Mental Practice in Music Memorization: An Ecological-Empirical Study

Abstract

The present study aims to systematically describe mental practice (MP) in music memorization, with regard to individual differences in the use of different MP strategies and their performance outcomes. Sixteen pianists were studied while they memorized piano pieces. Each subject memorized two pieces, either via MP or physical practice (PP). In order to keep the setting as ecologically valid as possible within the experimental setup, we allowed subjects to freely apply their preferred MP strategies with the exception of physically playing a real piano. Practice and performances were video documented and expert rated; practice strategies were reported in researcher-developed questionnaires. The use of MP alone led to successful music learning. MP combined with PP produced results that were indistinguishable from those following PP alone. Pitch imagery and structural analysis were associated with better post-MP performance. Results are discussed in the frame of expert memory theory (Chase & Simon, 1973; Chaffin, Logan, & Begosh, 2009) and practical implications for musicians are provided.

Full text

MENTAL PRACTICE IN MUSIC MEMORIZATION:
ANECOLOGICAL-EMPIRICAL STUDY
NICOLO
`FRANCESCO BERNARDI
University for Music, Theatre and Media, Hannover,
Germany & Milano-Bicocca University, Milano, Italy
ALEXANDER SCHORIES
University for Music, Theatre and Media, Hannover,
Germany
HANS-CHRISTIAN JABUSCH
University for Music, Theatre and Media, Hannover,
Germany & University of Music Carl Maria von Weber,
Dresden, Germany
BARBARA COLOMBO
Catholic University of Sacred Heart, Brescia, Italy
ECKART ALTENMU
¨LLER
University for Music, Theatre and Media, Hannover,
Germany
T
HE PRESENT STUDY AIMS TO SYSTEMATICALLY
describe mental practice (MP) in music memorization,
with regard to individual differences in the use of dif-
ferent MP strategies and their performance outcomes.
Sixteen pianists were studied while they memorized
piano pieces. Each subject memorized two pieces, either
via MP or physical practice (PP). In order to keep the
setting as ecologically valid as possible within the exper-
imental setup, we allowed subjects to freely apply their
preferred MP strategies with the exception of physically
playing a real piano. Practice and performances were
video documented and expert rated; practice strategies
were reported in researcher-developed questionnaires.
The use of MP alone led to successful music learning.
MP combined with PP produced results that were
indistinguishable from those following PP alone. Pitch
imagery and structural analysis were associated with
better post-MP performance. Results are discussed in
the frame of expert memory theory (Chase & Simon,
1973; Chaffin, Logan, & Begosh, 2009) and practical
implications for musicians are provided.
Received: July 24, 2011, accepted May 25, 2012.
Key words: mental practice, mental imagery, music
memorization, music performance, musicians’ injuries
M
ENTAL PRACTICE
(MP)
IS GENERALLY
defined as a technique by which someone
with the intent to practice creates a mental
representation of a preconceived idea or action in order
to enhance performance (van Meer & Theunissen,
2009). The focus of the present study is on performance
enhancing rehearsal strategies, which must be distin-
guished from other techniques of emotional and mental
training for performance preparation (e.g., relaxation
training, meditation, visualization of prescribed images;
see Connolly & Williamon, 2004, for a review).
MP has been investigated as a potentially useful
practice technique in different fields, including athle-
tics (Feltz & Landers, 1983), stroke rehabilitation
(Zimmermann-Schlatter, Schuster, Puhan, Siekierka,
& Steurer, 2008) and music (Cahn, 2008; Theiler &
Lippman, 1995). Converging evidence from different
fields has shown that MP has a moderate and significant
impact on performance, and that the effects of MP are
weaker than the effects of physical practice (PP)
(Gabrielsson, 1999). The efficacy of MP increases when
the task involves cognitive or symbolic skills and when
the subject has expertise in the specific task’s domain
(Driskell, Copper, & Moran, 1994). Moreover, several
studies have shown that proper combinations of MP
and PP may lead to results that are close to or equal
to those obtained in PP alone (Feltz, Landers, & Becker,
1988; Kopiez, 1990). Music performance serves as an
excellent model for studying MP because it is made up
of tasks with complex cognitive elements (to a greater
extent than in sport performance). At the same time, for
a musician, performance is something that is natural to
provide (thus, being ecologically valid) and for a scientist
the performance is an output that can be objectively
measured. Yet, up to now, few research studies have
specifically tested the effectiveness of MP in music per-
formance. In the field of music performance, MP has
been used and taught at least since the contribution of
the well-known piano teacher Karl Leimer and his most
famous pupil Walter Gieseking (Barry & McArthur,
1994; Leimer & Gieseking, 1998; McMillan, 2005).
According to Leimer, through the use of MP ‘‘ ...the
piece can be perfectly performed and this in a most
astonishingly shorter time’’ (Leimer & Gieseking,
Music Perception,VOLUME 30, ISSUE 3, PP. 275–290, ISSN 0730-7829, ELECTRONIC ISSN 1533-8312. ©2013 BY THE REGENTS OF THE UNIVERSITY OF CALIFORNIA ALL
RIGHTS RESERVED.PLEASE DIRECT ALL REQUESTS FOR PERMISSION TO PHOTOCOPY OR REPR ODUCE ARTICLE CONTENT THROUGH THE UNIVERSITY OF CALIFORNIA PRESS ’S
RIGHTS AND PERMISSIONS WEBSITE,HTTP://WWW.UCPRESSJOURNALS.COM/REPRINTINFO.ASP.DOI:10.1525/MP.2012.30.3.275
Mental Practice in Music Memorization 275

1998). MP techniques for musicians include conducting
a formal analysis of the score, listening to a recording of
the piece, forming auditory imagery of the pitches,
imagining movement (visually and/or kinaesthetically)
and using visual imagery of the score (Klo¨ppel, 1996;
Orloff-Tschekorsky, 1996).
Overall, results of empirical studies on musicians
show that MP is more effective than no practice and
not as effective as PP in terms of both objective mea-
surements (e.g., correctness of notes) and expressive
features (Highben & Palmer, 2004; Lim & Lippman,
1991). Still, it has also been shown that MP alone may
lead to the same plastic changes in the motor system as
those occurring with the acquisition of the skill by
repeated PP (Pascual-Leone et al., 1995). MP used along
with an auditory model showed better results than MP
alone (Lim & Lippman, 1991; Theiler & Lippman,
1995). The combination of MP and PP appeared to
be particularly effective in the field of music, as shown
by several experiments using different tasks and
instrumentalist groups (Cahn, 2008; Coffman, 1990;
Kopiez, 1990; Ross, 1985; Theiler & Lippman, 1995).
Depending on the task, the use of MP and PP led to
performances that were as close to, or even indistin-
guishable from, those following PP alone (Theiler &
Lippman, 1995).
However, all these results come from highly con-
trolled experimental situations that have imposed
severe constraints on the practice situation and/or the
subjects. First, subjects were always forced to use a spe-
cific MP strategy that was chosen by the experimenter
independently from task and individual-related fea-
tures. In fact, MP has been reduced to: a) an analytical
pre-study of the score, or listening to a recording of the
piece followed by analytical study (Rubin-Rabson,
1937); b) auditory plus kinesthetic imagery (Cahn
2008; Ross, 1985); c) imagery of sounds while pressing
silent keys, or imagery of the feeling of the movements
while actually hearing the sounds, or auditory plus
kinesthetic imagery in the absence of any feedback
(Highben & Palmer, 2004); d) visual plus kinesthetic
imagery with an auditory model (Coffman, 1990); or
e) visual plus auditory plus kinesthetic imagery, with
or without an auditory model (Coffman, 1990; Lim &
Lippman, 1991; Ross, 1985; Theiler & Lippman, 1995).
Second, subjects were often asked to practice in unnat-
ural situations: Depending on the study, subjects had to
avoid MP strategies other then the one prescribed
(Highben & Palmer, 2004). For example, they had to
avoid any overt movement of the hand/fingers (Cahn,
2008; Coffman, 1990; Ross, 1985) or humming (Lim &
Lippman, 1991; Ross, 1985). In some studies MP
included a recorded version of the piece played at fixed
time intervals (Theiler & Lippman, 1995) or even played
continuously for the entire practice time (Coffman,
1990; Highben & Palmer, 2004; Lim & Lippman,
1991). In other studies, participants had a very limited
time window (e.g., three minutes) to implement their
MP (Cahn, 2008; Coffman, 1990) or had to practice the
piece a fixed number of times without stopping or cor-
recting mistakes (Coffman, 1990; Highben & Palmer,
2004; Ross, 1985).
The above mentioned constraints were introduced in
order to improve experimental control, and these con-
trolled studies yielded valuable and converging results,
indeed. However, these constraints may have signifi-
cantly altered the MP processes from what they are in
a musician’s daily life, thereby producing partially
ambiguous or biased results. Encouraging or forcing
a musician to use a specific MP strategy potentially
raises the following problems:
I. In methodological terms, it is often not clear why
the experimenter has a priori decided to select one
strategy and not another. Inclusions and exclusions
of MP strategies in the previously mentioned stud-
ies seem to be partially guided by the author’s
implicit or explicit assumptions about which strat-
egies constitute the very core of MP. Yet there is no
consensus among scientists about the exact nature
of these core processes. Furthermore, such a debate
would not be particularly relevant for the musician,
who appreciates MP as a heterogeneous and flexi-
ble tool.
II. Such experiments do not take into account how
much the selected strategy conforms to each sub-
ject’s preferences, habits, and abilities (individual-
related features). For example, both Lim and
Lippman (1991) and Highben and Palmer
(2004) speculated about the existence of individ-
ual MP-related features that could have an
important role in the way MP was applied. In
fact, many of Lim and Lippman’s subjects
expressed the desire for more freedom in the
application of MP, while Highben and Palmer
documented an association between individual
performers’ musical memory and imagery skills
(auditory vs. motor).
III. Maintaining the same strategy throughout the
whole practice process does not allow the subjects
to flexibly change their approach in order to opti-
mize the learning process or manage specific ele-
ments of the task that might benefit from different
MP tactics (task-related features). Comparing the
276 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller

performance of guitarists and vocalists, Theiler and
Lippman (1995) concluded that ‘‘features of MP
regimen should be adjusted to accommodate par-
ticular applications, because different attributes
may be optimal for various physical and musical
endeavors.’’
IV. The demand of using only one MP strategy to the
exclusion of all others appears to be unrealistic for
musicians. This can be nicely illustrated by taking
the underlying brain mechanisms into account.
Several neuroimaging studies have described the
tight and automatic coupling between auditory,
visual, and motor networks in the brains of musi-
cians (Haslinger et al., 2005) and even of naı¨ve
subjects following just 20 minutes of piano training
(Bangert & Altenmu
¨ller, 2003). It is not likely that
all of these tight and long-term developed connec-
tions can be effectively interrupted by the subject’s
conscious will. Lim and Lippman’s (1991) results
underlie this hypothesis, since ‘‘all subjects found it
almost impossible to separate the kinaesthetic
image from the visualization of ‘hands on key-
board’; two of the subjects claimed that musicians
automatically listen to music and feel their instru-
ment tactually’’ (p. 27).
Consequently, the experimental control of a study
design that focuses on a specific MP strategy remains
doubtful, especially when attempting to identify indi-
vidual differences in: a) how effectively undesired MP
strategies have been switched off; b) the cognitive cost
paid for this switching off; c) the practice strategy actu-
ally used; and d) preferences, habits, and abilities in the
strategy actually applied.
The present study was primarily designed to address
these limitations. We therefore designed an ‘‘open’’ MP
condition in which the participants were completely
free to use any MP strategy they desired without any
constraint and within an amount of time close to the
one indicated in past research as the most effective for
MP (Driskell et al., 1994). We allowed participants, for
example, to move fingers and hands without the
instrument, since this is a commonly used MP strat-
egy. Furthermore, this strategy is explicitly recom-
mended in practical guides to MP in sports (Rushall,
1991). Thus, a higher degree of ecological validity
enables the present investigation to ask questions,
such as:
I. Are certain MP patterns more effective than others?
II. Is MP more effective when it is freely used than when
it is constrained, as in previous studies? Converging
evidence has shown that the effectiveness of skill
learning can be enhanced if the learner is given some
control over the practice conditions (Chen, Hendrick,
& Lidor, 2002; Chiviacowsky & Wulf, 2002; Janelle,
Barba, Frehlich, Tennant, & Cauraugh, 1997).
For the present investigation, a music memorization
paradigm was chosen because of its potential implica-
tions in both applied and theoretical terms. To begin
with, memorization is a major issue for musicians; play-
ing from memory is often a source of anxiety and mem-
orization still tends to be seen as an individual and
mysterious process, in which each person has to find
his or her own method (Ginsborg, 2002). A deeper and
systematic knowledge of the means to improve music
memorization abilities could therefore have practical
applications in the field.
Moreover, Chaffin and collaborators have shown that
the main principles described by expert memory theory
in classic domains (e.g., chess, see Chase & Simon, 1973;
digit strings, see Thompson, Cowan, & Friedman, 1993)
rule the behavior of expert musical memorists also, spe-
cifically a) the meaningful encoding of novel material;
b) the use of a well-learned retrieval structure; and c)
extended practice to decrease the time needed for
retrieval from long-term memory (Chaffin, Logan, &
Begosh, 2009; Ericsson & Kintsch, 1995). Of particular
interest is the finding that music performance from
memory relies heavily on structural and linguistic
memory, in addition to auditory and motor memory.
First, expert musicians rely on schematic knowledge
already stored in memory to organize information
into larger chunks. Second, they use a retrieval
scheme to organize the cues that provide access to
the chunks of information in long-term memory.
Musicians’ schematic structures can be identified as
familiar patterns such as chords, scales, and arpeg-
gios, while the formal structure of the music conve-
niently provides a ready made hierarchical
organization to serve as a retrieval scheme (Chaffin
& Imreh, 2002). These conclusions come from stud-
iesinwhichthepianowasalwaysavailabletothe
musician; therefore, it would be interesting to see
how memorizing strategies are applied and adapted
when the physical feedbacks of the real action are not
available. For example, movement simulation has
been shown to support operations such as letter
recall from memory in deaf children (Locke & Locke,
1971) or abacus calculation in expert abacus opera-
tors when the abacus was not available (Hatano &
Miyake,1977).Interestingly,thisformof‘‘transi-
tional’’ sensory representation appeared linked with
Mental Practice in Music Memorization 277

the degree of expertise, with more advanced subjects
being able to avoid it (Hatano & Miyake, 1977).
Based on these considerations, we generated hypoth-
eses about the features and strategies that are more
predictive of successful performance. Considering that
our subjects had to prepare a memorized performance
of a relatively simple piece of music using MP, we
hypothesized that: 1) Structural/formal analysis of the
piece would be associated with better performance. 2)
Pitch imagery would be strongly connected with perfor-
mance results; in fact, in MP pitch imagery is likely to be
a core operation, not a transitional one, acting as a pre-
requisite for structural analysis or other higher-order
operations. Auditory memory-based operations might
become less relevant when the piano is available. 3)
Motor and visual imagery should play a minor role. It
would also be possible to observe differences in the role
of motor and visual imagery between subjects with low
vs. high motor/visual imagery capabilities. 4) Subjects’
listening to the auditory model would be associated with
poorer performance. Compared with past studies
showing the advantage of auditory model-supported
MP vs. MP without this kind of support, in the present
study subjects could choose whether to listen to the
recording or not. Following the reasoning of Hatano
and Miyake (1977), it seems likely that more advanced
subjects would not ‘‘waste’’ resources in this transi-
tional operation. 5) For the same reasons, the subjects’
moving of fingers should be associated with poorer
performance.
Method
PARTICIPANTS
Sixteen right-handed pianists (8 males, 8 females) were
recruited on a volunteer basis from the University of
Music and Drama, Hannover, Germany. They had
a mean age of 26 +4 years (range ¼18 to 36) and
they had at least 15 years of individual piano instruction
(mean ¼20 +4; range ¼15 to 26). In an initial
questionnaire (see Results), all subjects reported being
familiar with MP skills and strategies.
MATERIALS
The first half of two Domenico Scarlatti sonatas of com-
parable length and difficulty were selected (see Appen-
dix for excerpts from the score of the two pieces,
Supplementary Figure 1 and Supplementary Figure 2).
The pieces were slightly modified to have the highest
degree of comparability without altering the original
musical context. As a result of these manipulations, the
excerpt of the C major sonata (K 72) included a total
number of 387 notes, while the excerpt of the A major
(K 113) included 385; both pieces included 19 bars,
48 four-semiquaver groups in the right hand, 20 four-
semiquaver groups in the left hand (a four-semiquaver
group consists of four successive semiquaver notes). In
both pieces, eight of these four-semiquaver groups were
identically repeated. In the C major sonata excerpt there
were 49 octave notes in the left hand and 54 in the A
major sonata excerpt. Although there was a high
similarity in the general structural form, the two
pieces still had several subtle differences; for example,
the complexity of the four-semiquaver groups was
slightly higher for the A major sonata excerpt. These
differences were preserved to avoid interferences
between the two pieces, but were leveled due to bal-
anced assignment to the two practice conditions. Dur-
ing performance, as well as during PP, subjects played
on a Wersi Digital Piano CT2 (Halsenbach, Germany)
using the standard piano timbre. All practice sessions
and performances were video documented by a digital
video camera from a standardized position that
revealed the profile of the pianists.
Day 1
Scarlatti n.1
(or n. 2)
TMQ
TMQ TMQ TMQ
TMQ
TMQ
30 minutes MP Memory
Performance
Memory
Performance
10 minutes
MP + PP
Memory
Performance
Memory
Performance
30 minutes PP 10 minutes PP
TMQ TMQ
Day 2
Scarlatti n.2
(or n. 1)
FIGURE 1. Experimental design.
Note.
MP ¼Mental Practice. PP ¼Physical Practice. TMQ ¼Ten Minute Questionnaire, administered during practice, aimed at describing the mental
strategies being used.
278 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller

DESIGN
Each subject was asked to learn the two pieces, one via
MP (see the Appendix for a list of abbreviations) and
the other via PP, on two different days (see Figure 1 for
a brief summary of the design). After the practice ses-
sion, the piece had to be performed by memory. Because
of the focus on the memorization paradigm, we selected
two highly comparable music excerpts that had no spe-
cific technical difficulties (which would have brought an
uncontrolled source of variability between subjects);
such pieces would have been too easily sight-read in
a non-memory performance task. A certain degree of
experimental control was therefore maintained in this
component of our design.
The assignment of the two pieces to each condition
(MP or PP) was counterbalanced between the subjects
so that half of the sample studied the A major sonata
excerpt using MP and the C major sonata using PP, and
the other half studied the A major sonata using PP and
the C major sonata using MP. Half of the sample had
the MP trial on the first day; the other half on the
second day. The time interval between the two study
days was on average 5 days. Subjects were randomly
assigned to their respective protocols.
PROCEDURE
Before entering the study, subjects confirmed that they
did not know the two musical pieces and filled out
a questionnaire assessing their familiarity with MP
strategies. This questionnaire was divided into six sec-
tions regarding registry information, MP-abilities, MP-
habits, MP strategies for music memorization, external
resources, and solfe
`ge. Apart from registry information,
each section contained statements (e.g., ‘‘Is mental visu-
alization of the score of a piece useful for you in the
memorization of the music?’’) that had to be rated on
a Likert scale (‘‘1’’ ¼‘‘never’’/‘‘not at all’’/‘‘absolutely
not’’ to ‘‘10’’ ¼‘‘always’’/‘‘perfectly’’/‘‘absolutely’’).
Subjects using MP on the first day were allowed some
time to freely familiarize themselves with the MIDI
piano before the start of the experiment. This was done
to avoid unexpected discomfort when playing on an
unfamiliar instrument following MP. For both, MP and
PP, a metronome indicated a speed of 80 bpm prior to
the start of the practice session. Subjects were asked to
adjust the tempo of their final performance to this
speed. During MP, subjects were seated comfortably in
front of a table with the score of the piece to be studied
and a pencil. Instructions for the use of MP were as
follows: ‘‘You can freely use whatever practice method
you prefer, except for physically playing a real piano.’’ A
MIDI recording of the piece was also available to the
subjects, who were free to listen to it and to pause and
resume it as many times as they wished. Subjects were
allowed to write on the score as well as to move their
fingers.
During PP subjects sat in front of the MIDI piano.
Instructions for the use of PP were as follows: ‘‘We ask
you to focus on physically practicing the piece, ignoring
any mental images you have as you practice. Do not to
stop to mentally rehearse the music and avoid formal
analysis of the piece.’’ In our view, the definition of PP
as a control condition represents a crucial point of
methodology. In past research, some studies (Lim &
Lippman, 1991; Theiler & Lippman, 1995) required
subjects to avoid imagery operations during PP, assum-
ing PP as a ‘‘pure’’ task, complementary to MP. Other
studies (Ross, 1985; Coffman, 1990; Cahn, 2008) did not
provide any definition of PP, leaving unclear to what
extent imagery operations were allowed or not. Only
Lim and Lippman (1991) controlled the degree of imag-
ery operations actually implemented during PP, finding
that ‘‘when given auditory or visual information, or
when making actual movements, it was impossible to
form an image.’’ Even if empirically supported, this
result is somewhat disputable: The authors reported
that in their own study ‘‘the sophistication of the sub-
jects appears to have been overestimated’’ and that
a ‘‘more sophisticated screening of participants may
be necessary, in pursuit of those having highly devel-
oped or better practiced imaging skills’’ (Lim & Lipp-
man, 1991). In the present study, the ‘‘pure PP’’ control
task was employed in order to have past research as
a clear reference point. To assess how often participants
relied on different imagery and mental strategies during
MP as well as during PP, we developed a short ques-
tionnaire (see below) that was administered from time
to time during both MP and PP.
Under both conditions—that is, using MP and PP—
subjects had 30 min (Phase 1-3, each phase with a 10-
min duration) to study the respective piece; subse-
quently, they had to perform the piece on the MIDI
piano by memory twice. Following these two perfor-
mances, subjects had 10 min (Phase 4) to continue
studying the same piece. Subjects who had previously
studied the piece using MP were now free to combine
mental strategies with real piano playing (MPþPP);
subjects who had previously studied the piece using
physical practice were asked to keep on practicing in
the same way, thus avoiding mental rehearsal, imagery,
or formal analysis (PPþPP). Finally, the subjects again
performed by memory twice. For all conditions, subjects
were not forced to memorize the whole piece; while
performing, they were free to play as far as they could,
Mental Practice in Music Memorization 279

but they were explicitly asked to give a performance
coherent with the score, thus avoiding improvisation,
repetitions, or jumping between different bars of the
piece.
MEASUREMENTS
During all practice conditions, after each phase (i.e.,
every 10 min) subjects were asked to fill out a short
questionnaire (Ten Minute Questionnaire; TMQ) doc-
umenting the mental strategies they may have used.
Subjects had to rate on a Likert-scale from ‘‘1’’ (‘‘not
at all’’) to ‘‘5’’ (‘‘very often’’) how often they used the
following strategies: mentally hearing the sound of
notes, mentally feeling the movement of fingers/hands,
mentally visualizing the movements of fingers/hands,
mentally visualizing the score, harmonic analysis of the
piece, rhythmical analysis of the piece, melodic analysis
of the piece.’’ The TMQ was validated by running a pre-
test with a sample of 38 professional musicians (the
expertise as professional musician always being greater
than 4 years). Participants were asked to rate each item
of the TMQ questionnaire assessing both clarity and
relevance of each question on a 5-point scale (‘‘1’’ ¼
‘‘very low’’ to ‘‘5’’ ¼‘‘very high’’). Overall clarity and
usefulness of the instrument was investigated as well.
A questionnaire similar to TMQ, without the harmonic/
rhythmical/melodic questions, was administered after
the performance, to reconstruct which mental strategies
were used while performing. Following the last perfor-
mance after Phase 4, a short interview was conducted to
reconstruct—this time by free recall—which strategies
were used during the 40 min of practice and how thor-
oughly the piece had been formally analyzed. All parti-
cipants confirmed that the pieces did not contain any
sequence difficult to play and that the main challenge of
the task was just to memorize the notes.
At the end of the session, a test for musical auditory
imagery was administered. We developed a test based
on the task described by Highben and Palmer (2004) to
investigate auditory imagery in mental practice
research: Participants were shown the score of a sin-
gle-line melody (9-12 pitches) and simultaneously
heard a melody, which was the same as the notated
melody or had a difference of one pitch. The stimuli
were adjusted by making the one-note difference
a change of 1-2 semitones; the total number of changes
that moved up or down in pitch were balanced. Twelve
of the 16 melodies presented had a one-note difference.
The 16 melodies were played via loudspeakers, and sub-
jects were asked to identify any pitch differences
between the melodies presented by loudspeaker and the
ones presented on the scores. The internal consistency
of this test was assessed by computing Cronbach’s alpha
on the data collected in a pretest with a sample of 20
musicians (the expertise as professional musician
always being greater than 4 years).
Individual differences in mental imagery were tested
by administering the standardized questionnaires
USOIMM77 (Antonietti & Colombo, 1996-1997),
Motor Imagery Questionnaire-Revised (Hall & Martin,
1997), and Verbal-Visual Strategies Questionnaire
(Antonietti & Giorgetti, 1996). USOIMM77 was devel-
oped to assess the spontaneous occurrence of mental
visualization in thinking; the Motor Imagery
Questionnaire-Revised was developed to examine kin-
esthetic and visual movement imagery ability; the
Verbal-Visual Strategies Questionnaire was developed
to measure the cognitive disposition to use visual or
verbal thinking strategies.
Two performances were recorded after each practice
period to control for the variability in individual fluc-
tuations that emerged in the pretest phase of the exper-
iment, with some subjects performing better at the first
recording due to short-term memory resources, and
others at the second due to initial disorientation. The
better performance of the two was selected for further
evaluation. This selection was done according to the
ratio score (see below) computed for both perfor-
mances. The first recording was found to be the best
in four subjects for MP, in three subjects for PP, in 8
subjects for MPþPP, and in 12 subjects for PPþPP.
Note-by-note recordings of the performances were
acquired with a MIDI piano. Error detection analysis
was done manually by the first author. Wrong notes
were defined as any notes not corresponding to the
prescribed note on the original score; an omitted note,
as well as an undesired additional note was treated as
a wrong note. MIDI data were used to compute two
objective parameters of performance: 1) the absolute
number of notes played and 2) the ratio between the
number of wrong notes and the total number of notes
played. The ratio score represents performance accu-
racy, scaled by the length of the performance. This
allows discriminating between subjects who made a sim-
ilar number of errors but played a different extent of the
piece. For example, a subject who played 10 wrong
notes out of 100 total notes would have a better (that
is, lower) ratio score (ratio ¼0.1), compared with
another subject who made the same number of mistakes
while playing only 50 notes of the piece (ratio ¼0.2).
DVD recordings of the performances were indepen-
dently evaluated by three professional musicians (one
pianist and piano teacher, one pianist, one professional
flutist and amateur pianist). The professional
280 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller

experience of these evaluators ranged from 16 to 40
years in their fields. Raters were blind as to which prac-
tice condition preceded the recorded performances and
were provided with the written scores. All performances
were rated on four dimensions: 1) correctness of notes;
2) articulation and phrasing; 3) dynamics and expres-
sion; 4) global score. The first three features are typically
examined during piano performance auditions and
competitions, and have been used in past research on
MP (Theiler & Lippman, 1995); an additional ‘‘global
score’’ was collected to incorporate all aspects of music
performance. Raters judged these dimensions on
a Likert scale ranging from ‘‘1’’ (‘‘poor’’) to ‘‘7’’ (‘‘excel-
lent’’). For the ‘‘correctness of notes’’ dimension, raters
were asked to take into consideration not only the cor-
rectness according to the score (already computed in
the ratio), but also how well the notes, even wrong
notes, fit the context. The ‘‘global score’’ dimension was
independent of the quantity of music played. Recorded
videos of the 30 min of MP (Phase 1-3) were used to
quantify the time each subject spent in the following
overt behaviors: 1) moving the fingers only; 2) singing
only; 3) listening to the audio reproduction of the piece
only; 4) moving the fingers while singing; 5) moving the
fingers while listening to the audio track; 6) total time
moving the fingers (even if other overt operations were
going on); 7) total time singing (even if other overt
operations were going on); and 8) total time listening
to the audio track (even if other overt operations were
going on). The time spent on these operations was
expressed in seconds. This quantification was done by
the first author.
STATISTICAL ANALYSES
Statistical analyses were run on SPSS 15.0. Normal dis-
tribution of the variables was assessed by means of the
Kolmogorov-Smirnov test. In consideration of the small
sample size, a conservative pvalue of .10, instead of .05,
was assumed for this assessment. For each of the four
dimensions of the rating, a measure of inter-rater reli-
ability was obtained by averaging the Pearson correla-
tion coefficient from each possible pair of raters,
a method already used in past research on MP (Lim
& Lippman, 1991; Theiler & Lippman, 1995). Separate
repeated-measures analyses of variance (ANOVA) were
conducted for each performance parameter, in order to
evaluate differences in performance between MP, PP,
MPþPP and PPþPP. Post hoc tests were computed
using the Bonferroni correction for multiple compari-
sons. Partial eta squared (Z
2p
) was assumed as a measure
of effect size. Potential associations between individual
features (initial questionnaires, imagery test) or
strategic data (overt behavior, TMQ) and performance
scores were assessed by the Pearson correlation coeffi-
cient. For the Pearson correlation coefficient, the .05
level of significance was assumed to be two-tailed for
all variables, with the exception of the score on the
musical auditory imagery test in relation to practice
outcomes. The use of imagery during MP and PP was
compared by means of two-sample dependent t-tests.
Imagery during PP was compared with the theoretical
‘‘pure’’ absence of mental imagery by means of one-
sample t-tests, with ‘‘1’’ as the test value; we therefore
compared subjects’ actual answers with the mean of an
hypothetical group of subjects that successfully avoided
mental imagery while doing PP (‘‘1’’ ¼‘‘not at all’’
answer on the TMQ). The validity of the TMQ was
assessed in a one-sample t-test by comparing the scores
obtained in the validation questionnaire with ‘‘3’’ as the
test value. We therefore tested whether subjects’ actual
answers were significantly shifted toward the high
values, compared with the mean of a hypothetical group
of subjects that rated the questionnaire as just ‘‘inter-
mediately’’ relevant and clear. Except for normal distri-
bution, all analyses were evaluated at the .05 level of
significance.
Results
The average level of agreement between the three raters
was .79 for correctness of notes, .63 for articulation and
phrasing, .58 for dynamics and expression, and .75 for
global score. Compared with previous studies (Lim &
Lippman, 1991; Theiler & Lippman, 1995), inter-rater
reliability was considered high enough to warrant aver-
aging the three raters’ independent judgments for each
performance. Each of the four dimensions of rating
showed a high level of correlation with the others (aver-
age correlation for correctness of notes: r¼.84; articu-
lation and phrasing: r¼.84; dynamics and expression: r
¼.80; global score: r¼.85); these dimensions were
therefore collapsed into a single ‘‘rating’’ indicator by
summing up the score on each dimension for each sub-
ject (rating range: 4-28). This procedure was done to
decrease the possibility for Type I error variable-wise,
considering the relatively small sample size. All vari-
ables showed normal distribution, with the exception
of the following MP strategies: formal analysis as
reported in the preliminary questionnaire, and auditory
imagery as reported in the preliminary questionnaire
and in the TMQ. For these variables, the nonparametric
Spearman’s rho was used instead of the Pearson corre-
lation coefficient. The two sonatas did not show differ-
ences in their degree of difficulty in any of the four
Mental Practice in Music Memorization 281

conditions (MP, PP, MPþPP, PPþPP: p> .05 for num-
ber of notes, for ratio and for rating). The TMQ valida-
tion process provided support for its validity: Each of
the seven items received a rating that was significantly
better than ‘‘intermediate’’ for both clarity and relevance
(one-sample t-test: all p< .001; all means > 3.87). The
overall clarity and relevance of the questionnaire also
received a satisfying evaluation, clarity: 4.86 +.83; t(37)
¼31.74, p< .001; relevance: 4.05 +.94; t(37) ¼26.20, p
< .001. Concerning the musical auditory imagery test, it
appeared to have a good internal consistency (Cron-
bach’ alpha ¼.77). All items appeared to be worthy
of retention: The greater increase in alpha would come
from deleting item 1, but removal of this item would
increase alpha only by .02. All items correlated with the
total scale to a good degree (lower r¼.31).
QUALITATIVE EVALUATION OF SUBJECTS’ BACKGROUND
On the initial questionnaire, subjects most frequently
reported the MP strategies of mentally hearing the
sound of notes and using formal analysis (on average
mentally hearing: 8.8 +1.9/10; formal analysis: 7.7 +
2.1/10); visualizing movements was the least commonly
used (on average 5.9 +2.9/10), mentally feeling move-
ments and visualizing the score occupied an intermedi-
ate position (on average feeling movements: 6.9 +2.8/
10; score visualization: 6.8 +2.9/10). Despite the high
level of proficiency in piano playing and knowledge of
MP skills, none of the subjects reported relying on MP
as a major practice strategy.
MP
After 30 min of using MP, subjects were able to per-
form by memory on average 242 +110 notes (range:
112-387), corresponding to 63 +28%of the piece.
The mean ratio (ratio wrong notes/total notes played)
score was .17 +.17 (range: 0 - .62). Mean scores for
number of notes and ratio are reported in Table 1.
Rating scores from the expert raters are reported in
Figure 2.
MP VS. PP
The analysis on the number of notes revealed significant
main effects for the practice condition, F(3, 12) ¼11.51,
p< .001, Z
2p
¼.43, power ¼.99. Post hoc tests revealed
that fewer notes were played after subjects used MP
compared with the other conditions (PP: p¼.02;
MPþPP: p¼.03; PPþPP: p¼.004); no significant
differences could be seen in the notes count between
MPþPP and PP, or between MPþPP and PPþPP
(p> .05). The analysis of the ratio also revealed a signif-
icant main effect for the practice condition, F(3, 12) ¼
7.99, p< .001, Z
2p
¼.34, power ¼.98. Post hoc tests
showed a poorer (i.e., higher) ratio score for MP com-
pared with PP (p¼.04) and MPþPP (p¼.02); a similar
tendency could be observed for PPþPP also (p¼.056),
which did not reach statistical significance (possibly
because of greater fluctuations in ratio scores after
PPþPP, compared with the other conditions). No sig-
nificant differences could be observed between the ratio
values of MPþPP and PP, or between those of MPþPP
and PPþPP (p> .05). Analysis of the ratings revealed
significant main effects for the practice condition,
F(3, 12) ¼26.21, p< .001, Z
2p
¼.64, power ¼1. Post
hoc tests revealed lower ratings after the subjects’ use of
MP compared with all other conditions (p< .001). No
significant differences emerged between PP and MPþPP
(p> .05), while a qualitatively small but significant dif-
ference could be observed between MPþPP and PPþPP
(p¼.02). PPþPP scores on all three dimensions were
possibly conditioned by a ceiling effect, given the non-
significant differences between PP and PPþPP (p> .05).
TABLE 1. Performance Scores for the Four Different Study
Conditions.
Practice
condition Number of notes
Ratio wrong notes /
total notes
MP 242 +110 .17 +.17
PP 326 +101 .08 +.11
MPþPP 319 +96 .07 +.08
PPþPP 349 +86 .04 +.04
Note: MP ¼mental practice; PP ¼physical practice.
MP
0
5
10
15
20
25
PP MP+PP PP+PP
Rating
FIGURE 2. Expert raters’ scores.
Note.
MP ¼Mental practice; PP ¼Physical practice. Expert raters
scored performance with reference to: i) correctness of notes, ii)
articulation & phrasing, iii) dynamics & expression, iv) global
evaluation. The four dimensions were collapsed into a single Rating
indicator, shown in this figure. MP alone is significantly less effective
compared with PP. However, with the addition of a short session of PP,
MP resulted as effective as continued PP.
282 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller

MP: STRATEGIES AND OUTCOMES
The main association between MP strategies and post-
MP outcomes are summarized in Table 2. In the next
paragraphs, results will be displayed for each of the MP
component we evaluated, following the order of the
hypotheses we outlined in the Introduction.
FORMAL ANALYSIS
Stronger reliance on formal analysis for music memo-
rization, as reported in the initial questionnaire, resulted
in better post-MP rating score (Spearman’s rho, two-
tailed: r¼.77, p< .001), and a similar tendency was
observed for the number of notes also (r¼.47, p¼.06).
Self-reports of harmonic, melodic, or rhythmic analy-
sis during the 30 min of MP (TMQ data) did not show
any association with post-MP performance during MP
(p>.05).
AUDITORY IMAGERY
The score on the musical auditory imagery test was
positively related to post-MP performance: Subjects
with more developed aural skills were able to play more
notes (Pearson correlation, one-tailed: r¼.45, p¼.04),
had a better ratio score (r¼-.43, p¼.048), and got
a higher rating score (r¼.45, p¼.04). Pitch imagery as
a means to support music memorization, as reported in
the initial questionnaire, was positively correlated
with post-MP rating score (Spearman’s rho, two-tailed:
r¼.65, p¼.007) and number of notes (r¼.53, p¼
.033), with a similar tendency for ratio (r¼-.48, p¼.06).
Pitch imagery as a strategy reported during the 30 min of
MP (TMQ data) was also correlated with a better rating
score (Spearman’s rho, two-tailed: r¼-.50, p¼.049) and
greater number of notes played (r¼.51, p¼.045).
MOTOR IMAGERY
On the basis of the initial questionnaire and TMQ data,
no associations could be detected between the strategy
of motor imagery and performance scores.
VISUAL IMAGERY
Subjects who relied more often on movement visualiza-
tion, as reported while using MP (TMQ data), gave
a poorer post-MP performance according to the number
of notes (Pearson correlation, two-tailed: r¼-.54, p¼
.03). No interactions could be found between the score
on the Motor Imagery Questionnaire-Revised ‘‘visual’’
subtest and the use of visual imagery of movements from
the TMQ in predicting post-MP performance.
AUDITORY MODELS
The strategy of frequently listening to experts’ perfor-
mances, as reported in the initial questionnaire, was
associated with a better ratio score (Pearson correlation,
two-tailed: r¼-.64, p¼.008). On the contrary, time
spent listening to the auditory model (without moving
fingers or singing) during the 30 min of MP was
associated with a poorer post-MP rating score (r¼-.58,
p¼.018). Moreover, the subjects who reported in the
initial questionnaire to rely on mentally hearing the
sound of notes listened less to the audio recording
during MP (Pearson correlation, two-tailed: r¼-.58,
p¼.018).
OTHER OVERT BEHAVIORS
No associations could be detected between the time
spent moving fingers or singing and performance out-
comes (p>.05)
MENTAL IMAGERY DURING MP AND PP
Figure 3 shows the mean frequency of reliance on dif-
ferent mental strategies during practice. Consistent with
the instructions given, subjects reported (TMQ data)
having used MP strategies significantly less during PP
compared with MP (MP vs. PP, dependent t-test
for mentally hearing the sound of notes: t¼16.35,
p< .001; mentally feeling movements: t¼15.69, p<
.001; mentally visualizing movements: t¼8.61, p< .001;
mentally visualizing the score: t¼3.34, p¼.001;
TABLE 2. Mental Practice Strategies and Outcomes.
Number of notes Rating score Ratio wrong notes /Total notes
Formal analysis (PRE-Q) þ
Auditory imagery (PRE-Q) þþ
Auditory imagery (TMQ) þþ
Visual imagery – motor (TMQ) –
Auditory models (PRE-Q) þ
Auditory models (overt) –
Note: þindicates that a higher score on the item is related with a better post-MP performance; - indicates that a higher score on the item goes
together with a worse post-MP performance. PRE-Q ¼Preliminary questionnaire; TMQ ¼Ten Minute Questionnaire, administered during
mental practice; overt ¼overt behavior during mental practice, quantified from the video documentation.
Mental Practice in Music Memorization 283

harmonic analysis: t¼3.52, p¼.001; melodic analysis:
t¼4.48, p< .001; rhythmic analysis: t¼5.73, p< .001).
Reliance on mental strategies was reduced in the PP
condition also during the last 10 min of practice, when
all subjects were allowed to physically play the instru-
ment (MPþPP vs. PPþPP, mentally hearing the sound
of notes: t¼2.52, p¼.017; mentally feeling move-
ments: t¼5.80, p< .001; mentally visualizing move-
ments: t¼2.35, p¼.025; harmonic analysis: t¼2.81,
p¼.013; rhythmic analysis: t¼3.87, p¼.002). How-
ever, these findings also indicate that mental practice
processes were always active during PP, although sub-
jects were asked to avoid them and although they were
trying to comply with this request. In fact, differences
between the actual use of mental strategies during PP
and the theoretical ‘‘pure’’ absence of these strategies
(TMQ score ¼‘‘1’’, ‘‘not at all’’) are constantly signifi-
cant (one sample t-test for PP Phase 1-3: mentally hear-
ing the sound of notes: t¼6.13; mentally feeling
movements: t¼5.14; mentally visualizing movements:
t¼4.85; mentally visualizing the score: t¼6.14; har-
monic analysis: t¼7.12; melodic analysis: t¼11.15;
rhythmic analysis: t¼7.50; all p< .001).
Discussion
A sample of pianists practiced to perform from memory
two pieces of music of comparable length and difficulty.
One piece was practiced by means of MP and the other
by PP.
The comparison between post-MP and post-PP per-
formances showed that MP alone allowed a level of
proficiency between 40%and 60%of that achieved by
PP. Moreover, combining an intense mental practice
(30 min) with a relatively short physical practice session
(10 min) led to results almost indistinguishable from
those following 30 min of continued PP. These results
are of immediate interest to musicians willing to:
a) optimize the time available for practicing, b) have
a deeper comprehension and stronger mental represen-
tation of the pieces they are practicing, or c) avoid mas-
sive physical practice and thus prevent playing-related
disorders. All of these goals can be achieved by mental
and physical practice properly combined, without a sig-
nificant loss in terms of performance.
In a previous investigation involving behavioral and
neurophysiologic recordings, MP alone resulted in sig-
nificant learning, but did not result in as much perfor-
mance improvement as PP alone (Pascual-Leone et al.,
1995). In fact, at the end of five days of practice, the
mental practice group’s performance was at the same
level as that of the group that had used PP for only three
days. However, the plastic changes in the motor system
following the use of MP alone were the same as those
occurring by repeated PP; moreover, after a single PP
session, the MP group’s performance improved to the
level of the group using PP for five days. As was sug-
gested by Jackson (Jackson, Lafleur, Malouin, Richards,
& Doyon, 2001), part of the behavioral improvement
seen due to MP may be latent, and would thus become
evident after the musician engaged in minimal physical
practice. Mental practice could thus have a preparatory
effect on the task, which increases the efficiency of sub-
sequent physical training (Kopiez, 1990). With specific
regard to music memorization, it is also interesting to
look at the study on MP by Lim and Lippman (1991),
which employed a rating system by expert judges very
similar to the one presented in our study. The average
ratings reported in Lim and Lippman’s study closely
matches those reported in our study, with subjects rat-
ing MP in the range of 3-4 on the 7-point Likert scale in
both studies. Similarities also can be seen when com-
paring measurements of quantity of music that could be
recalled after MP. Both ours and Highben and Palmer’s
(2004) study report that MP yielded 75%of the result
following PP (however, the way music memorization is
quantified in the two studies is not identical). We can
therefore conclude that while present and past studies
differed in terms of the degree of their ecological valid-
ity, they rendered a very similar picture considering the
effectiveness of MP. Two explanations could account
for this result. First, depending on subjects’ aural skills
and familiarity with MP, the lack of clear instructions
on how to use MP may have led expert subjects to
Pitch
imagery
1
2
3
4
5
Frequency
MP
PP
“Pure” PP
Motor
imagery
Visual
imagery-
movements
Visual
imagery-
score
Harmonic
analysis
FIGURE 3. Mental strategies used during practice.
Note.
Averaged reliance on different mental strategies in different
practice condition, derived from averaged TMQ data. All mental
strategies are used maximally during mental practice (MP). However,
also physical practice (PP) is accompanied by imagery strategies,
significantly more than one would expect in a hypothetical practicing
condition with no use of imagery (“pure” PP).
284 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller

completely display their abilities, and hesitant subjects
to get confused and misled. This could be particularly
relevant when considering that none of our subjects
reported systematically using MP while practicing,
a detail that supports the training of subjects in MP
before they are tested in experimental studies. Second,
this result may be due to the fact that past studies also
could not completely rule out the subjects’ free use of
MP, since the ‘‘undesired’’ cognitive operations implied
in MP are unlikely to have been effectively switched off
during the experiment. In fact, the TMQ data showed
an aspect that has been partially neglected in previous
research: PP intrinsically implies MP processes, despite
the subjects’ honest attempt to avoid them. From a cog-
nitive point of view, MP appears to be an automatic
rather than voluntary strategy used when facing a musi-
cal task. A musician can only partially regulate the
degree of the ongoing MP, and in any case, it seems
impossible to turn it off completely. In this respect, our
data are in line with neurophysiological investigations
showing automatic auditory-motor coactivation in
musicians’ brains during musical tasks (Bangert &
Altenmu
¨ller, 2003; Haslinger et al., 2005). Thus,
a dichotomy between MP and PP (or between different
forms of MP), which in practical terms seems evident,
appears to be rather arbitrary in terms of cognitive and
neurophysiological processes. This has already been
recognized by psychoneuromuscolar theorists, who
have shown physiological activity in the form of elec-
tromyographic action potentials as a result of mental
simulation of movements (Grouios, 1992; Hinshaw,
1991). Particular care to this ambiguity should be paid
in the research context in which PP is often considered
a control condition for MP and vice versa. Nevertheless,
contrasting MP and PP still holds a practical signifi-
cance for musicians. The proper experimental design
of this contrast requires the scientist to be conscious
of these tight links; ‘‘pure’’ PP, obtained by asking sub-
jects to avoid mental imagery and rehearsal, appears to
be an invalid control condition.
The present investigation has provided a novel con-
tribution in understanding the relation between indi-
vidual differences in practice strategies and practice
outcome. In fact, the adoption of an ecologically valid
perspective allowed for the first time to directly address
the question, which strategies effectively support mem-
orization in the absence of the physical instrument?
Based on the assumptions of the expert memory theory
(Ericsson & Kintsch, 1995) and its revision for the musi-
cal domain (Chaffin et al., 2009), we have put forward
specific hypotheses for each component of MP that can
now be discussed in light of the data.
Formal analysis as a MP strategy was expected to be
associated with better performance. This was found to
be true, but only to a certain degree. We found a signif-
icant association between the general habit of using
formal analysis for music memorization, as reported
in the initial questionnaire and performance results.
This likely happens because building a formal structure
of the piece allows the performer to use a retrieval, hier-
archical scheme that results in a better organization of
both practice and memory (Chaffin & Imreh, 1997). In
this respect, formal analysis would facilitate the organi-
zation of the material to be remembered in chunks of
information (Miller, 1956), that could be later recalled
as units, thus optimizing the encoding and the retrieval
of memories. The use of the formal structure of a piece
to organize practice and aid memory is a standard rec-
ommendation of piano pedagogues (Leimer & Giesek-
ing, 1998; Sandor, 1981); consistently, experts and
advanced performers appear to utilize analytical strate-
gies frequently, while inexperienced performers do not
(Hallam, 1997; Williamon & Valentine, 2002). How-
ever, no direct associations could be found between the
actual implementation of formal analysis during the
experiment and performance outcomes. Three interre-
lated explanations can be proposed for this unexpected
finding: First, particularly for experienced musicians,
formal analysis may occur as a background, semi-
unconscious process that actually shapes the way the
pianist reads and practices music without necessarily
becoming a deliberate and explicit tactic. As such, it is
more likely to be consistently reported as a general atti-
tude in an initial questionnaire as opposed to a TMQ-
like formulation. Second, the time constraints we imple-
mented could have further influenced the choice of
practice strategies toward implicit formal analysis, so
the subjects could spare the most time for directly prac-
ticing the retrieval from memory. It is possible that in
a more relaxed setting, as the one described by Chaffin
and Imreh (2002), different choices would have been
reported. Third, at a conscious level, the use of formal
analysis might have been obscured by the use of
another, closely connected mental strategy: pitch
imagery.
In fact, pitch imagery was expected to be a central
element of effective MP. This hypothesis was confirmed
by a tightly interconnected array of results, showing that
better post-MP performance was achieved by subjects
who a) had more developed aural skills, b) reported
a general reliance on pitch imagery to aid music mem-
orization, and c) reported the actual use of pitch imag-
ery during the experiment. The other forms of mental
simulation – motor and visual imagery – showed no
Mental Practice in Music Memorization 285

association with effective performance (or even a nega-
tive relation in the case of mental visualization). These
observations lead to the conclusion that effective mem-
orization of a piece of music by mental practice requires
the mental representation of how the music sounds.
Indeed, it is most likely that this key information serves
as raw material for building a higher-level, hierarchical
representation such as the structural one. These empir-
ical findings from a novel experimental setup substan-
tiate practical teaching and pedagogical literature (e.g.,
Gordon, 1997; Leimer & Geiseking, 1998) as well as
previous experimental results with a different method-
ological background (Highben & Palmer, 2004). The
potential of pitch imagery in mental rehearsal can be
also appreciated when considering how it has been
shown to engage the brain. Neuroimaging studies have
shown that neural activity within regions of the second-
ary auditory cortex can occur in the absence of sounds,
and this likely mediates the phenomenological experi-
ence of imagining music (Zatorre, 2007).
Representations of how the music looks or feels seem
more epiphenomenal, being more likely to distract than
to empower music memorization. Future studies will
determine whether a different set of strategies might
be predictive of optimal performance for tasks with
stronger emphasis on the motor, rather than on the
cognitive aspects of piano playing (e.g., playing in a fast
tempo, solving technical difficulties).
Turning to the role of external, auditory models, the
literature we reviewed in the introduction has shown
that external, transitional formats of representation are
helpful to a certain extent (Theiler & Lippman, 1995).
However, experts in a certain field tend not to use them
(Hatano & Miyake, 1977). A very similar picture
emerges from the present results: The habit of enriching
internal representation by collecting, reviewing, and
matching external enlightening models may have
a long-term impact on the ability to reproduce such
models on one’s own. In this way, when faced with the
situation, one can rely on already developed and stored
representation, without wasting resources on external
models that are likely to be more time consuming, less
integrated with the other ongoing processes, and that
possibly even conflict with one’s own models (Lim &
Lippman, 1991). We can hypothesize that the higher the
level of internal auditory/structural representation, the
less effective an auditory external model will be, and
vice versa. Interestingly, while these considerations
apply well to auditory models, they seem not to account
for finger movements. No relation could be observed
between the strategy of simulating piano playing on the
table and practice outcomes. Even more surprisingly,
continuously moving fingers was the strategy used by
many of the more—as well as less—proficient subjects
in our sample. For some, moving fingers seemed to lead
to stable and reliable traces that physically shaped
a robust structural comprehension of the piece; for
other subjects, it seemed a blind, mechanical shortcut
that produced a blurred and weak performance. These
results may open a debate that may be answered more
conclusively with a larger sample. In any case, the pres-
ent results challenge the validity of studies that force
subjects toward predetermined ways of mental practice.
Altogether, these findings have practical implications
for the way musicians could rehearse or memorize new
repertoire when the instrument is not available: First,
imagery of the sounds should be a default operation,
a foundation on which other operations rest. Second,
analyzing the structure of the piece in terms of har-
monic relations, melodic phrases, and rhythmic struc-
tures is another key component of effective MP. This
applies to formal analysis not just as a deliberate,
explicit operation, but also as a background process that
runs parallel to the auditory/motor rehearsal. Third,
reliance on external models can be an effective way to
support practice, as long as it is clear that the goal is to
build up an auditory/structural mental representation
that holds even when the model is no longer present.
Finally, motor and visual cues as well as overt finger
tapping could be of occasional help, depending on the
specific nature of the piece and the subject’s preferences.
However, they alone do not provide a reliable founda-
tion for mental study, and in some cases they might
even become a source of distraction.
Another question that prompted the present study
was whether there are common MP profiles. From
a qualitative overview of overt and covert practice strat-
egies, mental practice for music memorization appeared
to exist in two different forms. On the one hand, there
was a mostly ‘‘mental’’ use of MP. Subjects who applied
this type of MP were mainly focused on their internal
processes, sometimes aiding their mental representa-
tions by listening to the auditory model. They showed
almost no overt behavior, and they were focused on the
abstract-formal analysis of the piece, without spending
particular effort on the imagery of the precise move-
ments. On the other hand, a more physical form of
MP was observed. Subjects who applied this type of
MP frequently supported their internal operations with
overt behaviors such as moving fingers and/or singing,
and balancing their internal operations between formal
analysis and movement imagery. Interestingly, our data
do not show an advantage of one category over the
other, as the main discriminating factor between these
286 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller

groups—finger movements—was not associated with
practice outcomes. Both kinds of patterns may lead per
se to good or poor results, and probably each musician
should find a personal, optimal balance between the
two, also depending on the task.
While providing new insights into the mechanisms
and possible applications of MP, this study presents the
following limitations: a) The subjects’ selection.
Although all our subjects were familiar with MP, none
relied on it as a major practice strategy; b) The small
sample size. Results await to be confirmed by a larger
sample; c) The task. To perform a novel piece of music by
memory after a short practice session is not a very com-
mon task, as already noted by Theiler and Lippman
(1995). This problem is likely to interact with our sub-
jects’ lack of expertise with MP, since even those who
really had some MP skills probably never used them
when faced with this kind of task; d) ‘‘Pure’’ PP as a con-
trol condition for MP. Although we were aware that
inclusion of ‘‘pure’’ PP in our experimental design could
have presented methodological problems, these problems
had never been addressed or quantified in previous stud-
ies. The present investigation sought to demonstrate
these problems and, thus, eliminate a source of ambiguity
for future research. This intention, however, limits the
ecological validity of our control condition, keeping
a component of unnaturalness in our investigation.
Such limitations may interestingly guide the next
steps for further research on MP. First, one could train
subjects in MP prior to the study. The quantification of
MP skills could be related in diaries for daily-use report
as well as measured objectively (e.g., musical auditory
imagery test, solfe
`ge, sight-reading, improvisation, ear-
learning). A detailed entry file would allow researchers
to keep even low-experienced subjects and to make sep-
arate analyses for subjects with different skill levels.
Second, an exploratory, qualitative survey aimed to
describe how and when musicians typically apply MP
could precede a novel experiment employing more
widely used tasks in an ecologically valid setting. Third,
MP could be contrasted with ‘‘ecologically valid’’ PP
(not assumed to exclude MP processes). Greater effort
could be used to describe MP processes: TMQ reliability
could be improved by randomizing the order of the item
and reversing the scales to better disrupt memory for the
previous rating assigned; think aloud procedures could
also be considered. Finally, designs with free MP-PP com-
binations could be developed, giving, for example,
percentage-time for each condition as the only constraint.
Developing research in these directions will be of
great interest in both theoretical and applied frame-
works. The results from the present, methodologically
‘‘open’’ study represent an encouraging step toward
a deeper comprehension of the mechanisms by which
memorization of complex tasks can be improved. No
less important, conclusions from this line of research
could have a direct impact on the possibility for musi-
cians to better cope with health risk factors and to pro-
mote their own well-being.
Author Note
We would like to thank Professor Reinhard Kopiez,
Professor Louis Lippman, and the anonymous
reviewers for providing insightful comments at various
stages of this project. We also thank Michael Grossbach
for technical assistance.
Correspondence concerning this article should be
addressed to Nicolo
`Francesco Bernardi, Department
of Psychology, University of Milano-Bicocca, Building
U6 - Room 3170c, Piazza dell’Ateneo Nuovo 1, 20126
Milano, Italy. E-mail: bernardi.nf@gmail.com
References
ANTONIETTI A., & COLOMBO B. (1996-1997). The spontaneous
occurrences of mental visualization in thinking. Imagination,
Cognition, and Personality, 16, 415-428.
ANTONIETTI,A.,&GIORGETTI, M. (1996). A study of some
psychometric properties of the Verbalizer-Visualizer
Questionnaire. Journal of Mental Imagery,20, 59-68.
BANGERT,M.,&ALTENMU
¨LLER, E. (2003). Mapping perception
to action in piano practice: A longitudinal DC-EEG-study.
BMC Neuroscience, 4, 26-36.
BARRY,N.H.,&MCARTHUR, V. (1994). Teaching practice
strategies in the music studio: A survey of applied music tea-
chers. Psychology of Music, 22, 44-55.
CAHN, D. (2008). The effects of varying ratios of physical and
mental practice, and task difficulty on performance of a tonal
pattern. Psychology of Music, 36, 179-191.
CHAFFIN,R.,&IMREH, G. (1997). ‘Pulling teeth and torture’:
Musical memory and problem solving. Thinking and
Reasoning, 3, 315-336.
CHAFFIN,R.,&IMREH, G . (2002). Practicing perfection: Piano per-
formance as expert memory. Psychological Science,13,342-349.
CHAFFIN,R.,LOGAN,T.R.,&BEGOSH, K. T. (2009). Performing
from memory. In S. Hallam, I. Cross, & M. Thaut (Eds.), The
Oxford handbook of music psychology (pp. 352-363). Oxford,
UK: Oxford University Press.
Mental Practice in Music Memorization 287

CHASE,W.,&SIMON, H. (1973). Perception in chess. Cognitive
Psychology, 4, 55-81.
CHEN,D.D.,HENDRICK J. L., & LIDOR R. (2002). Enhancing
self-controlled learning environments: The use of self-
regulated feedback information. Journal of Human Movement
Studies, 43, 69-86.
CHIVIACOWSKY S., & WULF, G. (2002). Self-controlled feed-
back: Does it enhance learning because performers get
feedback when they need it? Research Quarterly of Exercise
and Sport, 73, 408-415.
COFFMAN, D. D. (1990). Effects of mental practice, physical
practice, and knowledge of results in piano performance.
Journal of Research in Music Education, 38, 187-196.
CONNOLLY,C.,&WILLIAMON, A. (2004). Mental skills training.
In A. Williamon (Ed.), Musical excellence: Strategies and
techniques to enhance performance (pp. 221-245). Oxford, UK:
Oxford University Press.
DRISKELL,J.E.,COPPER,C.,&MORAN, A. (1994). Does mental
practice enhance performance? Journal of Applied Psychology,
79, 481-492.
ERICSSON,K.A.,&KINTSCH, W. (1995). Long-term working
memory. Psychological Review, 102, 211-245.
FELTZ,D.L.,&LANDERS, D. (1983). The effects of mental
practice on motor skill learning and performance: A meta-
analysis. Journal of Sport Psychology, 5, 25-57.
FELTZ,D.L.,LANDERS,D.M.,&BECKER, B. J. (1988). A revised
meta-analysis of the mental practice literature on motor skill
learning. In D. Druckman & J. A. Swets (Eds.), Enhancing
human performance: Issues, theories, and techniques (Part III,
pp 1-65). Washington, DC: National Academy Press.
GABRIELSSON, A. (1999). Music performance. In D. Deutsch
(Ed.), The psychology of music (2nd ed., pp. 501-602). San
Diego, CA: Academic Press.
GINSBORG, J. (2002). Singing by heart: Memorization strategies
for the words and music of songs. In J. W. Davidson (Ed.), The
music practitioner: Exploring practices and research in the
development of the expert music performer, teacher and listener
(pp. 149-160). London, UK: Ashgate Press.
GORDON, E. E. (1997). A music learning theory for newborn and
young children. Chicago, IL: GIA Publications.
GROUIOS, G. (1992). Mental practice: A review. Journal of Sport
Behavior, 15, 42-59.
HALL,C.R.,&MARTIN, K . A. (1997). Measuring movement
imagery abilities: A revision of the Movement Imagery
Questionnaire. Journal of Mental Imagery, 21, 143-154.
HALLAM, S. (1997). The development of memorization strategies
in musicians: Implications for education. British Journal of
Music Education, 14, 87-97.
HASLINGER,B.,ERHARD,P.,ALTENMU
¨LLER,E.,SCHROEDER,U.,
BOECKER,H.,&CEBALLOS-BAUMANN, A. O. (2005). Transmodal
sensorimotor networks during action observation in profes-
sional pianists. Journal of Cognitive Neuroscience, 17, 282-293.
HATANO,G.,&MIYAKE, Y . (1977). Performance of expert aba-
cus operators. Cognition, 5, 57-71.
HIGHBEN,Z.,&PALMER, C. (2004). Effects of auditory and motor
mental practice in memorized piano performance. Bulletin of
the Council for Research in Music Education, 159,58-65.
HINSHAW, K. E. (1991). The effects of mental practice on motor
skill performance: Critical evaluation and meta-analysis.
Imagination, Cognition and Personality, 11, 3-35.
JACKSON,P.L.,LAFLEUR,M.F.,MALOUIN,F.,RICHARDS,C.,&
DOYON, J. (2001). Potential role of mental practice using
motor imagery in neurologic rehabilitation. Archives of
Physical Medicine and Rehabilitation, 82, 1133-1141.
JANELLE,C.M.,BARBA,D.A.,FREHLIC H,S.G.,TENNANT,L.K.,&
CAURAUGH, J. H. (1997). Maximising performance effectiveness
through videotape replay and a self-controlled learning envi-
ronment. Research Quarterly of Exercise and Sport, 68, 269-279.
KLO
¨PPEL, R. (1996). Mentales training fu¨r musiker [Mental
training for musicians]. Kassel, Germany: Bosse.
KOPIEZ, R. (1990). Der Einfluss kognitiver strukturen auf das
Erlernen eines Musikstu¨cks am Instrument [The influence of
cognitive structures on learning a piece of music on an
instrument]. Frankfurt am Main, Germany: Peter Lang.
LEIMER,K.,&GIESEKI NG, W. (1998). Modernes Klavierspiel
[Modern piano-playing] (27th ed.). Mainz, Germany: Schott.
LIM,S.,&LIPPMAN, L. G. (1991). Mental practice and memo-
rization of piano music. The Journal of General Psychology,
118, 21-30.
LOCKE,J.,&LOCKE, V . (1971). Deaf children’s phonetic, visual
and dactylic coding in a grapheme recall task. Journal of
Experimental Psychology, 89, 142-146.
MCMILLAN, J. (2005). Strategies for memorising. ISM Music
Journal,December, 268-272. http://www.jennymacmillan.co.
uk/Memorising%20for%20ISM.pdf
MILLER, G. A. (1956). The magical number seven, plus or minus
two: Some limits on our capacity for processing information.
Psychological Reviews, 63, 81-97.
ORLOFF-TSCHEKORSKY, T. (1996). Mentales Training in der
musikalischen Ausbildung [Mental training in musical educa-
tion]. Aarau: Musikedition Nepomuk.
PASCUAL-LEONE,A.,NGUYET,D.,COHEN,L.G.,BRASIL-NETO,
J. P., CAMMAROTA,A.,&HALLETT, M. (1995). Modulation of
muscle responses evoked by transcranial magnetic stimulation
during the acquisition of new fine motor skills. Journal of
Neurophysiology, 74, 1037-1045.
ROSS, S. L. (1985). The effectiveness of mental practice in
improving the performance of college trombonists. Journal of
Research in Music Education, 33, 221–30.
RUBIN-RABSON, G. (1937). The influence of analytical pre-study
in memorizing piano music. Archives of Psychology, 20, 3-53.
RUSHALL, B. S . (1991). Imagery training in sports: A handbook for
athletes, coaches and sport psychologists. Spring Valley, CA:
Sports Science Associates.
288 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller

SANDOR, G. (1981). On piano playing: Motion, sound, expression.
New York: Schirmer Books.
THEILER,A.M.,&LIPPMAN, L . G. (1995). Effects of mental
practice and modeling on guitar and vocal performance.
Journal of General Psychology, 122, 329–343.
THOMPSON,C.P.,COWAN,T.M.,&FRIEMAN, J. (1993).
Memory search by a memorist. Hillsdale, NJ: Erlbaum.
VAN MEER,P.J.,&THEUNISSEN, N . C. M. (2009). Prospective
educational applications of mental simulation: A meta-review.
Educational Psychology Review, 21, 93-112.
WILLIAMON,A.,&VALENTINE,E.(2002).Theroleofretrieval
structures in memorizing music. Cognitive Psychology, 44,
1-32.
ZATORRE, R. J. (2007). There’s more to auditory cortex than
meets the ear. Hearing Research, 229, 24-30.
ZIMMERMANN -SCHLATTER,A.,SCHUSTER,C.,PUHAN., A.M.,
SIEKIERKA,E.,&STEURER , J. (2008). Efficacy of motor imagery
in post-stroke rehabilitation: A systematic review. Journal of
NeuroEngineering and Rehabilitation, 14, 5-8.
Appendix
TEN MINUTE QUESTIONNAIRE (TMQ)
The following statements describe strategies that a musi-
cian might use while using mental practice.
According to the way you practiced, rate each state-
ment from 1 to 5 using the following scale:
1¼never
2¼seldom
3¼sometimes
4¼often
5¼very often
1. How often did you imagine the sound of notes?
2. How often did you imagine the feeling of the move-
ment of your fingers or hand?
3. How often did you visualize in your mind the move-
ment of your fingers or hand?
4. How often did you visualize in your mind the music
score?
5. How often did you analyze the harmonic structure of
the piece?
6. How often did you analyze the melodic structure of
the piece?
7. How often did you analyze the rhythmic structure of
the piece?
SUPPLEMENTARY FIGURE 1. Excerpt from Sonata K 72.
Mental Practice in Music Memorization 289

SUPPLEMENTARY FIGURE 2. Excerpt from Sonata K 113.
290 Nicolo
`Francesco Bernardi, Alexander Schories, Hans-Christian Jabusch, Barbara Colombo, & Eckart Altenmu¨ller