YNIMG-10029; No.

of pages: 8; 4C: 3, 5
NeuroImage xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

NeuroImage
journal homepage: www.elsevier.com/locate/ynimg

1 Experience with an amputee modulates one's own sensorimotor response during

2 action observation

F
Q1 3 Sook-Lei Liew a, b,⁎, 1, Tong Sheng a, c, Lisa Aziz-Zadeh a, b, c

O
a
4 Brain and Creativity Institute, University of Southern California, 3620A McClintock Avenue, Los Angeles, CA 90089-2921, United States
b
5 Division of Occupational Science and Occupational Therapy, University of Southern California, 3620A McClintock Avenue, Los Angeles, CA 90089-2921, United States
c
6 Neuroscience Graduate Program, University of Southern California, 3620A McClintock Avenue, Los Angeles, CA 90089-2921, United States

O
7

a r t i c l e i n f o a b s t r a c t

R
8
9 Article history: Observing actions performed by others engages one's own sensorimotor regions, typically with greater activ- 22

P
10 Accepted 17 December 2012 ity for actions within one's own motor abilities or for which one has prior experience. However, it is unclear 23
11 Available online xxxx how experience modulates the neural response during the observation of impossible actions, beyond one's 24
13
12
14 own abilities. Using fMRI, we scanned typically-developed participants as they observed actions performed 25
15 Keywords:
by a novel biological effector (the residual limb of a woman born without arms) and a familiar biological ef-
D 26
16 Action understanding
17 Visual experience
fector (a hand). Participants initially demonstrated greater activity in the bilateral inferior and superior pari- 27
18 Body representation etal cortices when observing actions made by the residual limb compared to the hand, with more empathic 28
participants activating the right inferior parietal lobule, corresponding to the posterior component of the ac-
E
19 Shared circuits 29
20 Action observation network tion observation network, more strongly. Activity in the parietal regions may indicate matching the kinemat- 30
21 Mirror neurons ics of a novel effector to one's own existing sensorimotor system, a process that may be more active in more 31
T

empathic individuals. Participants then received extended visual exposure to each effector, after which they 32
showed little difference between activation in response to residual limb compared to hand actions, only in 33
the right superior parietal lobule. This suggests that visual experience may attenuate the difference between
C

34
how residual limb and hand actions are represented using one's own body representations, allowing us to 35
flexibly map physically different others onto our own body representations. 36
E

© 2012 Elsevier Inc. All rights reserved. 37

41 39
38
R

40

42 Introduction activity when observing individuals more similar to ourselves (Buccino 56

R

et al., 2004; Molnar-Szakacs et al., 2007) or actions with which we are 57

43 When observing actions performed by another person, regions of more familiar (Calvo-Merino et al., 2005; Cross et al., 2006). Thus, we 58
one's own sensorimotor system become active, engaging specific corti- may utilize our own motor representations to help understand other's
O

44 59
45 cal motor representations that correspond to the observed body parts actions (Keysers and Gazzola, 2007). How then do we understand ac- 60
46 (Buccino et al., 2001; Fadiga et al., 1995). This ‘motor resonance’ be- tions made with a body that differs from our own? 61
C

47 tween observed actions and one's own motor representation occurs in Recent studies demonstrate that we may also engage motor regions 62
48 a network of regions in the inferior frontal gyrus, ventral premotor cor- when observing actions beyond our own abilities (Aziz-Zadeh et al., 63
49 tex and inferior parietal lobule collectively referred to as the action ob- 2011; Liew et al., 2011). However, these studies used actions or effec- 64
N

50 servation network (AON; Caspers et al., 2010; Rizzolatti and Craighero, tors that were visually familiar to the observer, and prior studies have 65
51 2004). This network in humans may be related to mirror neurons, shown that visual experience, even without motor practice, can still 66
U

52 which were discovered in macaque monkeys as neurons that fired allow one to incorporate novel actions into one's own motor repertoire. 67
53 both when the monkey performed and observed actions (Gallese et For instance, individuals demonstrate increased sensorimotor activity 68
54 al., 1996; Rizzolatti et al., 1996). Evidence suggests that our own when observing actions they have either performed or seen before 69
55 motor experiences affect how we activate this network, with increased (Cross et al., 2009). Even motor imagery training of phantom limbs in 70
amputees led to changes in perceived ownership and kinematic quali- 71
⁎ Corresponding author at: Brain and Creativity Institute, University of Southern ties of the phantom limb, suggesting that even imagined training of a 72
California, 3620A McClintock Avenue, Los Angeles, CA 90089-2921, United States. non-existent limb can affect sensorimotor representations for that 73
Fax: + 1 213 821 3099. limb (Moseley and Brugger, 2009). However, what remains to be ex- 74
E-mail addresses: sliew@usc.edu, liews@mail.nih.gov (S.-L. Liew), tsheng@usc.edu plored is how we process actions made by individuals with novel 75
(T. Sheng), lazizzad@usc.edu (L. Aziz-Zadeh).
1
Present address: Human Cortical Physiology and Stroke Neurorehabilitation Sec-
body parts that we do not have and have not seen or imagined before, 76
tion, National Institute of Neurological Disorders and Stroke, NIH, 10 Center Drive, Bg and the role of experience in modulating these responses. These ques- 77
10 Rm 7D50, Bethesda, MD 20892. Fax: +1 301 402 7010. tions hold important implications not only for the scientific community 78

1053-8119/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.neuroimage.2012.12.028

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028
 2 S.-L. Liew et al. / NeuroImage xxx (2013) xxx–xxx

79 but also for our increasingly diverse society. In 2007, over 1.7 million in- Catch trials consisted of a red frame outlining an image of a hand, 140
80 dividuals in the United States alone had limb differences such as ampu- presented for 2 s, indicating that participants should press the button 141
81 tations (Center, 2011), and many more have other uncommon physical boxes they were given. Participants were informed that the button- 142
82 differences. Such individuals cite perceived social stigma as a major bar- press stimuli were to ensure they paid careful attention to the stimuli 143
83 rier to participating in their communities, affecting their quality of life throughout each run. A fixation cross was presented during rest trials 144
84 (Frank, 2000; Murray, 2009). Given that the average typically devel- with a randomly varied duration between 2 and 8 s. 145
85 oped individual has limited exposure to individuals with physical differ-
86 ences, can experience change how we represent bodies unlike our Visual exposure runs 146
87 own? To provide participants with increased visual exposure to both 147
88 To answer these questions, we used functional magnetic resonance hand and residual limb actions, visual exposure stimuli consisted of 148
89 imaging (fMRI) to scan participants who had no prior experience 16-second blocks of short 4-second video clips of different actions of 149
90 interacting with or observing with individuals with amputations (Nov- each effector, cropped to provide more of the body and context for 150
91 ices) as they observed a woman with bilateral above-elbow arm ampu- each of the actions. Hand visual exposure observations included actions 151
92 tations perform actions with her residual upper limb, which extend such as a hand and arm twisting off a bottle cap and threading a needle 152

F
93 several inches past her shoulder, and a typically developed woman per- through a piece of fabric, while residual limb visual exposure actions in- 153
94 form the same actions with her hand as a control. During the scanning cluded using the residual limb to push objects and using the residual 154

O
95 session, individuals passively observed actions made by both effectors, limb plus mouth to manipulate a pencil. These videos were intended 155
96 then received extended visual exposure to each, and then were re- to provide subjects with visual experience of each effector in a more 156
97 scanned during action observation. We hypothesized that after visual natural, realistic setting and to to provide the most context for the 157

O
98 exposure to the novel effector (e.g., residual limb), individuals would unique kinematics of each effector. Thus, actions during this run dif- 158
99 demonstrate activation in the same regions they engaged when observ- fered slightly between typically developed and congenital amputee ac- 159
ing a familiar effector (e.g., hand), resulting in no difference between re- tresses in order to highlight the kinematic range of each effector. fMRI

R
100 160
101 sidual limb and hand actions in the post-experience scan. data was collected during this run but excluded from the current anal- 161
ysis as it falls beyond the scope of the current investigation. 162

P
102 Materials and methods
AON localizer run 163
103 Participants D In order to identify neural regions that were active both during ac- 164
tion observation and during action execution, participants performed 165
104 Nineteen healthy, typically-developed participants (9 females, 10 one AON localizer run at the end of the scanning session. This entailed 166
105 males; mean± SD= 24.8 ±4.8 years) who had minimal to no prior ex- observing 3-second videos of hands picking up objects (e.g., keys, a 167
E
106 perience with individuals with amputations as assessed by a self-report mug), still images of a hand next to an object, rest trials with a fixa- 168
107 questionnaire were recruited for this study. In specific, participants tion cross, and action execution trials. Action execution trials were 169
T

108 were asked to state how much contact they have had with individuals cued by a red box flashing briefly for 500 ms before a static image 170
109 with amputations, if any, and to specify the amount (daily, weekly, of a hand was presented for the remaining 2500 ms. This cued partic- 171
C

110 monthly, yearly or less) and type of contact (friend, family member, ac- ipants to perform a basic hand action (see below) for the duration of 172
111 quaintance, or other). Individuals who had less than yearly contact with the clip. 173
112 an acquaintance or stranger were included in the study. Due to techni-
E

113 cal difficulties, data from three novice participants was incomplete and Task design and procedure 174
114 excluded from the study, resulting in 16 novice participants total.
R

115 Amount of experience for all participants was briefly quantified during Participants were provided short training runs outside of the scan- 175
116 the initial screening and further elaborated upon with an extensive be- ner, prior to the scanning session (for a complete schematic of the ex- 176
117 havioral questionnaire after the fMRI scanning procedure. Detailed perimental paradigm, see Fig. S1). For action observation runs, 177
R

118 questions were not asked prior to the fMRI experiment to avoid biasing participants were asked to watch the actions performed on the screen 178
119 participants to the goal of the study. All participants were right-handed, and pay attention to the movements and actions that they saw. They 179
O

120 had normal or corrected-to-normal vision, and were safe for fMRI. Writ- were further informed that at the end of each run, they might be 180
121 ten informed consent was obtained from all participants before inclu- asked “What was the last action you saw?” During the action observa- 181
122 sion in the study. This study was approved by the University of tion training, they were shown 4 clips each of hand actions and residual 182
C

123 Southern California Institutional Review Board and was performed in limb actions to familiarize them with the format of stimulus presenta- 183
124 accordance with the 1964 Declaration of Helsinki. tion and the effectors they would be seeing in the scanner, and to 184
train them to respond to the catch trials. An additional motive of these
N

185
125 Stimuli clips was to lessen any initial emotional or attentional effects the unfa- 186
miliar effector may evoke. Action observation runs consisted of 16 trials 187
U

126 Action observation runs of each condition, plus a randomly varied rest period between 2 and 8 s. 188
127 Stimuli consisted of 2-second video clips of goal-oriented, goal- The trial order of the design was then optimized using a genetic algo- 189
128 matched actions performed by an individual with physical differences rithm (Wager and Nichols, 2003). The total run time was 5 min and 190
129 using her upper residual limb and typically developed women using 36 seconds (168 TRs), for each of 4 runs. For the analyses, the first 191
130 their hands (see Fig. 1 for the experimental scanning paradigm and ac- two action observation runs (PRE) and the second two action observa- 192
131 tion observation example run). Hand action observation (HAO) and re- tion runs (POST) were averaged for a total of 32 trials per condition in 193
132 sidual limb action observation (RLAO) clips both contained the same set each analysis. 194
133 of actions (e.g., flip book page, squeeze binder clip, press down book After two action observation runs (PRE), there was a visual expo- 195
134 crease, flick rubber ball), performed by the right hand or right residual sure run, during which participants observed longer video clips (16 s 196
135 limb respectively. These actions were chosen because they were each) consisting of several actions, followed by longer rest trials with 197
136 goal-oriented and possible for both the actress with hands and the ac- a fixation cross (12 s). Following this run, participants observed an- 198
137 tress with residual limbs to perform using only one effector. Control other two action observation runs (POST). 199
138 stimuli consisted of still images of the hand (hand still; HS) or upper re- For the AON localizer run, participants were asked to watch the 200
139 sidual limb (residual limb still; RLS) and were also presented for 2 s. actions performed on the screen and pay attention to the movements 201

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028
 S.-L. Liew et al. / NeuroImage xxx (2013) xxx–xxx 3

F
O
O
R
Fig. 1. Experimental scanning paradigm and example action observation run. A) The experimental scanning session included the acquisition of a T1-weighted anatomical image,
followed by the pre-exposure action observation runs, the visual exposure run, then the post-exposure action observation runs (identical to pre-exposure action observation

P
runs), and finally an action observation network localizer, during which participants both observed and executed right hand actions. Additional full experimental paradigm de-
tails can be found in Fig. S1. B) An example action observation run, which included observing hand actions, residual limb actions, hand still images, residual limb still images, and
rests of random duration in an event-related design. Each run was presented twice in the pre-exposure session, and twice in the post-exposure session.
D
202 and actions that they saw. When they saw the red frame indicating an optimized event-related design, with 30 trials of each observation con- 237
203 action execution trial, they were asked to move their right hand as dition (AO, S) and 15 trials of action execution, as execution tends to 238
E
204 though picking up a wine glass several times, for the duration of the produce a more robust signal, thus requiring less trials. This run lasted 239
205 clip (generally 2–3 times). They were further instructed to remain 492 s (246 TRs). A complete schematic of the experimental design can 240
T

206 still for all other times. Once in the scanner, participants were moni- be found in Fig. S1. 241
207 tored for extraneous movements via an MRI-safe mirror placed next
C

208 to the scanner bed, which allowed for monitoring of hand movements
Image acquisition 242
209 from the control room. Non-task-related movements were monitored
210 by an experimenter during the stimulus presentation, with no subject
E

All images were acquired using a Siemens MAGNETOM Trio 3 T MRI 243
211 demonstrating non-task-related movement greater than 1% of trials.
scanner with standard head coil. A high resolution T1-weighted ana- 244
212 After the scanning session, participants completed a demographics
tomical volume was acquired from each participant (176 slices,
R

245
213 questionnaire with in-depth questions about their prior experiences
256 × 256 × 208 matrix with a spatial resolution of 1 × 1 × 1 mm, TR = 246
214 with individuals with physical differences and impressions of the ob-
1950 ms, TE= 2.26 ms, FOV= 256 mm; flip angle = 9°). Functional 247
served videos. They also responded on a 10-point Likert scale to ques-
R

215
volumes were acquired while participants performed the action obser- 248
216 tions about their familiarity with both residual limb and hand actions,
vation, visual exposure, and AON localizer runs. Thirty-seven axial slices 249
217 whether they felt watching the videos helped them to understand
of functional images covering the whole brain were acquired using a
O

250
218 hand/residual limb actions better, and whether they felt watching the
gradient-echo echo-planar pulse sequence (64 × 64× 37 matrix with a 251
219 videos made them any more likely to interact with an individual
spatial resolution of 3.5 × 3.5× 3.5 mm, TR= 2000 ms, TE = 30 ms, 252
220 with hands/residual limbs (1= very unfamiliar; unhelpful; unlikely to
C

FOV= 224 mm, flip angle = 90°). Functional volumes were acquired 253
221 interact; 10= very familiar; helpful; likely to interact). Finally, partici-
using Siemens' prospective acquisition correction (PACE) technique 254
222 pants also completed the Interpersonal Reactivity Index (IRI; (Davis,
for motion correction in which head movements were calculated by 255
N

223 1983)), a self-report behavioral measure of both cognitive and emo-

comparing successively acquired volumes and were corrected online 256
224 tional empathy.
(Thesen et al., 2000). 257
U

225 Scanning procedure Data processing and analyses 258

226 The images were presented through a projector onto a rear- Functional data processing was carried out using FEAT (FMRI Expert 259
227 projection screen attached to the head coil and located above the sub- Analysis Tool) Version 5.98, part of FSL (FMRIB's Software Library, www. 260
228 ject's head. As described previously, participants' actions were moni- fmrib.ox.ac.uk/fsl). The following pre-statistics processing were applied 261
229 tored via an MRI-safe mirror placed adjacent to the scanner bed, to individual subjects: motion correction using MCFLIRT (Jenkinson et 262
230 allowing the experimenter to observe movements from the control al., 2002), slice-timing correction using Fourier-space time-series 263
231 room. The experiment utilized an event-related design in the Action phase-shifting; non-brain removal using BET (Smith, 2002), spatial 264
232 Observation runs in which all conditions (HAO, RLAO, HS, RLS, rest) smoothing using a Gaussian kernel of FWHM 5 mm, grand-mean inten- 265
233 were evenly distributed across 4 runs, which lasted 336 s (168 TRs). sity normalisation of the entire 4D dataset by a single multiplicative fac- 266
234 The Visual Exposure run utilized a block design with 16-second blocks, tor, and highpass temporal filtering (Gaussian-weighted least-squares 267
235 counterbalanced across participants, with 12-second rests in between straight line fitting, with sigma= 65.0 s) (Jenkinson et al., 2002; 268
236 and lasted 464 seconds (232 TRs). The AON Localizer run utilized an Smith, 2002). For each subject, a time-series statistical analysis was 269

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028
 4 S.-L. Liew et al. / NeuroImage xxx (2013) xxx–xxx

270 carried out using FILM GLM with local autocorrelation correction (MT/V5 into V1). In the direct contrast between Residual Limb and 331
271 (Woolrich et al., 2001). Z (Gaussianised T/F) statistic images were then Hand Action Observation (RLAO >HAO), there was activity in the bilat- 332
272 thresholded at p= 0.001 (uncorrected), and registered to a high resolu- eral IPL including the supramarginal gyrus, postcentral gyrus, and 333
273 tion standard space image (2 × 2 ×2 mm3 Montreal Neurological Insti- extending into the superior parietal lobules and the posterior middle 334
274 tute (MNI) space) using FLIRT (FSL's Linear Image Registration Tool) temporal gyrus (MT/V5; see Fig. 2). Hand versus Residual Limb Action 335
275 (Jenkinson and Smith, 2001; Jenkinson et al., 2002). A second-level anal- Observation (HAO > RLAO) generated activity in the bilateral occipital 336
276 ysis to average across the two runs in PRE and POST conditions respec- poles (BA 17/18) only (see Fig. 2). 337
277 tively was carried out using a fixed effects model, by forcing the random
278 effects variance to zero in FLAME (FMRIB's Local Analysis of Mixed Ef- Action observation POST-visual exposure 338
279 fects) (Beckmann et al., 2003; Woolrich, 2008; Woolrich et al., 2004). After the Visual Exposure run, participants viewing Residual Limb 339
280 Group-level analyses were then completed using FLAME stage 1, Action Observation versus Residual Limb Still images (RLAO >RLS) acti- 340
281 which employed a mixed effects model that includes both fixed effects vated the bilateral inferior and superior parietal lobules and bilateral oc- 341
282 and random effects from cross session/subject variance (Beckmann et cipital cortices from MT/V5 extending into V1, similar to the pre-visual 342
283 al., 2003; Woolrich, 2008; Woolrich et al., 2004). Z (Gaussianised T/F) exposure runs. Hand Action Observation versus Hand Still images 343

F
284 statistic images at this level were thresholded using clusters deter- (HAO>HS) resulted in a similar pattern of activity with clusters of ac- 344
285 mined by Z> 2.3 and a (corrected) cluster significance threshold of tivity in bilateral superior parietal regions, left inferior parietal lobule, 345

O
286 P = 0.05 (Worsley, 2001). To test for an interaction between observing right posterior superior temporal sulcus into the inferior parietal lobule, 346
287 Residual Limb versus Hand actions, between PRE versus POST condi- and strong bilateral occipital activation (MT/V5 into V1). Residual Limb 347
288 tions, lower-level feat directories for each subject were entered into a versus Hand Action Observation (RLAO > HAO) in the POST run demon- 348

O
289 2 × 2 repeated measures ANOVA, with factors of Effector (Residual strated activity in the right superior parietal lobule and bilateral poste- 349
290 Limb, Hand) and Experience (Pre, Post). rior middle temporal gyri (MT/V5; see Fig. 2). In contrast, Hand versus 350
Region of interest analyses were also performed for a priori re- Residual Limb Action Observation (HAO > RLAO) again generated activ-

R
291 351
292 gions in the AON (bilateral IFG/PMv and IPL). These regions were de- ity in the bilateral occipital poles (BA 17/18) only (see Fig. 2). 352
293 fined by the overlap between action observation and action execution

P
294 during the AON localizer run. As this overlap included many regions ANOVA results 353
295 outside of the canonical AON, including large occipital and temporal The 2 × 2 repeated measures ANOVA with factors of Effector (Re- 354
296 lobe clusters, we further constrained the functional overlap data sidual Limb, Hand) and Experience (Pre, Post) demonstrated a signif-
D 355
297 with anatomical definitions of the AON based on the probabilistic icant simple effect, with Pre > Post RLAO activating the right inferior 356
298 Harvard–Oxford atlas, using the bilateral IFG/PMv and IPL. The parietal lobule, with none of the other conditions (Pre > Post HAO, 357
299 resulting AON ROIs are displayed in Fig. S2. IFG and ventral premotor Post > Pre RLAO, Post > PRE HAO) demonstrating significant effects. 358
E
300 regions were combined into one region of interest as prior meta- There was also a main effect of RLAO, also in the right inferior parietal 359
301 analyses of the AON suggest that both comprise the frontal compo- lobule, overlapping with the simple effect (see Fig. S3, Table S3 for ac- 360
T

302 nent of the AON (Van Overwalle and Baetens, 2009). Percent signal tivation patterns and coordinates). There was no significant interac- 361
303 change (%SC) for the observation of each effector (HAO/RLAO) com- tion effect between effector and experience. 362
C

304 pared to the control still image (HS/RLS) was then extracted using
305 Featquery in FSL and contrasted in paired t-tests. Finally, correlation Correlations 363
306 analyses between the %SC values from the ROI analyses and scores
E

307 on the Interpersonal Reactivity Index empathy questionnaire and Percent signal change in AON regions of interest positively correlat- 364
308 the demographics questionnaire were run in SPSS (Release Version ed with several scores on the Interpersonal Reactivity Index empathy 365
R

309 18.0, © SPSS, Inc., 2009, Chicago, IL, www.spss.com). subscales. Participants in the PRE condition demonstrated a significant 366
positive correlation between activity in the R IPL during RLAO > RLS 367
310 Results with both the empathic concern (IRI EC; r = .66, p = .006; see Fig. 3) 368
R

and the perspective taking (IRI PT) subscales (r= .54, p = .03). This 369
311 Behavioral results also occurred at the R IPL for the contrast RLAO>HAO with empathic 370
O

concern (IRI EC; r = .51, p = .04). There were no significant correlations 371
312 Participants rated being significantly more likely to interact with for novices in the POST condition with empathy. 372
313 an individual with an upper residual limb after watching the videos
C

314 of residual limbs and hands, as compared to likeliness to interact Discussion 373
315 with an individual with typically developed hands (RL: 5.69 ± 2.52,
H: 4.43 ± 2.61, t = -2.21, p = .043). In addition, the videos helped par- The current study examined how we process actions made by bodies
N

316 374
317 ticipants to understand residual limb actions significantly more than different from our own and how this pattern of activation changes with 375
318 they helped participants to understand hand actions (RL: 8.13 ± experience. Novice participants, who had little to no prior experience 376
U

319 1.93, H: 5.188 ± 2.95, t = -5.74, p b .00005). with amputees, observed an individual with residual limbs and an indi- 377
vidual with hands perform actions, before and after visual exposure 378
320 fMRI results during an fMRI session. The results show that novices initially engage 379
their own sensorimotor regions, particularly in the right inferior parie- 380
321 Action observation PRE-visual exposure tal lobule, more when observing individuals with different bodies, de- 381
322 During the PRE-visual exposure action observation runs, Residual spite the fact that the observed effector has not been seen before and 382
323 Limb Action Observation versus Residual Limb Still images (RLAO >RLS) lacks corresponding motor representations in the observer's body. In 383
324 activated the right dorsal and ventral premotor cortices, the bilateral in- addition, more empathic novices demonstrated an increased sensori- 384
325 ferior and superior parietal lobules, and the bilateral lateral occipital motor response when observing residual limbs than less empathic indi- 385
326 cortices from the posterior middle temporal gyrus (MT/V5) into the me- viduals. Visual experience introduced during the experiment led to an 386
327 dial lingual gyri (BA 17/18; see SI Tables 1–3 for all coordinates of signif- attenuation of the difference in neural signal when viewing residual 387
328 icant clusters of activation). Hand Action Observation versus Hand Still limb actions after experience as compared to hands. This evidence sug- 388
329 images (HAO > HS) generated a similar pattern of activity in bilateral in- gests that moderate amounts of visual experience can reduce differ- 389
330 ferior and superior parietal lobules and the bilateral occipital cortices ences between how body parts one does and does not have are 390

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028
 S.-L. Liew et al. / NeuroImage xxx (2013) xxx–xxx 5

F
O
O
R
P
D
E
T

Fig. 2. fMRI results when participants observe residual limb and hand actions for the first time (PRE) and after visual exposure (POST). TOP — PRE. ORANGE: Residual Limb > Hand
activated sensorimotor regions, including the bilateral inferior and superior parietal lobules, and occipital regions. BLUE: Hand > Residual Limb generated significant activation in
C

the bilateral occipital poles. BOTTOM — POST. ORANGE: Residual Limb > Hand activated the right superior parietal lobule and occipital regions, including MT/V5. BLUE:
Hand > Residual Limb generated significant activation in the bilateral occipital poles. All results thresholded at Z > 2.3, p b 0.05 (cluster corrected for multiple comparisons).
E
R
R
O
C
N
U

Fig. 3. Correlation between percent signal change in right inferior parietal lobule during residual limb observation and empathic concern scores. Percent signal change in the R IPL
when novices observed Residual Limb Action Observation> Residual Limb Still (RLAO > RLS) in the PRE condition correlated with Empathic Concern scores on the Interpersonal
Reactivity Index (r= .66, p = .006). Correlations also occurred between the R IPL and the Perspective Taking subscale the IRI, and the R IPL and Empathic Concern during
RLAO >HAO.

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028
 6 S.-L. Liew et al. / NeuroImage xxx (2013) xxx–xxx

391 represented. Taken together, these results demonstrate the flexible en- Expanding on this literature, it appears that the AON not only is 454
392 gagement of one's own sensorimotor regions in representing novel and more active when observing novel versus familiar effectors, but it is 455
393 experienced actions during observation. also more active when observing a novel effector during the perfor- 456
mance of a familiar goal-directed action. This finding is in complement 457
to recent work, which shows that observing novel compared to familiar 458
394 Increased sensorimotor activity for the novel effector gestures also increases activation in one's own parietal regions (Liew et 459
al., 2011). In this case, the goal of the action – the end hand-state creat- 460
395 Observing the novel effector activated one's own sensorimotor re- ing the gesture – is unfamiliar, but the effector (e.g., hand) is familiar. 461
396 gions more than observing the familiar effector. This challenges the tra- Another study demonstrated increased activation of both premotor 462
397 ditional notion that we activate our own sensorimotor regions more for and parietal regions when observing robot-like compared to human- 463
398 actions within our motor repertoire, as found in expert ballet dancers like movements (Cross et al., 2011), where both goal of the action (a ro- 464
399 watching ballet compared to capoeria or individuals watching dance botic pose versus a human pose) and the kinematics of the effector are 465
400 moves they had rehearsed versus not rehearsed (Calvo-Merino et al., novel. Finally, an additional study found activity in premotor, parietal, 466
401 2005; Cross et al., 2006, 2009). These studies concluded that action ob- and dorsolateral prefrontal regions when observing novel versus famil- 467

F
402 servation engages one's existing motor repertoire to understand others' iar guitar patterns, which the authors suggested reflects an increased ef- 468
403 actions, with more activation when observing actions for which one has fort towards observational learning (Vogt et al., 2007). We propose that 469

O
404 a more detailed motor representation. However, here we find the oppo- similar processes are occurring here, with participants mapping the ob- 470
405 site, with more activation when observing actions for which one has less served novel actions or observed novel effectors, onto their own senso- 471
406 of a motor representation. This occurs despite the fact that the novel ac- rimotor representations, possibly though through covert imitation, to 472

O
407 tions are not only beyond one's current motor abilities, but also beyond better learn and understand the kinematics and affordances of the 473
408 one's future capabilities, as they are performed by a body part that one novel limb. In support of this, the behavioral results demonstrate that 474
does not possess. novices report understanding the residual limb better after observing

R
409 475
410 Increased activation occurred in both the right inferior and superi- the video stimuli—a finding that warrants more rigorous exploration 476
411 or parietal lobules when Novices observed the novel versus familiar in future studies. Overall, this data suggests that one's own sensorimo- 477

P
412 effector, with the right inferior parietal lobule appearing robustly. To- tor regions are flexibly engaged to support action understanding of 478
413 gether, these regions are associated with a multitude of tasks, includ- many different types of actions and effectors, both within one's realm 479
414 ing directing attention, spatial integration of visuomotor information, of expertise and beyond what one has ever seen before.
D 480
415 and encoding the kinematics and orientation of movements within
416 body space (Colby and Goldberg, 1999; Oztop et al., 2006; Rizzolatti Increased right parietal activity correlates with empathy in novices 481
417 et al., 1997). We propose that novice individuals engage their own
E
418 sensorimotor regions in an attempt to match the unfamiliar kinemat- If understanding another's actions relies in part on our ability to 482
419 ics of the observed, novel actions to a corresponding body part in engage our own corresponding sensorimotor regions, then individ- 483
T

420 their own motor schema as one way of better understanding it. Im- uals who are more empathic may also engage their own body repre- 484
421 portantly, while the observed actions and effector were novel, the sentations more when observing others. This has been demonstrated 485
C

422 goal of the actions were familiar (e.g., flipping a book page), and in a number of studies, with greater sensorimotor activity correlating 486
423 thus participants may have a motor schema for how they themselves with higher scores on trait empathy, such as perspective taking 487
424 might perform the action. They may thus use their existing knowl- (Aziz-Zadeh et al., 2010; Gazzola et al., 2006; Kaplan and Iacoboni, 488
E

425 edge of motor kinematics, based on their own bodily experiences, to 2006). These studies suggest that more empathic individuals engage 489
426 map the affordances of the residual limb onto their own bodies. In their own motor regions more than less empathic individuals when 490
R

427 doing so, they may dissect the kinematics of each part of the effector perceiving others' actions. Similarly, here we find that participants 491
428 onto the closest match in their own body, matching the novel limb ei- who score higher on either cognitive or affective measures of empa- 492
429 ther to the part they themselves would use to achieve the goal of the thy demonstrate more right inferior parietal activation when observ- 493
R

430 action or matching it to the closest kinematically-similar effector — in ing residual limb actions. This correlation is especially notable when 494
431 this case, matching the residual limb to one's own lower or upper arm novices initially view the novel effector, and is no longer significant 495
O

432 (Arbib et al., 2009; Bonaiuto and Arbib, 2010). However, stronger ev- once the effector becomes more familiar. This suggests that our ability 496
433 idence to dissociate how exactly the novel limb is being mapped onto to map the actions of unfamiliar, dissimilar others onto our own sen- 497
434 one's own body would necessitate finer mapping of the sensorimotor sorimotor representations may support our capacity to understand 498
C

435 system in each observer, using a localizer to identify representations and empathize with individuals who are different from ourselves. 499
436 for various points along the limb, which may be a promising area of Put another way, the more empathic an individual is, the more he 500
future inquiry. or she may attempt to represent very different bodies onto their
N

437 501
438 Prior research suggests that this kinematic match does not need to own, as though literally “putting oneself in another's shoes.” 502
439 be perfect, however. Motor training of incongruent motor perfor-
U

440 mance and action observation (e.g., observing finger flexion while Visual experience represents novel effectors similar to familiar ones 503
441 performing finger extension; Catmur et al., 2007), observation of bio-
442 mechanically impossible actions (Romani et al., 2005), and even ob- After visual exposure to the novel limb, the contrast between ob- 504
443 servation of actions performed by a tool versus by a hand (Van Elk serving residual limb actions compared to hand actions only yielded 505
444 et al., 2011), can all still result in greater motor excitability for the activity in the right superior parietal and occipitotemporal regions, 506
445 paired or observed effector, regardless of whether it is possible or im- with none of the robust AON regions active as found in the PRE con- 507
446 possible, a biological effector or a tool. These studies suggest that sen- dition. This suggests that after sufficient experience with an observed 508
447 sorimotor activation is not effector-specific and that activation of effector, one may effectively map the kinematics of a novel effector 509
448 one's own motor representations can be elicited by the observation onto one's own existing system. Subsequent observation of the effec- 510
449 of body movements or tools that are beyond our own motor reper- tor thus no longer requires extensive parietal activation to generate a 511
450 toire. Similar to these prior studies, our results suggest that activation kinetic representation of it, and can thus be represented onto similar 512
451 of the AON may occur during goal-directed movements in which the regions as more familiar effectors — in this case, such as hand. The 513
452 goal of the action is familiar to the observer, such as squeezing a bind- right superior parietal activation seen in the POST condition may be 514
453 er clip or turning a book page, even if the effector is not. indicative of participants updating their internal model of the 515

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028
 S.-L. Liew et al. / NeuroImage xxx (2013) xxx–xxx 7

516 different effector, fine-tuning their model with specific kinematic pa- the Dana and David Dornsife Neuroimaging Center; and the Brain and 576
517 rameters observed in the second session (Iacoboni and Zaidel, 2004; Creativity Institute. 577
518 Wolpert et al., 1998). In addition, the different effector may also
519 evoke more visual activation, particular in regions of the lateral occip-
520 ital cortex associated with biological movement. In the observed Acknowledgments 578
521 study, there was no interaction between effector and experience,
522 suggesting that the reduction in signal when observing the residual We thank Dr. Michael Arbib, Dr. Jonas Kaplan, Dr. Florence Clark, 579
523 limb actions after experience may also occur when observing hand Dr. Ann Neville-Jan, and Mona Sobhani for insightful discussions 580
524 actions, although to a lesser degree. and Mustafa Seckin, Alicia Johnson, and Ahra Roh for their assistance 581
525 Altogether, this suggests that the visual experience introduced in with conducting this study. 582
526 the experiment allows the viewer to develop similar neural patterns
527 for novel, different bodies as for familiar, similar ones. Experience Appendix A. Supplementary data 583
528 with an observed effector and observed actions may thus modulate
529 how much we use our own sensorimotor representations to map Supplementary data to this article can be found online at http:// 584

F
530 the kinematics of other's actions. dx.doi.org/10.1016/j.neuroimage.2012.12.028. 585

O
531 Future directions
References 586
532 Experience may modulate the sensorimotor response by driving

O
Arbib, M.A., Bonaiuto, J.B., Jacobs, S., Frey, S.H., 2009. Tool use and the distalization of 587
533 attention to both the more novel effector and the more ‘expert’ ac- the end-effector. Psychol. Res. 73, 441–462. 588
534 tions, and then to both effectors equally after novelty wears off, either Aziz-Zadeh, L., Sheng, T., Gheytanchi, A., 2010. Common premotor regions for the per- 589
through visual or real-life experience. Studies directly examining the ception and production of prosody and correlations with empathy and prosodic 590

R
535
ability. PLoS One 5, e8759. 591
536 effects of attention on motor resonance during action observation Aziz-Zadeh, L., Sheng, T., Liew, S.L., Damasio, H., 2011. Understanding otherness: the 592
537 suggest that selective attention increases motor resonance (Bach et neural bases of action comprehension and pain empathy in a congenital amputee. 593

P
538 al., 2007; Chong et al., 2009). Here we propose that the observed ef- Cereb. Cortex http://dx.doi.org/10.1093/cercor/bhr139. 594
Bach, P., Peatfield, N.A., Tipper, S.P., 2007. Focusing on body sites: the role of spatial at- 595
539 fects are the result of a complex interplay between 1) attention and
tention in action perception. Exp. Brain Res. 178, 509–517. 596
540 2) one's existing motor abilities based on prior experiences. It is likely DBeckmann, C.F., Jenkinson, M., Smith, S.M., 2003. General multilevel linear modeling for 597
541 that these two factors modulate sensorimotor activity together, and group analysis in FMRI. Neuroimage 20, 1052–1063. 598
Bonaiuto, J., Arbib, M.A., 2010. Extending the mirror neuron system model, II: What did 599
542 this proposed interplay between experience and attention may also
I just do? A new role for mirror neurons. Biol. Cybern. 102, 341–359. 600
543 explain the apparent contradiction between these findings and prior Buccino, G., Binkofski, F., Fink, G.R., Fadiga, L., Fogassi, L., Gallese, V., Seitz, R.J., Zilles, K.,
E
601
544 research. In particular, sensorimotor activity may increase for expert Rizzolatti, G., Freund, H.J., 2001. Action observation activates premotor and parietal 602
545 or practiced actions partially because individuals tend to pay more at- areas in a somatotopic manner: an fMRI study. Eur. J. Neurosci. 13, 400–404. 603
Buccino, G., Lui, F., Canessa, N., Patteri, I., Lagravinese, G., Benuzzi, F., Porro, C.A., 604
T

546 tention to actions within their own interests. Conversely, at other Rizzolatti, G., 2004. Neural circuits involved in the recognition of actions per- 605
547 times, we may pay more attention to things that are novel or different formed by nonconspecifics: An FMRI study. J. Cogn. Neurosci. 16, 114–126. 606
C

548 to us — such as a new body part, robotic movements, or uncommon Calvo-Merino, B., Glaser, D.E., Grezes, J., Passingham, R.E., Haggard, P., 2005. Action ob- 607
servation and acquired motor skills: an FMRI study with expert dancers. Cereb. 608
549 hand gestures (Cross et al., 2011; Liew et al., 2011). Future research Cortex 15, 1243–1249. 609
550 may explore the intricate interplay between our prior experiences Caspers, S., Zilles, K., Laird, A.R., Eickhoff, S.B., 2010. ALE meta-analysis of action obser- 610
E

551 and our current attentional focus in driving the sensorimotor re- vation and imitation in the human brain. Neuroimage 50, 1148–1167. 611
Catmur, C., Walsh, V., Heyes, C., 2007. Sensorimotor learning configures the human 612
552 sponse during when observing others’ actions. mirror system. Curr. Biol. 17, 1527–1531. 613
R

Center, N.L.L.I., 2011. Limb Loss FAQ's. 614

553 Conclusion Chong, T.T., Cunnington, R., Williams, M.A., Mattingley, J.B., 2009. The role of selective 615
attention in matching observed and executed actions. Neuropsychologia 47, 616
R

786–795. 617
554 The current work suggests that we understand other people's ac- Colby, C.L., Goldberg, M.E., 1999. Space and attention in parietal cortex. Annu. Rev. 618
555 tions by engaging our own motor regions in order to gain a simulated Neurosci. 22, 319–349. 619
O

Cross, E.S., Hamilton, A.F., Grafton, S.T., 2006. Building a motor simulation de novo: 620
556 form of motor experience, simply through observation. Here we ex-
observation of dance by dancers. Neuroimage 31, 1257–1267. 621
557 amine how experience modulates activity in neural regions associat- Cross, E.S., Kraemer, D.J.M., Hamilton, A.F.C., Kelley, W.M., Grafton, S.T., 2009. Sensitiv- 622
558 ed with action perception. While prior research suggested that we ity of the action observation network to physical and observational learning. Cereb. 623
C

559 activate our own sensorimotor regions more for actions we have Cortex 19, 315. 624
Cross, E.S., Liepelt, R., de, C.H.A.F., Parkinson, J., Ramsey, R., Stadler, W., Prinz, W., 2011. 625
560 practiced or experienced ourselves, the work presented here provides Robotic movement preferentially engages the action observation network. Hum. 626
N

561 evidence that we may also activate certain components of our own Brain Mapp. http://dx.doi.org/10.1002/hbm.21361. 627
562 sensorimotor regions, particularly in the right parietal lobule and Davis, M.H., 1983. Measuring individual differences in empathy: evidence for a multi- 628
dimensional approach. J. Pers. Soc. Psychol. 44, 113–126. 629
563 AON, when observing goal-directed movements that we have not Fadiga, L., Fogassi, L., Pavesi, G., Rizzolatti, G., 1995. Motor facilitation during action
U

630
564 seen before and cannot perform ourselves. Visual experience may observation: a magnetic stimulation study. J. Neurophysiol. 73, 2608–2611. 631
565 support a similar response to both possible actions and impossible ac- Frank, G., 2000. Venus on wheels: Two decades of dialogue on disability, biography, 632
and being female in America. University of California Press, Los Angeles, CA, p. 284. 633
566 tions as they become more familiar. Altogether, these findings dem- Gallese, V., Fadiga, L., Fogassi, L., Rizzolatti, G., 1996. Action recognition in the premotor 634
567 onstrate the flexibility of the human action observation network in cortex. Brain 119, 593–609. 635
568 representing actions beyond our own abilities and suggest a further Gazzola, V., Aziz-Zadeh, L., Keysers, C., 2006. Empathy and the somatotopic auditory 636
mirror system in humans. Curr. Biol. 16, 1824–1829. 637
569 role of these regions in acquiring and generating internal motor pat- Iacoboni, M., Zaidel, E., 2004. Interhemispheric visuo-motor integration in humans: the 638
570 terns for new, even impossible, actions through observation alone. role of the superior parietal cortex. Neuropsychologia 42, 419–425. 639
Jenkinson, M., Smith, S., 2001. A global optimisation method for robust affine registra- 640
tion of brain images. Med. Image Anal. 5, 143–156. 641
571 Funding
Jenkinson, M., Bannister, P., Brady, M., Smith, S., 2002. Improved optimization for the 642
robust and accurate linear registration and motion correction of brain images. 643
572 This work was supported by the National Science Foundation Grad- Neuroimage 17, 825–841. 644
573 uate Research Fellowship (grant number 2009072048 to S.-L.L.); the Kaplan, J.T., Iacoboni, M., 2006. Getting a grip on other minds: mirror neurons, inten- 645
tion understanding, and cognitive empathy. Soc. Neurosci. 1, 175–183. 646
574 University of Southern California Provost's PhD Fellowship (to S.-L.L.); Keysers, C., Gazzola, V., 2007. Integrating simulation and theory of mind: From self to 647
575 the Division of Occupational Science and Occupational Therapy; social cognition. Trends Cogn. Sci. 11, 194–196. 648

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028
 8 S.-L. Liew et al. / NeuroImage xxx (2013) xxx–xxx

649 Liew, S.L., Han, S., Aziz-Zadeh, L., 2011. Familiarity modulates mirror neuron and Thesen, S., Heid, O., Mueller, E., Schad, L.R., 2000. Prospective acquisition correction for 673
650 mentalizing regions during intention understanding. Hum. Brain Mapp. 32, head motion with image-based tracking for real-time fMRI. Magn. Reson. Med. 44, 674
651 1986–1997. 457–465. 675
652 Molnar-Szakacs, I., Wu, A.D., Robles, F.J., Iacoboni, M., 2007. Do you see what I mean? Van Elk, M., Van Schie, H.T., Bekkering, H., 2011. Imitation of hand and tool actions is 676
653 Corticospinal excitability during observation of culture-specific gestures. PLoS effector-independent. Exp. Brain Res. 214, 539–547. 677
654 One 2, e626. Van Overwalle, F., Baetens, K., 2009. Understanding others' actions and goals by mirror 678
655 Moseley, G.L., Brugger, P., 2009. Interdependence of movement and anatomy persists and mentalizing systems: a meta-analysis. Neuroimage 48, 564–584. 679
656 when amputees learn a physiologically impossible movement of their phantom Vogt, S., Buccino, G., Wohlschläger, A.M., Canessa, N., Shah, N.J., Zilles, K., Eickhoff, S.B., 680
657 arm. Proc. Natl. Acad. Sci. U. S. A. 106, 18798–18802. Freund, H.J., Rizzolatti, G., Fink, G.R., 2007. Prefrontal involvement in imitation 681
658 Murray, C.D., 2009. Being like everybody else: the personal meanings of being a pros- learning of hand actions: effects of practice and expertise. Neuroimage 37, 682
659 thesis user. Disabil. Rehabil. 31, 573–581. 1371–1383. 683
660 Oztop, E., Imamizu, H., Cheng, G., Kawato, M., 2006. A computational model of anterior Wager, T.D., Nichols, T.E., 2003. Optimization of experimental design in fMRI: a general 684
661 intraparietal (AIP) neurons. Neurocomputing 69, 1354–1361. framework using a genetic algorithm. Neuroimage 18, 293–309. 685
662 Rizzolatti, G., Craighero, L., 2004. The mirror-neuron system. Annu. Rev. Neurosci. 27, Wolpert, D.M., Goodbody, S.J., Husain, M., 1998. Maintaining internal representations: 686
663 169–192. the role of the human superior parietal lobe. Nat. Neurosci. 1, 529–533. 687
664 Rizzolatti, G., Fadiga, L., Gallese, V., Fogassi, L., 1996. Premotor cortex and the recogni- Woolrich, M., 2008. Robust group analysis using outlier inference. Neuroimage 41, 688
665 tion of motor actions. Cogn. Brain Res. 3, 131–141. 286–301. 689
666 Rizzolatti, G., Fogassi, L., Gallese, V., 1997. Parietal cortex: from sight to action. Curr. Woolrich, M.W., Ripley, B.D., Brady, M., Smith, S.M., 2001. Temporal autocorrelation in 690
Opin. Neurobiol. 7, 562–567. univariate linear modeling of FMRI data. Neuroimage 14, 1370–1386.

F
667 691
668 Romani, M., Cesari, P., Urgesi, C., Facchini, S., Aglioti, S.M., 2005. Motor facilitation of the Woolrich, M.W., Behrens, T.E., Beckmann, C.F., Jenkinson, M., Smith, S.M., 2004. 692
669 human cortico-spinal system during observation of bio-mechanically impossible Multilevel linear modelling for FMRI group analysis using Bayesian inference. 693

O
670 movements. Neuroimage 26, 755–763. Neuroimage 21, 1732–1747. 694
671 Smith, S.M., 2002. Fast robust automated brain extraction. Hum. Brain Mapp. 17, Worsley, K.J., 2001. Statistical analysis of activation images. Functional MRI: an intro- 695
672 143–155. duction to methods. 696

O
697
698

R
P
D
E
T
C
E
R
R
O
C
N
U

Please cite this article as: Liew, S.-L., et al., Experience with an amputee modulates one's own sensorimotor response during action
observation, NeuroImage (2013), http://dx.doi.org/10.1016/j.neuroimage.2012.12.028