ORIGINAL RESEARCH

published: 08 March 2019

doi: 10.3389/fnins.2019.00182

Music and Visual Art Training

Modulate Brain Activity in Older
Adults
Claude Alain 1* , Aline Moussard 2 , Julia Singer 1 , Yunjo Lee 1 , Gavin M. Bidelman 3 and
Sylvain Moreno 4
1
Rotman Research Institute, Baycrest Centre for Geriatric Care, Toronto, ON, Canada, 2 Centre de Recherche de l’Institut
Universitaire de Gériatrie de Montréal, Université de Montréal, Montréal, QC, Canada, 3 Institute for Intelligent Systems –
School of Communication Sciences and Disorders, The University of Memphis, Memphis, TN, United States, 4 Digital Health
Hub, School of Engineering Science, Simon Fraser University, Surrey, BC, Canada

Cognitive decline is an unavoidable aspect of aging that impacts important

behavioral and cognitive skills. Training programs can improve cognition, yet precise
characterization of the psychological and neural underpinnings supporting different
training programs is lacking. Here, we assessed the effect and maintenance (3-month
follow-up) of 3-month music and visual art training programs on neuroelectric brain
activity in older adults using a partially randomized intervention design. During the pre-,
post-, and follow-up test sessions, participants completed a brief neuropsychological
assessment. High-density EEG was measured while participants were presented with
auditory oddball paradigms (piano tones, vowels) and during a visual GoNoGo task.
Edited by: Neither training program significantly impacted psychometric measures, compared to
Gerard Francisco, a non-active control group. However, participants enrolled in the music and visual art
University of Texas Health Science
training programs showed enhancement of auditory evoked responses to piano tones
Center at Houston, United States
that persisted for up to 3 months after training ended, suggesting robust and long-
Reviewed by:
Lutz Jäncke, lasting neuroplastic effects. Both music and visual art training also modulated visual
University of Zurich, Switzerland processing during the GoNoGo task, although these training effects were relatively
Cyrille Magne,
Middle Tennessee State University, short-lived and disappeared by the 3-month follow-up. Notably, participants enrolled
United States in the visual art training showed greater changes in visual evoked response (i.e., N1
*Correspondence: wave) amplitude distribution than those from the music or control group. Conversely,
Claude Alain
those enrolled in music showed greater response associated with inhibitory control
[email protected]
over the right frontal scalp areas than those in the visual art group. Our findings
Specialty section: reveal a causal relationship between art training (music and visual art) and neuroplastic
This article was submitted to
Auditory Cognitive Neuroscience,
changes in sensory systems, with some of the neuroplastic changes being specific to
a section of the journal the training regimen.
Frontiers in Neuroscience
Keywords: aging, music training, art training, brain plasticity, executive functions, ERPs
Received: 31 October 2018
Accepted: 15 February 2019
Published: 08 March 2019
INTRODUCTION
Citation:
Alain C, Moussard A, Singer J,
Aging is associated with structural and functional changes in prefrontal, parietal, and medial
Lee Y, Bidelman GM and Moreno S
(2019) Music and Visual Art Training
temporal regions, which has been linked to enhanced sensory evoked responses in auditory
Modulate Brain Activity in Older and visual areas as well as deficits in attention, memory and executive functions (Dempster,
Adults. Front. Neurosci. 13:182. 1992; Alain and Woods, 1999; West and Alain, 2000; Bugaiska et al., 2007; Kropotov et al.,
doi: 10.3389/fnins.2019.00182 2016; Aljondi et al., 2018). Some theories emphasize age-related declines in inhibitory processes

Frontiers in Neuroscience | www.frontiersin.org 1 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

(Hasher and Zacks, 1988; Knight, 1994; West, 1996), which comparing it with that of another artistic activity would allow for
involve the ability to regulate incoming sensory input and a greater understanding of brain plasticity and potential transfer
behavior so as to override internal predispositions or external mechanisms in the aged brain.
lures to accomplish more appropriate actions or behaviors A cognitively demanding form of training that could be
(Diamond, 2013). Age-related declines in attentional regulation considered equally as engaging as musical activity is visual
and inhibitory control (May et al., 1999; Erber, 2013) result art. Evidence from neuroimaging studies suggests that activity
in negative consequences for perceptual and cognitive skills elicited during visual art engagement, including both visual object
(Gazzaley et al., 2007; Lustig et al., 2007), which may be learning (Op de Beeck and Baker, 2010) and hand-related motor
reflected in the overall responsiveness (amplitude) of sensory activities (Draganski et al., 2004), induces neuroplastic changes in
evoked responses (Bidelman et al., 2017). Age-related declines brain areas associated with spatial-reasoning skills (Pollmann and
in inhibitory control may also be associated with impairments von Cramon, 2000). Visual art training has also been associated
in language comprehension (Pichora-Fuller et al., 1995, 2017) with structural differences in areas of the brain pertaining to fine
and can lead to an increase in socially inappropriate behavior. motor control and procedural memory between artists and non-
Given the rapid expansion of the aging population in modern artists (Chamberlain et al., 2014). Therefore, there is reason to
society, there is increasing need to identify and develop believe that visual art training, like musical training, may promote
effective interventions that reduce, halt, or even reverse declines cognitive improvements in older adults. Importantly, visual art
in inhibitory control to preserve daily cognitive abilities training provides a means to assess whether neuroplastic changes
in older adults. associated with music training are specific to music training itself
Recent studies have shown that lifelong engagement in or whether they reflect non-training specific effects associated
musical activities might help maintain the brain in a younger with engagement in art programs in general.
state (Rogenmoser et al., 2018), slow cognitive declines (Hanna- In this study, we investigated the short-term (3-month) impact
Pladdy and Gajewski, 2012; Parbery-Clark et al., 2012; Zendel and of two forms of engaging training – music and visual art
Alain, 2012; Bidelman and Alain, 2015), and preserve or even instruction – on older adults’ perceptual and cognitive functions
enhance neural functioning in old age (Parbery-Clark et al., 2011; using a three group intervention design (music, visual art and
Bidelman et al., 2014; Zendel and Alain, 2014). For example, a no-contact control). We also measured maintenance of the
older musicians show improved performance compared to non- effects with a follow-up testing 3 months after the cessation of
musicians in several cognitive areas including spatial memory, training. All participants completed a brief neuropsychological
processing speed, and cognitive flexibility (Parbery-Clark et al., assessment to ensure that our groups were comparable before
2011), as well as inhibitory control (Moreno et al., 2014; training. We used the same battery after training and at follow-
Moussard et al., 2016). A positive relationship between years of up to explore whether music and visual art training impacted
musical engagement and cognitive performance further implied cognitive functioning as measured with neuropsychological tests.
experience-dependent changes in older adults’ behavioral skills We anticipated that participants in the music group would show
(Parbery-Clark et al., 2011, 2012; Bidelman and Alain, 2015). higher cognitive functioning than those in the active (visual
Several developmental studies have been conducted art) and passive control groups. We anticipated that participants
investigating the effect of learning how to sing or to play in the music group would show higher cognitive functioning
a musical instrument over a period of time. Using expert after training than those in the control group. We also expected
and intervention designs, results have shown benefits on potential improvements following visual art training, as an
different types of cognitive processes such as verbal processing exploratory hypothesis.
(Moreno et al., 2009; Seither-Preisler et al., 2014), intelligence In addition, we recorded neuroelectric event-related potentials
(Schellenberg, 2004), reading (Moreno et al., 2011), inhibitory (ERPs) using auditory oddball paradigms and during a visual
processing (Moreno et al., 2011), auditory processing (Moreno GoNoGo task at each session (pre, post, and follow-up),
and Farzan, 2015; Tierney and Kraus, 2015) and on general to evaluate training-related neuroplastic changes in sensory
brain development (Hyde et al., 2009). However, evidence to processing and executive functions. These paradigms were
date demonstrating benefits of musical training in older adults chosen because prior studies have shown neuroplastic changes
has often been correlational, not causal. Very few studies have associated with music training on these tasks (e.g., Shahin
investigated, in longitudinal designs, whether engagement in et al., 2003; Zendel and Alain, 2009; Moussard et al., 2016).
musical activity can improve cognitive functions. Although We hypothesized that each form of art training would impact
evidence suggests that short-term musical training (e.g., piano sensory processing in the trained modality paralleling training-
lessons) can improve executive functioning and working memory related benefits reported in children (e.g., Fujioka et al., 2006;
(Bugos et al., 2007; Seinfeld et al., 2013), prior studies often suffer Moreno et al., 2011, 2014). That is, music training was expected
from methodological limitations (e.g., lack of an active control to impact the N1 and P2 waves of the auditory ERPs as previously
group, lack of randomization for group assignment), and have found in younger populations (e.g., Tremblay et al., 2001;
not measured the effects of the intervention on brain activity. Reinke et al., 2003; Alain et al., 2010). We also anticipated
Thus, it remains unclear whether engagement in musical that music training would improve listeners’ ability to notice
activities would yield neuroplastic changes in brain activity, changes in auditory stimuli as indexed by the mismatch
and whether these changes could be long-lasting. Examining negativity (MMN), which has also been observed in young
neural activity prior to, and after short-term music training and adults (Tervaniemi et al., 1997; Nikjeh et al., 2009). We also

Frontiers in Neuroscience | www.frontiersin.org 2 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

hypothesized that visual art training would have greater impact Scale of Intelligence-Second Edition (WASI-II FSIQ4; p = 0.51;
than music training on visual sensory evoked responses such as music: M = 114.2, SD = 10.5; visual art: M = 115.8, SD = 11;
the N1 and P2 waves at parietal-occipital sites. control: M = 111.8, SD = 13.9). Participants were screened
Furthermore, we expected to observe neuroplastic for amusia and other auditory or musical deficits that could
enhancements in brain mechanisms subserving inhibitory have interfered with the study using the Musical Ear Test
control in the music group relative to the visual art group and (MET, Wallentin et al., 2010). All three groups showed similar
age-matched no-training controls. Following music training, scores at baseline psychometric assessment (all p-values > 0.1).
changes were expected in the N2, and P3 complex of the ERPs After 3 months, 15 participants from the music group and 14
during the visual GoNoGo task as previously found in younger participants from the visual art group returned for follow-up
(Brydges et al., 2014; Moreno et al., 2014) and older (Moussard testing. These two subgroups remained similar in age (p = 0.57),
et al., 2016) populations. As shown in previous studies in education (p = 0.31), and had comparable intelligence on WASI-
young adults, music training induced enhancements in right II FSIQ4 (p = 0.39) at pre-test. The study received approval from
hemisphere activity (Moreno et al., 2014). Thus, we hypothesized the Baycrest Research Ethics Committee, and all participants
that the music group would show an increase in ERP amplitude provided written informed consent.
over the right hemisphere after training. Prior research has
also shown long lasting benefits of music training that persist Study Design
years after the training has ended (White-Schwoch et al., 2013). This longitudinal study consisted of four phases: pre-test,
Hence, we expected brain and behavioral enhancements to 3-month training, post-test, and 3-month follow-up test
persist at 3-month follow-up. This finding would establish that (Figure 1). During the 3 months between post-test and follow-
relatively engaging art training could produce lasting effects in up, participants did not engage in formal music or visual
neural and behavioral function in older adults well after the arts activities. At pre-test, post-test, and follow-up sessions,
cessation of instruction. participants were tested individually and were blind to our
hypotheses. After the pre-test, participants were assigned to either
MATERIALS AND METHODS music or visual art training in a pseudorandom manner to equate
pre-training differences between groups on intelligence scores
Participants and background demographic measures (gender, age, and years
Sixty healthy older adults with limited prior musical or visual of education). An additional passive control group was further
art training were recruited from the Greater Toronto Area. recruited in order to collect data to distinguish potential training
For pre- and post-training phases, five participants were lost effects from test–retest effects.
to attrition, and two due to technical problems during EEG
recording, resulting in 17 participants in the music group (three Training Curricula
males), 19 in the visual art group (two males), and 17 in the no- The trained participants received group classroom instruction
contact control group (three males). The groups did not differ and activities in their respective training by a professional teacher
in age (p = 0.82; music: M = 67.7, SD = 5.8 years; visual art: (i.e., two teachers: a music and a visual art teacher) at the Royal
M = 68.9, SD = 6 years; control, M = 68.5, SD = 6 years), Conservatory of Music in Toronto for 3 months (36 1-h sessions,
years of formal education (p = 0.55; music: M = 16.4, SD = 2.6 three times per week). The music group was engaged in music
years; visual art: M = 17.2, SD = 2.3 years; control: M = 16.9, making using body percussion, voice, and non-pitched musical
SD = 1.5 years), or intelligence on the Wechsler Abbreviated instruments. They also learned basic music theory as well as

FIGURE 1 | Schematic of the study design.

Frontiers in Neuroscience | www.frontiersin.org 3 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

melody and harmony concepts through the singing of simple EEG Recording and Data Analysis
canons. The visual art group learned basic drawing and painting To probe training-related changes in brain activity supporting
techniques, analyzed work of famous artists, and created original perceptual and executive control processes, we measured
paintings of landscape, still-life, and self-portrait. All materials neuroelectric activity (electroencephalogram, EEG) using
were provided for them (i.e., instruments for music lesson and auditory oddball paradigms and during a visual GoNoGo task.
all the material for drawing and painting in the visual art lesson). The same stimuli were used at each of the three testing sessions
(pre, post, follow-up) while the order of trials was randomized
Procedure across participants and sessions.
Testing (EEG and psychometric tests) took place in the laboratory
on two different days and lasted approximately 1.5–2 h per Auditory oddball paradigms
session. For EEG testing, recordings took place in an acoustically Auditory processing was assessed via the N1, P2 and mismatch
and electrically shielded room. negativity (MMN) recorded before and after training using two
oddball sequences consisting of a either music or speech sound
Psychometric Assessment contrast. For the music condition, two synthesized piano tones,
Wechsler Abbreviated Scale of Intelligence-Second Edition Eb4 (F0 = 314 Hz) and D4 (F0 = 294 Hz) served as the standard
(WASI-II) and deviant tokens, respectively. The two notes differed only
WASI-II assesses intelligence using four different subtests in pitch and were otherwise matched in acoustic characteristics
including measures of verbal comprehension: (1) Vocabulary and including overall amplitude and duration (500 ms). For the
(2) Similarities; and measures of perceptual reasoning: (3) Block speech condition, two French vowels as produced by a native
Design and (4) Matrix Reasoning. A composite score (FSIQ4) was female speaker, /u/ and /ou/, functioned as the standard and
then calculated to reflect global IQ. deviant stimulus, respectively. The two speech tokens had similar
duration (280 ms), mean voice fundamental frequency (F0:
Forward and Backward Word Span ∼240 Hz), amplitude, and first/third formant frequencies; only
The Word Span task measured verbal working memory, with their second formant differed to yield the two distinct vowel
one forward and one backward condition. The experimenter read timbres (/u/: ∼1,850 Hz; /ou/: ∼750 Hz). Presentation order
sequences of words with a 1-s pause between words. Word lists was pseudorandom such that at least one standard stimulus
were organized into sequence lengths (spans) of two to eight. preceded a deviant. Stimuli were presented with a stimulus
There were two versions of the test that were counterbalanced onset asynchrony of 750 ms and delivered binaurally via ER-3
across testing sessions. Each sequence had two trials. Testing insert earphones (Etymotic Research) at an intensity of 80 dB
continued until the experimenter recorded all recalled words sound pressure level. A total of 510 standards and 90 deviants
from each list. A total span score was calculated for each (i.e., 85/15% ratio) were collected for both the speech and
condition as the total number of correctly recalled items across music conditions.
all trials (maximum word span score = 70).
Visual GoNoGo task
Stroop Test
Participants were presented with white or purple geometric
A paper version with three subtests of the Stroop test was used
triangle or squares at the center of the screen located at about
to measure cognitive inhibition and processing speed (Stroop,
1 m from the participant. Before each trial, a white fixation cross
1935). Subtest 1 involved reading color words printed in black
appeared on a black background for a variable duration (500–
font while subtest 2 included naming the color of squares. Subtest
1,000 ms), then a geometric shape appeared in the center of
3 required naming the color in which the word was printed while
the screen for 500 ms. Participants were instructed to press the
the written color name differed (i.e., incongruent condition).
right mouse button in response to white shapes (80% probability)
Response time was measured for each subtest. The interference
and to withhold responding to purple shapes (20% probability).
score was calculated by subtracting response time on subtest 3
The experiment consisted of 200 trials (160 Go and 40 NoGo
from that of subtest 2. The interference effect was also reflected
trials). A practice block of 20 trials was used to familiarize
by the number of mistakes at subtest 3.
participants with the task. During the task, participants did
Computerized Peabody Picture Vocabulary Test (PPVT) not receive feedback on their performance. Accuracy rates were
The Peabody Picture Vocabulary Test (PPVT) was used to recorded for Go and NoGo trials, and reaction times were
measure receptive vocabulary (Dunn and Dunn, 2007). In this recorded for Go trials.
test, participants saw four pictures on the screen while the
EEG recording and data processing
computer said a word. Their task was to click on the picture that
EEG was recorded from 66 scalp electrodes using a BioSemi
best characterized the meaning of the word. We used the number
Active Two acquisition system (BioSemi V.O.F., Amsterdam,
of correct responses in the analysis.
Netherlands). The electrode montage was according to the
Digit symbol (subtest of the WAIS-R) BioSemi electrode cap based on the 10/20 system and included
Participants were given 2 min to copy symbols below numbers a common mode sense active electrode and driven right leg
according to a coding key where a number corresponded to a passive electrode serving as ground. Ten additional electrodes
specific symbol. The test was scored as the number of correctly were placed below the hair line (both mastoid, both pre-auricular
completed symbols. points, outer canthus of each eye, inferior orbit of each eye,

Frontiers in Neuroscience | www.frontiersin.org 4 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

two facial electrodes) to monitor eye movements and to cover et al., 2003; Moreno et al., 2014; Moussard et al., 2016).
the whole scalp evenly. The latter is important because we For the oddball paradigms, the effects of training on N1
used an average reference (i.e., the average of all scalp EEG and P2 deflections was assessed during the 90–130 ms and
channels as the reference for each EEG channel) for ERP 170-210 ms interval at fronto-central scalp sites (F1, Fz, F2,
analyses. Neuroelectric activity was digitized continuously at rate FC1, FCz, FC2, C1, Cz, C2). The automatic change detection
of 512 Hz with a bandpass of DC-100 Hz, and stored for offline was quantified using the difference waves between standard
analysis. Offline analyses were performed using Brain Electrical and deviant. The peak amplitude and latency was quantified
Source Analysis software (BESA, version 6.1; MEGIS GmbH, during the 100–250 ms window using the same electrode
Gräfelfing, Germany). clusters as the N1 and P2 waves. For the GoNoGo task, the
Continuous EEGs were first digitally filtered with 0.5 Hz processing of go and nogo stimuli was assessed during the
high-pass (forward, 6 dB/octave) and 40 Hz low-pass filters 165–205 ms interval (i.e., N1) and the 210–280 ms interval
(zero phase, 24 dB/octave). For the auditory evoked potentials, at parieto-occipital electrodes from the left (PO7, P5, P7) and
the analysis epoch consisted of 100 ms of pre-stimulus activity right hemisphere (PO8, P6, P8). The effects of music and
and 500 ms of post-stimulus activity time-locked to sound art training on the ability to inhibit a motor response on
onset. For the GoNoGo, the analysis epoch consisted of the nogo trials was quantified by comparing mean amplitude
200 ms of pre-stimulus activity and 1,000 ms of post-stimulus for the 375–475 ms interval measured over the left (F1, F3,
activity time-locked to the onset of the visual stimuli. EEG AF3) and right (F2, F4, AF4) frontal scalp regions. These
segments contaminated by blink and saccade were corrected electrode clusters were chosen because music training in young
using BESA. Both eye blinks and lateral movements were adults has been shown to modulate visual GoNoGo ERP
first identified in the continuous EEG and then modeled amplitude over the right hemisphere (Moreno et al., 2014).
using artifact correction with a surrogate model. After ocular Thus, we anticipated that music activities in older adults
correction, traces were then scanned for artifacts and epochs would also show changes in ERP amplitude over the right
including deflections exceeding 120 µV were marked and hemisphere after training.
excluded from the analysis. The remaining epochs were The effects of art training on auditory evoked responses
averaged according to electrode position, trial type (e.g., Go, (i.e., N1, P2, MMN) were assessed using a mixed model
NoGo), and session (i.e., pre-training, post-training, follow-up). ANOVA using groups (control, music, visual art) as between-
Each average was baseline-corrected with respect to the pre- subject factor, and sequence type (music vs. speech) and
stimulus interval. stimulus type (standard, deviant) as the within-subject
For the oddball paradigms, the proportion of trials included in factors. The effects of art training on visual evoked responses
the auditory ERPs ranged from 283 to 510 trials for the standard recorded during the GoNoGo task was assessed using a
stimuli and from 48 to 98 for deviant stimuli. The number of mixed model ANOVA with group (control, music, visual
trials was comparable for the pre- (Standard: M = 487; Deviant: art) as the between-subject factors, and stimulus type
M = 85) and post-training sessions (Standard: M = 484; Deviant: (Go, NoGo trials) and hemisphere (right, left) as within-
M = 86), and was similar for controls (Standard: M = 493; subject factors. We used the Benjamini–Hochberg method
Deviant: M = 88), music (Standard: M = 478; Deviant: M = 84) to adjust the familywise p-value for multiple comparisons
and visual art (Standard: M = 484; Deviant: M = 85) groups. For with q = 0.1, m = 160 (i.e., number of p-values) and p = 0.05
the GoNoGo task, the number of trials included in the visual (Hochberg and Benjamini, 1990), resulting in a p-value of 0.025
ERPs ranged from 78 to 160 trials for the Go condition and 17 to for significance.
40 trials for the NoGo condition. As for the oddball paradigms,
the number of trials was comparable before (Go: M = 141; NoGo:
M = 34) and after (Go: M = 144; NoGo: M = 35) training, and RESULTS
was similar for control (Go: M = 148; NoGo: M = 36), music (Go:
M = 141; NoGo: M = 33), and visual art (Go: M = 139; NoGo: Psychometric Assessment: Training
M = 34) groups. Effects
The analyses of performance at neuropsychological test revealed
Statistical analyses a group × session interaction on the Stroop subtest 2 [color
Training effects on the psychometric measures were assessed naming speed; F(2,50) = 6.16, p = 0.004, η2 = 0.20], where
using repeated measures ANOVAs. We focused on interactions the music group improved in naming speed between pre- and
that involved group and session as factors. Although they were post-assessment [t(16) = 3.32, p = 0.004, Figure 2]. There
not central to the goal of the present study, we also report was also a main effect of session for the Digit Symbol task,
other main effects and interactions for sake of completeness. For with better performance at post-test [compared to pre-test;
the psychometric assessments, we used the Benjamini–Hochberg F(1,50) = 15.98, p < 0.000, η2 = 0.24], but no interaction
method to adjust the familywise p-value for multiple comparisons with group. There were no other significant differences in
(q = 0.1, m = 30, and p = 0.05), resulting in a p-value of 0.005 performance between pre- and post-training session. In the music
for significance. group, a follow-up analysis showed that the gain in response
The ERP analyses focused on pre-defined time windows speed was maintained 3 months after training (post-test vs.
and electrode clusters motivated by prior studies (e.g., Reinke follow-up contrast, p > 0.1).

Frontiers in Neuroscience | www.frontiersin.org 5 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

FIGURE 2 | Mean reaction time at the Stroop subtest 2 (color naming) for the music, art, and control groups at pre-test, post-test and follow-up.

FIGURE 3 | Group mean event-related potentials elicited by standard and deviant piano tones as well as the corresponding difference wave (i.e., MMN). Here and
throughout, negativity is plotted upward. Waveforms are from the midline fronto-central electrode (FCz).

Effects of Music and Visual Art Training prototypical deflections at ∼60, ∼110, and ∼195 ms after
on Auditory Processing sound onset (i.e., P1-N1-P2). The MMN was isolated as the
Figures 3, 4 show group mean auditory ERPs elicited by standard difference in neural activity between standard and deviant
and deviant stimuli as well as the corresponding difference waves (oddball) stimuli. One participant from the music group and
in both sequences (piano tones and vowels) at pre- and post- two from the art group were excluded from the analyses because
training sessions, as well as from the subset of participants who data from one of the two sequences were unavailable due to
completed the follow-up. The scalp-recorded ERPs comprised technical problems.

Frontiers in Neuroscience | www.frontiersin.org 6 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

FIGURE 4 | Group mean event-related potentials elicited by standard and deviant vowel stimuli as well as the corresponding difference wave (i.e., MMN). Waveforms
are from the midline fronto-central electrode (FCz).

Training Effects on Auditory ERPs significant [F(1,47) = 52.19, p < 0.001, η2 = 0.53], reflecting
First, we examined whether music and visual art training larger N1 amplitude for deviant than standard stimuli. The
modulated early cortical processing indexed by the N1 and P2 interaction between session and stimulus type was also significant
waves. The ANOVA on the N1 mean amplitude (90–130 ms) [F(1,47) = 12.82, p = 0.001, η2 = 0.21], which was due to
yielded a significant group × session × sequence type interaction greater difference between standard and deviant at post-test
[F(2,47) = 4.30, p = 0.019, η2 = 0.16]. To better understand than at pre-test.
this three-way interaction, we examined the effects of training For the P2 mean amplitude (175–215 ms), the group × session
as a function of sequence type. For the piano tones, both was not significant, nor was the group × session × stimulus
music and art groups showed larger N1 amplitude at post- type. However, the group × session × sequence type interaction
test than pre-test (p < 0.005 in both cases). For the control approached the corrected significance level [F(2,47) = 3.63,
group, there was no difference in N1 amplitude between pre- p = 0.034, η2 = 0.13]. For the piano tones, both music and visual
test and post-test. For the vowel stimuli, there was no significant art groups showed larger P2 amplitude at post-test than pre-test
difference in N1 amplitude between pre-test and post-test (p < 0.005 in both cases). For the control group, there was no
(p > 0.19 in all three groups). Thus, while the N1 amplitude difference in P2 amplitude between pre-test and post-test. For the
elicited by vowel stimuli was little affected by training, the N1 vowel stimuli, there was no significant difference in P2 amplitude
generated by piano tones was larger at post-test than pre-test between sessions (p > 0.65 in all three groups). As for the N1
in the music and visual art training groups (Figures 3, 4). amplitude, the P2 wave elicited by vowel stimuli was little affected
The interaction between group × session × stimulus type was by training while the P2 generated by piano tones was larger
not significant. at post-test than pre-test in the music and visual art training
We also observed a main effect of session [F(1,47) = 10.15, groups (Figures 3, 4).
p = 0.003, η2 = 0.18], and a session × sequence type interaction The omnibus ANOVA on the P2 mean amplitude also
[F(1,47) = 6.06, p = 018, η2 = 0.11], with greater increased revealed a main effect of session [F(1,47) = 6.66, p = 0.013,
in N1 amplitude at post-test for the piano tone than for η2 = 0.12], and an interaction between session and sequence
the vowel stimuli. The main effect of stimulus type was type [F(1,47) = 12.80, p < 0.001, η2 = 0.21]. The main effect

Frontiers in Neuroscience | www.frontiersin.org 7 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

of stimulus type was significant [F(1,47) = 44.50, p < 0.001, effect of session was significant [F(2,50) = 10.29, p < 0.001,
η2 = 0.49], with the P2 wave being more negative for deviant than η2 = 0.29]. Pairwise comparisons revealed larger P2 amplitude
for standard stimuli. The sequence × stimulus type interaction at post-test and follow-up than at pre-test (p < 0.001 in both
was also significant [F(1,47) = 21.81, p < 0.001, η2 = 0.32], cases). There was no difference in P2 amplitude between post-
reflecting greater deviance-related activity for the vowel stimuli test and follow-up (p = 0.51). For vowel stimuli, the main effect of
than for the piano tones. session was not significant (F < 1) nor was the group × session
interaction (F < 1).
Training Effects on MMN In summary, music and visual art training modulated auditory
Figures 3, 4 show the mean MMN across groups. For the processing, which was retained 3 months after training end.
MMN latency, the group × session interaction was not
significant [F(2,47) = 1.70, p = 0.193, η2 = 0.07], nor was the Effects of Music and Visual Art Training
group × sequence type interaction (F < 1). The three-way on ERPs During Visual Go NoGo Task
group × session × sequence type interaction was not significant Behavioral Data
[F(2,47) = 1.80, p = 0.177, η2 = 0.071]. However, the ANOVA All three groups showed ceiling performance on Go trials with
yielded a main effect of sequence type [F(1,47) = 49.06, p < 0.001, few (if any) false positives on NoGo trials (see Table 1). There was
η2 = 0.51], with the MMN elicited by piano tones (M = 150 ms, no difference between the groups nor between the pre- and post-
SE = 2.5 ms) peaking earlier than the MMN elicited vowels test sessions. The group × session interaction was not significant
(M = 182 ms, SE = 4.1 ms). The main effect of session was also for accuracy or response time measures.
significant [F(1,47) = 16.42, p < 0.001, η2 = 0.23], with earlier
peak latency at post- than pre-test. Training Effects on Visual ERPs
For the MMN peak amplitude, the group × session interaction Figure 5 shows group mean visual ERPs elicited during Go
was not significant [F(2,47) = 2.55, p = 0.089, η2 = 0.10], nor was and NoGo trials as well as the corresponding difference
the group × sequence type (F < 1). However, the MMN was waves at parietal-occipital sites. The scalp-recorded visual ERPs
larger for vowel than piano tones [F(1,47) = 49.22, p < 0.001, comprised prototypical deflections at ∼100, ∼185, and ∼325 ms
η2 = 0.51]. All other main effects and interactions were not after stimulus onset (i.e., P1-N1-P2). The difference wave
significant. Thus, neither music nor visual art training had a between Go and NoGo trials reveals processing associated with
significant effect on the automatic change detection process as suppressing the motor response.
measured by the MMN. We first examined whether music and visual art training
modulated early visual cortical processing indexed by the N1
Follow-Up Retention and P2 waves. The ANOVA on the N1 mean amplitude (165–
Three months after the end of training, a subgroup of 205 ms) yielded a group × session × hemisphere was significant
participants (music N = 15, visual art N = 14) returned for [F(2,50) = 4.00, p = 0.024, η2 = 0.14]. To better understand
follow-up assessment and evaluation of long-term retention this three-way interaction, we examined the effects of training
effects. One participant in each group was excluded from the as a function of hemisphere. For the left parietal-occipital areas,
analysis because he/she was missing one of the experimental the music and visual art group showed comparable visual N1
conditions due to technical problems. The remaining sample amplitude before and after training (p > 0.5 in both cases)
comprised 14 from the music group and 13 from the whereas the N1 in the control group was larger at post-test than
visual art group. pre-test (p < 0.01). For the right parietal-occipital areas, the
The ANOVA on the N1 mean amplitude with session (pre, music and control group showed comparable N1 amplitude at
post, follow-up) as between-subject factor and sequence (piano pre-test and post-test, while the N1 amplitude was significantly
tones, vowels) and stimulus type (standard, deviant) as within- enhanced in the visual art group (p = 0.003). While the N1
subject factors, yielded a significant session × sequence type amplitude measured over the left hemisphere was little affected
interaction [F(2,50) = 4.47, p = 0.016, η2 = 0.15]. To better by training, the N1 recorded over the right hemisphere was larger
understand this interaction, we performed separate ANOVAs at post-test than pre-test in the visual art group only. No other
for piano and vowel stimuli. For piano tones, the ANOVA interactions involving group and sessions were significant.
revealed a main effect of session [F(2,50) = 11.33, p < 0.001, The omnibus ANOVA also showed that the N1 wave was
η2 = 0.31]. Pairwise comparisons revealed a significant increase larger at post-test than pre-test [F(1,50) = 6.21, p = 0.016,
in N1 amplitude at post-test and follow-up relative to pre- η2 = 0.11]. The main effect of stimulus type was also significant
test (p < 0.001 in both cases). There was no difference in N1 [F(1,50) = 18.78, p < 0.001, η2 = 0.27], with larger N1 amplitude
amplitude between post-test and follow-up (p = 0.144). The during the NoGo than Go trials. The latter could reflect an
group × session interaction was not significant (F < 1), nor attention-related negativity that overlaps the N1 wave rather than
was the group × session × stimulus type [F(2,50) = 1.67, a modulation of the N1 response.
p = 0.192, η2 = 0.31]. For vowel stimuli, the main effect of session For the visual P2 mean amplitude (305–345 ms), there
was not significant nor was the group × session interaction was no significant interaction between group and session, nor
(F < 1 in both cases). As for the N1, the ANOVA on the was the group × session × hemisphere interaction significant
P2 mean amplitude yielded a session × sequence interaction [F(2,50) = 1.35, p = 0.269, η2 = 0.05]. The main effect of
[F(2,50) = 6.27, p = 0.004, η2 = 0.20]. For piano tones, the main session was not significant [F(1,50) = 2.23, p = 0.142, η2 = 0.04].

Frontiers in Neuroscience | www.frontiersin.org 8 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

TABLE 1 | Behavioral results on the GoNoGo task.

Group Session Group mean Error (%)

response time (ms)
and (SE)

Music Pre-training 444 (17) 2.19 (0.62)

Post-training 427 (15) 2.56 (0.63)
Follow-up 431 (10) 3.33 (0.62)
Visual art Pre-training 434 (16) 3.33 (0.59)
Post-training 436 (14) 3.17 (0.59)
Follow-up 443 (15) 5.50 (2.36)
Control Pre-training 435 (15) 1.71 (0.60)
Post-training 439 (17) 1.47 (0.61)

FIGURE 5 | Group mean event-related potentials elicited during the visual GoNoGo task over the right parietal scalp area (i.e., electrode P6). The iso-contour maps
(view from the back) show amplitude distribution for the mean voltage between 210 and 280 ms. From left to right, the maps show the distribution at pre-test,
post-test, and follow-up, respectively.

However, the session × hemisphere interaction was significant component at parietal-occipital sites that peak at about
[F(1,50) = 4.14, p = 0.047, η2 = 0.08]. While the P2 measured 230–240 ms after stimulus onset. The mean amplitude
over the left hemisphere was comparable between pre- and post- for 210–280 was used for left (PO7, P5, P7) and right
training sessions, the P2 recorded over the right parieto-occipital hemisphere (PO8, P6, P8).
area was larger at post-test. The main effect of group was not significant [F(2,50) = 1.41,
p = 0.253, η2 = 0.05], nor was the group × session
NoGo-related effects interaction (F < 1). However, there was a significant
The visual ERPs elicited during the NoGo trials showed a group × session × hemisphere interaction [F(2,50) = 6.17,
negative displacement, which could be accounted for by p = 0.004, η2 = 0.19]. This was due to a shift in the response
attention-related effects superimposed on the N1 and P2. laterality between the first and second session in the visual art
This is best illustrated by the difference wave between Go group (Figure 6). That is, at pre-test, the ERP amplitude was
and NoGo trials. This difference wave revealed a negative greater over the left parietal-occipital sites whereas after visual

Frontiers in Neuroscience | www.frontiersin.org 9 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

FIGURE 6 | Group mean visual event-related potential amplitudes over left (PO7, P5, P7) and right (PO8, P6, P8) parieto-occipital areas.

art training the ERPs showed a more symmetric amplitude better understand this four-way interaction, we performed
distribution over the left and right hemispheres. In the control separate ANOVAs for each group. In the music group, the
and music group, the ERP amplitude was larger over the left and main effect of condition was significant [F(1,16) = 8.19,
right hemisphere during the pre- and post-training sessions. p = 0.011, η2 = 0.339], so was the session × condition
There was a main effect of session [F(1,50) = 6.21, p = 0.016, interaction [F(1,16) = 8.67, p = 0.010, η2 = 0.351], with greater
η2 = 0.11], with larger amplitude at post-test. The main effect increased in ERP amplitude at post-test for the NoGo than
of condition was also significant [F(1,50) = 124.47, p < 0.001, Go condition. The condition × hemisphere was significant
η2 = 0.71], which reflected enhanced negativity generated by the [F(1,16) = 13.31, p = 0.002, η2 = 0.454], reflecting greater
NoGo trials compare to the Go trials. difference between Go and NoGo trials over the left hemisphere.
The session × hemisphere was not significant (F < 1). In the
Follow-up retention visual art group, the main effect of condition was significant
We examined whether the changes in laterality persisted [F(1,18) = 21.18, p < 0.001, η2 = 0.541]. The session × condition
3 months after the end of training. An ANOVA comparing was also significant [F(1,18) = 8.038, p = 0.011, η2 = 0.309],
ERPs recorded in all three session in the art group revealed a with greater increased in ERP amplitude at post-test for the
session × hemisphere interaction [F(2,26) = 6.02, p = 0.013]. NoGo than Go condition. The session × hemisphere trended
Separate ANOVAs for the left and right hemisphere yielded toward significance [F(1,18) = 3.51, p = 0.077, η2 = 0.163].
a main effect of session only for ERPs recorded over the However, the session × condition × hemisphere was significant
right parietal-occipital area [right hemisphere: F(2,26) = 4.35, [F(1,18) = 5.78, p = 0.027, η2 = 0.243]. This was due to greater
p = 0.029; left hemisphere: F < 1]. Pairwise comparisons revealed difference in ERP amplitude between Go and NoGo trials over
enhanced negativity at post-test relative both pre-test and follow- the left hemisphere after training. In the control group, the main
up (p < 0.05 in both cases). The ERPs at follow-up did not differ effect of condition was significant [F(1,16) = 17.12, p = 0.001,
from those obtained at pre-test (p = 0.693). As during the pre-test, η2 = 0.517]. However, neither the main effect of session or
the ERPs recorded at follow-up were larger over the left than the the session × condition interaction were significant (F < 1
right hemisphere. In summary, art training was associated with in both cases). The condition × hemisphere was significant
increased early attention-related effects over the right parietal- [F(1,16) = 29.04, p < 0.001, η2 = 0.645], reflecting greater
occipital area, which was not retained 3 months after training. difference between Go and NoGo trials over the left hemisphere.
The other main effects or interactions were not significant.
ERPs Associated With Response Inhibition The ANOVA on the mean amplitude (375–475 ms) yielded
The NoGo trials generated a modulation that peaked at about a main effect of session [F(1,50) = 4.37, p = 0.042, η2 = 0.08],
425 ms after stimulus onset at fronto-central sites (Figure 7). and a main effect of condition [F(1,50) = 43.59, p < 0.001,
The group × session × condition × hemisphere interaction η2 = 0.466]. That is, the ERP amplitude was more positive at
was significant [F(2,50) = 4.24, p = 0.020, η2 = 0.15]. To post-test than at pre-test, and more positive for NoGo trials than

Frontiers in Neuroscience | www.frontiersin.org 10 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

FIGURE 7 | Group mean visual event-related potential elicited during the visual GoNoGo task over the right frontal scalp region (i.e., electrode F2). The iso-contour
maps (view from the top) show amplitude distribution for the mean voltage between 375 and 475 ms. From left to right, the maps show the distribution at pre-test,
post-test, and follow-up, respectively.

for Go trials. The session × condition interaction was significant changes in older adults. In the present study, we also observed
[F(1,50) = 12.50, p = 0.001, η2 = 0.20]. This reflected greater some training-specific effects using auditory oddball and visual
difference between Go and NoGo trials (NoGo ERP effect) at GoNoGo paradigms.
post-test than pre-test. The condition × hemisphere interaction
was also significant [F(1,50) = 62.18, p < 0.001, η2 = 0.55]. This Arts Training and Auditory Processing
was due to larger NoGo responses over the left compared to right
In both groups, we observed increased N1 and P2 amplitude
frontal scalp region.
after training, which was retained at least 3 months after the
Follow-up retention training ended. Although enhanced N1 and P2 amplitude has
In both music and visual art groups, the NoGo ERP effect was been observed after training in several studies (e.g., Tremblay
slightly reduced at follow-up compared to post-test, but the et al., 2001; Reinke et al., 2003), this is, to our knowledge, the first
difference was not statistically significant. Moreover, the NoGo study showing training-related increased in N1 and P2 amplitude
ERP effect at follow-up did not differ from the effects observed at to untrained stimuli in older adults. In the present study, the
pre-test. These results suggest that some training effects remained effect of training was more pronounced for the piano tones.
three months after the cessation of training. However, this finding This difference in training-related N1 effect between piano tones
should be interpreted with caution because of the small sample and vowels could be related to familiarity with the material
size at follow-up and the lack of difference between pre-test and prior to study. Prior research has shown increased N1 and
follow-up session. P2 amplitude with increased familiarity (Sheehan et al., 2005;
Alain et al., 2015). For most participants, speech sounds such
as vowels are over-learned and highly familiar stimuli leaving
DISCUSSION less opportunity to observe neuroplastic effects. In comparison,
the piano tones used in the present study were more novel
Using a partially randomized intervention design, our study thereby benefiting more from training. The music group may
reveals changes in older adults’ brain responses following thus have benefited from exposure to musical sounds; as for the
short-term arts training (music and visual art engagement). visual art group, it is possible that such training may predispose
As such, we provide important new evidence suggesting a the brain to be more receptive to new material than over-
causal relationship between these activities and neuroplastic learned stimuli.

Frontiers in Neuroscience | www.frontiersin.org 11 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

Training-specific changes were not observed in the auditory in NoGo-P3 in left frontal sites actually increased hemispheric
MMN as response latency peaked earlier at post-test than pre- symmetry following training. This finding is interesting because
test in all groups including controls. This suggests that exposure such results could be interpreted as a compensatory process
to stimuli at pre-test may be sufficient to facilitate the change in cognitive aging. Neuroimaging work has shown that high-
detection process. Further research is needed to examine whether performing older adults recruit additional resources and engage
more specific training using auditory discrimination tasks would prefrontal cortex bilaterally during demanding tasks (Cabeza,
yield greater changes in older adults’ MMN responses. 2002; Cabeza et al., 2002; Grady, 2012; Du et al., 2016). It is
possible that training for both groups improved functioning of
Arts Training and Visual Processing left frontal brain regions that had declined with age. In Go
Participants engaged in visual art training showed differences trials, the music group showed enhanced P3 amplitude at post-
in early attention-related brain processing at posterior parietal- training compared to the visual art and control groups. P3 may
occipital scalp regions. This suggests that visual art training reflect closure of the inhibition processing of an overt response
may improve processing of visual features. No such changes in (Gajewski and Falkenstein, 2013) or the ongoing evaluation of
early visual responses were observed in the control group or in an intention to inhibit (Liotti et al., 2005). Thus, we infer that
participants involved in the music training program. Given that music training improves supervisory mechanisms that act to
the visuals art group showed changes in both auditory and visual ensure desired responses and reinforce interference monitoring
neural processing, these findings reveal an important property (Moreno et al., 2014; Moreno and Farzan, 2015).
about the nature of brain plasticity in aging, showing cross- Training-specific effects were expressed in P2 and N2 waves,
modality transfer of benefits (here from visual art training to in which we observed seemingly opposite effects for each training
auditory processing). group: music students showed increased P2 and decreased N2
amplitudes after training whereas the visual art group showed
Arts Training and Cognitive Control decreased P2 and increased N2 amplitudes (no-contact controls
Following a short 3-month music or visual art training program, showed no changes). In our previous study of younger adults
functional brain changes in inhibitory control were observed undergoing musical instruction (Moreno et al., 2014), larger P2
in both training groups. While both forms of training offered was also observed in the music group. Taken alongside present
similar improvements (i.e., enhanced P3), notably, they also findings, this suggests P2 is a common marker of plasticity in
yielded training-specific changes (i.e., differential changes in N1 the brain’s inhibitory control across the lifespan. The P2 wave
and P2 for visual art versus music groups). These findings are is thought to index the ability to construct a representation of
consistent with earlier reports of training-driven modulations the current task context and the associated behavioral response
in prefrontal cortical activity of older adults (Voss et al., 2010; in the early stages of processing (Gajewski and Falkenstein,
Anguera et al., 2013). 2013). Electrophysiological evidence suggests that this early
At the behavioral level, the music group improved in color aspect of visual processing has links to higher cognitive functions
naming speed (subtest of the Stroop task) between pre- and by facilitating stronger internal representations of behaviorally
post-assessment. This effect of music training could reflect relevant stimuli (Moreno et al., 2014; Moreno and Farzan, 2015).
improvement in processing speed. Decline in processing speed Here, the larger P2 amplitude in the music group may reflect
has been linked to general cognitive decline (Sliwinski and earlier processing and stronger representation of stimuli and
Buschke, 1999) and may have major consequences for the the appropriate response (or non-response) pairing strengthened
lives of older adults. Indeed, processing speed plays a critical through training. Our results also showed a reduced N2 after
role in everyday activities by facilitating important cognitive music training, consistent with previous work in young adults
functions like learning and long-term memory, comprehension, (Moreno et al., 2014). N2 has been described as a marker of
decision making, and planning (West, 1996; Baddeley, 2012). inhibition and conflict detection/attention load (Nieuwenhuis
Previous correlational studies have reported a link between et al., 2003; Burle et al., 2004; Falkenstein, 2006). We interpret this
higher cognitive skills and musicianship (Hanna-Pladdy and effect in conjunction with the increased P2 amplitude such that
MacKay, 2011; Hanna-Pladdy and Gajewski, 2012; Amer et al., music training facilitates dissociation of desired and undesired
2013). Our study extends these earlier findings by demonstrating stimulus-response planning reflected by increased P2, and this
a causal relationship between musical training and processing enhanced effect of P2 subsequently reduces the need for cognitive
speed in older adults. This finding offers hope for improving control processes reflected by decreased N2 (Bokura et al.,
the aging process using music intervention strategies that aim to 2001; Nieuwenhuis et al., 2003; Donkers and van Boxtel, 2004;
strengthen cognitive skills. Kok et al., 2004).
Our results also revealed changes in the neural correlates On the other hand, the opposite pattern observed in the
of inhibitory control during a GoNoGo task following only visual art training group may suggest that this form of training
3 months of training. In NoGo trials, both training groups, as reduced early perceptual demands (reduced P2). This is plausible
compared to the control group, showed increased and protracted given that our GoNoGo was a visual paradigm (same domain as
P3 amplitude over the left hemisphere at post-test while the music visual art training). Subsequent enhancements in later N2 may
group also showed similar effects over the right hemisphere. reflect improved post-perceptual conflict detection and response
This is consistent with a prior study in older expert musicians inhibition. Our findings agree well with existing evidence
(Moussard et al., 2016). It is noteworthy that the enhancement suggesting an impact of visual art instructions on the ventral

Frontiers in Neuroscience | www.frontiersin.org 12 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

visual pathway (Draganski et al., 2004; Op de Beeck and Baker, older adults’ brain function following short-term music and
2010; Pollmann and von Cramon, 2000) as well as higher-order visual arts training. Although the neuroplastic changes remain
cognitive brain networks (Chamberlain et al., 2014). Overall, modest, especially for behavioral measures, our results offer
the specific nature of changes induced by music and visual art clear causal evidence that the aged brain is more plastic than
instructions indicate that these divergent forms of training offer traditionally thought and suggests new possibilities for cognitive
unique influences on brain function as well as enhancements to training and rehabilitation. Future studies should use longer
domain-general skills outside the direct scope of training (i.e., training regimens to explore transfer effects to everyday life.
inhibitory control). With the growth of the aging population in modern society,
there is a pressing need to find remedies which counteract
Limitations cognitive decline, improve the aging process, and ultimately
A limitation of the current study is that our sample showed a reduce health-related costs. Our study establishes that both
gender imbalanced, which could affect the generalizability of our visual art and music programs might be effective, engaging, and
findings. All three groups comprised a much larger proportion cost-effective solutions to boost older adults’ brain plasticity.
of women than men. In older women, the decrease in estrogen These programs could improve essential skills, bolster inhibitory
level that occurs post-menopause may affect cognitive functions. control, and ultimately improve the quality of life for older adults.
Further research should ensure a more balanced representation of
men and women and also monitor more closely whether women
participants are receiving hormonal replacement therapy. ETHICS STATEMENT
In the present study, the passive control group was added
to assess test–retest effects at post-test (i.e., an improvement The experimental protocol was approved by the Human Research
that would be due to repeating the same battery of tests a and Ethics Committee at Baycrest Centre, Toronto, ON, Canada.
second time). This allowed us to better identify the nature and
specificity of improvements at post-test. The purpose of the
follow-up assessment was to test whether these training-related AUTHOR CONTRIBUTIONS
improvements (or neural changes) were maintained after training
has stopped. Given that the control group did not receive any CA and SM designed the experiments. YL, AM, and GB
training, it was not justified to test for maintenance of training- performed the experiments and collected the data. CA, AM,
related benefits/changes for this group. Although the control GB, YL, and JS analyzed the data. CA, AM, GB, YL, and SM
group showed little changes in performance and brain activity, interpreted results of experiments. CA, AM, and SM drafted the
we cannot rule out the possibility that differences would have manuscript. All authors edited and revised the manuscript, and
emerged later due to either repeated testing or normal age-related approved the final version of manuscript.
decline. However, we should note that our study did not aim to
directly test whether art training slows down cognitive decline: a
much longer follow-up would have been required for this purpose FUNDING
(as healthy older adults do not usually show measurable declines
after 3 months). Future longitudinal studies should consider This research was supported by a grant (MOP 106619) from
adding additional testing for the control group(s) as well as the Canadian Institutes of Health Research (CIHR) to CA,
longer follow-up. the Ministry of Economic Development and Innovation of
Finally, although quite intensive (three sessions per week), Ontario to SM, the Natural Sciences and Engineering Research
the training program was relatively short (3 months). This Council (RGPIN-06196-2014 to SM), and the National Institute
may have limited the potential benefits of music training on on Deafness and Other Communication Disorders of the NIH
behavioral performance. Future studies could consider longer (R01DC016267 to GB).
periods of training, in order to better reflect the real experience
of learning music.
ACKNOWLEDGMENTS
CONCLUSION We would like to thank our music and visual art teachers, Donna
Takacs and the Royal Conservatory of Music, Madeline Chan,
Collectively, our findings expand current knowledge of Courtney Smith, Buddhika Bellana, Tristan Watson, as well as all
neuroplasticity in aging by demonstrating dynamic changes in the participants in our study.

REFERENCES Alain, C., and Woods, D. L. (1999). Age-related changes in processing auditory
stimuli during visual attention: evidence for deficits in inhibitory control and
Alain, C., Campeanu, S., and Tremblay, K. (2010). Changes in sensory evoked sensory memory. Psychol. Aging 14, 507–519.
responses coincide with rapid improvement in speech identification Alain, C., Zhu, K. D., He, Y., and Ross, B. (2015). Sleep-dependent neuroplastic
performance. J. Cogn. Neurosci. 22, 392–403. doi: 10.1162/jocn.2009. changes during auditory perceptual learning. Neurobiol. Learn. Mem. 118,
21279 133–142. doi: 10.1016/j.nlm.2014.12.001

Frontiers in Neuroscience | www.frontiersin.org 13 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

Aljondi, R., Szoeke, C., Steward, C., Yates, P., and Desmond, P. (2018). A decade Fujioka, T., Ross, B., Kakigi, R., Pantev, C., and Trainor, L. J. (2006). One year of
of changes in brain volume and cognition. Brain Imaging Behav. doi: 10.1007/ musical training affects development of auditory cortical-evoked fields in young
s11682-018-9887-z [Epub ahead of print]. doi: 10.1007/s11682-018-9887-z children. Brain 129(Pt 10), 2593–2608. doi: 10.1093/brain/awl247
Amer, T., Kalender, B., Hasher, L., Trehub, S. E., and Wong, Y. (2013). Do Gajewski, P. D., and Falkenstein, M. (2013). Effects of task complexity on ERP
older professional musicians have cognitive advantages? PLoS One 8:e71630. components in Go/Nogo tasks. Int. J. Psychophysiol. 87, 273–278. doi: 10.1016/
doi: 10.1371/journal.pone.0071630 j.ijpsycho.2012.08.007
Anguera, J. A., Boccanfuso, J., Rintoul, J. L., Al-Hashimi, O., Faraji, F., Janowich, J., Gazzaley, A., Rissman, J., Cooney, J., Rutman, A., Seibert, T., Clapp, W., et al.
et al. (2013). Video game training enhances cognitive control in older adults. (2007). Functional interactions between prefrontal and visual association
Nature 501, 97–101. doi: 10.1038/nature12486 cortex contribute to top-down modulation of visual processing. Cereb. Cortex
Baddeley, A. (2012). Working memory: theories, models, and controversies. Annu. 17(Suppl. 1), i125–i135. doi: 10.1093/cercor/bhm113
Rev. Psychol. 63, 1–29. doi: 10.1146/annurev-psych-120710-100422 Grady, C. (2012). The cognitive neuroscience of ageing. Nat. Rev. Neurosci. 13,
Bidelman, G. M., and Alain, C. (2015). Musical training orchestrates coordinated 491–505. doi: 10.1038/nrn3256
neuroplasticity in auditory brainstem and cortex to counteract age-related Hanna-Pladdy, B., and Gajewski, B. (2012). Recent and past musical activity
declines in categorical vowel perception. J. Neurosci. 35, 1240–1249. predicts cognitive aging variability: direct comparison with general lifestyle
doi: 10.1523/JNEUROSCI.3292-14.2015 activities. Front. Hum. Neurosci. 6:198. doi: 10.3389/fnhum.2012.00198
Bidelman, G. M., Lowther, J. E., Tak, S. H., and Alain, C. (2017). Mild cognitive Hanna-Pladdy, B., and MacKay, A. (2011). The relation between instrumental
impairment is characterized by deficient brainstem and cortical representations musical activity and cognitive aging. Neuropsychology 25, 378–386.
of speech. J. Neurosci. 37, 3610–3620. doi: 10.1523/JNEUROSCI.3700-16.2017 doi: 10.1037/a0021895
Bidelman, G. M., Villafuerte, J. W., Moreno, S., and Alain, C. (2014). Age-related Hasher, L., and Zacks, R. T. (1988). “Working memory, comprehension, and aging:
changes in the subcortical-cortical encoding and categorical perception of a review and a new view,” in The Psychology of Learning and Motivation, ed.
speech. Neurobiol. Aging 35, 2526–2540. doi: 10.1016/j.neurobiolaging.2014.05. G. H. Bower (New York, NY: Academic Press. ), 193–225.
006 Hochberg, Y., and Benjamini, Y. (1990). More powerful procedures for multiple
Bokura, H., Yamaguchi, S., and Kobayashi, S. (2001). Electrophysiological significance testing. Stat. Med. 9, 811–818. doi: 10.1002/sim.4780090710
correlates for response inhibition in a Go/NoGo task. Clin. Neurophysiol. 112, Hyde, K. L., Lerch, J., Norton, A., Forgeard, M., Winner, E., Evans, A. C., et al.
2224–2232. doi: 10.1016/S1388-2457(01)00691-5 (2009). The effects of musical training on structural brain development: a
Brydges, C. R., Fox, A. M., Reid, C. L., and Anderson, M. (2014). Predictive validity longitudinal study. Ann. N. Y. Acad. Sci. 1169, 182–186. doi: 10.1111/j.1749-
of the N2 and P3 ERP components to executive functioning in children: a 6632.2009.04852.x
latent-variable analysis. Front. Hum. Neurosci. 8:80. doi: 10.3389/fnhum.2014. Knight, R. T. (1994). “Attention regulation and human prefrontal cortex,” in Motor
00080 and Cognitive Functions of the Prefrontal Cortex, ed. A.-M. Thierry (Berlin:
Bugaiska, A., Clarys, D., Jarry, C., Taconnat, L., Tapia, G., Vanneste, S., et al. (2007). Springer-Verlag), 160–173. doi: 10.1007/978-3-642-85007-3_11
The effect of aging in recollective experience: the processing speed and executive Kok, A., Ramautar, J. R., De Ruiter, M. B., Band, G. P., and Ridderinkhof, K. R.
functioning hypothesis. Conscious Cogn. 16, 797–808. doi: 10.1016/j.concog. (2004). ERP components associated with successful and unsuccessful stopping
2006.11.007 in a stop-signal task. Psychophysiology 41, 9–20. doi: 10.1046/j.1469-8986.2003.
Bugos, J. A., Perlstein, W. M., McCrae, C. S., Brophy, T. S., and Bedenbaugh, 00127.x
P. H. (2007). Individualized piano instruction enhances executive functioning Kropotov, J., Ponomarev, V., Tereshchenko, E. P., Muller, A., and Jancke, L. (2016).
and working memory in older adults. Aging Ment. Health 11, 464–471. doi: Effect of aging on erp components of cognitive control. Front. Aging Neurosci.
10.1080/13607860601086504 8:69. doi: 10.3389/fnagi.2016.00069
Burle, B., Vidal, F., and Bonnet, M. (2004). Electroencephalographic nogo Liotti, M., Pliszka, S. R., Perez, R., Kothmann, D., and Woldorff, M. G. (2005).
potentials in a no-movement context: the case of motor imagery in humans. Abnormal brain activity related to performance monitoring and error detection
Neurosci. Lett. 360, 77–80. doi: 10.1016/j.neulet.2004.02.034 in children with ADHD. Cortex 41, 377–388. doi: 10.1016/S0010-9452(08)
Cabeza, R. (2002). Hemispheric asymmetry reduction in older adults: the 70274-0
HAROLD model. Psychol. Aging 17, 85–100. doi: 10.1037/0882-7974.17.1.85 Lustig, C., Hasher, L., and Zacks, R. T. (2007). “Inhibitory deficit theory: recent
Cabeza, R., Anderson, N. D., Locantore, J. K., and McIntosh, A. R. (2002). developments in a “new view”,” in Inhibition in Cognition, eds D. S. Gorfein
Aging gracefully: compensatory brain activity in high-performing older adults. and C. M. MacLeod (Washington, DC: American Pschological Association),
Neuroimage 17, 1394–1402. doi: 10.1006/nimg.2002.1280 145–162.
Chamberlain, R., McManus, I. C., Brunswick, N., Rankin, Q., Riley, H., and May, C. P., Hasher, L., and Kane, M. J. (1999). The role of interference in memory
Kanai, R. (2014). Drawing on the right side of the brain: a voxel-based span. Mem. Cogn. 27, 759–767. doi: 10.3758/BF03198529
morphometry analysis of observational drawing. Neuroimage 96, 167–173. doi: Moreno, S., Bialystok, E., Barac, R., Schellenberg, E. G., Cepeda, N. J., and Chau, T.
10.1016/j.neuroimage.2014.03.062 (2011). Short-term music training enhances verbal intelligence and executive
Dempster, F. N. (1992). The rise and fall of the inhibitory mechanism: toward function. Psychol. Sci. 22, 1425–1433. doi: 10.1177/0956797611416999
a unified theory of cognitive development and aging. Dev. Rev. 12, 45–75. Moreno, S., and Farzan, F. (2015). Music training and inhibitory control: a
doi: 10.1016/0273-2297(92)90003-K multidimensional model. Ann. N. Y. Acad. Sci. 1337, 147–152. doi: 10.1111/
Diamond, A. (2013). Executive functions. Annu. Rev. Psychol. 64, 135–168. nyas.12674
doi: 10.1146/annurev-psych-113011-143750 Moreno, S., Marques, C., Santos, A., Santos, M., Castro, S. L., and Besson, M.
Donkers, F. C., and van Boxtel, G. J. (2004). The N2 in go/no-go tasks (2009). Musical training influences linguistic abilities in 8-year-old children:
reflects conflict monitoring not response inhibition. Brain Cogn. 56, 165–176. more evidence for brain plasticity. Cereb. Cortex 19, 712–723. doi: 10.1093/
doi: 10.1016/j.bandc.2004.04.005 cercor/bhn120
Draganski, B., Gaser, C., Busch, V., Schuierer, G., Bogdahn, U., and May, A. Moreno, S., Wodniecka, Z., Tays, W., Alain, C., and Bialystok, E. (2014). Inhibitory
(2004). Neuroplasticity: changes in grey matter induced by training. Nature 427, control in bilinguals and musicians: event related potential (ERP) evidence
311–312. doi: 10.1038/427311a for experience-specific effects. PLoS One 9:e94169. doi: 10.1371/journal.pone.
Du, Y., Buchsbaum, B. R., Grady, C. L., and Alain, C. (2016). Increased activity in 0094169
frontal motor cortex compensates impaired speech perception in older adults. Moussard, A., Bermudez, P., Alain, C., Tays, W., and Moreno, S. (2016). Life-long
Nat. Commun. 7:12241. doi: 10.1038/ncomms12241 music practice and executive control in older adults: an event-related potential
Dunn, L. M., and Dunn, D. M. (2007). Peabody Picture Vocabulary Test, 4th Edn. study. Brain Res. 1642, 146–153. doi: 10.1016/j.brainres.2016.03.028
London: Pearson Education. Nieuwenhuis, S., Yeung, N., van den Wildenberg, W., and Ridderinkhof, K. R.
Erber, J. T. (2013). Aging and Older Adulthood. New Jersey, NJ: Wiley-Blackwell. (2003). Electrophysiological correlates of anterior cingulate function in a go/no-
Falkenstein, M. (2006). Inhibition, conflict and the Nogo-N2. Clin. Neurophysiol. go task: effects of response conflict and trial type frequency. Cogn. Affect. Behav.
117, 1638–1640. doi: 10.1016/j.clinph.2006.05.002 Neurosci. 3, 17–26. doi: 10.3758/CABN.3.1.17

Frontiers in Neuroscience | www.frontiersin.org 14 March 2019 | Volume 13 | Article 182

Alain et al. Art Training and Aging

Nikjeh, D. A., Lister, J. J., and Frisch, S. A. (2009). Preattentive cortical-evoked Stroop, J. R. (1935). Studies of interference in serial verbal reactions. J. Exp. Psychol.
responses to pure tones, harmonic tones, and speech: influence of music 18, 643–662. doi: 10.1037/h0054651
training. Ear Hear. 30, 432–446. doi: 10.1097/AUD.0b013e3181a61bf2 Tervaniemi, M., Ilvonen, T., Karma, K., Alho, K., and Naatanen, R. (1997). The
Op de Beeck, H. P., and Baker, C. I. (2010). The neural basis of visual object musical brain: brain waves reveal the neurophysiological basis of musicality in
learning. Trends Cogn. Sci. 14, 22–30. doi: 10.1016/j.tics.2009.11.002 human subjects. Neurosci. Lett. 226, 1–4. doi: 10.1016/S0304-3940(97)00217-6
Parbery-Clark, A., Anderson, S., Hittner, E., and Kraus, N. (2012). Musical Tierney, A., and Kraus, N. (2015). Neural entrainment to the rhythmic structure of
experience offsets age-related delays in neural timing. Neurobiol. Aging music. J. Cogn. Neurosci. 27, 400–408. doi: 10.1162/jocn_a_00704
33:1483e1-e4. doi: 10.1016/j.neurobiolaging.2011.12.015 Tremblay, K., Kraus, N., McGee, T., Ponton, C., and Otis, B. (2001). Central
Parbery-Clark, A., Strait, D. L., Anderson, S., Hittner, E., and Kraus, N. (2011). auditory plasticity: changes in the N1-P2 complex after speech-sound training.
Musical experience and the aging auditory system: implications for cognitive Ear Hear. 22, 79–90. doi: 10.1097/00003446-200104000-00001
abilities and hearing speech in noise. PLoS One 6:e18082. doi: 10.1371/journal. Voss, M. W., Prakash, R. S., Erickson, K. I., Basak, C., Chaddock, L., Kim, J. S.,
pone.0018082 et al. (2010). Plasticity of brain networks in a randomized intervention trial of
Pichora-Fuller, K., Alain, C., and Schneider, B. (2017). “Older adults at the cocktail exercise training in older adults. Front. Aging Neurosci. 2:32. doi: 10.3389/fnagi.
party,” in The Auditory System at the Cocktail Party, eds J. C. Middlebrook, 2010.00032
J. Z. Simon, A. N. Popper, and R. R. Fay (Berlin: Springer), 227–259. doi: Wallentin, M., Nielsen, A. H., Friis-Olivarius, M., Vuust, C., and Vuust, P. (2010).
10.1007/978-3-319-51662-2_9 The musical ear test, a new reliable test for measuring musical competence.
Pichora-Fuller, M. K., Schneider, B. A., and Daneman, M. (1995). How young Learn. Individ. Differ. 20, 188–196. doi: 10.1016/j.lindif.2010.02.004
and old adults listen to and remember speech in noise. J. Acoust. Soc. Am. 97, West, R., and Alain, C. (2000). Age-related decline in inhibitory control contributes
593–608. doi: 10.1121/1.412282 to the increased Stroop effect observed in older adults. Psychophysiology 37,
Pollmann, S., and von Cramon, D. Y. (2000). Object working memory and 179–189. doi: 10.1111/1469-8986.3720179
visuospatial processing: functional neuroanatomy analyzed by event-related West, R. L. (1996). An application of prefrontal cortex function theory to cognitive
fMRI. Exp. Brain Res. 133, 12–22. doi: 10.1007/s002210000396 aging. Psychol. Bull. 120, 272–292. doi: 10.1037/0033-2909.120.2.272
Reinke, K. S., He, Y., Wang, C., and Alain, C. (2003). Perceptual learning modulates White-Schwoch, T., Woodruff Carr, K., Anderson, S., Strait, D. L., and Kraus, N.
sensory evoked response during vowel segregation. Brain Res. Cogn. Brain Res. (2013). Older adults benefit from music training early in life: biological
17, 781–791. doi: 10.1016/S0926-6410(03)00202-7 evidence for long-term training-driven plasticity. J. Neurosci. 33, 17667–17674.
Rogenmoser, L., Kernbach, J., Schlaug, G., and Gaser, C. (2018). Keeping brains doi: 10.1523/JNEUROSCI.2560-13.2013
young with making music. Brain Struct. Funct. 223, 297–305. doi: 10.1007/ Zendel, B. R., and Alain, C. (2009). Concurrent sound segregation is enhanced
s00429-017-1491-2 in musicians. J. Cogn. Neurosci. 21, 1488–1498. doi: 10.1162/jocn.2009.
Schellenberg, E. G. (2004). Music lessons enhance IQ. Psychol. Sci. 15, 511–514. 21140
doi: 10.1111/j.0956-7976.2004.00711.x Zendel, B. R., and Alain, C. (2012). Musicians experience less age-related decline in
Seinfeld, S., Figueroa, H., Ortiz-Gil, J., and Sanchez-Vives, M. V. (2013). Effects of central auditory processing. Psychol. Aging 27, 410–417. doi: 10.1037/a0024816
music learning and piano practice on cognitive function, mood and quality of Zendel, B. R., and Alain, C. (2014). Enhanced attention-dependent activity in the
life in older adults. Front. Psychol. 4:810. doi: 10.3389/fpsyg.2013.00810 auditory cortex of older musicians. Neurobiol. Aging 35, 55–63. doi: 10.1016/j.
Seither-Preisler, A., Parncutt, R., and Schneider, P. (2014). Size and synchronization neurobiolaging.2013.06.022
of auditory cortex promotes musical, literacy, and attentional skills in children.
J. Neurosci. 34, 10937–10949. doi: 10.1523/JNEUROSCI.5315-13.2014 Conflict of Interest Statement: The authors declare that the research was
Shahin, A., Bosnyak, D. J., Trainor, L. J., and Roberts, L. E. (2003). Enhancement of conducted in the absence of any commercial or financial relationships that could
neuroplastic P2 and N1c auditory evoked potentials in musicians. J. Neurosci. be construed as a potential conflict of interest.
23, 5545–5552. doi: 10.1523/JNEUROSCI.23-13-05545.2003
Sheehan, K. A., McArthur, G. M., and Bishop, D. V. (2005). Is discrimination Copyright © 2019 Alain, Moussard, Singer, Lee, Bidelman and Moreno. This is an
training necessary to cause changes in the P2 auditory event-related brain open-access article distributed under the terms of the Creative Commons Attribution
potential to speech sounds? Brain Res. Cogn. Brain Res. 25, 547–553. doi: 10. License (CC BY). The use, distribution or reproduction in other forums is permitted,
1016/j.cogbrainres.2005.08.007 provided the original author(s) and the copyright owner(s) are credited and that the
Sliwinski, M., and Buschke, H. (1999). Cross-sectional and longitudinal original publication in this journal is cited, in accordance with accepted academic
relationships among age, cognition, and processing speed. Psychol. Aging 14, practice. No use, distribution or reproduction is permitted which does not comply
18–33. doi: 10.1037/0882-7974.14.1.18 with these terms.

Frontiers in Neuroscience | www.frontiersin.org 15 March 2019 | Volume 13 | Article 182