ORIGINAL PAPER
Mental Imagery for Full and Upper Human Bodies: Common
Right Hemisphere Activations and Distinct Extrastriate
Activations
Olaf Blanke
•
Silvio Ionta
•
Eleonora Fornari
•
Christine Mohr
•
Philippe Maeder
Received: 3 November 2009 / Accepted: 12 February 2010
Ó Springer Science+Business Media, LLC 2010
Abstract The processing of human bodies is important
in social life and for the recognition of another person’s
actions, moods, and intentions. Recent neuroimaging
studies on mental imagery of human body parts suggest
that the left hemisphere is dominant in body processing.
However, studies on mental imagery of full human bodies
reported stronger right hemisphere or bilateral activations.
Here, we measured functional magnetic resonance imaging
during mental imagery of bilateral partial (upper) and full
bodies. Results show that, independently of whether a full
or upper body is processed, the right hemisphere (temporo-
parietal cortex, anterior parietal cortex, premotor cortex,
bilateral superior parietal cortex) is mainly involved in
mental imagery of full or partial human bodies. However,
distinct activations were found in extrastriate cortex for
partial bodies (right fusiform face area) and full bodies (left
extrastriate body area). We propose that a common brain
network, mainly on the right side, is involved in the mental
imagery of human bodies, while two distinct brain areas in
extrastriate cortex code for mental imagery of full and
upper bodies.
Keywords Mental rotation Body Neuroimaging
Out-of-body Neuropsychology fMRI
Introduction
Human bodies provide a particularly rich source of visual
social information. Behavioural, neuropsychological, and
neuroimaging studies suggest that processing human bodily
stimuli involves brain regions that are at least partially
different from those sub-serving the processing of non-
corporeal objects (Ionta et al. 2010; Wraga et al. 2005;
Creem et al. 2001; Bonda et al. 1995; Parsons 1987a, b).
With respect to neural mechanisms, neuropsychological
findings suggest that the left hemisphere might be domi-
nant for the processing of body parts (Schwoebel and
Coslett 2005; Guariglia et al. 2002; Sirigu et al. 1991;
Ogden 1985), although own body illusions and deficits in
corporeal awareness have been linked primarily to the right
hemisphere (Blanke and Mohr 2005; Berlucchi and Aglioti
1997). Investigating the differences between different
modalities of mental imagery (i.e., motor vs. visual), it has
been suggested that left brain regions are more involved in
motor imagery, whereas right brain regions are more
involved in visual imagery (Sirigu and Duhamel 2001;
Tomasino and Rumiati 2004). This has been extended by
neuroimaging work studying mental imagery in healthy
subjects and revealing a left-hemisphere dominance in
O. Blanke (&) S. Ionta
Laboratory of Cognitive Neuroscience, Brain-Mind Institute,
Ecole Polytechnique Fe
´
de
´
rale de Lausanne (EPFL), Station 19,
1015 Lausanne, Switzerland
e-mail: olaf.blanke@epfl.ch
O. Blanke C. Mohr
Department of Neurology, University Hospital, Geneva,
Switzerland
E. Fornari
Department of Radiology, CIBM-CHUV Unit, Centre
Hospitalier Universitaire Vaudois and University of Lausanne,
Lausanne, Switzerland
C. Mohr
Department of Experimental Psychology, University of Bristol,
Bristol, UK
P. Maeder
Department of Radiology, Centre Hospitalier Universitaire
Vaudois and University of Lausanne, Lausanne, Switzerland
123
Brain Topogr
DOI 10.1007/s10548-010-0138-x
body processing, showing the activation of superior pari-
etal lobule (SPL) and cortex at the intraparietal sulcus (IPS)
during mental imagery for human body parts (Overney and
Blanke 2008; Overney et al. 2005; de Jong et al. 2001;
Bonda et al. 1995). Others reported bilateral parietal acti-
vations (Kosslyn et al. 1998; Parsons et al. 1995). Acti-
vations at the temporo-parietal junction (TPJ) are generally
bilateral or have been found mainly in the right (Blanke
et al. 2005) or the left hemisphere (Zacks et al. 1999).
Finally, the majority of the studies that employed stimuli
depicting human body parts such as hands or arms reported
stronger left parietal activations (Overney and Blanke
2008; Overney et al. 2005; de Jong et al. 2001; Bonda et al.
1995), whereas studies depicting full bodies revealed more
bilateral (Zacks et al. 1999, 2002) or right parietal and/or
temporo-parietal activations (Blanke et al. 2005).
Given the important roles played by neural body rep-
resentations in mental imagery in the processing of human
bodies, we investigated the neural differences during
mental imagery of partial (upper) bodies and full bodies
using fMRI. Two main reasons guided this choice. First,
body stimuli in previous studies were generally presented
as a single left or right arm or hand, whereas studies on
mental imagery of full human bodies used bilateral body
stimuli. Using bilateral stimuli of the upper part of the
human body (Reed et al. 2006) would allow comparison
with mental imagery of bilateral stimuli of the full body.
Second, our study was motivated by differences in neural
coding for upper and full human bodies as suggested by
clinical evidence from neurological patients with illusory
own body perceptions such as autoscopic hallucinations
and out-of-body experiences (Blanke et al. 2005; Blanke
and Metzinger 2009). During both illusions patients expe-
rience seeing a second own body in extrapersonal space.
During out-of-body experiences the illusory body is usually
perceived as a full body, whereas during autoscopic hal-
lucination it is mostly the upper part of the body that is
perceived (Brugger 2002; Blanke and Mohr 2005). We
therefore adapted a classical mental imagery task (Parsons
1987a) that had already been employed in electrical neu-
roimaging work (Blanke et al. 2005; Zacks et al. 1999) and
designed the upper body based on patients with autoscopic
allucinations.
Materials and Methods
Subjects
Fourteen healthy volunteers (7 male; mean age 28 ±
2.9 years) participated in the study. Handedness was
evaluated using the Oldfield-Edinburgh questionnaire
(Oldfield 1971). Thirteen participants were right-handed
(scores between ?0.8 and ?1) and one participant was left
handed (score of -0.9). All participants had normal or
corrected-to-normal vision, and no history of neurological
or psychiatric disorders as indicated by a self-report.
Participants gave written informed consent prior to inclu-
sion in the study, which have been approved by the Ethical
Committee of the University Hospital of Lausanne
(Switzerland). The procedure was performed in accordance
with the ethical standards laid down in the Declaration of
Helsinki 1964.
Stimuli
Stimuli were modified from stimuli previously used in
mental own body transformation tasks (OBT-task) by
others (Parsons 1987a; Zacks et al.
1999) as well as our
group (Blanke et al. 2005; Mohr et al. 2006). Stimuli
consisted of schematic human figures that could be facing
toward or away from the participant (Fig. 1). Front- and
back-facing figures had the same outline and differed only
in the rendering of the figure’s clothing and the presence of
a face or of the back of a head (Fig. 1a). The figure’s hands
were marked such that one hand appeared to be wearing a
grey glove and a black ring at the wrist. This indication of
side could appear on the right or on the left hand. The task
performed with such stimuli was called OBT-task for full
bodies (OBT
f
-task). In addition participants performed the
OBT-task with other visual stimuli, taken from the same
schematic human figure but consisting only of the upper
part of the body. Again, front- and back-facing figures had
the same outline and differed in the rendering of the fig-
ure’s clothing and the presence of a face or of the back of a
head (Fig. 1b). The figure’s ears were marked such that one
ear was grey and had a black earring. This indication
of side could appear on the right or on the left ear. The
Fig. 1 Stimuli. The eight stimuli used in both the OBT-task and the
LAT-task. Within rows a and b, the first and the third figure are
examples of back-facing stimuli, the second and the fourth figure are
examples of front-facing stimuli. Correct responses in the OBT-task
are indicated below each figure
Brain Topogr
123
OBT-task with this kind of stimuli was called OBT-task for
upper bodies (OBT
u
-task). The presentation duration and
size of stimulus, as well as the interstimulus interval were
identical to the OBT
f
-task. Full and partial body stimuli
were controlled for overall luminance.
Stimuli appeared for 1000 ms in the center of the
computer screen (±6.0°9±6.0° of visual angle). An LCD
projector with a refresh rate of 75 Hz displayed the stimuli.
The projector was equipped with a photographic zoom lens
projecting images onto a translucent screen in a custom-
made mirror box positioned inside the magnet. The mirror
box was designed to minimize light reflection. The inter-
stimulus interval was 1000 ms (equal to stimulus duration).
The choice of the body part to mark in each type of
stimulus (hand for full bodies and ear for upper bodies) was
chosen in order to keep the target across stimuli as similar
as possible in size and in position within the visual fields
across conditions.
Procedure
In both tasks, participants were asked to make right-left
judgements of the schematic full or upper human figure
after having imagined themselves to be in the figure’s body
position and visuo-spatial perspective. In the OBT
f
-task,
participants were instructed to indicate whether the marked
hand of the figure would be their right or left hand. In the
OBT
u
-task, participants were instructed to indicate whether
the marked ear of the figure would be their right or left ear.
In both tasks participants were instructed to respond with a
button press on a serial response box as fast and accurate as
possible and to always perform the mental transformation
of their body prior to giving the response. At the end of the
experiment participants were asked to report if they could
perform the OBT task as suggested or if they used another
strategy. Responses were given with the right hand. Left
judgements were indicated by a button press of the index
finger and right judgements by a button press of the middle
finger. Participants were previously trained on the task on a
computer and were asked to answer with their preferred
hand on a keyboard. All the participants chose to answer
with the right hand. This strategy was accepted because it
has been shown that there is a left hand advantage in
behavioural data during self-recognition task (Keenan et al.
1999) and we were concerned not to bias behavioural
results towards one condition. Moreover, since participants
were scanned during one single session, it was not possible
to counter-balance hands within participants across
sessions.
In two control conditions (lateralization task or LAT-
task), the same visual stimuli (full bodies, upper bodies)
were used, but participants decided whether the indicated
hand or ear was on the right or the left side of the computer
screen (no imagined change in own body position and
visuo-spatial perspective). These two conditions (LAT
f
task;
LAT
u
-task) were carried out in order to dissociate central
mechanisms of OBT from those due to the mere perception
of the human body and right-left judgements. Responses
were also given with the right hand. Again, left judgements
were indicated by a button press of the index finger and right
judgements by a button press of the middle finger.
All participants performed a training session before
being scanned. Blocks of 20 presentations of randomized
images of different stimuli (upper bodies or full bodies)
and different tasks (LAT-task or OBT-task) were alternated
with 16 s of resting state during which only a fixation point
was displayed. Each cycle described was repeated four
times.
Analysis of Behavioural Data
Repeated measures ANOVAs were performed on the mean
reaction times for correct responses with Task (OBT- vs.
LAT-task), Body (Upper body vs. full body), and Orien-
tation (back-facing vs. front-facing) as main factors. Post-
hoc comparisons were carried out using LSD tests.
Recording and Analysis of fMRI Data
BOLD fMRI acquisitions were performed with a head coil
on a 1.5 T Siemens Magnetom Vision system equipped for
echoplanar imaging. The participant’s head was cushioned
in the coil with a vacuum beanbag to prevent motion.
Functional MRI images were acquired with an EPI gradient
echo T2*-weighted sequence (FA 90, TE 66, pixel size
3.75 9 3.75 mm, acquisition time 1.7 s, 16 slices of 5 mm
with a gap of 1 mm) with a TR = 4 s for a total of 20
acquisitions for each stimulus. fMRI pre-processing steps,
conducted with SPM2 (Wellcome Department of Cognitive
Neurology, London, UK), included realignment of intra-
session acquisitions to correct for head movement, nor-
malization to a standard template (Montreal Neurological
Institute template, MNI) to minimize inter-participant
morphological variability, and convolution with an isotro-
pic Gaussian kernel (FWHM = 9 mm) to increase signal-
to-noise ratio. Single participant analysis was performed
according to the General Linear Model. The signal drift
across acquisitions was removed with high-pass filter and
global signal changes by proportional scaling. Statistical
parametrical maps of the contrasts of interest were com-
puted for each subject as input values for the group sta-
tistics based on Random Field Theory. In particular,
the inferential statistics included a repeated measures
ANOVA, and T-tests (Knyazeva et al. 2006; Klo
¨
ppel et al.
2007; Henson et al. 2004). Only voxels with the height
threshold set at P \ 0.01 (F [ 5.19) were considered
Brain Topogr
123
significant in the F-tests performed to obtain ANOVA
results. Post-hoc comparisons were then tested with the
corresponding paired T-tests between conditions, thres-
holded for peak height at P \0.001 (T [ 3.85), in order to
determine which specific comparison supported the effect.
In each analysis, the extent threshold k [ 30 contiguous
voxels, larger than the minimum number of voxels
expected per cluster (Friston et al. 2003), were applied to
SPMs. Corrections for multiple comparisons were used at a
cluster level [P (corrected) \ 0.05]. From the clusters
showing a significant P (corrected) value both in an
ANOVA effect and in post-hoc paired T-test, we then
extracted the mean percent signal change and variance for
each condition in order to show the direction of the effect
and the contribution of each separate condition. A sagittal
T1-weighted 3D gradient-echo sequence (MPRAGE), 128
slices (with voxel size of 1 9 1 9 1.25 mm), was also
acquired as structural basis for brain segmentation and
surface reconstruction.
Results
Behaviour
Statistical analysis revealed a significant main effect of Task
[F(1,13) = 3464.30; P \ 0.001; Fig. 2] with longer RTs
observed in the OBT-task (632.1 ± 146.6 ms) than the
LAT-task (408.9 ± 101.6 ms; Zacks et al. 1999; Blanke
et al. 2005). The main effect of Orientation
[F(1,13) = 21.12, P \0.001]) revealed that reaction times
were longer for front-facing figures (526.6 ± 160.0 ms)
than for back-facing figures (509.2 ± 158.2 ms). The
Task by Orientation interaction was also significant
[F(1,13) = 4.69; P \ 0.03]. Post-hoc analysis revealed that
the OBT-task took longer than the LAT-task if the figure was
seen as front-facing (645.0 ± 146.9 ms) as compared to
back-facing (619.3 ± 146.3 ms). This front-back difference
was specific for the OBT-task (for the LAT-task, front-facing
figures: 412.8 ± 112.1 ms, back-facing figures: 408.3
±
88.8 ms). Post-hoc analysis showed the significant differ-
ence between responses for front-facing and back-facing
stimuli for the OBT-task (P \0.001), but not for the LAT-
task (P [ 0.05). There was no significant main effect of
Body on reaction times [full bodies: 521.1 ± 162.8 ms;
upper bodies: 515.3 ± 168.1 ms; F(1,13) = 2.46, P =
0.12]. With respect to error rates, statistical analysis revealed
a significant main effect of Task [F(1,13) = 16.9,
P \ 0.001] accounted for by the better performance with the
LAT-task (99%) with respect to the OBT-task (95%), but not
for Body or Orientation or any interaction (all P [ 0.3).
A general issue regarding mental imagery tasks is to
control whether participants really perform the requested
task. We think that our participants performed mental
imagery as requested for several reasons. First, partici-
pants were repeatedly instructed to respond as fast and
precisely as possible, but to always perform the mental
transformation of their body prior to giving the response.
Second, at the end of the experiment all subjects were
asked to report how they performed the OBT task and
whether this was as instructed. Four subjects who
reported an alternative strategy (i.e., inversion of the
response for the front-facing figures or rotation of the
visually presented body around the z-axes) were exclu-
ded from the study. Third, our behavioural results
showed significant differences between responses to
front-facing and back-facing stimuli in the OBT-task.
This differed from the LAT-task in which such differ-
ences were not found.
Fig. 2 Behavioural data. Behavioural data during OBT-task and
lateralization task for full bodies and upper bodies. Solid lines
represent the reaction time for front-facing figures while dotted lines
for back-facing. Plots show the reaction time mean values ±2
standard errors
Brain Topogr
123
fMRI
First we will describe the brain regions showing a task-
related effect independent from the stimulus type, then
those regions showing a stimulus-related effect and finally
those regions showing an interaction between the factors
task and stimulus. Each described region was tested with
the corresponding T-test supporting the effect and the mean
value of percent signal change within the cluster and its
variance for each condition were extracted and displayed as
plot bars.
Main Effect of Task
The BOLD signal differed significantly in several brain
regions when comparing OBT-task with LAT-task
regardless of the stimulus type (Fig. 3a).
At the right TPJ a significant cluster was localized at the
junction of the posterior superior temporal gyrus with the
angular gyrus (x, y, z = 48, -42, 18; Fig. 3b). Post-hoc
comparisons showed that both OBTf versus LATf
(T = 4.52) and OBTu versus LATu (T = 4.05) comparison
supported the effect although with different strength. Also
the right dorsal premotor cortex (x, y, z = 51, 12, 21;
Fig. 3c) responded more strongly in the OBT-task than in
the LAT-task for both stimuli (full bodies, T = 6.31; upper
bodies, T = 7.02). No similar activations were found in the
left premotor cortex. The BOLD response in parietal cortex
showed two distinct locations of activation. The posterior
peak was in the superior parietal lobule and found in
both hemispheres [x, y, z (left) =-9, -78, 48; x, y, z
(right) = 9, -75, 51; Fig. 3d]. Both clusters were sup-
ported by OBTf versus LATf [T(left) = 7.55, T(right) =
5.97] and OBTu versus LATu [T(left) = 4.65, T(right) =
6.02] contrasts. The anterior peak was in the middle/ante-
rior part of the intraparietal sulcus [x, y, z (right) = 42,
-48, 39; x, y, z (left) =-48, -45, 36; Fig. 3e]. Although
this region also appeared bilaterally, its increase of
Fig. 3 fMRI Data: Main effect
of task. Surface 3D display of
the comparison between OBT
and LAT-task regardless of
whether full or upper bodies
were shown. Color bar
represents F statistical values
(a). For four regions the BOLD
response is shown for the four
experimental conditions (OBTf,
LATf, OBTu, LATu) separately
and expressed in mean percent
of signal change ± standard
deviation for significant
clusters. These areas are: right
temporo-parietal cortex (b),
right premotor cortex (c), right
and left superior parietal lobule
(d), and right and left
intraparietal sulcus area (e).
The location of each cluster of
interest is shown in a
representative slice, centered in
its center of gravity. Both in the
cluster display and in the
plotbars, red refers to clusters in
the right hemisphere, blue, in
the left
Brain Topogr
123
