J Neurophysiol 122: 490 – 499, 2019.
First published June 5, 2019; doi:10.1152/jn.00036.2019.
RESEARCH ARTICLE Neural Circuits
Effects of cerebellar transcranial direct current stimulation on the cognitive
stage of sequence learning
Hannah K. Ballard,1 James R. M. Goen,2 Ted Maldonado,2 and Jessica A. Bernard1,2
1
Texas A&M Institute for Neuroscience, Texas A&M University, College Station, Texas; and 2Psychological and Brain
Sciences Department, Texas A&M University, College Station, Texas
Submitted 16 January 2019; accepted in final form 4 June 2019
Ballard HK, Goen JR, Maldonado T, Bernard JA. Effects of
cerebellar transcranial direct current stimulation on the cognitive stage
of sequence learning. J Neurophysiol 122: 490 – 499, 2019. First
published June 5, 2019; doi:10.1152/jn.00036.2019.—Though the
cerebellum has been previously implicated in explicit sequence learn-
ing, the exact role of this structure in the acquisition of motor skills is
not completely clear. The cerebellum contributes to both motor and
nonmotor behavior. Thus, this structure not only may contribute to the
motoric aspects of sequence learning but may also play a role in the
cognitive components of these learning paradigms. Therefore, we
investigated the consequence of both disrupting and facilitating cer-
ebellar function using high-definition transcranial direct current stim-
ulation (tDCS) before the completion of an explicit motor sequence
learning paradigm. Using a mixed within- and between-subjects
design, we employed cathodal (n ⫽ 21) and anodal (n ⫽ 23) tDCS
(relative to sham), targeting the lateral posterior cerebellum, to tem-
porarily modulate function and investigate the resulting effects on the
acquisition of a sequential pattern of finger movements. Results
indicate that cathodal stimulation has a positive influence on learning
while anodal stimulation has the opposite effect, relative to sham.
Though the cerebellum is presumed to be primarily involved in motor
function and movement coordination, our results support a cognitive
contribution that may come into play during the initial stages of
learning. Using tDCS targeting the right posterior cerebellum, which
communicates with the prefrontal cortex via closed-loop circuits, we
found polarity-specific effects of cathodal and anodal stimulation on
sequence learning. Thus, our results substantiate the role of the
cerebellum in the cognitive aspect of motor learning and provide
important new insights into the polarity-specific effects of tDCS in
this area.
NEW & NOTEWORTHY The cerebellum contributes to motor and
cognitive processes. Investigating the cognitive contributions of the
cerebellum in explicit sequence learning stands to provide new in-
sights into this learning domain, and cerebellar function more gener-
ally. Using high-definition transcranial direct current stimulation, we
demonstrated polarity-specific effects of stimulation on explicit se-
quence learning. We speculate that this is due to facilitation of
working memory processes. This provides new evidence supporting a
role for the cerebellum in the cognitive aspects of sequence learning.
cerebellum; cognition; motor skill; sequence learning; tDCS
Address for reprint requests and other correspondence: H. Ballard, Texas
A&M Institute for Neuroscience, Texas A&M University, College Station, TX
77843 (e-mail: hannah_ballard@tamu.edu).
490 0022-3077/19 Copyright © 2019 the American Physiological Society
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
INTRODUCTION
The acquisition of motor skills is important for day-to-day
functioning. Movement coordination is necessary for the op-
eration of routine tasks, such as driving a car, riding a bike,
playing a musical instrument, playing sports, etc. As such, the
performance of these everyday tasks is dependent on the ability
to successfully learn and execute motor skills. Investigating the
neural underpinnings of sequence learning, therefore, is essen-
tial for enhancing our knowledge and understanding of motor
processes and allowing for effective protocols designed for
improvement or rehabilitation of skilled motor behavior.
As suggested by Karni et al. (1998), sequence learning skills
develop over time across distinct learning phases, beginning
with fast, or early, learning and ending with slow, or late,
learning. During the early learning phase, initial improvement
in motor performance is observed. Additional improvement
continues, and eventually, during the later phase of learning
after extended practice, individuals reach a state of automatic-
ity in performance. Distinct brain structures have been impli-
cated in each sequence learning phase, contributing to the
differences seen in skill progression (Doyon et al. 1997). The
cerebellum is particularly active during the early learning
phase when procedural memories are first being formed
whereas the striatum takes over during later learning phases
when performance improvements begin to level off (Bernard et
al. 2013; Doyon et al. 2018; Ungerleider et al. 2002). Impor-
tantly, however, it may be that the cerebellum also contributes
to the initial skill acquisition stages via its role in cognitive
processing. A growing literature has indicated that the cerebel-
lum contributes to cognition (e.g., Schmahmann and Sherman
1998; Stoodley and Schmahmann 2009; Stoodley 2012), pre-
sumably through closed-loop circuits with the prefrontal cortex
(cf. Strick et al. 2009).
In addition to the motor nodes associated with motor learn-
ing, nonmotor regions of the brain also contribute to skill
acquisition. As noted in a recent review by Doyon et al. (2018),
cognitive processes play a role in the acquisition of new motor
sequences. In functional neuroimaging investigations of motor
adaptation, the dorsolateral prefrontal cortex was active in
regions associated with cognitive function, specifically work-
ing memory, during the initial stage of learning (Anguera et al.
2010, 2012). Similarly, activation in the lateral prefrontal
cortex, commonly implicated in cognitive processing, has also
been demonstrated during explicit sequence learning (Aizen-
www.jn.org
 TDCS EFFECTS ON SEQUENCE LEARNING
stein et al. 2004; Eliassen et al. 2001; Honda et al. 1998; Sakai
et al. 1998; Schendan et al. 2003). Furthermore, working
memory has been associated with the formation of motor
chunks during the learning of a new sequence (Bo and Seidler
2009). Given that the cerebellum plays a role in nonmotor
processing (e.g., Bernard and Seidler 2013, 2014; Guell et al.
2018; Schmahmann and Sherman 1998; Stoodley 2012), in
addition to contributing to the internal representation of motor
responses during early sequence learning (Spampinato and
Celnik 2018), this structure may also be involved in the
cognitive aspects of initial skill acquisition.
One way to investigate the role of the cerebellum in explicit
motor sequence learning is to consider the effects of noninva-
sive brain stimulation on skill acquisition, using methods such
as transcranial direct current stimulation (tDCS). With this
technique, brain activity in a targeted area can be temporarily
disrupted or facilitated, and the resulting impact on behavior
can be observed (e.g., Nitsche and Paulus 2000; Nitsche et al.
2008; Takano et al. 2011). At present, several studies have
used tDCS to examine the role of different motor circuit nodes,
including the cerebellum, in motor learning (e.g., Block and
Celnik 2013; Buch et al. 2017; Cantarero et al. 2015; Foerster
et al. 2013; Galea et al. 2011; Hardwick and Celnik 2014; Pope
and Miall 2012; Shah et al. 2013). These studies have estab-
lished that the cerebellum plays a significant role in adaptation
learning, but a clear understanding of the cerebellum’s involve-
ment in explicit sequence learning, specifically, is still lacking.
Therefore, though we know that the cerebellum is important
for learning, investigating its contributions from a more cog-
nitive circuit perspective with this particular type of task is
necessary for extending current insights. Furthermore, high-
definition (HD) methods of tDCS (HD-tDCS) have now be-
come available, allowing for better regional targeting of the
stimulation. Notably, to our knowledge, HD-tDCS has not yet
been employed to investigate the cerebellum in motor learning.
This technique is particularly useful for investigating the cog-
nitive contributions of the cerebellum in motor sequence learn-
ing as it allows us to better target the nonmotor aspects of this
structure.
Here, we targeted the lateral posterior cerebellum, primarily
in the region of Crus I and Crus II, using cathodal and anodal
HD-tDCS in two experiments. Notably, these regions are
associated with prefrontal cortical networks in humans (Ber-
nard et al. 2012, 2014; Krienen and Buckner 2009; O’Reilly et
al. 2010; Salmi et al. 2010) and in nonhuman primates (Kelly
and Strick 2003), in addition to being implicated across brain
imaging studies of nonmotor behavior (E et al. 2014; Stoodley
and Schmahmann 2009). Our goal was to investigate differ-
ences in motor sequence learning after purported inhibition
with cathodal stimulation or excitation with anodal stimulation
(Nitsche and Paulus 2000), relative to sham. We predicted that
cathodal stimulation targeting the lateral posterior cerebellum
would negatively impact motor sequence learning relative to
sham whereas anodal stimulation to the same cerebellar area
would facilitate motor learning, due to the impact on cognitive
function and associated cognitive circuits. These predictions
are consistent with findings from recent sequence learning
research using the traditional two electrode pad stimulation
technique rather than the HD-tDCS approach employed here
(Ehsani et al. 2016; Shimizu et al. 2017; Wessel et al. 2016).
Thus, we conducted two independent experiments with a
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
491
mixed within- and between-subjects design. For the within-
subjects comparison, we employed cathodal and sham stimu-
lation in experiment 1 and anodal and sham in experiment 2.
We then compared across the cathodal and anodal groups in a
between-subjects manner to get a complete idea of the effect of
cerebellar stimulation on motor skill learning.
MATERIALS AND METHODS
Experiment 1
Participants. Twenty-eight healthy young adults participated in the
first experiment, which consisted of two HD-tDCS sessions (cathodal
and sham) that were counterbalanced in order. Two of those individ-
uals did not return for the second session of the study, one individual
was not right-handed as per our screening criteria, and complete data
was not collected from one individual due to technical difficulties.
Additionally, three individuals were excluded from analyses for hav-
ing task performance scores below three standard deviations from the
group mean. This particular criterion was implemented as an a priori
cutoff to exclude behavioral outliers. In investigating these outliers,
the three individuals excluded had accuracy scores of 52% or lower.
Indeed, one individual did not respond in time on as many as 88% of
trials within a given block, suggesting that those excluded based on
behavioral scores failed to comply with task instructions and may
have performed poorly due to inattention or lack of motivation. As
such, our final sample included 21 subjects (15 female, mean age 18.6
yr ⫾ 0.7 SD, range 18 –20 yr). None of the subjects had any history of
medical or neurological disease, and, as assessed by the Edinburgh
Handedness Inventory, all were right-handed (mean score 96.8 ⫾ 10.0
SD). Subjects were screened for exclusion criteria associated with
tDCS (Nitsche et al. 2008), and none were taking medication that
could potentially impact the central nervous system (e.g., psychiatric
medications such as antidepressants, anxiolytics, neuroleptics, narcot-
ics, stimulants, and analgesics). Subjects were recruited and scheduled
through the Texas A&M University Psychology Subject Pool and
received course credit for their participation. Participation was limited
to individuals that were tDCS naive. Prior to the initiation of study
procedures, all subjects provided written, informed consent. Our
protocol was submitted to and approved by the Institutional Review
Board at Texas A&M University.
Procedure. Participants underwent two sessions, exactly 1 wk
apart, in which they received both cathodal and sham tDCS in a
within-subjects design. We used a single-blind procedure wherein
participants were not told which condition they were receiving in a
given session. The order of stimulation condition (cathodal or sham)
was counterbalanced across all participants. Stimulation was admin-
istered with a Soterix MxN HD-tDCS system (Soterix, New York,
NY) using a montage of eight stimulation electrodes. The electrode
array and current intensities used for cathodal cerebellar stimulation
are presented in Table 1 and the associated modeled current flow is
depicted in Fig. 1. Different intensities were used for each electrode
location to effectively administer 2.0 mA of net current. The stimu-
lation montage was obtained using HD-Targets software that provides
specific electrode locations to more focally stimulate a targeted region
of interest (Datta et al. 2016; Huang et al. 2017). In this case, our
target was the right lateral posterior cerebellum. In determining the
optimal stimulation montage, we chose a “spiral out” approach for this
experiment, which is equivalent to cathodal stimulation in a traditional
two electrode pad montage. A “spiral” approach is the terminology
used to describe any montage containing multiple stimulation elec-
trodes and distinguishes the direction of current flow. During each
session, an initial stimulation of 0.1 mA was administered for 1 min
to allow the current to break through the scalp and establish an
adequate connection. Once resistance levels of 50 k⍀ or lower were
achieved across all electrodes, the full stimulation period began,
492 TDCS EFFECTS ON SEQUENCE LEARNING
Table 1. Electrode array and current intensities for cerebellar
cathodal stimulation
Location Current
1 P2 0.1135
2 PO10 ⫺0.1551
3 O10 ⫺0.3618
4 EX3 0.1956
5 EX4 ⫺1.4832
6 EX5 0.4066
7 EX12 0.22
8 EX14 0.7886
C NK1 0.2757
Current, specific intensity of electrical current applied at a given location;
Location, particular locations for electrode placement, determined by the 10-20
system. 1– 8, Stimulation electrodes; C, return electrode.
lasting 20 min. In the course of each condition, current gradually
increased during the first 30 s. During active stimulation, the current
remained steady for the entire span of the stimulation period. In the
sham, or placebo, condition, subjects underwent the standard setup for
HD-tDCS but only experienced current during the first and last 30 s of
the stimulation period to allow for attenuation to occur and to simulate
the general experience associated with active stimulation. During this
30-s period, the current slowly ramped up and then quickly returned
to zero. This sham procedure has been implemented to avoid the
subjects’ detection of the stimulation condition as attenuation to the
sensation of stimulation occurs relatively quickly under cathodal (or
anodal) protocols.
To assess the influence of tDCS on motor skill learning, we used an
explicit motor sequence task modeled after Kwak et al. (2012). We
chose to use this type of task to pinpoint the cognitive aspects of
procedural learning, specifically working memory-mediated strate-
gies, that are primarily engaged in explicit paradigms. Indeed, inves-
tigations using functional magnetic resonance imaging (fMRI) have
revealed greater activation in cognitive and working memory related
areas during explicit, rather than implicit, learning (Grafton et al.
1995; Jenkins et al. 1994; Rauch et al. 1995; Schendan et al. 2003). In
the task, participants were instructed to place their left middle, left
index, right index, and right middle fingers over the numbers 1, 2, 3,
and 4 on the computer keyboard, respectively. The task was bimanual
to allow for translation to the MRI environment for follow-up work;
this technique has been implemented in multiple other studies inves-
tigating sequence learning (Berner and Hoffmann 2008; Bhakuni and
Mutha 2015; Bo and Seidler 2009; Kwak et al. 2010, 2012; Sun et al.
2007). Participants were then presented with four rectangles in the
center of the screen, corresponding to the instructed finger locations
from left to right. In each trial, a different rectangle was shaded black,
and participants were instructed to press the button matching the
location of that shaded rectangle as quickly as possible. To account for
baseline motor function, random blocks were included in between
groups of sequence blocks throughout the task. Thus, each participant
completed 6 random blocks with 18 trials each and 9 sequence blocks
with 36 trials each for a total of 15 blocks and 432 trials (Kwak et al.
2012). Random blocks were delineated with an “R” before stimulus
presentation, while sequence blocks were preceded by an “S.” Thus,
the nature of the upcoming trials was announced before each block as
the task was intended to be clearly explicit. Relatedly, our paradigm
did not include extensive practice or subsequent retention sessions to
avoid the recruitment of implicit learning processes in the develop-
ment of automaticity. Each stimulus was presented for 200 ms and
participants were allowed 800 ms to respond before the next stimulus
appeared. The block design also mirrored that of Kwak et al. (2012)
with the following order: R-S-S-S-R-R-S-S-S-R-R-S-S-S-R. To ac-
count for practice effects, participants were presented with a new
sequence pattern during the second experimental session. Each se-
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
quence comprised six elements and did not contain any trills or
sequential numbers. The following sequences were used in our design:
2-4-1-3-2-4 and 3-2-4-1-2-4; these sequences did not evidently differ
in complexity as there was no significant difference in distribution
between sessions, according to the Kolmogorov-Smirnov test (D ⫽
0.182, P ⫽ 0.423).
Data analysis. Performance on the motor sequence learning task
was measured using average total accuracy (ACC) per sequence block
and mean reaction time (RT) for correct responses during both
random and sequence blocks. Accuracy is presented as the percent
correct out of the total number of trials for each sequence block. Total
trials included correct, incorrect (e.g., commission errors), and missed
responses (e.g., omission errors). To remain consistent with the
current sequence learning literature (Kwak et al. 2012) and factor in
timing-specific effects on skill acquisition, we analyzed accuracy
through a learning phase perspective. The first three sequence blocks
during the task were concatenated and represented the early learning
phase whereas the middle three blocks represented the intermediate
phase, and the last three blocks represented the late learning phase, in
the context of our study setup. We chose to concatenate across
sequence blocks in this manner to get at the more broad changes in
motor skill learning that may occur over the course of the task, as our
paradigm was relatively brief. This approach was adopted instead of
modeling across each individual block to simplify the analyses and
avoid excessive follow-up comparisons that may complicate the
interpretation of results, as other studies have similarly done (Bo and
Seidler 2009; Kwak et al. 2010, 2012). However, in the grand scheme
of the sequence learning literature, our design still reflects an early
learning paradigm that focuses on initial skill acquisition as extensive
practice or subsequent experimental sessions using the same sequence
were not included. In recognition of the overall brief nature of the
task, we did not find it necessary to normalize scores relative to initial
performance as baseline learning was not expected to influence later
learning over the course of the task with only nine sequence blocks,
total, being executed. For RT, missed trials were excluded and we
used correct trials only from both random and sequence blocks when
calculating averages. Repeated-measures ANOVAs and paired-sam-
ple t-tests were performed to investigate differences in performance
on the motor sequence learning task between cathodal and sham
stimulation. All analyses were conducted using SPSS Statistics.
Experiment 2
Participants. Thirty-five healthy young adults participated in the
second experiment, consisting of two HD-tDCS sessions (anodal and
sham) that were counterbalanced across subjects. Individuals in this
experiment were completely separate from those in experiment 1, and,
Fig. 1. Modeled current flow for cerebellar stimulation using Soterix HDTar-
gets software. Color bar denotes field intensity (V/m) on a scale of 0 – 0.21
(bottom to top). L, left; R, right; F, front; B, back.
 TDCS EFFECTS ON SEQUENCE LEARNING
as such, none of our subjects participated in both experiments result-
ing in two completely independent data sets. Three of those individ-
uals did not return for the second session of the study, one individual
did not undergo tDCS due to potential interactions with a medication,
complete data was not collected from one individual due to technical
difficulties, and one individual did not complete a full session due to
discomfort from stimulation. Additionally, six individuals were ex-
cluded from final analyses for having task performance scores below
three standard deviations from the group mean. The same a priori
cutoff from experiment 1 was applied here to uniformly exclude
behavioral outliers. Those excluded performed the task with less than
46% accuracy. As such, our final sample included 23 subjects (15
female, mean age 18.6 yr ⫾ 0.9 SD, range 18 –21 yr). None of the
subjects had any history of medical or neurological disease and all
were right-handed (mean score 93.9 ⫾ 12.7 SD). Subjects were
screened for the same exclusion criteria as in experiment 1 (Nitsche et
al. 2008), and none were taking medication that could potentially
impact the central nervous system. Subjects were again recruited and
scheduled through the Texas A&M University Psychology Subject
Pool and received an equal amount of course credit as those in the first
experiment. Participation was again limited to individuals that had not
previously completed other tDCS studies. Prior to the initiation of
study procedures, all subjects provided written, informed consent, and
the protocol was submitted to and approved by the Institutional
Review Board at Texas A&M University.
Procedure. All procedures employed in the second experiment
were identical to those described for experiment 1, with the exception
of the polarity of stimulation. Participants underwent two sessions,
exactly 1 wk apart, in which they received both anodal and sham
tDCS in a within-subjects design. We again used a single-blind
procedure wherein participants were not told which condition they
were receiving, and the order of stimulation condition was counter-
balanced. The electrode array and current intensities used for anodal
cerebellar stimulation are presented in Table 2 and the associated
modeled current flow is identical to that of Fig. 1, but the flow of
current is in the opposite direction. This was achieved using a “spiral
in” approach in this experiment, equivalent to anodal stimulation in a
traditional two electrode pad system, again, determining the direction
of current flow. Notably, the electrode placements and current inten-
sities are identical to those in experiment 1 to replicate the stimulation
design with 2.0 mA of current applied to the lateral posterior cere-
bellum; however, the polarity of the current has been flipped. The
same stimulation procedures from the first experiment were imple-
mented in the second to avoid the subjects’ detection of stimulation
condition and maintain consistency in our experimental design.
The motor sequence learning paradigm directly paralleled that from
experiment 1 (Kwak et al. 2012), and all parameters were identical.
Participants were presented with a new sequence pattern during the
second session of this experiment, and each sequence comprised six
Table 2. Electrode array and current intensities for cerebellar
anodal stimulation
Location Current
1 P2 ⫺0.1135
2 PO10 0.1551
3 O10 0.3618
4 EX3 ⫺0.1956
5 EX4 1.4832
6 EX5 ⫺0.4066
7 EX12 ⫺0.22
8 EX14 ⫺0.7886
C NK1 ⫺0.2757
Current, specific intensity of electrical current applied at a given location;
Location, particular locations for electrode placement, determined by the 10-20
system. 1– 8, Stimulation electrodes; C, return electrode.
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
493
elements that did not contain any trills or sequential numbers. The
same sequences used in the first experiment were employed here as
well. As we had all new participants in experiment 2, there was no
concern with individuals already having learned the sequence.
Data analysis. As in experiment 1, performance was measured by
average total accuracy (ACC) per sequence block and mean correct
reaction time (RT) during both random and sequence blocks. Addi-
tionally, scores for each measure were calculated identically to those
used in the first experiment. Repeated-measures ANOVAs and paired
sample t-tests were performed to investigate performance differences
between anodal and sham stimulation.
RESULTS
Experiment 1
Reaction time. A 2 ⫻ 2 repeated-measures ANOVA (block
type by stimulation condition), collapsing across all random
and sequence blocks, respectively, revealed a significant main
effect of block type, F(1,20) ⫽ 108.23, P ⬍ 0.001, ␩2p ⫽ 0.844,
indicating that responses were significantly faster on sequence
blocks compared with random, as expected. However, there
was no significant block type by stimulation condition inter-
action, F(1,20) ⫽ 0.144, P ⫽ 0.71, ␩2p ⫽ 0.001, and no signif-
icant main effect of stimulation condition, F(1,20) ⫽ 1.92, P ⫽
0.18, ␩2p ⫽ 0.087, indicating that RT did not differ between
cathodal and sham stimulation. The difference in RT between
block types indicates that learning took place as anticipating
the upcoming key presses allowed for faster responses (Fig. 2).
In fact, t-tests reveal a largely significant difference in reaction
time between the first (S2) and last (S14) sequence blocks (P ⬍
0.001), further indicative of learning.
Accuracy. A 3 ⫻ 2 repeated-measures ANOVA (learning
phase by stimulation condition), on sequence blocks only,
revealed a significant learning phase by stimulation condition
interaction, F(2,40) ⫽ 3.35, P ⬍ 0.05, ␩2p ⫽ 0.143. However,
there was no significant main effect of learning phase,
F(2,40) ⫽ 0.919, P ⫽ 0.41, ␩2p ⫽ 0.044, or stimulation condi-
tion, F(1,20) ⫽ 1.58, P ⫽ 0.22, ␩2p ⫽ 0.073. Follow-up paired
samples t-tests to further unpack this interaction indicated that
there is a marginally significant difference between cathodal
and sham stimulation during the late phase of learning,
t(20) ⫽ 2.07, P ⫽ 0.051. Accuracy during the early [t(20) ⫽
0.119, P ⫽ 0.901] and middle [t(20) ⫽ 0.522, P ⫽ 0.607]
phases of learning did not differ with stimulation. While
individuals showed similar levels of performance during the
first two phases of the task, performance was better (though
only marginally significant) in the late phase, indicating an
interaction between stimulation and performance depending on
the specific phase of learning (Fig. 3).
Experiment 2
Reaction time. A 2 ⫻ 2 repeated-measures ANOVA (block
type by stimulation condition), collapsing across all random
and sequence blocks, revealed a significant main effect of
block type on RT, F(1,22) ⫽ 77.78, P ⬍ 0.001, ␩2p ⫽ 0.78,
wherein RT was significantly slower during the random blocks
as compared with the sequence blocks. However, there was no
significant main effect of stimulation condition (anodal versus
sham), F(1,22) ⫽ 0.27, P ⫽ 0.61, ␩2p ⫽ 0.012, or an interaction
between block type and stimulation condition, F(1,22) ⫽
0.104, P ⫽ 0.74, ␩2p ⫽ 0.005, consistent with experiment 1.
494 TDCS EFFECTS ON SEQUENCE LEARNING

Reaction Time by Block

Average Correct Reaction Time (msec)

450

400

350

Fig. 2. Group mean of correct reaction time, excluding missed

300
trials, across all blocks in both cathodal and sham conditions
for cerebellar stimulation. Error bars indicate standard error
(SE). R, random block; S, sequence block. 250

200
Cathodal Sham

150
R1 S2 S3 S4 R5 R6 S7 S8 S9 R10 R11 S12 S13 S14 R15
Block

This indicates that participants perform the task more quickly
when presented with sequence trials in comparison to random,
as would be expected when movements are repeated in a
sequential manner and said sequence is learned (Fig. 4). Fur-
thermore, follow-up t-tests reveal a significant difference in
reaction time between the first and last sequence blocks for this
experimental group (P ⬍ 0.001). However, though reaction
time improves over the course of the sequence blocks, accu-
racy does not improve, presumably due to the effects of
stimulation on task performance.
Accuracy. A 3 ⫻ 2 repeated-measures ANOVA (learning
phase by stimulation condition) revealed a significant learning
phase by stimulation condition interaction (Fig. 3), F(2,44) ⫽
3.43, P ⬍ 0.05, ␩2p ⫽ 0.135, and a significant main effect of
learning phase, F(2,44) ⫽ 3.34, P ⬍ 0.05, ␩2p ⫽ 0.132. The
main effect of stimulation condition was approaching the
standard cutoff for statistical significance, F(1,22) ⫽ 3.65, P ⫽
0.069, ␩2p ⫽ 0.142. Follow-up paired samples t-tests to inves-
tigate this interaction revealed that there was no significant
accuracy difference between sham and anodal stimulation
during early learning [t(22) ⫽ ⫺0.353, P ⫽ 0.73], though there
was a significant difference during the intermediate phase
[t(22) ⫽ ⫺2.069, P ⫽ 0.05], and a difference that was ap-
proaching significance during the late phase of learning
[t(22) ⫽ ⫺1.980, P ⫽ 0.06]. In both the intermediate and late
phases, performance following anodal stimulation was worse
than that following sham stimulation; however, performance
did not differ during early learning, driving the interaction.
Thus, though participants responded more quickly, they did so
with poorer accuracy over time: performance worsened as
assessed by accuracy.
Comparing Cathodal and Anodal Stimulation
To follow up on the results from experiments 1 and 2 and get
a better idea of the effect of stimulation type on explicit motor
sequence learning, we conducted additional between-subjects
analyses on the two independent groups. These statistical
procedures were carried out as a priori analyses aimed at
accomplishing a comprehensive perspective on the nature of
the cerebellum’s role in motor learning by comparing the
stimulation conditions. In all cases, we used 3 ⫻ 2 (learning
phase by stimulation group) mixed-model ANOVAs. Because
significant stimulation effects were limited to the accuracy
measure, we focused our analyses on accuracy. First, we
investigated group differences on the sham condition only, to
determine whether or not there were any baseline learning
differences between the two experimental groups. Critically,
there was no significant main effect of group, F(1,42) ⫽ 0.796,
P ⫽ 0.377, ␩2p ⫽ 0.019. Furthermore, there was no significant
interaction between learning phase and group, F(2,84) ⫽
0.441, P ⫽ 0.645, ␩2p ⫽ 0.010, though the main effect of phase
was approaching significance, F(2,84) ⫽ 2.73, P ⫽ 0.071,

100
Accuracy by Learning Phase
Sequence Trials Only
Average Total Accuracy (%)

95
Fig. 3. Group means of total accuracy, missed trials included,
across early, intermediate, and late sequence learning phases
in cathodal, sham-experiment 1, sham-experiment 2, and an-
odal conditions for cerebellar stimulation. Error bars indicate
SE. Data from experiment 1 is presented with solid lines, 90 Cathodal
while that from experiment 2 is presented with dashed lines. Sham 1
Anodal
Sham 2

85
Early Intermediate Late

Learning Phase

J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org

Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.

Average Correct Reaction Time (msec)
450
400
350
300
250
200
Anodal
150
R1 S2 S3 S4
␩2p ⫽ 0.061. Therefore, in the absence of stimulation, partici-
pants exhibited similar degrees of performance on the sequence
trials across both sham groups.
Next, we directly compared performance between the two
groups during active stimulation (cathodal compared with
anodal). This revealed a significant learning phase by group
interaction, F(2,84) ⫽ 4.35, P ⬍ 0.05, ␩2p ⫽ 0.094, as well as a
significant main effect of group, F(1,42) ⫽ 6.79, P ⬍ 0.05,
␩2p ⫽ 0.139. The main effect of learning phase was approach-
ing significance, F(2,84) ⫽ 2.53, P ⫽ 0.085, ␩2p ⫽ 0.057. As
seen in Fig. 3, participants perform significantly better on the
sequence learning paradigm after cathodal cerebellar stimula-
tion compared with anodal. Thus, stimulation has polarity-
specific effects on our explicit motor sequence learning para-
digm.
Furthermore, we calculated difference scores relative to
sham (i.e., ACC cathodal group – ACC sham experiment 1
group and ACC anodal group – ACC sham experiment 2
group) for both experiments, so as to compare across groups
relative to baseline. This exploratory, post hoc analysis was
implemented to provide information about the magnitude of
learning achieved within each group after active stimulation
when different conditions were compared both directly and
also in consideration of baseline performance after sham stim-
ulation. Importantly, as described above, accuracy perfor-
mance between the groups in the absence of stimulation did not
differ, justifying our use of the sham scores as controls for this
specific measure. First, there was a significant learning phase
by stimulation group interaction, F(2,84) ⫽ 5.42, P ⬍ 0.01,
␩2p ⫽ 0.114. There was also a significant main effect of group,
F(1,42) ⫽ 4.88, P ⬍ 0.05, ␩2p ⫽ 0.104, though the main effect
of learning phase was not significant, F(2,84) ⫽ 1.04, P ⫽
0.36, ␩2p ⫽ 0.024. Following up on these analyses, independent
samples t-tests revealed that the interaction is driven by a
significant difference in performance relative to sham during
late learning specifically, t(42) ⫽ 2.59, P ⫽ 0.01, and a signif-
icant difference during the middle phase, t(42) ⫽ 2.02, P ⫽
0.05, though there were no differences during early learning,
t(42) ⫽ 0.35, P ⫽ 0.728. This provides further support for
polarity-specific effects of cerebellar stimulation on explicit
motor sequence learning, even when normalized relative to
baseline (sham stimulation). Cathodal stimulation is resulting
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
TDCS EFFECTS ON SEQUENCE LEARNING 495
Reaction Time by Block
Fig. 4. Group mean of correct reaction time, excluding
missed trials, across all blocks in both anodal and sham
conditions for cerebellar stimulation. Error bars indicate SE.
R, random block; S, sequence block.
Sham
R5 R6 S7 S8 S9 R10 R11 S12 S13 S14 R15
Block
in improvement whereas anodal is making performance worse
over time, specifically when scores from each active condition
are normalized to sham and subsequently compared.
DISCUSSION
Applying HD-tDCS to the lateral posterior cerebellum be-
fore the learning of a new motor sequence may have helped
facilitate initial skill acquisition when using cathodal stimula-
tion (relative to sham), whereas anodal stimulation negatively
impacted performance (relative to sham). Across the learning
phases, accuracy increased with cathodal stimulation (though
these increases were quite small) but decreased with anodal
when comparing between experiments. Critically, performance
across the two experiments and stimulation types significantly
differed, suggesting a polarity-specific effect of stimulation on
accuracy during explicit motor sequence learning. However,
we did not see any effects on RT other than differences in
speed between block types, wherein responses were signifi-
cantly faster during sequence blocks relative to random blocks,
across both experiments. These findings and their implications
are discussed, in turn, below.
With respect to performance accuracy, we found that cath-
odal cerebellar stimulation provides a benefit relative to sham.
Specifically, this was driven by improved performance toward
the later stages of our paradigm, though only marginally
significant. Our results suggest that cathodal stimulation is
more favorable than no stimulation at all, but the modest nature
of these within-subject findings from experiment 1 do not
conclusively indicate that we have directly facilitated learning.
Rather, we may have prevented any potential fatigue or atten-
tional declines that may occur during the end of the task.
Nonetheless, cathodal cerebellar stimulation positively im-
pacted sequence learning. Conversely, anodal stimulation had
a negative influence on skill acquisition. While performance
was nearly identical to that of the sham condition during the
first few blocks of learning (here, the early phase), perform-
ance after anodal stimulation was significantly worse later in
the task. These effects were considerably more pronounced in
the anodal group compared with the cathodal, but the between-
subject analyses further supported the idea that a differential
influence on learning is observed across the two conditions. As
such, these results indicate polarity-specific effects of tDCS on
explicit motor sequence learning.
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
496 TDCS EFFECTS ON SEQUENCE LEARNING
As this paradigm is only performed once and without ex-
tensive practice, these effects take place during the overall
earlier stages of explicit sequence learning when cognitive
resources, such as working memory, are put to use. Stimulation
was specifically applied to a cognitive region of the cerebel-
lum, namely the lateral posterior cerebellum (roughly includ-
ing Crus I and II), which has been implicated in working
memory processes (Bernard and Seidler 2013, 2014; Guell et
al. 2018; Schmahmann and Sherman 1998; Stoodley 2012).
This may, therefore, explain the early influence on motor skill
acquisition as the motor pattern is actively rehearsed and held
at the forefront of the mind in light of explicit information.
Furthermore, it is possible that our stimulation montage ex-
erted some downstream effects on the prefrontal cortex as the
cerebellum maintains closed-loop connections with this cogni-
tive area (Bernard et al. 2012, 2014; Krienen and Buckner
2009; O’Reilly et al. 2010; Salmi et al. 2010). We speculate
that the diverging impacts on accuracy that emerged over the
course of the task after stimulation are due to these networks
and functions. That is, cathodal stimulation of the lateral
posterior cerebellum may have “primed” or activated network
circuits involved in working memory, allowing them to be
more effectively utilized when cognitive demands are higher in
these initial skill acquisition phases of motor sequence learn-
ing. It is possible that our approach may have influenced the
cognitive processes at play more directly (Desmond et al.
2005; Ferrucci and Priori 2014; Oliveri et al. 2007), but there
is also evidence to suggest that cerebellar stimulation induces
functional network changes that may underlie the observed
effects as well (Halko et al. 2014). Cathodal stimulation may
have facilitated working memory resources and the associated
cerebellar circuits that communicate with the cerebral cortex.
In contrast, anodal stimulation may have inhibited these cere-
bellar cortical regions and circuits, making it more difficult for
participants to engage cognitive resources, presumably verbal
working memory (cf. Bo and Seidler 2009), during this overall
earlier learning period.
Modulating, and specifically improving, motor learning is an
area of great interest, given the potential implications that
success in this regard would have for rehabilitation after injury
or infarct. For a broad review on the effects of tDCS on motor
learning, we refer readers to Buch et al. (2017). Pertinent to our
work here, there have been several studies targeting the cere-
bellum specifically (e.g., Block and Celnik 2013; Cantarero et
al. 2015; Foerster et al. 2013; Galea et al. 2011; Hardwick and
Celnik 2014; Shah et al. 2013), though these investigations
have primarily focused on adaptation tasks. As such, our
investigation here using explicit sequence learning presents a
novel extension of this literature as a whole. Across multiple
investigations of adaptation learning, results suggest that an-
odal tDCS can facilitate motor learning (Block and Celnik
2013; Galea et al. 2011; Hardwick and Celnik 2014), though
cathodal stimulation has also been shown to facilitate learning
in an ankle adaptation task (Shah et al. 2013). More recently,
work from Cantarero et al. (2015) included both cathodal and
anodal conditions, during which individuals performed a task
involving the production of isometric force with a design
that includes sequencing components. Notably, they found
polarity-specific effects, wherein anodal stimulation improved
performance, while cathodal stimulation had no effect on
learning.
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
In the present study, we found polarity-specific effects of
stimulation. However, the directionality of these findings dif-
fered from the larger literature using tDCS to modulate motor
learning. Our results suggest that cathodal stimulation may
have had a small facilitatory effect, whereas anodal stimulation
had a negative effect on learning, as measured by accuracy.
While at first this stands in contrast to the given literature, there
are several key considerations. First, the adaptation work has
largely relied on online tDCS stimulation (Block and Celnik
2013; Cantarero et al. 2015; Galea et al. 2011). That is,
stimulation was applied during performance of the learning
paradigm. Cantarero et al. (2015) suggest that this online
stimulation may aid in the processing of motor errors in the
cerebellar cortex. Second, our investigation took advantage of
an HD-tDCS system, whereas work to date has primarily relied
on two electrode stimulation pads (Ehsani et al. 2016; Wessel
et al. 2016; Shimizu et al. 2017). In addition, these studies have
employed different types of stimuli and made use of task
designs distinct from our own, as the ultimate intention of these
investigations was to examine motor retention rather than skill
acquisition as is our objective in the current study. Thus,
differences in the underlying methodology and experimental
aims may contribute to these discrepancies as well. However,
Shah et al. (2013) did see comparable facilitation with both
anodal and cathodal stimulation, and a recent meta-analysis of
the effects of cerebellar tDCS suggests that while stimulation
impacts behavior (both motor and cognitive) across multiple
studies, there is little support to suggest that these effects are
polarity specific (Oldrati and Schutter 2018). Finally, the un-
derlying physiology of the cerebellum is also of note.
In an investigation of the impact of cerebellar tDCS on
cognition, Pope and Miall (2012) demonstrated that cathodal
stimulation resulted in behavioral improvements, while anodal
stimulation had no impact on behavior. This is broadly con-
sistent with our findings here, and our goal was to ultimately
modulate cognitive networks and systems. Notably, Pope and
Miall (2012) suggest that this seemingly counterintuitive result
is due to the inhibitory action of cerebellar Purkinje cells on the
deep cerebellar nuclei, which are the primary output regions of
this structure. Because Purkinje cells have an inhibitory effect
on the deep cerebellar nuclei, presumably the dentate nucleus
in this particular investigation, cathodal stimulation may de-
crease the action of the Purkinje cells and release this inhibition
(Pope and Miall 2012). Reduced inhibition may in turn result
in facilitation of cerebello-cortical networks.
This is also consistent with the idea of cerebellar-brain
inhibition. With transcranial magnetic stimulation, stimulation
to the cerebellum results in inhibition over the primary motor
cortex (Popa et al. 2010), and when tDCS is used, there are
comparable polarity-specific effects (Galea et al. 2009). Cath-
odal tDCS to the cerebellum decreases this inhibition, while
anodal stimulation increases it. While this work has focused on
the primary motor cortex, we speculate that the comparable
circuitry with the prefrontal cortex acts in a similar manner.
Thus, in the present study, we suggest that stimulation to the
lateral posterior cerebellum results in either an increase in or
release from inhibition on prefrontal cortical regions and net-
works depending on stimulation polarity, paralleling what was
suggested by Pope and Miall (2012). This would then result in
either negative impacts or relative improvement on accuracy
during explicit motor sequence learning. Targeting these cer-
 TDCS EFFECTS ON SEQUENCE LEARNING
ebello-thalamo-frontal circuits may have inhibited or facili-
tated the use of working memory resources, which in turn
influenced accuracy during the learning of a motor pattern.
Future work incorporating functional neuroimaging methodol-
ogies, however, is needed to further investigate this idea.
More broadly, our findings further underscore the impor-
tance of the lateral posterior cerebellum in nonmotor behavior
and suggest that the cerebellum may also contribute to the
cognitive aspects of initial skill acquisition. In addition to
contributing to the control of movements necessary to com-
plete the task, we suggest that cerebellar networks associated
with cognitive processing are also used during learning. Prior
work has linked motor sequence learning with working mem-
ory capabilities (Bo and Seidler 2009), suggesting that those
with higher working memory capacity are better able to chunk
sequence elements. Notably, the cortical networks that typi-
cally support sequence learning include the basal ganglia,
prefrontal cortex, parietal cortex, and of course, primary motor
cortex (Doyon et al. 2002; Galea et al. 2011; Hikosaka et al.
1999; Seidler 2006). While the cerebellum is also activated
during sequence learning, it is less prevalent, particularly as
compared with activation seen during adaptation learning par-
adigms (Anguera et al. 2010; Baizer et al. 1999; Imamizu et al.
2000, 2003). However, as previously noted, the region we
targeted with stimulation is associated with working memory
processes as well as language (Stoodley and Schmahmann
2009; Stoodley 2012). Stimulation may have therefore facili-
tated the cortical networks necessary for efficient chunk for-
mation (though we did not directly measure chunking here) and
allowed for the use of verbal working memory rehearsal
strategies during initial skill acquisition.
While our findings provide interesting new insights into the
role of the cerebellum in explicit sequence learning, there are
several limitations to consider. First, we stimulated the cere-
bellum and suggest that we may have facilitated activation in
the broader cerebello-prefrontal network associated with work-
ing memory (Bernard et al. 2012; Krienen and Buckner 2009;
Stoodley 2012; Stoodley and Schmahmann 2009); however,
this stimulation may have spread to other regions of the
cerebellum. Though HD-tDCS allows for improved targeting,
the stimulation does reach other areas that could have an effect
on performance (Fig. 1). And furthermore, the cerebellum is
also known to play a key role in timing (Ivry and Keele 1989;
Ivry et al. 2002; Spencer and Ivry 2013). While typically the
more medial regions of the cerebellum have been implicated in
this function (Ivry et al. 1988), again, the spread of the
stimulation effect may have helped facilitate timing. Improved
event and movement timing may, in turn, result in improved
accuracy in initial skill learning. To further test our suggestion
that improved working memory impacts explicit sequence
learning, stimulation of the prefrontal cortex in future work is
warranted as well. Second, we suggest that this working
memory facilitation may also aid in the chunking of motor
elements (Bo and Seidler 2009). However, we did not explic-
itly measure chunking as the employed sequences were rela-
tively short and our interstimulus interval was fairly small and
fixed, so this, therefore, remains speculative. Relatedly, indi-
viduals may have potentially approached the task with different
strategies in each experiment, but the lack of performance
differences across the two sham groups and the counterbalanc-
ing implemented in our design likely reduces this possibility.
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
497
Additionally, the long-term benefits of stimulation are not
clear. Our sequence learning paradigm was fairly brief, and we
did not measure retention over time. Thus, it would be bene-
ficial in future work to investigate more long-term differences
in performance after cathodal cerebellar stimulation compared
with anodal and relative to sham.
Due to the nature of working with participant subject pools,
subject attrition can be an issue. We therefore chose to conduct
two separate experiments with independent samples and then
compare across stimulation conditions as a fully within-sub-
jects design is difficult to accomplish with the recruitment
opportunities available, though such a design would be opti-
mal. This constraint also led us to exclude a considerable
number of subjects that did not complete a second session so
that we could conduct our analyses on a full, balanced data set.
Furthermore, several subjects were excluded due to poor per-
formance and a lack of compliance with instructions. In these
cases, participants were likely relatively unmotivated and did
not perform the task appropriately, resulting in poor perfor-
mance beyond what would be expected from variability within
the sample. Moreover, our subjects skew female, and this is an
artifact of the sample from which our participants were se-
lected. Though we are not able to reliably test any possible
gender effects due to power constraints, this may have im-
pacted our findings. Finally, while we generally see that anodal
cerebellar tDCS negatively impacts performance, the extent to
which this is the case is widely variable, as demonstrated by
the large error bars seen in Fig. 3. This may, in part, be due to
a considerable degree of variability in responsiveness to stim-
ulation as the impact on performance can fluctuate with respect
to individual differences (Chew et al. 2015; Li et al. 2015).
Conclusions
Together, these findings provide key new insights into the
role of the cerebellum in explicit sequence learning. Our results
suggest that stimulating the lateral posterior cerebellum with a
cathodal approach benefits initial skill acquisition, relative to
both sham and anodal stimulation. In addition, our results
suggest that, compared with both sham and cathodal stimula-
tion, applying anodal tDCS to this particular region of the
cerebellum negatively impacts explicit motor sequence learn-
ing. Therefore, these findings advance our understanding of the
role of the cerebellum in explicit sequence learning as we
suggest that we may have influenced cognitive processes in
verbal working memory, often relied on during initial learning.
Furthermore, exploring techniques that may improve motor
skill acquisition is valuable for the development of new ther-
apeutic targets for neurological disorders (e.g., Parkinson’s
disease), as well as enhancing restorative therapy after stroke
or other injuries impacting motor function and reducing nor-
mative motor declines with advanced age.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the authors.
AUTHOR CONTRIBUTIONS
J.A.B. conceived and designed research; H.K.B., J.R.G., and T.M. per-
formed experiments; H.K.B. analyzed data; H.K.B. and J.A.B. interpreted
results of experiments; H.K.B. prepared figures; H.K.B. drafted manuscript;
498 TDCS EFFECTS ON SEQUENCE LEARNING
T.M. and J.A.B. edited and revised manuscript; J.A.B. approved final version
of manuscript.
REFERENCES
Aizenstein HJ, Stenger VA, Cochran J, Clark K, Johnson M, Nebes RD,
Carter CS. Regional brain activation during concurrent implicit and explicit
sequence learning. Cereb Cortex 14: 199 –208, 2004. doi:10.1093/cercor/
bhg119.
Anguera JA, Bernard JA, Jaeggi SM, Buschkuehl M, Benson BL, Jennett
S, Humfleet J, Reuter-Lorenz PA, Jonides J, Seidler RD. The effects of
working memory resource depletion and training on sensorimotor adapta-
tion. Behav Brain Res 228: 107–115, 2012. doi:10.1016/j.bbr.2011.11.040.
Anguera JA, Reuter-Lorenz PA, Willingham DT, Seidler RD. Contribu-
tions of spatial working memory to visuomotor learning. J Cogn Neurosci
22: 1917–1930, 2010. doi:10.1162/jocn.2009.21351.
Baizer JS, Kralj-Hans I, Glickstein M. Cerebellar lesions and prism adap-
tation in macaque monkeys. J Neurophysiol 81: 1960 –1965, 1999. doi:10.
1152/jn.1999.81.4.1960.
Bernard JA, Dean DJ, Kent JS, Orr JM, Pelletier-Baldelli A, Lunsford-
Avery JR, Gupta T, Mittal VA. Cerebellar networks in individuals at ultra
high-risk of psychosis: impact on postural sway and symptom severity. Hum
Brain Mapp 35: 4064 – 4078, 2014. doi:10.1002/hbm.22458.
Bernard JA, Peltier SJ, Wiggins JL, Jaeggi SM, Buschkuehl M, Fling BW,
Kwak Y, Jonides J, Monk CS, Seidler RD. Disrupted cortico-cerebellar
connectivity in older adults. Neuroimage 83: 103–119, 2013. doi:10.1016/
j.neuroimage.2013.06.042.
Bernard JA, Seidler RD. Relationships between regional cerebellar volume
and sensorimotor and cognitive function in young and older adults. Cere-
bellum 12: 721–737, 2013. doi:10.1007/s12311-013-0481-z.
Bernard JA, Seidler RD. Moving forward: age effects on the cerebellum
underlie cognitive and motor declines. Neurosci Biobehav Rev 42: 193–207,
2014. doi:10.1016/j.neubiorev.2014.02.011.
Bernard JA, Seidler RD, Hassevoort KM, Benson BL, Welsh RC, Wiggins
JL, Jaeggi SM, Buschkuehl M, Monk CS, Jonides J, Peltier SJ. Resting
state cortico-cerebellar functional connectivity networks: a comparison of
anatomical and self-organizing map approaches. Front Neuroanat 6: 31,
2012. doi:10.3389/fnana.2012.00031.
Berner MP, Hoffmann J. Effector-related sequence learning in a bimanual-
bisequential serial reaction time task. Psychol Res 72: 138 –154, 2008.
doi:10.1007/s00426-006-0097-8.
Bhakuni R, Mutha PK. Learning of bimanual motor sequences in normal
aging. Front Aging Neurosci 7: 76, 2015. doi:10.3389/fnagi.2015.00076.
Block H, Celnik P. Stimulating the cerebellum affects visuomotor adaptation
but not intermanual transfer of learning. Cerebellum 12: 781–793, 2013.
doi:10.1007/s12311-013-0486-7.
Bo J, Seidler RD. Visuospatial working memory capacity predicts the orga-
nization of acquired explicit motor sequences. J Neurophysiol 101: 3116 –
3125, 2009. doi:10.1152/jn.00006.2009.
Buch ER, Santarnecchi E, Antal A, Born J, Celnik PA, Classen J, Gerloff
C, Hallett M, Hummel FC, Nitsche MA, Pascual-Leone A, Paulus WJ,
Reis J, Robertson EM, Rothwell JC, Sandrini M, Schambra HM,
Wassermann EM, Ziemann U, Cohen LG. Effects of tDCS on motor
learning and memory formation: a consensus and critical position paper.
Clin Neurophysiol 128: 589 – 603, 2017. doi:10.1016/j.clinph.2017.01.004.
Cantarero G, Spampinato D, Reis J, Ajagbe L, Thompson T, Kulkarni K,
Celnik P. Cerebellar direct current stimulation enhances on-line motor skill
acquisition through an effect on accuracy. J Neurosci 35: 3285–3290, 2015.
doi:10.1523/JNEUROSCI.2885-14.2015.
Chew T, Ho K-A, Loo CK. Inter- and intra-individual variability in response
to transcranial direct current stimulation (tDCS) at varying current intensi-
ties. Brain Stimul 8: 1130 –1137, 2015. doi:10.1016/j.brs.2015.07.031.
Datta A, Krause MR, Pilly PK, Choe J, Zanos TP, Thomas C, Pack CC.
On comparing in vivo intracranial recordings in non-human primates to
predictions of optimized transcranial electrical stimulation. Conf Proc IEEE
Eng Med Biol Soc 2016: 1774 –1777, 2016. doi:10.1109/EMBC.2016.
7591061.
Desmond JE, Chen SHA, Shieh PB. Cerebellar transcranial magnetic stim-
ulation impairs verbal working memory. Ann Neurol 58: 553–560, 2005.
doi:10.1002/ana.20604.
Doyon J, Gabitov E, Vahdat S, Lungu O, Boutin A. Current issues related
to motor sequence learning in humans. Curr Opin Behav Sci 20: 89 –97,
2018. doi:10.1016/j.cobeha.2017.11.012.
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
Doyon J, Gaudreau D Jr, Laforce R Jr, Castonguay M, Bédard PJ,
Bédard F, Bouchard JP. Role of the striatum, cerebellum, and frontal lobes
in the learning of a visuomotor sequence. Brain Cogn 34: 218 –245, 1997.
doi:10.1006/brcg.1997.0899.
Doyon J, Song AW, Karni A, Lalonde F, Adams MM, Ungerleider LG.
Experience-dependent changes in cerebellar contributions to motor se-
quence learning. Proc Natl Acad Sci USA 99: 1017–1022, 2002. doi:10.
1073/pnas.022615199.
E KH, Chen SH, Ho MH, Desmond JE. A meta-analysis of cerebellar
contributions to higher cognition from PET and fMRI studies. Hum Brain
Mapp 35: 593– 615, 2014. doi:10.1002/hbm.22194.
Ehsani F, Bakhtiary AH, Jaberzadeh S, Talimkhani A, Hajihasani A.
Differential effects of primary motor cortex and cerebellar transcranial
direct current stimulation on motor learning in healthy individuals: a
randomized double-blind sham-controlled study. Neurosci Res 112: 10 –19,
2016. doi:10.1016/j.neures.2016.06.003.
Eliassen JC, Souza T, Sanes JN. Human brain activation accompanying
explicitly directed movement sequence learning. Exp Brain Res 141: 269 –
280, 2001. doi:10.1007/s002210100822.
Ferrucci R, Priori A. Transcranial cerebellar direct current stimulation
(tcDCS): motor control, cognition, learning and emotions. Neuroimage 85:
918 –923, 2014. doi:10.1016/j.neuroimage.2013.04.122.
Foerster A, Rocha S, Wiesiolek C, Chagas AP, Machado G, Silva E,
Fregni F, Monte-Silva K. Site-specific effects of mental practice combined
with transcranial direct current stimulation on motor learning. Eur J Neu-
rosci 37: 786 –794, 2013. doi:10.1111/ejn.12079.
Galea JM, Jayaram G, Ajagbe L, Celnik P. Modulation of cerebellar
excitability by polarity-specific noninvasive direct current stimulation. J
Neurosci 29: 9115–9122, 2009. doi:10.1523/JNEUROSCI.2184-09.2009.
Galea JM, Vazquez A, Pasricha N, de Xivry J-JO, Celnik P. Dissociating
the roles of the cerebellum and motor cortex during adaptive learning: the
motor cortex retains what the cerebellum learns. Cereb Cortex 21: 1761–
1770, 2011. doi:10.1093/cercor/bhq246.
Grafton ST, Hazeltine E, Ivry R. Functional mapping of sequence learning
in normal humans. J Cogn Neurosci 7: 497–510, 1995. doi:10.1162/jocn.
1995.7.4.497.
Guell X, Schmahmann JD, Gabrieli J, Ghosh SS. Functional gradients of
the cerebellum. eLife 7: e36652, 2018. doi:10.7554/eLife.36652.
Halko MA, Farzan F, Eldaief MC, Schmahmann JD, Pascual-Leone A.
Intermittent theta-burst stimulation of the lateral cerebellum increases func-
tional connectivity of the default network. J Neurosci 34: 12049 –12056,
2014. doi:10.1523/JNEUROSCI.1776-14.2014.
Hardwick RM, Celnik PA. Cerebellar direct current stimulation enhances
motor learning in older adults. Neurobiol Aging 35: 2217–2221, 2014.
doi:10.1016/j.neurobiolaging.2014.03.030.
Hikosaka O, Nakahara H, Rand MK, Sakai K, Lu X, Nakamura K,
Miyachi S, Doya K. Parallel neural networks for learning sequential
procedures. Trends Neurosci 22: 464 – 471, 1999. doi:10.1016/S0166-
2236(99)01439-3.
Honda M, Deiber MP, Ibáñez V, Pascual-Leone A, Zhuang P, Hallett M.
Dynamic cortical involvement in implicit and explicit motor sequence
learning. A PET study. Brain 121: 2159 –2173, 1998. doi:10.1093/brain/
121.11.2159.
Huang Y, Liu AA, Lafon B, Friedman D, Dayan M, Wang X, Bikson M,
Doyle WK, Devinsky O, Parra LC. Measurements and models of electric
fields in the in vivo human brain during transcranial electric stimulation.
eLife 6: e18834, 2017. doi:10.7554/eLife.18834.
Imamizu H, Kuroda T, Miyauchi S, Yoshioka T, Kawato M. Modular
organization of internal models of tools in the human cerebellum. Proc Natl
Acad Sci USA 100: 5461–5466, 2003. doi:10.1073/pnas.0835746100.
Imamizu H, Miyauchi S, Tamada T, Sasaki Y, Takino R, Pütz B, Yoshioka
T, Kawato M. Human cerebellar activity reflecting an acquired internal
model of a new tool. Nature 403: 192–195, 2000. doi:10.1038/35003194.
Ivry RB, Keele SW. Timing functions of the cerebellum. J Cogn Neurosci 1:
136 –152, 1989. doi:10.1162/jocn.1989.1.2.136.
Ivry RB, Keele SW, Diener HC. Dissociation of the lateral and medial
cerebellum in movement timing and movement execution. Exp Brain Res
73: 167–180, 1988. doi:10.1007/BF00279670.
Ivry RB, Spencer RM, Zelaznik HN, Diedrichsen J. The cerebellum and
event timing. Ann N Y Acad Sci 978: 302–317, 2002. doi:10.1111/j.1749-
6632.2002.tb07576.x.
Jenkins IH, Brooks DJ, Nixon PD, Frackowiak RS, Passingham RE. Motor
sequence learning: a study with positron emission tomography. J Neurosci
14: 3775–3790, 1994. doi:10.1523/JNEUROSCI.14-06-03775.1994.
 TDCS EFFECTS ON SEQUENCE LEARNING
Karni A, Meyer G, Rey-Hipolito C, Jezzard P, Adams MM, Turner R,
Ungerleider LG. The acquisition of skilled motor performance: fast and
slow experience-driven changes in primary motor cortex. Proc Natl Acad
Sci USA 95: 861– 868, 1998. doi:10.1073/pnas.95.3.861.
Kelly RM, Strick PL. Cerebellar loops with motor cortex and prefrontal
cortex of a nonhuman primate. J Neurosci 23: 8432– 8444, 2003. doi:10.
1523/JNEUROSCI.23-23-08432.2003.
Krienen FM, Buckner RL. Segregated fronto-cerebellar circuits revealed by
intrinsic functional connectivity. Cereb Cortex 19: 2485–2497, 2009. doi:
10.1093/cercor/bhp135.
Kwak Y, Müller MLTM, Bohnen NI, Dayalu P, Seidler RD. Effect of
dopaminergic medications on the time course of explicit motor sequence
learning in Parkinson’s disease. J Neurophysiol 103: 942–949, 2010. doi:
10.1152/jn.00197.2009.
Kwak Y, Müller MLTM, Bohnen NI, Dayalu P, Seidler RD. l-DOPA
changes ventral striatum recruitment during motor sequence learning in
Parkinson’s disease. Behav Brain Res 230: 116 –124, 2012. doi:10.1016/j.
bbr.2012.02.006.
Li LM, Uehara K, Hanakawa T. The contribution of interindividual factors
to variability of response in transcranial direct current stimulation studies.
Front Cell Neurosci 9: 181, 2015. doi:10.3389/fncel.2015.00181.
Nitsche MA, Cohen LG, Wassermann EM, Priori A, Lang N, Antal A,
Paulus W, Hummel F, Boggio PS, Fregni F, Pascual-Leone A. Trans-
cranial direct current stimulation: State of the art 2008. Brain Stimul 1:
206 –223, 2008. doi:10.1016/j.brs.2008.06.004.
Nitsche MA, Paulus W. Excitability changes induced in the human motor
cortex by weak transcranial direct current stimulation. J Physiol 527:
633– 639, 2000. doi:10.1111/j.1469-7793.2000.t01-1-00633.x.
O’Reilly JX, Beckmann CF, Tomassini V, Ramnani N, Johansen-Berg H.
Distinct and overlapping functional zones in the cerebellum defined by
resting state functional connectivity. Cereb Cortex 20: 953–965, 2010.
doi:10.1093/cercor/bhp157.
Oldrati V, Schutter DJLG. Targeting the human cerebellum with transcranial
direct current stimulation to modulate behavior: a meta-analysis. Cerebellum
17: 228 –236, 2018. doi:10.1007/s12311-017-0877-2.
Oliveri M, Torriero S, Koch G, Salerno S, Petrosini L, Caltagirone C. The
role of transcranial magnetic stimulation in the study of cerebellar cognitive
function. Cerebellum 6: 95–101, 2007. doi:10.1080/14734220701213421.
Popa T, Russo M, Meunier S. Long-lasting inhibition of cerebellar output.
Brain Stimul 3: 161–169, 2010. doi:10.1016/j.brs.2009.10.001.
Pope PA, Miall RC. Task-specific facilitation of cognition by cathodal
transcranial direct current stimulation of the cerebellum. Brain Stimul 5:
84 –94, 2012. doi:10.1016/j.brs.2012.03.006.
Rauch SL, Savage CR, Brown HD, Curran T, Alpert NM, Kendrick A,
Fischman AJ, Kosslyn SM. A PET investigation of implicit and explicit
sequence learning. Hum Brain Mapp 3: 271–286, 1995. doi:10.1002/hbm.
460030403.
Sakai K, Hikosaka O, Miyauchi S, Takino R, Sasaki Y, Pütz B. Transition
of brain activation from frontal to parietal areas in visuomotor sequence
J Neurophysiol • doi:10.1152/jn.00036.2019 • www.jn.org
Downloaded from www.physiology.org/journal/jn at UB Frankfurt (141.002.140.067) on August 5, 2019.
499
learning. J Neurosci 18: 1827–1840, 1998. doi:10.1523/JNEUROSCI.18-
05-01827.1998.
Salmi J, Pallesen KJ, Neuvonen T, Brattico E, Korvenoja A, Salonen O,
Carlson S. Cognitive and motor loops of the human cerebro-cerebellar
system. J Cogn Neurosci 22: 2663–2676, 2010. doi:10.1162/jocn.2009.
21382.
Schendan HE, Searl MM, Melrose RJ, Stern CE. An FMRI study of the role
of the medial temporal lobe in implicit and explicit sequence learning.
Neuron 37: 1013–1025, 2003. doi:10.1016/S0896-6273(03)00123-5.
Schmahmann JD, Sherman JC. The cerebellar cognitive affective syndrome.
Brain 121: 561–579, 1998. doi:10.1093/brain/121.4.561.
Seidler RD. Differential effects of age on sequence learning and sensorimotor
adaptation. Brain Res Bull 70: 337–346, 2006. doi:10.1016/j.brainresbull.
2006.06.008.
Shah B, Nguyen TT, Madhavan S. Polarity independent effects of cerebellar
tDCS on short term ankle visuomotor learning. Brain Stimul 6: 966 –968,
2013. doi:10.1016/j.brs.2013.04.008.
Shimizu RE, Wu AD, Samra JK, Knowlton BJ. The impact of cerebellar
transcranial direct current stimulation (tDCS) on learning fine-motor se-
quences. Philos Trans R Soc Lond B Biol Sci 372: 372, 2017. doi:10.1098/
rstb.2016.0050.
Spampinato D, Celnik P. Deconstructing skill learning and its physiological
mechanisms. Cortex 104: 90 –102, 2018. doi:10.1016/j.cortex.2018.03.017.
Spencer RMC, Ivry RB. Cerebellum and timing. In: Handbook of the
Cerebellum and Cerebellar Disorders, edited by Manto M, Schmahmann
JD, Rossi F, Gruol DL, Koibuchi N. Dordrecht: Springer Netherlands, 2013,
p. 1201–1219.
Stoodley CJ. The cerebellum and cognition: evidence from functional imaging
studies. Cerebellum 11: 352–365, 2012. doi:10.1007/s12311-011-0260-7.
Stoodley CJ, Schmahmann JD. Functional topography in the human cere-
bellum: a meta-analysis of neuroimaging studies. Neuroimage 44: 489 –501,
2009. doi:10.1016/j.neuroimage.2008.08.039.
Strick PL, Dum RP, Fiez JA. Cerebellum and nonmotor function. Annu
Rev Neurosci 32: 413– 434, 2009. doi:10.1146/annurev.neuro.31.060407.
125606.
Sun FT, Miller LM, Rao AA, D’Esposito M. Functional connectivity of
cortical networks involved in bimanual motor sequence learning. Cereb
Cortex 17: 1227–1234, 2007. doi:10.1093/cercor/bhl033.
Takano Y, Yokawa T, Masuda A, Niimi J, Tanaka S, Hironaka N. A rat
model for measuring the effectiveness of transcranial direct current stimu-
lation using fMRI. Neurosci Lett 491: 40 – 43, 2011. doi:10.1016/j.neulet.
2011.01.004.
Ungerleider LG, Doyon J, Karni A. Imaging brain plasticity during motor
skill learning. Neurobiol Learn Mem 78: 553–564, 2002. doi:10.1006/nlme.
2002.4091.
Wessel MJ, Zimerman M, Timmermann JE, Heise KF, Gerloff C, Hum-
mel FC. Enhancing consolidation of a new temporal motor skill by cere-
bellar noninvasive stimulation. Cereb Cortex 26: 1660 –1667, 2016. doi:10.
1093/cercor/bhu335.