Content uploaded by Roy Mukamel
Author content
All content in this area was uploaded by Roy Mukamel on Aug 21, 2016
Content may be subject to copyright.
1
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
www.nature.com/scientificreports
Activity in superior parietal cortex
during training by observation
predicts asymmetric learning levels
across hands
Ori Ossmy1,2 & Roy Mukamel1,2
A dominant concept in motor cognition associates action observation with motor control. Previous
studies have shown that passive action observation can result in signicant performance gains in
humans. Nevertheless, it is unclear whether the neural mechanism subserving such learning codes
abstract aspects of the action (e.g. goal) or low level aspects such as eector identity. Eighteen healthy
subjects learned to perform sequences of nger movements by passively observing right or left hand
performing the same sequences in egocentric view. Using functional magnetic resonance imaging
we show that during passive observation, activity in the superior parietal lobule (SPL) contralateral
to the identity of the observed hand (right\left), predicts subsequent performance gains in individual
subjects. Behaviorally, left hand observation resulted in positively correlated performance gains of
the two hands. Conversely right hand observation yielded negative correlation - individuals with high
performance gains in one hand exhibited low gains in the other. Such behavioral asymmetry is reected
by activity in contralateral SPL during short-term training in the absence of overt physical practice and
demonstrates the role of observed hand identity in learning. These results shed new light on the coding
level in SPL and have implications for optimizing motor skill learning.
It has long been proposed that motion is intrinsically linked to perception such that observing an action acti-
vates motor programs that are used to execute the same action1. During the last two decades, there has been an
increasing interest in identifying the neural underpinning of this perceptual-motor link2–5. Direct recordings in
monkeys6 and in humans7,8 demonstrate an overlapping neural representation of observed and executed actions
in frontal and parietal regions. Additionally, neuroimaging studies reveal that action observation evokes activity
within various regions traditionally associated with motor function such as the primary motor cortex (M1),
supplementary motor area (SMA), dorsal premotor cortex (dPMC), supramarginal gyrus, inferior frontal gyrus
(IFG) and the superior parietal lobe (SPL)9–11. Such evoked activity supports the involvement of these regions in
translating perceived actions performed by someone else into a motor plan that can be executed by the observer.
More specically, parieto-frontal regions, that are active during passive action observation, overlap with those
engaged during online imitation12–16. Together, these results suggest a perceptual-motor translation mechanism
in the brain, ideally suited for imitation.
Imitation is an important form by which individuals can learn how to perform an action from observing
others17–20. Indeed, action observation has been shown to facilitate subsequent executed movements and improve
skill performance21–27. e process of learning by observation, was extensively investigated24,28–30 and supports the
idea that the perceptual-motor translation mechanism, which matches observed behaviors with internal motor
representations, facilitates subsequent motor performance. Understanding this mechanism can provide a useful
account of how we acquire new behaviors, as well as correct our movements12,31,32.
To date, there is no consensus regarding the level of abstraction of what is extracted from observed actions
and encoded during the learning process (e.g. kinematics, action goal). Given that the neural representation dur-
ing physical practice is known to have a contra-lateral bias33,34, it is not clear whether during action observation
the perceptual-motor learning mechanism displays a similar bias. In their associative sequence learning theory,
1Sagol School of Neuroscience, Tel-Aviv University, Tel-Aviv 69978, Israel. 2School of Psychological Sciences, Tel-Aviv
University, Tel-Aviv 69978, Israel. Correspondence and requests for materials should be addressed to R.M. (email:
rmukamel@tau.ac.il)
Received: 11 February 2016
Accepted: 03 August 2016
Published: 18 August 2016
OPEN
www.nature.com/scientificreports/
2
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
Heyes and Ray argue that action observation engages similar learning processes as overt physical practice, and
hence depends on the identity of the observed eector31,35. is theory was supported by a series of empirical
studies in which participants performed a serial reaction time (SRT) task, and performance was facilitated only
when subjects used the same eector as the one they observed during passive training28,36–38. Such evidence are
taken to support the notion that the perceptual-motor learning mechanism uses an eector-dependent coordi-
nate frame. However, other recent studies suggest that the perceptual-motor learning mechanism can generalize
across eectors. In a behavioral study by Williams and Gribble, subjects passively observed videos of another
individual reaching to visual targets in a force eld with the right or le hand39. Results showed that subsequent
right hand performance was facilitated in a similar manner following observation sessions depicting right or le
hand. Another study demonstrates shorter reaction times following observation of the same action performed
with a dierent eector40. ese later results suggest that observed actions evoke neural representations of action
goals which are eector independent.
From a neural perspective, neuroimaging studies showed that activation in the parietal lobe is sensitive to
observed eectors in a somatotopic manner2 and also to visual perspectives (egocentric vs. allocentric)41. However
little is known as to how the identity of the observed eector modulates neural activity and how such modulation
relates to learning by observation. A study by Frey and Gerry showed that activity in the right intraparietal sulcus
predicts learning levels following training by observation. However, they used a bimanual sequential task and
therefore were not able to outline the role of hand identity (right/le) in the perceptual-motor learning process42.
e current study focuses on how learning by observation and its underlying neural mechanisms depend
on the identity of the observed effector (right/left hand). To this end we acquired whole-brain functional
magnetic resonance imaging (fMRI) data while healthy subjects were engaged in a short-term unimanual
learning-by-observation task.
Results
Eighteen healthy subjects were presented with visual input consisting of two virtual hands in egocentric view
(Fig.1a) while either the right or the le virtual hand performed a to-be-learned sequence of nger movements
(conditions ‘Obs-RH’ and ‘Obs-LH’ respectively). Performance levels on the sequence of nger movements were
evaluated before and aer training (see experimental design in Fig.1b). Le and right hand performance gains
(calculated as the accuracy index G; see Methods) were signicant in both training conditions – demonstrating
signicant learning by observation. Within hands, there was no signicant dierence between performance gains
in the Obs-RH and Obs-LH conditions at the group level (p = 0.55 and p = 0.4 for right and le hand respec-
tively). Regression analysis on individual subject data conrmed no signicant relationship between performance
gains and identity of observed hand. is was true both for le and right hands (See Fig.2a). We also compared
performance gains across hands within each observational training condition. At the group level, there was no
signicant dierence between hands, and this was true both following right or le hand observation (p = 0.68
and p = 0.55 respectively; see Fig.2b). However, regression analysis on individual subjects revealed a signicant
correlation between right and le hand performance gains which depended on the identity of observed hand
during training. Following le hand observation (‘Obs-LH’), performance gains across hands were positively
correlated. Subjects exhibiting high performance gains in the right hand also exhibited high performance gains
in the le hand (r = 0.47, p = 0.04; Fig.2b le panel). Conversely, following right hand observation (‘Obs-RH’),
correlation between the two hands was negative (r = − 0.54, p = 0.02; Fig.2b right panel). us subjects exhibiting
high performance gains in the right hand, exhibited low performance gains in the le hand and vice versa. Taken
together, these data demonstrate that although on average at the group level, the identity of observed hand does
not aect post-training performance gains, across individual subjects the identity of observed hand has an asym-
metric impact on short-term learning across the two hands.
Figure 1. Experimental design. (a) Subjects learned a novel sequence of nger movements by observing two
virtual hands while one of the hands performed the sequence. (b) Each subject completed two experimental
runs. In each run, a unique sequence of ve digits was presented together with a sketch of the mapped ngers.
Aer instructions, subjects performed the sequence using their right hand (RH) and their le hand (LH) for
initial evaluation of performance level. In the training stage subjects passively observed either the right or le
virtual hand performing the sequence. Aer training, subjects repeated the evaluation stage for assessment of
changes in performance level.
www.nature.com/scientificreports/
3
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
We also explored putative neural networks underlying these behavioral results. First, to examine global dif-
ferences in activation between le hand observation and right hand observation at the group level, we performed
a general linear model (GLM) analysis on the fMRI data obtained during the training stage. e multi-subject
map in Fig.3a demonstrates the high overlap between activation during le and right hand observation relative
to baseline (contrast Obs-LH > rest and contrast Obs-RH > rest, red and blue respectively). e direct contrast:
Obs-RH vs. Obs-LH, yielded an empty multi-subject map indicating no signicant dierence in the global acti-
vation levels for the two conditions in any particular voxel. ese results suggest that at the group level, observing
right or le hand nger movements engage spatially overlapping neural networks and at similar activation levels.
Next, we performed a regions of interest (ROI) analysis in individual subjects. In each subject, we dened brain
regions engaged during task execution (obtained from the pre-training test period; see multi subject map in Fig.3b
and Methods). For each subject and ROI, fMRI activity during the observational training period was compared
against the subsequent behavioral performance gains of each hand (see Methods). We found that across subjects,
activity in the right SPL during ‘Obs-LH’ training correlated with le (r = 0.72, p = 7·10−4) and right (r = 0.66,
p = 2·10−3) hand performance gains (See Fig.3c, le panel; Bonferroni corrected for the 16 regions examined).
is is compatible with our observation that across subjects, le and right hand performance gains are positively
correlated following training by le hand observation. In the remaining 14 ROIs, no signicant correlation values
were obtained. We also examined whether fMRI activity level in the same ROIs during ‘Obs-RH’ training period
corresponds to subsequent le or right hand performance gains. is time, out of all ROIs only activity in le SPL
exhibited a signicant positive correlation with right hand performance gain and a signicant negative correlation
with le hand performance gain (r = 0.68, p = 1.9·10−3 and r = − 0.71, p = 9.6·10−4 respectively; Bonferroni cor-
rected for multiple comparisons; see Fig.3c, right panel; for Montreal Neurological Institute (MNI) coordinates
of ROI centers in individual subjects see Table1). is is compatible with our behavioral results demonstrating
that le and right hand performance gains are negatively correlated following training by right hand observation.
ese results imply that in the absence of overt physical practice, the SPL in the hemisphere contralateral to the
identity of the observed hand plays an important role in the process of short term learning by observation.
Discussion
Observing and imitating others, has long been recognized as constituting a powerful learning strategy for
humans. Previous studies evaluating learning by observation report inconsistent results with respect to the level
of abstract representation coded by the perceptual-motor learning mechanism36,37,39,40. e present study was
Figure 2. Behavioral results. (a) Le hand performance gains were signicant relative to baseline following
right (Obs-RH; p = 0.002) and le (Obs-LH; p = 0.006) hand observation but not signicantly dierent between
the two observation condition. Similarly, right hand performance gains were signicant following right and le
hand observation but not signicantly dierent between the two observation conditions. (b) Following le hand
observation there was no signicant dierence in performance gains between right and le hands at the group
level. However, regression analysis on individual subject data revealed a signicant positive correlation between
le and right hand performance gains. Following right hand observation, there was no signicant dierence
between the two hands at the group level. However, regression analysis on individual subject data revealed a
signicant negative correlation.
www.nature.com/scientificreports/
4
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
designed to determine the eect of identity of the observed eector on learning by observation and trace its
underlying neural mechanism.
Behaviorally, the performance gains we found aer training support the idea that humans can indeed acquire
information about “how” to make movements by mere observation26. At the group level, we found that short
term motor learning by observation does not depend on the identity of the observed eector suggesting that the
perceptual-motor learning mechanism codes actions in an abstract manner. is is consistent with previous stud-
ies showing that subjects can acquire a new skill similarly following observed actions performed with dierent
eectors39,40.
However, closer inspection of individual subject data revealed an unexpected asymmetrical eect of the iden-
tity of observed hand. is asymmetry demonstrates that the identity of the observed eector does play a role
Figure 3. Correlation of neural activity with behavior. (a) Random eect multi-subject activation
map (N = 18) displaying signicant regions obtained from the GLM contrasts: Obs-LH > rest (red) and
Obs-RH > rest (blue) in the training stage (q(FDR) < 0.05). Purple regions correspond to overlapping regions
between the two contrasts. (b) ROI analysis. Random eect multi-subject activation map (N = 18) displaying
signicant regions obtained from the GLM contrasts right hand execution > rest and le hand execution > rest
obtained from the pre-training evaluation period (see Methods). For all analyses, ROIs were dened using
this contrast at the individual subject level. (c) Across all ROIs, the fMRI signal during training by observation
in two regions – right and le SPL - correlated signicantly with subsequent performance gains. Each color
denotes peak cluster of each subject dened from the localizer (described in panel b). During training by le
hand observation, activity in the right SPL correlated both with subsequent le and right hand performance
gains. During training by right hand observation, activity in the le SPL showed positive correlation with right
hand performance gain and negative correlation with le hand performance gain.
www.nature.com/scientificreports/
5
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
in short term learning and contributes to the discussion regarding the level of abstract representation during
learning by observation. e behavioral asymmetry we found might be a reection of well documented hemi-
spheric dierences in cortical hand representation43–45. Such cortical dierences are usually associated with hand
dominance33,46 and inter-limb dierences in motor control and learning47,48. Whether or not the asymmetry we
report is related to hand dominance is not known and a further study in le-handers is required to clarify this
point. Nonetheless, the current results do imply that at least in right-handers, le hand rather than right hand
visual input can lead to optimization of short term motor learning by observation.
At the neural level, our results provide strong evidence that in the absence of overt physical practice, the
bilateral SPL plays an important role in enhancing short-term learning during observation. e parietal cortex
has been previously found active during action observation, demonstrating its engagement in visuo-motor con-
trol and imitation learning6,13,24,49,50. Additionally, it has been suggested that the SPL represents low level aspects
(e.g. visual perspective) of the action rather than abstract aspects (e.g. action goal)12,41,51. We extend this view
to the context of learning. Although at the group level we did not nd signicant dierence in activation power
during le and right hand observation, at the individual subject level activation in SPL depended on the identity
of observed hand and correlated with subsequent performance gains. ese correlations during observational
training on a unimanual task are in agreement with a previous study showing that activity level in parietal regions
predicted subsequent performance gains in a bimanual problem solving task42. Taken together, the results empha-
size the key role of parietal regions in short-term learning by observation of complex action sequences.
Although in the current study we focus on the role of hand identity in learning by observation, the current
ndings may well have a bearing on the neural underpinning of motor imitation. Previous studies revealed a
set of parieto-frontal regions active during both observation and synchronous imitation13,14,52. We found simi-
lar regions during the observational training with no signicant dierence in activation level between observa-
tion of right and le hand movements at the group level. A previous imitation study by Aziz-Zadeh et al. also
demonstrated bilateral activations independent of the identity of the observed hand52. us, during imitation, the
perceptual-motor translation mechanism seems to form a neural representation of the observed action in both
hemispheres implying an abstraction with respect to hand identity. However, our nding of hemispheric laterality
in SPL’s prediction of subsequent performance gains in individual subjects, suggests that the learning process in
SPL is sensitive to the identity of the observed hand.
Finally, the importance of SPL in coding spatial aspects of movements is well established53,54. In the current
study we used a nger movement task that has a strong spatial component. erefore, additional research will be
necessary to assess whether the SPL involvement we report in learning generalizes to other types of tasks with a
less dominant spatial factor (e.g. temporal learning tasks). Another key question is to what extent the role of SPL
in learning observed actions holds true also for learning motor movements that are not in the repertoire of the
observer55–57. Future studies should go beyond learning how to combine known movement elements in a new
arrangement (e.g. order of nger movements), and examine the acquisition process of entirely new movements
(e.g. when children imitate adults58).
In summary, by examining not only the group level but also the individual subject level, we show asymmetric
performance gains across hands which depend on the identity of the observed hand. Furthermore, we show that
the SPL plays a signicant role in this asymmetric learning by observation. ese ndings have implications for
the debate regarding the coding level in SPL, and demonstrate that for the purpose of learning in both hands, le
rather than right hand observation is preferable.
Subject Le SPL coordinates Right SPL coordinates
1 [− 31 − 63 50] [24 − 65 49]
2 [− 22 − 74 44] [25 − 76 42]
3 [− 28 − 64 49] [17 − 65 54]
4 [− 18 − 65 54] [25 − 57 55]
5 [− 25 − 65 34] [25 − 65 43]
6 [− 32 − 51 62] [24 − 53 53]
7 [− 20 − 52 51] [23 − 56 50]
8 [− 34 − 51 67] [13 − 63 63]
9 [− 25 − 66 54] [25 − 66 50]
10 [− 24 − 59 61] [21 − 54 56]
11 [− 29 − 58 59] [30 − 57 56]
12 [− 34 − 64 56] [22 − 63 62]
13 [− 27 − 52 60] [22 − 60 62]
14 [− 31 − 51 49] [31 − 53 56]
15 [− 31 − 54 56] [29 − 54 57]
16 [− 29 − 63 47] [24 − 65 45]
17 [− 35 − 72 35] [28 − 66 41]
18 [− 23 − 55 55] [24 − 54 66]
Table 1. MNI locations of each subject’s peak voxels in the le and right SPL ROIs (Fig.3c).
www.nature.com/scientificreports/
6
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
Methods
Subjects. Eighteen healthy subjects (10 females, mean age: 27.4, range: 22–34 years), naïve to the purpose
of the study, participated aer providing written informed consent. All subjects were right handed according to
the Edinburgh Handedness Inventory, with normal or corrected-to-normal vision and no reported cognitive or
neurological decits. Subjects were recruited according to the standard safety criteria for fMRI studies and were
compensated by payment (55 NIS/hour). e study protocol was approved by the Ethics Committee of Tel-Aviv
University and the Helsinki committee at the Tel-Aviv Sourasky Medical Center, and carried out in accordance
with the approved guidelines.
Procedure and Task. During fMRI scans, subjects completed 5 consecutive experimental runs, in which
they learned to perform unique sequences of nger movements. During 3 runs, subjects physically practiced a
sequence of nger movements (results reported in a separate manuscript). In two runs, subjects passively learned
to perform a sequence of nger movements by observing two virtual hands on the screen while either the le
or the right virtual hand performed the sequence (Fig.1a). roughout these runs the subjects’ real hands were
immobile.
Fingers were numbered from index (1) to little nger (4) and subjects were instructed to learn a unique
sequence in each experimental run (one of 5 optional sequences: 4-1-3-2-4, 4-2-3-1-4, 3-2-4-1-3, 3-1-4-2-3, 2-1-
4-3-2). e subjects were instructed to learn the sequence and knew that they would be tested on it in pre- and
post- evaluation stages. Run order and the nger sequence associated with each run were counter balanced across
subjects. Subjects’ nger movements were monitored using motion-sensing MR-compatible gloves (5DT Data
Glove Ultra). We programmed a custom-build soware, based on the application programming interface pro-
vided by 5DT, to extract nger movements from 14 dierent sensors and control the presentation of the virtual
hands (http://www.5dt.com). is allowed us to verify that the subject’s hands are immobile during the train-
ing stage and also allowed us to yoke virtual hand movement presented on the screen to real hand movement
during the evaluation stages (see below). Small delays in feedback are unavoidable due to sampling rate of the
motion-sensing gloves (up to 16 ms) and the refresh rate of the screen (up to 16 ms). We veried that the soware
itself does not introduce additional delays (time duration for calculating ngers’ position and updating the hands
animation was less than 15 ms). erefore we estimate that delays in visual feedback were no longer than 3 refresh
rates in the worst case scenario. During experimental runs, virtual hands were presented with a black background
on a screen. Subjects lied supine with their arms to the side of their body and palms facing up. Subjects could not
see their real hands during the scans and viewed the screen through a tilted mirror mounted in front of their eyes.
In the beginning of each run (Fig.1b), subjects were presented with an instructions slide that depicted two
hand illustrations with numbered ngers and a 5-digit sequence underneath representing the sequence of nger
movements to be learned. e instructions slide was presented for 12 seconds and was followed by a pre-training
evaluation stage in which baseline performance level of each hand was separately assessed. During the evaluation,
subjects physically performed the required sequence with one hand repeatedly, as fast and as accurate as possible,
for 30 seconds (hand order evaluation right\le was counter balanced across all sessions). During the evaluation
stage, real-time visual feedback was provided by the virtual hands’ movements which were yoked to the subjects’
real hand movements.
Following the pre-training evaluation stage, a “Start Training” slide appeared for 9 seconds and cued the sub-
jects to the upcoming training stage. In the training stage subjects observed the le or right virtual hand perform-
ing the sequence (order of training conditions was counter balanced across subjects). Subject were instructed to
observe the screen and refrain from moving their hands. e pace of virtual hand movement during training was
constant across both conditions and was set based on the average pace of the subject during previous execution
sessions. In cases where the training by observation run was rst, the pace was set based on the average pace of
previous subjects (range across subject: 4–9 digit sequences per block). Each training block lasted 15 seconds and
was followed by a 9-second resting period, in which a blank yellow screen appeared. e training stage consisted
of 20 such training blocks (with overall experimental run duration lasting 8 minutes). Aer the training stage,
subject’s performance was evaluated again for 30 seconds in each hand. Similar to the pre-training evaluation,
subjects were instructed to repeatedly execute the sequence as fast and as accurately as possible.
Performance evaluation. In all evaluation stages, we calculated subject’s performance using the formula
(1) below:
=
−
+
__
__
G
pp
pp
(1)
post training pretraining
post training pretraining
where p is the number of times the subject performed a complete sequence with no errors. erefore, a positive
G index reects improvement in performance. We calculated this performance gain index for each subject, each
training condition (Obs-RH, Obs-LH), and each hand (le/right) - allowing us to compare improvement under
dierent motor trainings conditions.
Behavioral analysis. We used the data from the motion-detection gloves to verify that the subjects did not
move their ngers during the observational training conditions. Each sensor of the glove provided the angle of
each nger joint (sampling rate = 16 ms) and the subjects always started the training sessions with the hand in the
same orientation. e maximal angle of each nger during observational training was not signicantly dierent
from the maximal angle during rest. is was true for all ngers of both the right and the le hand (minimal
p = 0.42; two-tailed paired t-test; in the main text, unless otherwise stated, all signicance tests were two-tailed
www.nature.com/scientificreports/
7
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
paired t-test and all correlation tests were Pearson correlation). is analysis rules out an alternative explanation
that dierences in performance gains (across individuals or conditions) are due to subliminal physical movement
of the subjects during observation.
fMRI data acquisition. Blood oxygenation level dependent (BOLD) images were obtained on a 3 T General
Electric scanner with an 8 channel head coil located at the Tel-Aviv Souraski Medical Center, Tel-Aviv, Israel. An
echo-planar imaging sequence was used to obtain the functional data (39 ascending interleaved axial slices, 4 mm
thickness, slice gaps = 0; TR = 3000 ms; ip angle = 90°; TE = 30 ms; in-plane resolution = 1.72 × 1.72 mm; matrix
size = 128 × 128). In addition, anatomical reference was obtained by T1-weighted scan (voxel size = 1 × 1 × 1 mm)
for each subject.
fMRI Preprocessing. All fMRI data were processed using the BrainVoyager QX soware (version 2.6, Brain
Innovation, Maastricht, Netherlands; http://www.brainvoyager.com). Prior to statistical analysis, a preprocess-
ing procedure was performed on all functional images and included cubic spline slice-time correction, trilinear
3D motion correction, and high-pass ltering (above 0.006 Hz). In addition, we assessed head movements and
veried no scans contained head movement exceeding 2 mm in either direction. e 2D functional images were
co-registered to the anatomical images and the complete dataset was transformed into the Talairach coordinate
system for multi-subject comparisons59. Functional data of individual subjects was spatially smoothed (Gaussian
lter, FWHM 6 mm) prior to statistical analysis. We also transformed the dataset to Montreal Neurological
Institute coordinate system (MNI-305) in order to extract the individual subject peak coordinates that are pro-
vided in Table1.
ROI Analysis. Regions of interest (ROIs; see Fig.3b) were dened at an individual subject level. We used
general linear model (GLM) contrasts to reveal brain regions active during the pre-training evaluations collapsed
across all sessions of each hand: (Pre_Trainingright_hand > rest) and (Pre_Trainingle_hand > rest). e resulting maps
were corrected by controlling the False Discovery Rate (FDR60) and thresholded at q(FDR) < 0.05. For each sub-
ject, we dened a sphere, with maximum cluster size of 12 mm radius around the peak activation within each
anatomically dened region according to Mai et al. 1997. Regions revealed in this analysis include the right and
le pre-motor cortex (R-PMc/L-PMc), primary motor cortex (R- 1/L-M1), visual cortex (R-Visual/L-Visual),
post-central gyrus (R-PoG/L-PoG), superior parietal lobule (R-SPL/L-SPL), supplementary motor area
(R-SMA/L-SMA) and subcortical regions (R-lamus/L-alamus/R-Striatum/L-Striatum). Next, for each of the
eighteen subjects, we calculated separately in each ROI the average fMRI activity level across all signicant voxels
during training (either Obs-RH or Obs-LH; yielding a vector of 18 values for each ROI and training condition).
As the behavioral measure we took the corresponding performance gain (G) in each hand of each subject (See
Fig.3c).
References
1. James, W. e principles ofpsychology (Vol. 1). New Yor: Holt (1890).
2. Buccino, G. et al. Action observation activates premotor and parietal areas in a somatotopic manner: an fMI study. European
journal of neuroscience 13, 400–404, doi: 10.1111/j.1460-9568.2001.01385 (2001).
3. Blaemore, S.-J. & Decety, J. From the perception of action to the understanding of intention. Nature eviews Neuroscience 2,
561–567, doi: 10.1038/35086023 (2001).
4. Flanagan, J. . & Johansson, . S. Action plans used in action observation. Nature 424, 769–771, doi: 10.1038/nature01861 (2003).
5. Prinz, W. Perception and action planning. European journal of cognitive psychology 9, 129–154, doi: 10.1080/713752551 (1997).
6. izzolatti, G., Fogassi, L. & Gallese, V. Neurophysiological mechanisms underlying the understanding and imitation of action. Nat
ev Neurosci 2, 661–670, doi: 10.1038/35090060 (2001).
7. Muamel, ., Estrom, A. D., aplan, J., Iacoboni, M. & Fried, I. Single-neuron responses in humans during execution and
observation of actions. Current biology : CB 20, 750–756, doi: 10.1016/j.cub.2010.02.045 (2010).
8. Tremblay, C. et al. Action observation and execution Intracranial recordings in a human subject. Neurology 63, 937–938, doi: 10.
1212/ 01. WNL. 0000137111. 16767. C6 (2004).
9. Grezes, J. & Decety, J. Functional anatomy of execution, mental simulation, observation, and verb generation of actions: a meta
analysis. Human brain mapping 12, 1–19, doi: 10.1002/1097-0193(200101)12 (2001).
10. Tanaa, S. & Inui, T. Cortical involvement for action imitation of hand/arm postures versus nger congurations: an fMI study.
Neuroreport 13, 1599–1602 (2002).
11. Gazzola, V., Aziz-Zadeh, L. & eysers, C. Empathy and the somatotopic auditory mirror system in humans. Current biology 16,
1824–1829, doi: 10.1016/j.cub.2006.07.072 (2006).
12. Iacoboni, M. et al. Cortical mechanisms of human imitation. Science 286, 2526–2528, doi: 10.1126/science.286.5449.2526 (1999).
13. Buccino, G. et al. Neural circuits underlying imitation learning of hand actions: an event-related fMI study. Neuron 42, 323–334,
doi: 10.1016/S0896-6273(04)00181-3 (2004).
14. Macuga, . L. & Frey, S. H. Neural representations involved in observed, imagined, and imitated actions are dissociable and
hierarchically organized. NeuroImage 59, 2798–2807, doi: 10.1016/j.neuroimage.2011.09.083 (2012).
15. Higuchi, S., Holle, H., oberts, N., Eicho, S. B. & Vogt, S. Imitation and observational learning of hand actions: prefrontal
involvement and connectivity. NeuroImage 59, 1668–1683, doi: 10.1016/j.neuroimage.2011.09.021 (2012).
16. rüger, B. et al. Parietal and premotor cortices: Activation reects imitation accuracy during observation, delayed imitation and
concurrent imitation. NeuroImage 100, 39–50, doi: 10.1016/j.neuroimage.2014.05.074 (2014).
17. Meltzo, A. N. & Prinz, W. e imitative mind: Development, evolution and brain bases. Vol. 6 (Cambridge University Press, 2002).
18. Byrne, . W. Imitation as behaviour parsing. Philosophical Transactions of the oyal Society of London B: Biological Sciences 358,
529–536, doi: 10.1098/rstb.2002.1219 (2003).
19. Tomasello, M. & Call, J. Primate cognition. Vol. 24 (Oxford University Press, USA, 1997).
20. Bandura, A. Social foundations of thought and action: A social cognitive theory. (Prentice-Hall, Inc, 1986).
21. Brass, M., Beering, H. & Prinz, W. Movement observation aects movement execution in a simple response tas. Acta psychologica
106, 3–22, doi: 10.1016/S0001-6918(00)00024-X (2001).
22. ilner, J. M., Paulignan, Y. & Blaemore, S. J. An Interference Eect of Observed Biological Movement on Action. Current Biology
13, 522–525, doi: 10.1016/s0960-9822(03)00165-9 (2003).
www.nature.com/scientificreports/
8
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
23. Boutin, A., Fries, U., Panzer, S., Shea, C. H. & Blandin, Y. ole of action observation and action in sequence learning and coding.
Acta psychologica 135, 240–251, doi: 10.1016/j.actpsy.2010.07.005 (2010).
24. Cross, E. S., raemer, D. J., Hamilton, A. F., elley, W. M. & Graon, S. T. Sensitivity of the action observation networ to physical
and observational learning. Cerebral cortex 19, 315–326, doi: 10.1093/cercor/bhn083 (2009).
25. elly, S. W., Burton, A. M., iedel, B. & Lynch, E. Sequence learning by action and observation: evidence for separate mechanisms.
British journal of psychology 94, 355–372, doi: 10.1348/000712603767876271 (2003).
26. Mattar, A. A. & Gribble, P. L. Motor learning by observing. Neuron 46, 153–160, doi: 10.1016/j.neuron.2005.02.009 (2005).
27. Nojima, I., oganemaru, S., awamata, T., Fuuyama, H. & Mima, T. Action observation with inesthetic illusion can produce
human motor plasticity. e European journal of neuroscience 41, 1614–1623, doi: 10.1111/ejn.12921 (2015).
28. Heyes, C. M. & Foster, C. L. Motor learning by observation: evidence from a serial reaction time tas. e Quarterly journal of
experimental psychology. A, Human experimental psychology 55, 593–607, doi: 10.1080/02724980143000389 (2002).
29. Petrosini, L. et al. Watch how to do it! New advances in learning by observation. Brain esearch ev iews 42, 252–264, doi: 10.1016/
S0165-0173(03)00176-0 (2003).
30. Wulf, G., Chiviacowsy, S. & Lewthwaite, . Normative feedbac eects on learning a timing tas. esearch quarterly for exercise and
sport 81, 425–431, doi: 10.1080/02701367.2010.10599703 (2010).
31. Heyes, C. Causes and consequences of imitation. Trends in cognitive sciences 5, 253–261, doi: 10.1016/S1364-6613(00)01661-2 (2001).
32. Catmur, C., Walsh, V. & Heyes, C. Associative sequence learning: the role of experience in the development of imitation and the
mirror system. Philosophical Transactions of the oyal Society of London B: Biological Sciences 364, 2369–2380, doi: 10.1098/
rstb.2009.0048 (2009).
33. im, S.-G. et al. Functional magnetic resonance imaging of motor cortex: hemispheric asymmetry and handedness. Science 261,
615–617, doi: 10.1126/science.8342027 (1993).
34. Halsband, U. & Lange, . . Motor learning in man: a review of functional and clinical studies. Journal of physiology, Paris 99,
414–424, doi: 10.1016/j.jphysparis.2006.03.007 (2006).
35. ay, M., Dewey, D., ooistra, L. & Welsh, T. N. e relationship between the motor system activation during action observation and
adaptation in the motor system following repeated action observation. Human movement science 32, 400–411, doi: 10.1016/j.
humov.2012.02.003 (2013).
36. Osman, M., Bird, G. & Heyes, C. Action observation supports effector-dependent learning of finger movement sequences.
Experimental brain research 165, 19–27, doi: 10.1007/s00221-005-2275-0 (2005).
37. Bird, G. & Heyes, C. Eector-dependent learning by observation of a nger movement sequence. J Exp Psychol Hum Percept Perform
31, 262–275, doi: 10.1037/0096-1523.31.2.262 (2005).
38. Bird, G., Osman, M., Saggerson, A. & Heyes, C. Sequence learning by action, observation and action observation. British journal of
psychology 96, 371–388, doi: 10.1348/000712605X47440 (2005).
39. Williams, A. & Gribble, P. L. Observed effector-independent motor learning by observing. Journal of neurophysiology 107,
1564–1570, doi: 10.1152/jn.00748.2011 (2012).
40. Costantini, M. & Committeri, G. & Galati, G. Eector-and target-independent representation of observed actions: evidence from
incidental repetition priming. Experimental brain research 188, 341–351, doi: 10. 1007/ s00221-008-1369-x (2008).
41. Shmuelof, L. & Zohary, E. Mirror-image representation of action in the anterior parietal cortex. Nature neuroscience 11, 1267–1269,
doi: 10.1038/nn.2196 (2008).
42. Frey, S. H. & Gerry, V. E. Modulation of neural activity during observational learning of actions and their sequential orders. e
Journal of neuroscience: the ocial journal of the Society for Neuroscience 26, 13194–13201, doi: 10.1523/JNEUOSCI.3914-06.2006
(2006).
43. Volmann, J., Schnitzler, A., Witte, O. & Freund, H.-J. Handedness and asymmetry of hand representation in human motor cortex.
Journal of neurophysiology 79, 2149–2154 (1998).
44. Hammond, G. Correlates of human handedness in primary motor cortex: a review and hypothesis. Neuroscience & biobehavioral
reviews 26, 285–292, doi: 10.1016/S0149-7634(02)00003-9 (2002).
45. Cashmore, L., Uomini, N. & Chapelain, A. e evolution of handedness in humans and great apes: a review and current issues. J
Anthropol Sci 86, 7–35 (2008).
46. Amunts, ., Jänce, L., Mohlberg, H., Steinmetz, H. & Zilles, . Interhemispheric asymmetry of the human motor cortex related to
handedness and gender. Neuropsychologia 38, 304–312, doi: 10.1016/S0028-3932(99)00075-5 (2000).
47. Serrien, D. J., Ivry, . B. & Swinnen, S. P. Dynamics of hemispheric specialization and integration in the context of motor control.
Nature eviews Neuroscience 7, 160–166, doi: 10.1038/nrn1849 (2006).
48. Garry, M. I., amen, G. & Nordstrom, M. A. Hemispheric dierences in the relationship between corticomotor excitability changes
following a ne-motor tas and motor learning. Journal of neurophysiology 91, 1570–1578, doi: 10.1152/jn.00595.2003 (2004).
49. Culham, J. C. et al. Visually guided grasping produces fMI activation in dorsal but not ventral stream brain areas. Experimental
brain research 153, 180–189, doi: 10.1167/1.3.194 (2003).
50. Tracy, J. et al. egional brain activation associated with dierent performance patterns during learning of a complex motor sill.
Cerebral cortex 13, 904–910, doi: 10.1093/cercor/13.9.904 (2003).
51. Shmuelof, L. & Zohary, E. A mirror representation of others’ actions in the human anterior parietal cortex. The Journal of
neuroscience: the ocial journal of the Society for Neuroscience 26, 9736–9742, doi: 10.1523/JNEUOSCI.1836-06.2006 (2006).
52. Aziz-Zadeh, L., osi, L., Zaidel, E., Mazziotta, J. & Iacoboni, M. Lateralization of the human mirror neuron system. e Journal of
neuroscience : the ocial journal of the Society for Neuroscience 26, 2964–2970, doi: 10.1523/JNEUOSCI.2921-05.2006 (2006).
53. Galletti, C., Battaglini, P. P. & Fattori, P. Parietal neurons encoding spatial locations in craniotopic coordinates. Experimental brain
research 96, 221–229, doi: 10.1007/BF00227102 (1993).
54. Vandenberghe, ., Gitelman, D., Parrish, T. & Mesulam, M. Functional specicity of superior parietal mediation of spatial shiing.
NeuroImage 14, 661–673, doi: 10.1006/nimg.2001.0860 (2001).
55. Calvo-Merino, B., Grezes, J., Glaser, D. E., Passingham, . E. & Haggard, P. S eeing or doing? Inuence of visual and motor familiarity
in action observation. Current biology : CB 16, 1905–1910, doi: 10.1016/j.cub.2006.07.065 (2006).
56. Vogt, S. et al. Prefrontal involvement in imitation learning of hand actions: eects of practice and expertise. NeuroImage 37,
1371–1383, doi: 10.1016/j.neuroimage.2007.07.005 (2007).
57. Avanzino, L. et al. Action observation: mirroring across our spontaneous movement tempo. Scientic reports 5, 10325, doi: 10.1038/
srep10325 (2015).
58. Tizard, B. & Hughes, M. Young children learning. (John Wiley & Sons, 2008).
59. Talairach, J. & Tornoux, P. Co-planar stereotaxic atlas of the human brain: 3-dimensional proportional system: an approach to cerebral
imaging. (Georg ieme, 1988).
60. Benjamini, Y. & Hochberg, Y. Controlling the False Discovery ate - a Practical and Powerful Approach to Multiple Testing. J oy
Stat Soc B Met 57, 289–300 (1995).
Acknowledgements
is study was supported by the I-CORE Program of the Planning and Budgeting Committee and the Israel
Science Foundation (grant No. 51/11), e Israel Science Foundation (grants No. 1771/13 and 2043/13), and
www.nature.com/scientificreports/
9
Scientific RepoRts | 6:32133 | DOI: 10.1038/srep32133
Human Frontiers Science Project (HFSP) Career Development Award (CDA00078/2011-C) (R.M.); e Yosef
Sagol Scholarship for Neuroscience Research, e Israeli Presidential Honorary Scholarship for Neuroscience
Research, and the Sagol School of Neuroscience fellowship (O.O.). e authors thank A. Shuster and R. Gilron
for fruitful comments on the manuscript. e funders had no role in study design, data collection and analysis,
decision to publish, or preparation of the manuscript.
Author Contributions
R.M. and O.O. developed the study concept and designed the experiment; O.O. programmed and collected data;
O.O. carried out data analysis under supervision from R.M; O.O. and R.M. wrote the paper and approved the
nal version.
Additional Information
Competing nancial interests: e authors declare no competing nancial interests.
How to cite this article: Ossmy, O. and Mukamel, R. Activity in superior parietal cortex during training by
observation predicts asymmetric learning levels across hands. Sci. Rep. 6, 32133; doi: 10.1038/srep32133 (2016).
is work is licensed under a Creative Commons Attribution 4.0 International License. e images
or other third party material in this article are included in the article’s Creative Commons license,
unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license,
users will need to obtain permission from the license holder to reproduce the material. To view a copy of this
license, visit http://creativecommons.org/licenses/by/4.0/
© e Author(s) 2016
