An Intact Action-Perception Coupling Depends on the Integrity of the Cerebellum

Abstract

It is widely accepted that action and perception in humans functionally interact on multiple levels. Moreover, areas originally suggested to be predominantly motor-related, as the cerebellum, are also involved in action observation. However, as yet, few studies provided unequivocal evidence that the cerebellum is involved in the action perception coupling (APC), specifically in the integration of motor and multisensory information for perception. We addressed this question studying patients with focal cerebellar lesions in a virtual-reality paradigm measuring the effect of action execution on action perception presenting self-generated movements as point lights. We measured the visual sensitivity to the point light stimuli based on signal detection theory. Compared with healthy controls cerebellar patients showed no beneficial influence of action execution on perception indicating deficits in APC. Applying lesion symptom mapping, we identified distinct areas in the dentate nucleus and the lateral cerebellum of both hemispheres that are causally involved in APC. Lesions of the right ventral dentate, the ipsilateral motor representations (lobules V/VI), and most interestingly the contralateral posterior cerebellum (lobule VII) impede the benefits of motor execution on perception. We conclude that the cerebellum establishes time-dependent multisensory representations on different levels, relevant for motor control as well as supporting action perception. Ipsilateral cerebellar motor representations are thought to support the somatosensory state estimate of ongoing movements, whereas the ventral dentate and the contralateral posterior cerebellum likely support sensorimotor integration in the cerebellar-parietal loops. Both the correct somatosensory as well as the multisensory state representations are vital for an intact APC.

Full text

Behavioral/Cognitive
An Intact Action-Perception Coupling Depends on the
Integrity of the Cerebellum
Andrea Christensen,
1
Martin A. Giese,
1
Fahad Sultan,
2
Oliver M. Mueller,
3
Sophia L. Goericke,
4
Winfried Ilg,
1
*
and Dagmar Timmann
5
*
1
Section Computational Sensomotorics, Department of Cognitive Neurology, Hertie Institute for Clinical Brain Research, and Centre for Integrative
Neuroscience,
2
MRI Laboratory, Department of Cognitive Neurology, Hertie Institute for Clinical Brain Research, University Clinic Tu¨bingen, 72076
Tu¨bingen, Germany, and
3
Department of Neurosurgery,
4
Institute of Diagnostic and Interventional Radiology and Neuroradiology, and
5
Department of
Neurology, University of Duisburg-Essen, 45147 Essen, Germany
It is widely accepted that action and perception in humans functionally interact on multiple levels. Moreover, areas originally suggested
to be predominantly motor-related, as the cerebellum, are also involved in action observation. However, as yet, few studies provided
unequivocal evidence that the cerebellum is involved in the action perception coupling (APC), specifically in the integration of motor and
multisensory information for perception. We addressed this question studying patients with focal cerebellar lesions in a virtual-reality
paradigm measuring the effect of action execution on action perception presenting self-generated movements as point lights. We
measured the visual sensitivity to the point light stimuli based on signal detection theory. Compared with healthy controls cerebellar
patients showed no beneficial influence of action execution on perception indicating deficits in APC. Applying lesion symptom mapping,
we identified distinct areas in the dentate nucleus and the lateral cerebellum of both hemispheres that are causally involved in APC.
Lesions of the right ventral dentate, the ipsilateral motor representations (lobules V/VI), and most interestingly the contralateral poste-
rior cerebellum (lobule VII) impede the benefits of motor execution on perception. We conclude that the cerebellum establishes time-
dependent multisensory representations on different levels, relevant for motor control as well as supporting action perception. Ipsilateral
cerebellar motor representations are thought to support the somatosensory state estimate of ongoing movements, whereas the ventral
dentate and the contralateral posterior cerebellum likely support sensorimotor integration in the cerebellar-parietal loops. Both the
correct somatosensory as well as the multisensory state representations are vital for an intact APC.
Key words: action perception coupling; biological motion; cerebellum; lesion symptom mapping
Introduction
As humans, we interact with our world by moving ourselves and
perceiving others move. It is widely accepted that action and
perception are not two isolated domains but interact on multiple
levels. In consequence, it has been argued that similar mecha-
nisms are involved in the control of our own movements and in
understanding and anticipation actions of others (Gallese et al.,
1996;Prinz, 1997;Wolpert et al., 2003). Behavioral studies pro-
vide evidence for a bidirectional relationship of action execution
and action perception showing motor resonance effects while
observing movements (Fadiga et al., 1995) and effects of percep-
tual resonance, the influence of motor behavior on perception,
when executing actions (Casile and Giese, 2006;Schu¨tz-Bosbach
and Prinz, 2007). We recently showed that perceptual resonance
depends on temporal matching: Action detection is facilitated by
concurrent motor behavior of same actions if visual stimuli and
executed movements are synchronous. If the visual stimulus is
delayed with respect to the movement, biological motion detec-
tion is inhibited (Christensen et al., 2011).
Functional imaging studies additionally support the coupling
of action and perception, showing that several cortical motor
areas are also activated during action observation. The inferior
frontal gyrus, premotor cortex, supplemental motor area, area PF
in the supramarginal gyrus, as well as the cerebellum are consid-
ered to be part of this human action-observation network (De-
cety et al., 1997;Gazzola and Keysers, 2009;Kilner et al., 2009).
The cerebellum is well known to be involved in the coordina-
tion and fine-tuning of movements. As key mechanism for this
purpose, internal forward models have been proposed, predict-
ing sensory consequences of actions (Wolpert and Flanagan,
2001;Miall, 2003;Bastian, 2006;Ebner and Pasalar, 2008). Com-
Received Aug. 1, 2013; revised Feb. 25, 2014; accepted April 7, 2014.
Author contributions: A.C., M.A.G., W.I., and D.T. designed research; A.C., W.I., and D.T. performed research; A.C.,
F.S., O.M.M., S.L.G., W.I., and D.T. analyzed data; A.C., M.A.G., F.S., O.M.M., S.L.G., W.I., and D.T. wrote the paper.
This work was supported by the EC FP7 Projects TANGO Grant FP7-249858-TP3 and AMARSi Grant FP7-ICT-
248311, the Deutsche Forschungsgemeinschaft Grant GI 305/4-1, the German Federal Ministry of Education and
Research FKZ 01GQ1002A, and EU Training Network (ITN) ABC PITN-GA-011-290011. We thank Beate Brol for help
withthedataanalysis;MartinLo¨ffler for the setup illustration; Peter Thier, Thomas Haarmeier, and Claire Roether for
helpfuldiscussions; two anonymous reviewers for their helpful feedback on an earlier version of this manuscript; and
all participants of our study.
The authors declare no competing financial interests.
*W.I. and D.T. contributed equally to this work.
Correspondence should be addressed to Dr. Winfried Ilg, Section Computational Sensomotorics, Department of
Cognitive Neurology, Hertie Institute for Clinical Brain Research, and Centre for Integrative Neuroscience, Otfried
Mueller Str. 25, 72076 Tuebingen, Germany. E-mail: winfried.ilg@uni-tuebingen.de.
DOI:10.1523/JNEUROSCI.3276-13.2014
Copyright © 2014 the authors 0270-6474/14/346707-10$15.00/0
The Journal of Neuroscience, May 7, 2014 •34(19):6707– 6716 • 6707

bining such predictions with current sensory information, the
cerebellum is suggested to establish a state estimate, essential for
the control of movements (Miall et al., 2007) and motor adapta-
tion (Bastian, 2011;Izawa et al., 2012).
In addition to its key role in motor control, other studies have
shown an involvement of the cerebellum in sensory processing
(Gao et al., 1996) and perception (Ha¨ndel et al., 2009;Bastian,
2011), especially in tasks requiring precise timing (Ivry and Spen-
cer, 2004;O’Reilly et al., 2008).
First support for an involvement of the cerebellum in the
integration of action and perception is given by a study reporting
deficits of cerebellar patients in active force perception during
self-generated movements but preserved passive proprioception
(Bhanpuri et al., 2012).
These findings raise the question to which degree the cerebel-
lum is involved in a more general integration of motor informa-
tion and higher-order multimodal perception, including visual
action perception.
We directly addressed both the causal role and the specific
areas of the cerebellum for intact action-perception coupling.
We showed a critical influence of cerebellar lesions on the
perceptual resonance investigating patients with focal cerebel-
lar lesions compared with healthy controls. This was accom-
plished by examining the influence of movements on visual
sensitivity (i.e., noise tolerance based on d⬘values/signal de-
tection theory) in a virtual-reality paradigm, which enabled us
to display biological motion stimuli synchronously or asyn-
chronously to self-executed movements.
Materials and Methods
Participants. Seventeen patients (mean age, 28 years 4 months; 10 female,
7 male) with chronic focal lesions of the cerebellum after benign tumor
resection participated in the experiment: pilocystic astrocytoma WHO
Grade I (n⫽10), astrocytoma WHO Grade II (n⫽1), hemangioblas-
toma (n⫽5) or angioma (n⫽1). None of the patients received adjuvant
radiotherapy or chemotherapy. Patients showed mild to moderate ataxia
symptoms as examined by an experienced neurologist (D.T). Severity of
ataxia was rated using the International Cooperative Ataxia Rating Scale
(ICARS; for an overview of individual ataxia symptoms, see Table 1)
(Trouillas et al., 1997). In addition, we tested 17 control subjects,
matched in age, gender, and handedness (mean age, 27 years 9 months;
10 female, 7 male; 15 right handed). All healthy participants had normal
or corrected vision and no motor impairment influencing their arm move-
ments. They were naive with regard to the purpose of the study and received
payment for their participation. All patients and control subjects gave in-
formed consent before participation. The study had been approved by the
local institutional ethical review boards in Tu¨bingen and Essen.
Virtual-reality setup. We recorded observers’ movements using a Vi-
con MX motion capture system with six cameras at a sampling frequency
of 120 Hz. Six passively reflecting markers were attached to the partici-
pant’s right arm and left shoulder with double-sided adhesive tape. These
markers corresponded to the major joints and the centers of the adjacent
limbs (upper and lower arm). Commercial software was used to recon-
struct and label these markers with spatial 3D reconstruction errors ⬍1.5
mm. We developed custom software to access the marker positions in
real time and to generate the visual stimulus with a closed-loop delay of
⬃32 ms (for an illustration of the setup, see Fig. 1A). Thus, the visual
stimuli the participants observed were actually generated by their own
arm movements.
Visual stimuli. We presented waving arms as point light stimuli (Jo-
hansson, 1973) consisting of five black signal dots on a gray background.
These stimuli were embedded in a camouflaging noise mask. The noise
dots were created by “scrambling” (Cutting et al., 1988). The trajectory of
each noise dot was derived from a subset of previous waving movements,
presented at a randomized average position and start phase within the
movement cycle. Hence, the scrambling destroys the global structure of
the point light movements but preserves average local motion energy.
Two-thirds of the trials were signal trials that contained the point light
arm and the noise mask; in one-third of the trials, the target signal rep-
resented the actual movement of the observer, and one-third of the trials
presented a temporal modification of the movement. In the remaining
trials, only the noise mask was presented, in which the number of noise
dots was increased by five to match the number of dots in trials with and
without a target signal.
We determined the number of noise dots and thus the difficulty level
on an individual basis in a training session before the first test block
started. On average, the number of dots varied between 15 and 130 in
equidistant steps. The noise mask subtended 25.4 ⫻19.2° in visual angle.
The target arm was presented at random positions within the mask with
a size of 7.2 ⫻5.7°.
Experimental procedure. Subjects sat in front of a projection screen that
was placed 1.5 m ahead of them (Fig. 1A). Each experiment started with
a short training period during which the subjects familiarized themselves
with point light stimuli by watching a waving arm, which was based on
recorded movements from previous trials. Further, the maximum num-
ber of noise dots for the final experiment was assessed individually for
each participant using a staircase-like procedure. Subsequently, partici-
pants practiced to wave their arms with a frequency of ⬃30 full waving
cycles per minute using waving point light arms in the desired frequency
as visual feedback. A single waving movement lasted for ⬃2 s, approxi-
mately matching the presentation time of the visual stimuli. The desired
pacing throughout the experiment was later confirmed by kinematic
analysis.
To investigate the influence of action execution on visual-perception
performance, participants were asked to perform the same biological
motion detection task under two conditions: (1) without concurrent
motor activity (baseline condition) and (2) with concurrent motor ac-
tivity (test condition). Using real-time motion capture, we were able to
replay the participants’ own movements online as point light stimulus.
The visual stimulus in the test condition corresponded either to the
actual executed waving arm movement of the participant (synchronous)
or was a temporally altered version of it (asynchronous). In the asynchro-
nous condition, the visual stimulus was delayed by 700 ms with respect to
the executed movement.
The experiment comprised a sequence of 10 alternating test and base-
line blocks. Each block contained 42 stimuli and lasted for ⬃7 min. In the
Table 1. Patient information
a
Patient
no.
Age
(yr)
Time (yr)
since
surgery
ICARS
P
(maximum
34)
K
(maximum
52)
S
(maximum
6)
O
(maximum
8)
Total
(maximum
100)
CP 1 28 3 0 0 0 0 0
CP 2 22 5 1 0 0 1 2
CP 3 32 6 0 0 0 0 0
CP42117142629
CP 5 26 10 3 1 0 0 4
CP 6 21 12 3 2 0 1 6
CP 7 61 11 1 0 0 0 1
CP 8 18 13 3 4 0 0 7
CP 9 44 6 3 5 1 0 8
CP 10 36 5 0 0 0 0 0
CP 11 20 4 4 5 0 3 12
CP 12 23 14 4 5 0 0 9
CP 13 39 3 0 1 0 0 1
CP 14 23 8 0 0 0 0 0
CP 15 24 14 3 7 0 4 14
CP 16 19 16 3 5 0 6 14
CP 17 25 16 2 2 0 0 4
a
Clinical scores were rated using the ICARS score (Trouillas et al., 1997). The table lists the total ICARS scores and the
subscores for gait and posture (P), limb kinetics (K), speech (S), and oculomotor functions (O). Higher scores indicate
more severe ataxia. Maximum scores are given in parentheses. Clinical ataxia symptoms, as scored by the ICARS,
were not correlated with either the APCIdx or the baseline NTV. More detailed analyses of the relevant subscores of
the ICARS for kinetic disturbances also revealed no significant correlation (all Kendall’s tau ⬍0.3, all p⬎0.2 for the
ICARS score and subscores).
6708 •J. Neurosci., May 7, 2014 •34(19):6707– 6716 Christensen et al. •Action-Perception Coupling and the Cerebellum

test blocks, subjects simultaneously waved their right arm with the
trained frequency while observing the visual stimuli. Each stimulus lasted
for 2 s while the participants continuously moved their arms. Arm move-
ments of the target trials were recorded for kinematic analysis and for the
generation of the stimuli in the baseline blocks.
In the baseline blocks, subjects observed the same stimuli, this time
without executing any concurrent motor behavior.
The instruction for the detection task was the same during all blocks.
Participants had to detect any waving point light arm, regardless of
whether they thought the displayed limb to be associated with their own
motion or not. The participants verbally reported whether they had ob-
served any target arm movement in the stimulus or not, and the consec-
utive trial started immediately after the experimenter had entered the
response.
Eye movements were recorded using a Tobii ⫻120 Hz eyetracker to
ensure that no oculomotor deficits impaired participants in the action
observation.
Assessment of detection performance. Detection performance was as-
sessed applying signal detection theory individually for each subject, ev-
ery test, and corresponding baseline conditions. Each test and baseline
condition was assigned one noise tolerance value (NTV), indicating the
maximum number of noise dots in the camouflaging mask that would
lead to 75% of the optimal detection sensitivity. To calculate the NTV, in
a first step, we analyzed the hit and false alarm rates for the seven tested
noise levels individually to determine one corresponding d⬘value for
each noise level. For every condition (test or baseline ⫻synchronous or
asynchronous) the 7 d⬘values were fitted with a logistic function f. The
NTV was then defined as the number of noise dots that fulfilled the
equation: f(NTV) ⫽0.75 * max_d_prime, with max_d_prime ⫽4.65
(Macmillan and Creelman, 2004).
To quantify the direction and strength of the influence of motor exe-
cution on biological motion (BM) perception, we introduced as Interac-
tion index (IntIdx, see Table 2), the logarithm of the ratio of the NTVs for
the testing and the baseline condition. This index reflects the relationship
between noise tolerances in both conditions in a symmetric way. If the
NTV is bigger in the test condition than in the baseline condition, the
IntIdx is positive, indicating a facilitation of perception by concurrent
motor behavior. The IntIdx is zero in cases where both performances are
equal, and it is negative in cases of an inhibition
of perception due to the motor execution.
More importantly, if the NTV for the test con-
dition is twice as big as in the baseline condi-
tion, the resulting index has the same absolute
value as if the performance in the baseline con-
dition is twice as good as in the test condition.
Reflecting the influence of motor execution
on perception the IntIdx is susceptible to dual
task effects that might play a role when investi-
gating the behavior of cerebellar patients.
However, this dual task influence would cause
lower IntIdxs in both the synchronous and the
asynchronous test condition.
To quantify the overall action perception
coupling (APC) pattern in one single value that
(1) cancels dual task effects and (2) can be used
for the lesion symptom mapping, we combined
the single IntIdxs for the synchronous and the
asynchronous condition to one APC index
(APCIdx;Table 2). The APCIdx directly reveals
normal (positive) compared with disturbed
(negative) APC.
Kinematic analysis. Movement analysis was
based on the trajectories of the displayed mark-
ers of the target arm. For each participant, the
kinematic data of 60 randomly chosen individ-
ual trials were evaluated. We investigated
movement amplitude, velocity, and jerk. The
variability of each parameter was computed as
SD from the mean of those 60 trials. Movement
amplitude within one waving cycle was defined
as the maximum traveling distance of the hand as projected on the 2D
plane in the visual stimulus in degrees. Velocity was determined based on
the trajectories of the hand marker and the kinematic analysis confirmed
the desired pacing of the participant’s hand movements (average velocity
of 145.19 ⫾36.93 degrees per second). The third temporal derivative of
the hand-marker displacement served as a measure of jerk (Goldvasser et
al., 2001).
Lesion symptom mapping. Anatomical magnetic resonance (MR) im-
ages of all patients were acquired with a 1.5 Tesla Siemens scanner using
a 12-channel head coil (Siemens). A 3D sagittal volume of the entire
brain was obtained using a T1-weighted, MPRAGE sequence with a rep-
etition time of 2400 ms, 3.63 ms echo time, 160 slices, a field of view of
256 mm, and 1.0 ⫻1.0 ⫻1.0 mm
3
voxel size. All images were examined
by an experienced neuroradiologist, and extracerebellar pathology was
excluded. We manually traced cerebellar lesions on axial, sagittal, and
coronal slices of the non-normalized 3D MRI dataset and saved them as
regions of interest using the free MRIcro software (http://www.
mccauslandcenter.sc.edu/mricro/). All patients had chronic surgical le-
sions. Surgical lesions are clearly visible as dark (i.e., intensity of CSF)
areas on T1-weighted MR images. “Grayscale” default thresholds were
used in MRIcro. To investigate functional differences of both cerebellar
hemispheres, all regions of interest were defined as hemisphere-specific
and not flipped to only one hemisphere. Images were normalized using a
spatially unbiased infratentorial template of the cerebellum (SUIT; www.
icn.ucl.ac.uk/motorcontrol/imaging/suit.htm (Diedrichsen (2006)) with
the SUIT toolbox in SPM5 (http://www.fil.ion.ucl.ac.uk/spm/software/
spm5). Manual corrections, in case parts of the occipital cortex were in-
cluded in the masks resulting from the automatic segmentation, were
done with the help of CARET software (http://brainvis.wustl.edu/wiki/
index.php/Caret:About). We used the probabilistic atlas of the human
cerebellum (http://www.icn.ucl.ac.uk/motorcontrol/imaging/propatlas.
htm) (Diedrichsen et al., 2009) in MRIcron (http://www.mccauslandcenter.
sc.edu/mricro/mricron/) to define the affected lobules. Nomenclature of
cerebellar lobules was used, which has been introduced by Schmahmann
et al. (1999) and is closely related to Larsell’s nomenclature (Larsell,
1958). Affected cerebellar nuclei were defined with a newly developed
Figure 1. Behavioral setup and results. A, Experimental setup. The participant sat on a chair observing a point light stimulus
projected on a screen 1.5 m ahead of them while waving their right arm. Arm movements were recorded using an infrared-based
optical motion capture system with 6 cameras (3 are shown in the figure) for (1) online generation of the point light stimulus, (2)
replayof the same arm movementsin the baseline blocks,and (3) posthoc kinematic analysis.The target stimulusis highlighted for
illustration in green. Trajectories are indicated as lighter green traces. Noise dots were generated using spatiotemporal scrambling
and moved along arm and hand trajectories (indicated as arrows) as well. Gaze movements were recorded using a Tobii ⫻120 Hz
eyetracker to ensure that performance deficits on the action-observation task were not the result of oculomotor deficits. B,
Behavioral results of the action-perception coupling experiment. Bars represent the mean interaction indexes for healthy controls
(green) and patients (yellow). Controls benefited from their self-generated movements if the visual stimulus was displayed in
synchrony(⬃40 ms) with their executedmotion (green shadedarea). Introduction ofan artificial delay(⬃700 ms) ledto negative
interaction indices, indicating an inhibitory effect of self-motion on biological motion perception (pink shaded area). The patients
showed no such pattern of intact action-perception coupling. Error bars indicate SE. *p⬍0.05 (pairwise differences, ttest).
Christensen et al. •Action-Perception Coupling and the Cerebellum J. Neurosci., May 7, 2014 •34(19):6707– 6716 • 6709

probabilistic atlas of cerebellar nuclei (Diedrichsen et al., 2009,2011).
Figure 2 illustrates superposition of individual lesions for all patients.
To identify lesion areas associated with performance deficits, we used
voxel-based lesion symptom mapping with the Liebermeister test to test
for statistical significance at a threshold of p⬍0.05 using nonparametric
mapping software as part of MRICron (http://www.mccauslandcenter.
sc.edu/mricro/npm/) (Rorden et al., 2007). Only voxels damaged in at
least 10% of individuals (n⫽2) were considered. Strength of the associ-
ation is color-coded within each figure showing the lesion sides (see Figs.
5and 7). Brighter colors indicate higher z-values and thus stronger asso-
ciations of lesion sites and the behavioral measure.
In addition, subtraction analysis was performed in MRIcron. For each
lesioned voxel, the percentage of unaffected patients with a lesion in that
voxel was subtracted from the percentage of affected patients with a
lesion in that voxel (Karnath et al., 2002;Thieme et al., 2013). Affected
and unaffected patients were defined as outlined below. Voxels that were
at least 25% more likely to be lesioned in impaired patients were
considered.
Statistical analysis. Statistical testing was performed using SPSS
(http://www-01.ibm.com/software/de/stats20/). Statistical significance
is reported with respect to the
␣
⫽5% significance level. Behavioral
differences between conditions were analyzed using a repeated-measures
ANOVA with the within-subject factor temporal modification for con-
trol subjects and patients individually. To separate unaffected and af-
fected patients, we conducted a cluster analysis based on the k-means
algorithm with initial cluster centers c1 ⫽0.15 and c2 ⫽⫺0.15 on the
APC indices. To validate our subject classification, we ran an additional
repeated-measures ANOVA on the d⬘value with the within-subject fac-
tors noise level, temporal modification, experimental condition, and the
between-subject factor classified group.
We tested for the effect of subject group on the NTV, the eye move-
ment parameters, and the kinematic parameters applying the nonpara-
metric Kruskal–Wallis test. Paired comparisons between groups were
accomplished using the nonparametric Mann–Whitney Utest.
We further tested for associations between the APCIdx and various
other measures as the baseline perception performance, the ICARS score
and its subscores, the lesion volume, and the kinematic parameters cal-
culating the Kendall rank correlation coefficient.
Figure 2. Regions of cerebellar lesions in all patients overlapped in a single image. The
number of patients presenting lesions in the specific areas is color-coded. D, Dentate; I,
interposed.
Figure3. Classification results. Ellipses represent clustersresulting fromcluster-center anal-
ysis based on the APCIdx. Red ellipse represents the cluster of affected patients; blue ellipse
represents the cluster of unaffected patients. Filled symbols represent individual APCIdxs; open
symbols represent cluster centers (final cluster centers c1 ⫽0.13, c2 ⫽⫺0.2). Noteworthy, a
separation of affected and unaffected patients based on the performance of the healthy control
subjects led to the same classification. Range of healthy controls from 25 lower to 75 upper
percentile outlined as green box.
Table 2. Overview of different behavioral indices used to describe perception performance and to quantify APC
Abbreviation Formula Description Properties
NTV NTV ⫽arg( f⫽0.75 * max_d_prime) Number of noise dots in the mask with preserved biological
motion detection.
Higher NTV reflects better performance.
No relation between test and baseline condition (4 values
per participant).
IntIdx IntIdx ⫽ln
冉
NTVtest
NTVbaseline
冊
Quantifies the relationship between test and baseline
condition.
No influence of motor execution on perception: IntIdx ⫽0
Susceptible to dual task effects (two values per participant
and patient).
Facilitatory influence of motor execution on perception: IntIdx ⬎0
Inhibitory influence of motor execution on perception: IntIdx ⬍0
APCIdx APCIdx ⫽IntIdx
synchronous
⫺IntIdx
asynchronous
Unifies the IntIdx for both delay conditions resulting in one
value to quantify the APC that is needed for the lesion
symptom mapping.
Unaffected APC: APCIdx ⱖ0
Unaffected by dual task effects (one value per participant
and patient).
Affected APC: APCIdx ⬍0
6710 •J. Neurosci., May 7, 2014 •34(19):6707– 6716 Christensen et al. •Action-Perception Coupling and the Cerebellum

Results
Behavioral performance
All participants were well capable of performing the BM detec-
tion task in the baseline condition. The detection sensitivity in the
baseline condition, as measured by mean d⬘values, did not differ
between healthy controls and patients (controls: 3.20 ⫾0.6; pa-
tients: 3.28 ⫾0.4; p⫽0.708, Mann–Whitney Utest).
The interaction indices reflecting the influence of action exe-
cution on visual perception performance for both subject groups
are depicted in Figure 1B. Concurrent motor execution signifi-
cantly altered the visual perception performance of the controls
depending on the temporal matching of executed and observed
action: Controls benefitted from their self-generated motion if
the visual stimulus was displayed in synchrony; but if the shown
arm movement was delayed with respect to their own move-
ments, performance in the perceptual task dropped (Fig. 1B;
repeated-measures ANOVA: main effect temporal modification,
F
(1,16)
⫽5.842, p⫽0.028).
This pattern of the action perception coupling, described by a
substantial benefit from self-generated motion in the synchro-
nous condition and an inhibition of perception through motor
activity in the asynchronous condition, is reflected in a positive
APCIdx (Table 2).
Cerebellar patients as whole group showed no clear APC pat-
tern: neither a benefit nor an inhibition (i.e., the interaction in-
dices did not differ significantly from zero) (Fig. 1B;F
(1,16)
⫽
0.093, p⫽0.765).
Using k-means cluster analysis on the APCIdx, we defined two
clusters within the whole patient group. Eleven patients belong-
ing to the first cluster showed an APC pattern comparable with
the healthy controls, with a positive APCIdx. This group was
defined as the unaffected patients. The second cluster includes 6
patients with an abnormal APC (APCIdx ⬍0); this group was
defined the affected patients (Fig. 3).
We verified the validity of this cate-
gorization by analyses on the basic mea-
surement of detection sensitivity, the d⬘
value. Whereas healthy controls and un-
affected patients show a significant in-
crease in sensitivity for the synchronous
test condition compared with the asyn-
chronous test condition, affected pa-
tients do not show such a difference
(Fig. 4). Further, a repeated-measure
ANOVA revealed an interaction effect
of the factors temporal modification,
experimental condition, and classified
group with an estimated effect size of
ŋ
2
⫽0.107 (F
(1,32)
⫽3.84, p⫽0.059).
As basis for the lesion symptom map-
ping, we used the APCIdx as one compact
description of the behavioral effect.
To ensure that the APC results were
not confounded by a general detection
deficit in the affected patients, we tested
for group differences in the baseline
performance. Nonparametric significance
testing failed to detect a group difference
in the baseline NTV between the affected
patients, unaffected patients, and controls
(Kruskal–Wallis test, p⫽0.112; all paired
comparisons using the Mann–Whitney U
test revealed no significant differences).
However, given the small group size of only 6 affected patients, a
difference between patients and controls cannot be completely
ruled out with a significance test yielding a pvalue of 0.112.
Nevertheless, the absence of a significant correlation between
baseline performance and the APCIdx for patients (Kendall’s
tau ⫽0.224, p⫽0.215) makes it rather unlikely that a low APC is
predominantly the result of a general detection deficit.
In summary, the behavioral results for the controls confirm
our previous results (Christensen et al., 2011), revealing in
healthy individuals a modulation of BM perception by concur-
rent motor behavior: compared with baseline, visual detection
was facilitated by synchronous motor behavior and inhibited by
asynchronous motor behavior. In cerebellar patients, two sub-
groups were defined. One group of patients was unaffected,
showing an APC pattern comparable with the controls, whereas
the group of the affected patients showed an inverse pattern of
results.
Notably, the patients defined as affected with a diminished
APC showed a normal BM detection in the baseline condition.
We further ruled out possible confounds because of the severity
of clinical ataxia symptoms.
Controlling for the specificity of the APC deficit
To confirm the specificity of the identified APC deficit, we con-
ducted several control experiments and analyses. Because all vi-
sual stimuli were created online based on the individual arm
movements of the participants, the individual stimuli could have
differed in their appearance and detection complexity. To rule
out that these differences might have confounded the results con-
cerning the APC performance, we tested post hoc in a separate
experiment for differences in the detectability of arm movements
of control subjects and patients. Therefore, we conducted an ad-
ditional purely visual perception experiment to control for dif-
ferences in the detectability of stimuli created from arm
Figure 4. Mean d⬘values for the test conditions showing the biological motion stimuli synchronous or asynchronous to the
executed arm movements. Controls as well as unaffected patients show an increased sensitivity for synchronous stimuli compared
with asynchronous stimuli (within-subject comparison). Affected patients do not show such sensitivity difference between con-
ditions. *p⬍0.05 (pairwise differences, ttest). ***p⬍0.01 (pairwise differences, ttest). Error bars indicate SE.
Christensen et al. •Action-Perception Coupling and the Cerebellum J. Neurosci., May 7, 2014 •34(19):6707– 6716 • 6711

movements of affected patients compared with stimuli generated
from movements of control subjects. The task followed the same
procedure as the baseline condition in the main experiment.
Twenty new naive subjects took part in this experiment. Com-
paring the detection performance in terms of percentage correct
detection of these new participants, we found no differences for
stimuli from controls and stimuli from affected patients (Wil-
coxon rank sum test, p⫽0.372).
Further, it could be possible that a disturbed eye movement
pattern might cause deficits in the visual detection performance.
We failed to observe any correlation between the oculomotor
subscore of the ICARS and a disturbed APC (Kendall’s tau ⬍0.3,
all p⬎0.2 for the oculomotor ICARS subscore). This lack of a
correlation between clinical scores for eye movements and the
APCIdx was further supported by investigations of the fixation
behavior of the participants. We analyzed the following: (1) the
number of saccades per nontarget baseline trials, (2) the mean
saccadic amplitude, and (3) the mean saccade duration as a mea-
sure of the eye movement behavior (Lackner and Mather, 1981).
Group comparisons revealed no differences in these parameters
and in the variability as measured by the SD of these parameters
between controls, affected, and unaffected patients (Kruskal–
Wallis test, all p⬎0.4). Moreover, the only 2 patients that showed
clinical oculomotor disturbances (deficits in smooth pursuit
movements and nystagmus, respectively) belonged to the unaf-
fected patient group. Thus, a disturbed eye movement pattern
could not explain the deficit in the APC.
These results underline the specificity of the APC deficit.
Lesion symptom mapping of the behavioral results
Lesion symptom mapping (LSM) revealed significant associa-
tions of specific cerebellar areas with a deficit in the APC. Areas
were the same based on descriptive subtraction analysis and sta-
tistical Liebermeister test. Lesions associated with an impaired
APC are shown in Figure 5A,B. The affected patients show lesions
in the right dentate nucleus (ipsilateral to the moved arm). The
maximum z-values for dentate lesions associated with the
APCIdx are located ventrally, but lesions extended into the dorsal
part of the dentate nucleus. In addition, the motor representa-
tions in lobule V and VI are correlated with a deficit in the APC.
Further, lesions in the right posterior lobules VIIIa, VIIIb, and IX
are significantly more likely in patients with a low APCIdx.
Notably, there was also a relationship between lesions of the
left posterior cerebellum and the APCIdx. Specifically, lesions of
the left Crus II and the left lobule VIIb were significantly related
to the group of affected patients compared with the unaffected
patients. Comparisons of the lesion volume showed no differ-
ences between affected and unaffected patients (p⫽0.43). Fur-
ther, there is no significant correlation between the lesion volume
and the APCIdx (Kendall’s tau: ⫺0.206, p⫽0.249).
Statistical testing for lesions associated with a general visual
detection performance as reflected by a baseline NTV below the
25th percentile of the control subjects revealed no cerebellar areas
that are significantly correlated with a perceptual deficit at a
threshold level of p⬍0.05.
Figure 5. Lesion symptom mapping in patients after benign tumor resection. Arabic num-
bers indicate ycoordinates. Latin numbers indicate cerebellar lobules. A, Colored regions rep-
resent lesions that are significantly correlated with a disturbed action-perception coupling
(p⬍0.05) according to the Liebermeister test of significance. Strength of the association is
color-coded. Brighter colors represent higher z-values. White line indicates a z-value of 1.65
(p⬍0.05). Lesions associated with a disturbed APC were identified in the ipsilateral motor
representations(lobules V and VI),the ventral dentate(z⫽1.79, x⫽18,y⫽⫺60,z⫽⫺43)
with extensions into the dorsal part of the dentate nucleus (z⫽1.79, x⫽18, y⫽⫺56, z⫽
⫺32),and in the leftCrus IIand lobuleVIIb. B,Subtraction analysisidentified thesame regions.
The right dentate, the ipsilateral motor representation, and the left Crus II are 33% more likely
lesioned in affected compared with unaffected patients. D, Dentate; I, interposed.
Figure 6. Variability of kinematic parameters for the different subject groups. *Significant
pairwise differences ( p⬍0.05, Mann–Whitney Utest). A, Variability of movement amplitude.
B, Variability of movement velocity.
6712 •J. Neurosci., May 7, 2014 •34(19):6707– 6716 Christensen et al. •Action-Perception Coupling and the Cerebellum

Motor behavior
To investigate the potential influence of movement disturbances
on the APC, we analyzed various kinematic parameters (e.g.,
movement amplitude, velocity, and jerk) as well as the variability
of those.
Because all participants were trained to wave their arms in the
same velocity, we did not find any differences in movement ve-
locity between participants. We found nonsignificant group dif-
ferences for the variability of the movement amplitude and
velocity (Kruskal–Wallis test, p⫽0.062 and p⫽0.06, respec-
tively; Fig. 6). In contrast, group comparisons between controls
and affected patients revealed significant differences in the vari-
ability of movement amplitude and velocity (Mann–Whitney U
test, p⫽0.044 and p⫽0.044, respectively), and comparisons
between affected and unaffected patients revealed a significant
difference in the variability of movement amplitudes (Mann–
Whitney Utest, p⫽0.037). This increased movement variability
indicates a deficit in performing rhythmic waving movements
with constant velocity and amplitude.
Disturbances of motor representations that cause such motor
deficits potentially also cancel beneficial influences on the APC.
However, we did not observe a clear correlation between the
variability in velocity and the APCIdx (Kendall’s tau ⫽⫺0.162,
p⫽0.365). The correlation between the variability of the move-
ment amplitude and the APCIdx reached a Kendall’s tau coeffi-
cient of ⫺0.294 but failed to reach significance ( p⫽0.099). These
results indicate that these motor representations are potentially
involved in the APC but cannot completely explain the observed
effects.
Lesion symptom mapping of motor behavior
Subtraction analysis and Liebermeister test revealed the same
areas associated with an increase of the variability in arm velocity
during execution of cyclic waving move-
ments (Fig. 7A,B). Lesions of the right
dentate nucleus, especially the dorsal part,
with extensions into the right interposed
nucleus, and with some extensions to the
right ventral dentate as well as lesions
within the motor representations of the
hand and arm in lobules V and VI were
significantly more likely in patients with
increased velocity variability. Further,
small parts of the vermal lobules VIIIA,
VIIIB, and IX, extending into right lobule
IX, showed also a significant relationship
with the increased velocity variability.
Notably, both a perturbed motor be-
havior and a disturbed APC are associated
with lesions in cerebellar motor represen-
tations (e.g., the interposed nucleus, dor-
sal dentate nucleus, and lobules V/VI)
(Fig. 7C). In addition, an intact APC de-
pends on the integrity of the more ventral
part of the dentate nucleus and the poste-
rior cerebellum, most prominently the left
Crus II, and left lobule VIIb (Fig. 7C).
Discussion
Using a novel experimental framework that
allowed us to quantify the effect of concur-
rent motor execution on BM perception, we
found specific deficits in the APC in patients
with focal cerebellar lesions.
We identified circumscribed areas within the cerebellum that
are causally involved in the APC on two different levels:
(1) On the level of a somatosensory state representation (ip-
silateral lobules V/VI) presumably within the cerebrocerebellar
motor circuits. Our results support the hypothesis that an intact
somatosensory state estimate of own movements is not only es-
sential for control but also facilitates the detection of synchro-
nous BM stimuli as we observed in controls.
(2) On the level of multisensory integration (lobule VII, ven-
tral dentate), including higher-order visual perception. Hence,
the posterior regions of the cerebellum are likely crucial parts
within the cerebellar–parietal loop for the integration of sensory
information from different modalities, potentially establishing a
corresponding multisensory state estimate.
Cerebellar involvement in an intact APC
Cerebellar cortex
On the level of the cerebellar cortex, an APC deficit was strongly
associated with lesions of the motor representation of the arm in
the right anterior and posterior cerebellar cortex (lobules V/VI, and
VIII). An increased variability in the motor behavior was associated
with lesions in lobule V/VI. Although both deficits, the reduced APC
as well as the increased motor variability, were associated with le-
sions in lobules V/VI, the specific regions are distinct from each
another (Fig. 7C). Lesions of patients with an APC deficit extended
more into the posterior part of lobule VI. These results are in line
with the hypothesis that lobule VI encodes sensory prediction
(Schlerf et al., 2012) that potentially are used to calculate and update
a somatomotor state estimate (Miall et al., 2007).
Remarkably, lesions of the left posterior cerebellum (Crus II,
lobule VIIb), contralateral to the arm movement, were associated
with a perturbed APC. Whereas the skeletomotor divisions of the
Figure 7. Lesion symptom mapping of motor behavior. A, Voxel-based lesion symptom mapping analysis shows regions with
a significant correlation with increased velocity variability ( p⬍0.05) according to the Liebermeister test of significance (strength
of association color-coded). Brighter colors represent higher z-values. The white line indicates a z-value of 1.65 ( p⬍0.05). Arabic
numbers indicate ycoordinates. Latin numbers indicate cerebellar lobules. B, Subtraction analysis. The right dentate and inter-
posednucleus as well as the lobulesV and VI are38% more likely,and the right lobuleIX is 50% morelikely lesioned inpatients with
increased velocity variability. C, Lesion symptom map overlay plot of lesions/regions specifically associated with APC deficits (in
red) and lesion sites that are strongly correlated with increased velocity variability (blue).
Christensen et al. •Action-Perception Coupling and the Cerebellum J. Neurosci., May 7, 2014 •34(19):6707– 6716 • 6713

cerebellum in the anterior lobe and the
intermediate zone of the posterior lobe
are related to ipsilateral movements
(Glickstein et al., 2011), data from electro-
physiology studies revealed, especially in
the more lateral posterior cerebellum
(Crus II), effector-side independent neu-
ral activations (Greger et al., 2004).
Dentate nucleus
The observed deficits in the APC as well as
the motor disturbances were both most
prominently associated with lesions in the
right dentate nucleus. A disturbed APC
was correlated with lesions of the ventral
dentate nucleus, whereas a perturbed mo-
tor behavior was strongly associated with
the dorsal dentate nucleus with extensions
into the ventral part and into the right in-
terposed nucleus.
Our results provide evidence for a
dorsal-to-ventral transition from simple
motor to higher-order motor and integra-
tive functions within the human dentate
as proposed by high-resolution imaging
studies (Ku¨per et al., 2011;Bernard et al.,
2013). These findings contribute to the
ongoing debate of the functional division
of the dentate into a motor (dorsal) and a
nonmotor (ventral) part (Strick et al.,
2009). They support recent hypotheses
that such a strict subdivision is an over-
simplification and that complex motor
function involves the ventral dentate.
Corticocerebellar networks
Figure 8 illustrates the corticocerebellar
connections as an overlay of the cerebellar
lesions from our study and the network
parceling from Buckner et al. (2011),
identified by resting-state analysis. Our
identified lesions in the cerebellar lobules
and the deep cerebellar nuclei that caused
an increased motor variability are connected to primary motor
networks in the cerebral cortex (M1).
Cerebellar areas associated with a disturbed APC are function-
ally connected to networks in frontal (inferior frontal gyrus, pre-
motor cortex), temporal (superior temporal sulcus), and parietal
areas (supramarginal gyrus, posterior parietal cortex [PPC]) in
the cerebral cortex (O’Reilly et al., 2010;Buckner et al., 2011;
Sokolov et al., 2012). These networks have been suggested to be
associated with sensorimotor integration (Yeo et al., 2011), ac-
tion observation (Gazzola and Keysers, 2009), updating one’s
own body image (Blakemore and Sirigu, 2003), and timed sen-
sory predictions (Bastian, 2006). Especially the PPC is known as a
key structure to integrate sensory cues, such as proprioceptive
and visuospatial information (Buneo and Andersen, 2006;Stein
and Stanford, 2008). A previous fMRI study argued in favor of
coactivations of the left posterior cerebellum and the right PPC
during the integration of visual and kinesthetic information
(Hagura et al., 2009). With the present lesion study, we were able
to show even a causal involvement of the left Crus II in these
integrative processes.
Summarizing our findings support the theory that the net-
work, including the cerebellum and the PPC, establishes internal
forward models to compute state estimates (Wolpert et al., 1998;
Miall, 2003;Bastian, 2006;Buneo and Andersen, 2006). We show
that disruptions of such state estimates lead to (1) larger variabil-
ity in motor behavior (Miall et al., 2007) and (2) disturbances of
the benefit from interactions of self-executed movements in BM
perception, as has been observed for healthy subjects (Miall et al.,
2006;Christensen et al., 2011).
Figure 8Cshows a hypothetical corticocerebellar loop for an
intact APC. The motor representation in lobule V/VI of the ipsi-
lateral cerebellum receives an efference copy of the motor com-
mand, sent out from cortical motor areas (Ebner and Pasalar,
2008). This efference copy, together with the proprioceptive feed-
back from the ongoing movements (Fuchs and Kornhuber, 1969;
Bloedel and Courville, 1981), and the somatosensory informa-
tion from the cortex are used to calculate a current somatosen-
sory state estimate and to predict the sensory consequences of the
motor behavior. The lobules V and VI transmit the information
to the ipsilateral dorsal dentate nucleus, which itself also receives
cerebral input via the pontine nuclei (Shinoda et al., 1992). The
Figure 8. Functional connectivity map. A, Overlay of lesion sides on the map of the human cerebellum based on the functional
connectivity of the cerebral cortex. Colors encode the 7 major networks as identified by Yeo et al. (2011). White stars represent
major lesions associated with a disturbed action-perception pattern plotted on the illustration of the 7 network parceling from
Buckneret al. (2011). White circles representlesion sides correlated withperturbed motor behavior. B,Cortical networks asdefined
by Yeo et al. (2011). [Reproduced and adapted with permission from the American Physiological Society.] C, Our proposed corti-
cocerebellar loop involved in intact action-perception coupling. Pathways between the cerebral cortex and the cerebellum are
routed via the pontine nuclei. These include pathways from the motor cortex, somatosensory cortex, BA5, PPC, and pathways from
the visual cortical areas (Glickstein et al., 1985). Motor and somatosensory information reaches lobules V and VI (Kelly and Strick,
2003). Visual information reaches lobules IXA and VIIIB (dorsal paraflocculus) but also targets Crus I and II via the dorsolateral parts
of the pontine nuclei (Glickstein et al., 1994). A similar route also exists for cortical information from the PPC (Leichnetz, 2001). The
cerebellar output from the anterior interposed and dorsal dentate reaches motor cortex (Kelly and Strick, 2003), whereas the
output from the posterior interposed and ventral dentate reaches the PPC (Prevosto et al., 2010). Commissural fibers connect
the PPC of both hemispheres (Van Essen et al., 1982). Corticocortical pathways also connect motor /somatosensory cortex and PPC
within the same hemisphere (Gharbawie et al., 2011).
6714 •J. Neurosci., May 7, 2014 •34(19):6707– 6716 Christensen et al. •Action-Perception Coupling and the Cerebellum

contralateral ventral dentate receives input from the right parietal
cortex via lobule VII (May and Andersen, 1986;Glickstein et al.,
1994,2011) and projects back to the PPC (Strick et al., 2009). In
addition to the well-known contralateral cerebrocerebellar con-
nections, recent electrophysiological experiments give evidence
for bilateral connections, especially of higher-order cortical areas
with the interposed nucleus and fastigial nucleus (Sultan et al.,
2012). These connections probably arise from the common bilat-
eral mossy fiber inputs to deep cerebellar nuclei via precerebellar
nuclei, such as the nucleus reticularis tegmentis pontis (Serapide
et al., 2002) and the lateral reticular nucleus (Wu et al., 1999).
Thus, one might speculate also that the cerebellar–parietal con-
nections involving brainstem nuclei are partly bilateral (Prevosto
et al., 2010). However, the dentate nucleus appears to have
smaller portions of bilateral mossy fiber inputs (Shinoda et al.,
1992;Serapide et al., 2002). Therefore, the more likely site of
integration would be via callosal fibers connecting the right and
left PPC (Van Essen et al., 1982). This would also explain the
additional involvement of the left PPC in temporally critical tasks
similar to the APC (Wiener et al., 2010;Vicario et al., 2013).
According to this network model, the integrity of the right
lobules V, VI, the left and right dentate nucleus, and the left Crus
I and II is of vital importance for an intact APC. Lesion symptom
mapping results revealed significant associations with a disturbed
APC for all of these areas but the left dentate nucleus. As a first
hint for its involvement, one affected patient presented lesions in
the left dentate while the right dentate and the left Crus II re-
mained unlesioned. Still, no final conclusion can be drawn on the
level of the left dentate nucleus because of the limitation of the
LSM approach that we can only make valuable statements about
areas with sufficient amount of lesions. Therefore, we cannot
exclude the importance of areas for which no patient presented
lesions. However, despite our small number of affected patients
and the multifocal and bilateral approach of our analysis, we find
robust and clear positive effects: For all of our identified regions
associated with the APC, at least 2 of the 6 affected patients and
more importantly, none of the unaffected patients presents a
lesion. Further, as a negative control, multiple areas are lesioned
in several patients but are not associated with the APC (e.g., left
lobules IV/V, right interposed, right lobule IV).
In conclusion, we showed a causal involvement of the cerebel-
lum in an intact APC on two different levels. First, lesions of the
right motor representation in lobules V/VI (together with the
dorsal dentate and adjacent interposed nucleus) likely cause a
disrupted somatosensory state estimate of ongoing movements.
Thus, patients with lesions in these areas show not only greater
movement variability but also a disturbed APC. Second, lesions
in the ventral dentate and in the left posterior cerebellum (Crus
II) are further disrupting the important connections between the
posterior cerebellum and the PPC. This hinders the sensorimotor
integration of higher-order visual motion perception and motor
information and thus causes a deficit in the APC. The present
results broaden the perspective that the cerebellum is merely in-
volved in perceptual processes concerning somatosensory as-
pects toward a more general view that the cerebellum integrates
predictions about different kinds of perceptual consequences of
actions, including predictions about vision and thus contribute
to higher-order cognition.
References
Bastian AJ (2006) Learning to predict the future: the cerebellum adapts
feedforward movement control. Curr Opin Neurobiol 16:645–649.
CrossRef Medline
Bastian AJ (2011) Moving, sensing and learning with cerebellar damage.
Curr Opin Neurobiol 21:596– 601. CrossRef Medline
Bernard JA, Peltier SJ, Benson BL, Wiggins JL, Jaeggi SM, Buschkuehl M,
Jonides J, Monk CS, Seidler RD (2013) Dissociable functional networks
of the human dentate nucleus. Cereb Cortex. Advance online publication.
Retrieved March 19, 2013. doi:10.1093/cercor/bht065. CrossRef Medline
Bhanpuri NH, Okamura AM, Bastian AJ (2012) Active force perception
depends on cerebellar function. J Neurophysiol 107:1612–1620. CrossRef
Medline
Blakemore SJ, Sirigu A (2003) Action prediction in the cerebellum and in
the parietal lobe. Exp Brain Res 153:239–245. CrossRef Medline
Bloedel JR, Courville J (1981) A review of cerebellar afferent systems. In:
Handbook of physiology, vol. II. Motor control (Brooks VB, ed), pp
735–830. Baltimore: Williams & Wilkins.
Buckner RL, Krienen FM, Castellanos A, Diaz JC, Yeo BT (2011) The orga-
nization of the human cerebellum estimated by intrinsic functional con-
nectivity. J Neurophysiol 106:2322–2345. CrossRef Medline
Buneo CA, Andersen RA (2006) The posterior parietal cortex: sensorimotor
interface for the planning and online control of visually guided move-
ments. Neuropsychologia 44:2594–2606. CrossRef Medline
Casile A, Giese MA (2006) Nonvisual motor training influences biological
motion perception. Curr Biol 16:69–74. CrossRef Medline
Christensen A, Ilg W, Giese MA (2011) Spatiotemporal tuning of the facil-
itation of biological motion perception by concurrent motor execution.
J Neurosci 31:3493–3499. CrossRef Medline
Cutting JE, Moore C, Morrison R (1988) Masking the motions of human
gait. Percept Psychophys 44:339–347. CrossRef Medline
Decety J, Gre`zes J, Costes N, Perani D, Jeannerod M, Procyk E, Grassi F, Fazio
F (1997) Brain activity during observation of actions. Influence of action
content and subject’s strategy. Brain 120:1763–1777. CrossRef Medline
Diedrichsen J (2006) A spatially unbiased atlas template of the human cer-
ebellum. Neuroimage 33:127–138. CrossRef Medline
Diedrichsen J, Balsters JH, Flavell J, Cussans E, Ramnani N (2009) A prob-
abilistic MR atlas of the human cerebellum. Neuroimage 46:39– 46.
CrossRef Medline
Diedrichsen J, Maderwald S, Ku¨ per M, Thu¨ rling M, Rabe K, Gizewski ER,
Ladd ME, Timmann D (2011) Imaging the deep cerebellar nuclei: a
probabilistic atlas and normalization procedure. Neuroimage 54:1786–
1794. CrossRef Medline
Ebner TJ, Pasalar S (2008) Cerebellum predicts the future motor state. Cer-
ebellum 7:583–588. CrossRef Medline
Fadiga L, Fogassi L, Pavesi G, Rizzolatti G (1995) Motor facilitation during
action observation: a magnetic stimulation study. J Neurophysiol 73:
2608–2611. Medline
Fuchs AF, Kornhuber HH (1969) Extraocular muscle afferents to the cere-
bellum of the cat. J Physiol 200:713–722. Medline
Gallese V, Fadiga L, Fogassi L, Rizzolatti G (1996) Action recognition in the
premotor cortex. Brain 119:593–609. CrossRef Medline
Gao JH, Parsons LM, Bower JM, Xiong J, Li J, Fox PT (1996) Cerebellum
implicated in sensory acquisition and discrimination rather than motor
control. Science 272:545–547. CrossRef Medline
Gazzola V, Keysers C (2009) The observation and execution of actions share
motor and somatosensory voxels in all tested subjects: single-subject anal-
yses of unsmoothed fMRI data. Cereb Cortex 19:1239–1255. CrossRef
Medline
Gharbawie OA, Stepniewska I, Kaas JH (2011) Cortical connections of
functional zones in posterior parietal cortex and frontal cortex motor
regions in New World monkeys. Cereb Cortex 21:1981–2002. CrossRef
Medline
Glickstein M, May JG 3rd, Mercier BE (1985) Corticopontine projection in
the macaque: the distribution of labelled cortical cells after large injections
of horseradish peroxidase in the pontine nuclei. J Comp Neurol 235:343–
359. CrossRef Medline
Glickstein M, Gerrits N, Kralj-Hans I, Mercier B, Stein J, Voogd J (1994)
Visual pontocerebellar projections in the macaque. J Comp Neurol 349:
51–72. CrossRef Medline
Glickstein M, Sultan F, Voogd J (2011) Functional localization in the cere-
bellum. Cortex 47:59– 80. CrossRef Medline
Goldvasser D, McGibbon CA, Krebs DE (2001) High curvature and jerk
analyses of arm ataxia. Biol Cybern 84:85–90. CrossRef Medline
Greger B, Norris SA, Thach WT (2004) Spike firing in the lateral cerebellar
Christensen et al. •Action-Perception Coupling and the Cerebellum J. Neurosci., May 7, 2014 •34(19):6707– 6716 • 6715

cortex correlated with movement and motor parameters irrespective of
the effector limb. J Neurophysiol 91:576–582. CrossRef Medline
Hagura N, Oouchida Y, Aramaki Y, Okada T, Matsumura M, Sadato N, Naito
E (2009) Visuokinesthetic perception of hand movement is mediated by
cerebro-cerebellar interaction between the left cerebellum and right pari-
etal cortex. Cereb Cortex 19:176–186. CrossRef Medline
Ha¨ ndel B, Thier P, Haarmeier T (2009) Visual motion perception deficits
due to cerebellar lesions are paralleled by specific changes in cerebro-
cortical activity. J Neurosci 29:15126–15133. CrossRef Medline
Ivry RB, Spencer RM (2004) The neural representation of time. Curr Opin
Neurobiol 14:225–232. CrossRef Medline
Izawa J, Criscimagna-Hemminger SE, Shadmehr R (2012) Cerebellar con-
tributions to reach adaptation and learning sensory consequences of ac-
tion. J Neurosci 32:4230– 4239. CrossRef Medline
Johansson G (1973) Visual perception of biological motion and a model for
its analysis. Percept Psychophys 14:201–211. CrossRef
Karnath HO, Himmelbach M, Rorden C (2002) The subcortical anatomy of
human spatial neglect: putamen, caudate nucleus and pulvinar. Brain
125:350–360. CrossRef Medline
Kelly RM, Strick PL (2003) Cerebellar loops with motor cortex and prefron-
tal cortex of a nonhuman primate. J Neurosci 23:8432–8444. Medline
Kilner JM, Neal A, Weiskopf N, Friston KJ, Frith CD (2009) Evidence of
mirror neurons in human inferior frontal gyrus. J Neurosci 29:10153–
10159. CrossRef Medline
Ku¨ per M, Dimitrova A, Thu¨ rling M, Maderwald S, Roths J, Elles HG, Gize-
wski ER, Ladd ME, Diedrichsen J, Timmann D (2011) Evidence for a
motor and a non-motor domain in the human dentate nucleus: an fMRI
study. Neuroimage 54:2612–2622. CrossRef Medline
Lackner JR, Mather JA (1981) Eye-hand tracking using afterimages: evi-
dence that sense of effort is dependent on spatial constancy mechanisms.
Exp Brain Res 44:138–142. Medline
Larsell O (1958) Lobules of the mammalian and human cerebellum Anat
Rec 130:329–330.
Leichnetz GR (2001) Connections of the medial posterior parietal cortex
(area 7 m) in the monkey. Anat Rec 263:215–236. CrossRef Medline
Macmillan NA, Creelman CD (2004) Detection theory: a user’s guide. Mah-
wah, NJ: Lawrence Erlbaum.
May JG, Andersen RA (1986) Different patterns of corticopontine projec-
tions from separate cortical fields within the inferior parietal lobule and
dorsal prelunate gyrus of the macaque. Exp Brain Res 63:265–278.
Medline
Miall RC (2003) Connecting mirror neurons and forward models. Neurore-
port 14:2135–2137. CrossRef Medline
Miall RC, Stanley J, Todhunter S, Levick C, Lindo S, Miall JD (2006) Per-
forming hand actions assists the visual discrimination of similar hand
postures. Neuropsychologia 44:966–976. CrossRef Medline
Miall RC, Christensen LO, Cain O, Stanley J (2007) Disruption of state es-
timation in the human lateral cerebellum. PLoS Biol 5:e316. CrossRef
Medline
O’Reilly JX, Mesulam MM, Nobre AC (2008) The cerebellum predicts the
timing of perceptual events. J Neurosci 28:2252–2260. CrossRef Medline
O’Reilly JX, Beckmann CF, Tomassini V, Ramnani N, Johansen-Berg H
(2010) Distinct and overlapping functional zones in the cerebellum de-
fined by resting state functional connectivity. Cereb Cortex 20:953–965.
CrossRef Medline
Prevosto V, Graf W, Ugolini G (2010) Cerebellar inputs to intraparietal
cortex areas LIP and MIP: functional frameworks for adaptive control of
eye movements, reaching, and arm/eye/head movement coordination.
Cereb Cortex 20:214–228. CrossRef Medline
Prinz W (1997) Perception and action planning. Eur J Cogn Psychol 9:129 –
154. CrossRef
Rorden C, Karnath HO, Bonilha L (2007) Improving lesion-symptom map-
ping. J Cogn Neurosci 19:1081–1088. CrossRef Medline
Schlerf J, Ivry RB, Diedrichsen J (2012) Encoding of sensory prediction er-
rors in the human cerebellum. J Neurosci 32:4913–4922. CrossRef
Medline
Schmahmann JD, Doyon J, McDonald D, Holmes C, Lavoie K, Hurwitz AS,
Kabani N, Toga A, Evans A, Petrides M (1999) Three-dimensional MRI
atlas of the human cerebellum in proportional stereotaxic space. Neuro-
image 10:233–260. CrossRef Medline
Schu¨ tz-Bosbach S, Prinz W (2007) Perceptual resonance: action-induced
modulation of perception. Trends Cogn Sci 11:349–355. CrossRef
Medline
Serapide MF, Zappala` A, Parenti R, Panto` MR, Cicirata F (2002) Laterality
of the pontocerebellar projections in the rat. Eur J Neurosci 15:1551–
1556. CrossRef Medline
Shinoda Y, Sugiuchi Y, Futami T, Izawa R (1992) Axon collaterals of mossy
fibers from the pontine nucleus in the cerebellar dentate nucleus. J Neu-
rophysiol 67:547–560. Medline
Sokolov AA, Erb M, Gharabaghi A, Grodd W, Tatagiba MS, Pavlova MA
(2012) Biological motion processing: the left cerebellum communicates
with the right superior temporal sulcus. Neuroimage 59:2824–2830.
CrossRef Medline
Stein BE, Stanford TR (2008) Multisensory integration: current issues from
the perspective of the single neuron. Nat Rev Neurosci 9:255–266.
CrossRef Medline
Strick PL, Dum RP, Fiez JA (2009) Cerebellum and nonmotor function.
Annu Rev Neurosci 32:413–434. CrossRef Medline
Sultan F, Augath M, Hamodeh S, Murayama Y, Oeltermann A, Rauch A,
Thier P (2012) Unravelling cerebellar pathways with high temporal pre-
cision targeting motor and extensive sensory and parietal networks. Nat
Commun 3:924. CrossRef Medline
Thieme A, Thu¨ rling M, Galuba J, Burciu RG, Go¨ ricke S, Beck A, Aurich V,
Wondzinski E, Siebler M, Gerwig M, Bracha V, Timmann D (2013)
Storage of a naturally acquired conditioned response is impaired in pa-
tients with cerebellar degeneration. Brain 136:2063–2076. CrossRef
Medline
Trouillas P, Takayanagi T, Hallett M, Currier RD, Subramony SH, Wessel K,
Bryer A, Diener HC, Massaquoi S, Gomez CM, Coutinho P, Ben Hamida
M, Campanella G, Filla A, Schut L, Timann D, Honnorat J, Nighoghos-
sian N, Manyam B (1997) International Cooperative Ataxia Rating Scale
for pharmacological assessment of the cerebellar syndrome: the Ataxia
Neuropharmacology Committee of the World Federation of Neurology.
J Neurol Sci 145:205–211. CrossRef Medline
Van Essen DC, Newsome WT, Bixby JL (1982) The pattern of interhemi-
spheric connections and its relationship to extrastriate visual areas in the
macaque monkey. J Neurosci 2:265–283. Medline
Vicario CM, Martino D, Koch G (2013) Temporal accuracy and variability
in the left and right posterior parietal cortex. Neuroscience 245:121–128.
CrossRef Medline
Wiener M, Turkeltaub P, Coslett HB (2010) The image of time: a voxelwise
meta-analysis. Neuroimage 49:1728–1740. CrossRef Medline
Wolpert DM, Flanagan JR (2001) Motor prediction. Curr Biol 11:R729–
R732. CrossRef Medline
Wolpert DM, Miall RC, Kawato M (1998) Internal models in the cerebel-
lum. Trends Cogn Sci 2:338–347. CrossRef Medline
Wolpert DM, Doya K, Kawato M (2003) A unifying computational frame-
work for motor control and social interaction. Philos Trans R Soc Lond B
Biol Sci 358:593–602. CrossRef Medline
Wu HS, Sugihara I, Shinoda Y (1999) Projection patterns of single mossy
fibers originating from the lateral reticular nucleus in the rat cerebellar
cortex and nuclei. J Comp Neurol 411:97–118. CrossRef Medline
Yeo BT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M,
Roffman JL, Smoller JW, Zo¨ llei L, Polimeni JR, Fischl B, Liu H, Buckner
RL (2011) The organization of the human cerebral cortex estimated by
intrinsic functional connectivity. J Neurophysiol 106:1125–1165. CrossRef
Medline
6716 •J. Neurosci., May 7, 2014 •34(19):6707– 6716 Christensen et al. •Action-Perception Coupling and the Cerebellum