010). Body schematics: On the role of the body
schema in embodied lexical–semantic representations.
Neuropsychologia,48(3), 774–781.
Rueschemeyer, S.-A., Rooij, D., Lindemann, O.,
Willems, R., & Bekkering, H. (2010). The function
of words: Distinct neural correlates for words denot-
ing differently manipulable objects. Journal of
Cognitive Neuroscience,22(8), 1844–1851.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12) 2327
EMBODIED GROUNDING OF MEMORY
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014

Saccuman, M. C., Cappa, S. F., Bates, E. A., Arevalo,
A., Della Rosa, P., Danna, M., & Perani, D.
(2006). The impact of semantic reference on word
class: An fMRI study of action and object naming.
NeuroImage,32, 1865–1878.
Schiller, D., Monfils, M.-H., Raio, C. M., Johnson, D.
C., LeDoux, J. E., & Phelps, E. A. (2010).
Preventing the return of fear in humans using recon-
solidation update mechanisms. Nature,463, 49–53.
Schütz, K., Schendzielarz, I., Zwitserlood, P., &
Vorberg, D. (2006). Nice wor_if you can get the
wor_: Subliminal semantic and form priming in frag-
ment completion. Consciousness and Cognition,16,
520–532.
Simmons, W. K., Ramjee, V., Beauchamp, M. S.,
McRae, K., Martin, A., & Barsalou, L. W. (2007).
A common neural substrate for perceiving and
knowing about color. Neuropsychologia,45,
2802–2810.
Smith, M. C. (1991). On the recruitment of semantic
information for word fragment completion:
Evidence from bilingual priming. Journal of
Experimental Psychology: Learning, Memory, and
Cognition,17(2), 234–244.
Snodgrass, J. G., & Feenan, K. (1990). Priming effects in
picture fragment completion: Support for the percep-
tual closure hypothesis. Journal of Experimental
Psychology: General,119(3), 276–296.
Stanfield, R. A., & Zwaan, R. A. (2001). The effect of
implied orientation derived from verbal context on
picture recognition. Psychological Science,12, 153–156.
Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese,
V., Danna, M., Scifo, P., …Perani, D. (2005).
Listening to action-related sentences activates
fronto-parietal motor circuits. Journal of Cognitive
Neuroscience,17, 273–281.
Tulving, E., Schacter, D. L., & Stark, H. A. (1982).
Priming effects in word-fragment completion are
independent of recognition memory. Journal of
Experimental Psychology: Learning, Memory, and
Cognition,8, 336–342.
Van Dam, W. O., Rueschemeyer, S.-A., & Bekkering,
H. (2010). How specifically are action verbs rep-
resented within the neural motor system: An fMRI
study. NeuroImage,53, 1318–1325.
Van Dam, W. O., Rueschemeyer, S.-A., Lindemann,
O., & Bekkering, H. (2010). Context effects in
embodied lexical–semantic processing. Frontiers in
Psychology,1, 150. doi:10.3389/fpsyg.2010.00150
Walker, M. P., Brakefield, T., Hobson, A., & Stickgold,
R. (2003). Dissociable stages of human memory con-
solidation and reconsolidation. Nature,425, 616–620.
Zimmer, H. D., & Engelkamp, J. (2003). Signing
enhances memory like performing actions.
Psychonomic Bulletin & Review,10(2), 450–454.
Zwaan, R. A., Stanfield, R. A., & Yaxley, R. H. (2002).
Language comprehenders mentally represent the
shapes of objects. Psychological Science,13, 168–171.
Zwaan, R. A., & Taylor, L. J. (2006). Seeing, acting,
understanding: Motor resonance in language compre-
hension. Journal of Experimental Psychology: General,
135, 1–11.
2328 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (12)
VAN DAM ET AL.
Downloaded by [Universitaetsbibliothek Potsdam] at 08:31 07 January 2014