University of Zurich
Zurich Open Repository and Archive
Winterthurerstr. 190
CH-8057 Zurich
http://www.zora.uzh.ch
Year: 2008
Role of dorsal and ventral stream development in biological
motion perception
Lichtensteiger, J; Loenneker, T; Bucher, K; Martin, E; Klaver, P
Lichtensteiger, J; Loenneker, T; Bucher, K; Martin, E; Klaver, P (2008). Role of dorsal and ventral stream
development in biological motion perception. Neuroreport, 19(18):1763-1767.
Postprint available at:
http://www.zora.uzh.ch
Posted at the Zurich Open Repository and Archive, University of Zurich.
http://www.zora.uzh.ch
Originally published at:
Neuroreport 2008, 19(18):1763-1767.
Role of dorsal and ventral stream development in biological
motion perception
Abstract
Little is known about the functional development of dorsal and ventral visual streams. The right
posterior superior temporal sulcus (pSTS) represents a pivotal point of the two streams and is involved
in the perception of biological motion. Here, we compared brain activity between children (aged 5-7
years) and adults (aged 20-32 years) while they were viewing point-light dot animations of biological
motion. Biological motion-related activation was found in right pSTS of adults, and in right fusiform
gyrus and left middle temporal lobe of children. Group comparisons revealed increased activity in pSTS
for adults and in fusiform gyrus for children. Only poorly performing children showed fusiform gyrus
activity. These findings indicate that pSTS functioning is not adult-like at the age of 6 years.
Article
The role of dorsal and ventral stream development in biological
motion perception
Running head title: Dorsal and ventral stream development
Janine Lichtensteiger1, Thomas Loenneker1,2, Kerstin Bucher1, Ernst Martin1, Peter Klaver1
1) MR Center, University Children’s Hospital, Zurich, Switzerland
2) Center for Integrative Human Physiology, University of Zurich, Zurich, Switzerland
Corresponding author:
Peter Klaver
MR Center
University Children’s Hospital
Steinwiesstr. 75
8032 Zurich
Switzerland
Phone: +41 44 2667906
Fax: +41 44 2667153
Email: [email protected]
Character count of abstract: 120 including spaces
Character count of article text: 16’188 characters including spaces; 1 figure, 1 table
-1-
Abstract
Little is known about the functional development of areas within dorsal and ventral visual
streams. The right posterior superior temporal sulcus (pSTS) represents a pivotal point of the
two streams and is involved in the perception of biological motion (BM). Here, we compared
brain activity between children (5-7 y) and adults (20-32 y) while they were viewing point-light
dot animations of BM. Adults showed BM related activation in right pSTS, whereas in children
right fusiform gyrus (FG) and left hMT+ was activated. Group comparisons revealed increased
activity in right pSTS for adults and in right FG for children. Only poorly performing children
showed FG activity. These findings indicate that dorsal stream and pSTS functioning is not
adult-like at age six.
Keywords: Development, ventral and dorsal stream, biological motion perception, functional
magnetic resonance imaging, children
-2-
Introduction
Perception of other individuals’ movements and intentions is essential for successful
interactions in a social environment. Adults are highly sensitive in identifying different actions
even in displays that contain reduced information about biological motion (BM). Beyond infancy
there are only a few behavioral studies comparing the sensitivity to BM of typically developing
children and adults. Young children have generally served as a control group for comparison
with a clinical population [1, 2]. The performance of typically developing children increases
linearly up to the age of five [3], and from this age on children perform in an adult-like manner
[1]. Differences between adults and children over 5 years of age in sensitivity to BM have been
observed only in more demanding BM recognition tasks, e.g. when point-light animations were
embedded within an array of dynamic noise dots [2, 4].
The extrastriate area most closely associated with the perception of BM is the posterior part of
the superior temporal sulcus (pSTS) [5-7]. Neuroimaging data provide evidence that BM
processing occurs at the confluence of dorsal and ventral visual stream areas [8] in which the
pSTS represents a pivotal point [9]. The number of neuroimaging studies investigating BM
perception in children is very limited and inconsistent. One study reported BM related neural
activity in the pSTS that correlated positively with age in 7-10 year-old children [10]. For the
same age group, activity in the pSTS related to perception of eye gaze shifts was reported in
another study [11]. No age related effects within the pSTS were found. Child and adult data
were not compared directly in either study. Studies on the functional development of the ventral
pathway indicate that face perception elicits ventral stream activity in both preschool (5 - 8y)
and school-age (9 - 11 y) children but only school-age children showed activation in the fusiform
face area (FFA) [12]. Moreover, functional activity shifts from being bilateral and more
widespread to a more localized profile [13]. These results also indicate that ventral stream
functioning is not mature at preschool age. Studies on the development of visual motion
processing report that both activation pattern [14, 15] and electrophysiological responses [15]
develop up to adolescence, suggesting that functional dorsal stream development also
undergoes substantial progression from infancy to adulthood.
-3-
In this study, we used event-related fMRI to examine for the first time the functional
development of extrastriate areas involved in BM processing by comparing the activation
patterns of 5-to-7-year-old children and adults. Based upon neuroimaging results [14] of dorsal
and ventral stream development we hypothesize that for BM processing in children as young as
those in our group functional activity within dorsal regions may be substituted for the
engagement of networks within the ventral stream, which are known to mature earlier [16]. With
accumulated experience and maturation, dorsal regions then become increasingly involved in
BM processing.
Materials and Methods
Participants
Twenty-one adults and 23 children were recruited from the local community. A local ethics
committee approved the study and all participants were treated according to the Declaration of
Helsinki. fMRI data from 10 children and 3 adults were excluded due to excessive head motion
(over 4mm in either x-, y- or z-plane) or technical problems. The final data set included 18
healthy adult volunteers (10 females) [mean age = 27.54 3.7 years] and 13 children (5 girls)
[mean age = 6.59 1.2 years].
Experimental design
Point-light BM animations depicting a front-view of three different human actions (walking,
jumping and waving) [BM] and scrambled versions of the same animations [SM] were used as
stimuli. Animations were also masked by noise dots representative of BM in snow [BM_s] and
equivalent scrambled motion [SM_s]. The inter-stimulus interval (ISI) was randomized between
1 and 13s and showed a fixation cross. The number of pictures varied between 36 and 76
depending on the depicted motion, but presentation time was, at 4s, the same for all stimuli. BM
sequences were taken from standardized point-light pictures applying 15 white point-light dots
presented on a black background [17]. Scrambled point-light images were created using a
custom Matlab® script by permuting and randomly shifting the starting position of the pointlights. Trajectories were kept intact, except that each point-light trajectory was randomly rotated
in 900 increments and/or mirrored. Rotation and reflectance of dots by scrambling additionally
-4-
disrupts local form information that may remain after spatial scrambling. Starting positions were
chosen within a region so that the total area encompassed by the figure was similar to that of
the BM figures. BM_s and SM_s were created by adding 33 randomly distributed dots of the
same size as in BM stimuli that were continuously shifted between frames with a velocity that
varied between 4.78°/s and 12°/s and a pseudo randomized offset angle between 0.23° and
1.8°. Presentation software (www.neurobs.com) was used to display the stimuli and collect
behavioural data. Subjects were instructed to discriminate moving persons from randomly
moving dots (finding “the man in the snowstorm”). They indicated their decision with a button
press of their right index and middle finger, respectively. All subjects were instructed to answer
as quickly and accurately as possible.
FMRI acquisition and analysis
All participants underwent a short behavioural practice session before the brain imaging session
to assure that task instructions were understood and BM could be discriminated from SM. To
present the stimuli an MR-compatible video goggle (MRI Audio/Video System, Resonance
Technology, Inc., Northridge, USA) was used. Participants indicated their decision on a
response box (Lumina LP-400, Cedrus Cooperation, San Pedro, USA).
BOLD sensitive images were collected on a 3.0 T General Electric MR-scanner using a
gradient-echo EPI-sequence with 32 slices that were aligned to the anterior and posterior
commissure (flip angle=50º, field of view=22 cm, matrix=64x64, slice thickness=2.7mm,
repetition time=2500 ms, echo-time=32 ms, 246 repetitions). A three dimensional T1-weighted
anatomical scan (FOV=230x198x158mm, matrix=224x192x132; TR=8.6 ms; TE=2.1 ms) of the
whole brain was acquired.
The two initial EPI scans were discarded to allow for T1 saturation effects. Using SPM2
(Wellcome Department of Cognitive Neurology, University College London Medical School,
London, UK; www.fil.ion.ucl.ac.uk/spm/) functional images were realigned and unwarped to
correct for motion artefacts. Following recent studies that have tested the possibility of
generalizing results across age for adults and 6 and 7 year old children we transferred the
functional and anatomical data of adults and children into a common stereotactic space [14, 18].
-5-
To this end, individual T1-weighted anatomical images were taken for co-registration with the
functional images. T1-weighted anatomical images were segmented and the normalization
parameters were estimated for grey matter images on standard grey matter template in
Montreal Neurological Institute (MNI) space. Both EPI (to 3mm3 voxels) and T1-weighted (to
1mm3) images were normalized. EPI images were then spatially smoothed using a 9 mm full
width at half maximum (FWHM) Gaussian kernel. The hemodynamic response was modelled by
a stick function (general linear model) to each stimulus presentation in each category convolved
with a canonical hemodynamic response function and its temporal derivative. Parameter images
for the covariates BM, BM_s, SM and SM_s were obtained in each subject while using a
temporal high-pass filter (cut-off 128s) and modelling temporal autocorrelation as an AR(1)
process. Incorrect and missed responses were modelled as a covariate of no interest. We
applied planned t-tests in a second level random-effect analysis for the contrast BM_all (BM +
BM_s) > SM_all (SM + SM_s). For within- and between-group comparison (adults > children,
children > adults) a threshold level of p<0.001 (k≥5, two-tailed, uncorrected) was applied to
report significant voxels.
Results
Behavioural Data
A mixed effect repeated measures analysis of variance (ANOVA) on reaction times [RT (ms)]
and accuracy rates [AR (%)] with within-subject factor condition (BM/SM) and noise revealed no
significant main effect or interaction of noise (AR: F1,29=1.27; p=0.27; RT: F1,29=0.64; p=0.43).
Therefore, noise conditions were collapsed in all subsequent analyses (BM_all/SM_all). The
analyses of RT revealed a significant effect of condition (F1,29=50.96; p<0.001) with shorter RTs
for BM_all than for SM_all in both groups and a trend toward interaction of condition and group
(F1,29=2.96; p<0.1). The difference between the RTs in BM_all and SM_all tended to be larger in
adults (mean RT BM_all=808.07; SD=142.75; SM_all=1036.68; SD=249.15) than in children
(mean RT BM_all=1364.02; SD=220; SM_all=1503.87; SD=249.15). There was also a
significant main effect of the factor group (F1,29=57.36; p<0.001), indicating faster RT across all
conditions for adults than for children. A significant main effect of the factor condition (F1,29
-6-
=7.38; p=0.01) of AR data indicated that both groups made more mistakes in SM_all than
BM_all. Additionally, children made significantly (F1,29 =23.33; p<0.001) more mistakes (mean
AR=90.38%; SD=1.43) than adults (mean AR=96.45%; SD=0.50). No other main effect or
interaction reached significance.
Functional magnetic resonance imaging results
Brain regions showing significant responses to BM_all > SM_all are listed separately in Table 1
for within- and between- group comparisons.
- insert Table 1 about here - insert FIG1 about here –
When BM_all and SM_all responses were compared in adults, enhanced neural activity was
found in the right anterior cingulate gyrus, pSTS, supramarginal gyrus and cuneus (see Table 1;
FIG 1). In children, we found enhanced activation patterns in right fusiform gyrus and left hMT+
for the same comparison. Group comparison revealed increased activation in adults compared
to children in right pSTS. The opposite contrast (children > adults) revealed a significant
difference in right nucleus caudatus and right fusiform gyrus.
To test whether task difficulty confounded group differences, we separated both groups into
good (AD_g; CH_g) and bad performers (AD_b, CH_b), on the basis of the median in AR. A
post-hoc one-way ANOVA analysis of the four groups was then performed for the contrast
BM_all>SM_all. Main effects and interactions of performance and age were masked by group
comparison (AD_g + AD_b > CH_g + CH_b). A liberal statistical threshold (p<0.05, uncorrected)
was used in order to prove non-significance. This analysis revealed an interaction of age and
performance in the right FG (x=39, y=-51, z=-12; Z=2.21; p<0.014). This interaction indicated
that only in poorly performing children the right FG was activated (Z=4.64, p<0.001), but not in
well-performing children, or in any adult (all p>0.05). No main effect of performance or other
significant interaction with performance was found in the right STS or any other area.
Discussion
The aim of the study was to investigate the neural development of BM perception in preschool
and school-age children and to clarify the extent to which successful processing (i.e. perception)
-7-
of BM depends on the development of dorsal stream functioning. In accordance with previous
studies we found that BM perception in adults [5-7] elicited task-specific activity in right pSTS –
the area most closely associated with BM perception, and a pivotal point of the dorsal and
ventral stream. Moreover, activation clusters were detected in right anterior cingulate gyrus and
cuneus. Compared with children, we found enhanced BM-related functional activity in right
pSTS in adults. During BM perception children exhibited increased activity in right FG and left
hMT+. However, in comparison to adults, children induced greater activation in right FG and
caudate nucleus. These findings suggest that perception of BM may activate the dorsal stream
and pSTS in adult subjects, whereas in children ventral regions are recruited for processing BM.
In addition, we found that effortful performance might contribute to activity in right FG because
only poorly performing children showed increased activity in this area. In contrast, performance
level seems not to affect activity within pSTS. The data therefore provide evidence for a more
protracted development of task-specific functioning of the dorsal system.
A possible explanation might be that the dorsal pathway hosts more complex functions than the
ventral pathway. Functions within the dorsal stream appear to rely on life-long learning
(plasticity) whereas in the ventral stream they might be more constant. For instance, during BM
perception, dorsal stream areas have to integrate information from sensory systems with
propioceptive feedback to generate “body-centred” representations for action. As the child’s
body is developing physically, important factors such as length and weight of limbs continue to
change. Thus, it may be more important to retain plasticity in the dorsal than in the ventral
stream (see [19]).
The present findings differ in part from previous studies. For example, our results do not agree
with the behavioural findings of Freire et al. [4], who reported differences in BM recognition
when distractors of different numbers of dots were used. Our study used a fixed and relatively
small number of random noise dots. This might explain why no behavioural differences between
BM and BM_s could be detected. Furthermore, it is difficult to compare our results to previous
imaging studies with children since no study made a direct comparison between children and
adults [10, 11]. Mosconi and colleagues used a gaze perception task and found pSTS activity in
-8-
children [11]. Carter and colleagues reported that age in 7-10 year old children correlated with
pSTS activity during BM perception [10]. Our data are in line with the latter findings and show
no significant activity in the pSTS related to BM perception in children that were about 2-3 years
younger. Furthermore, both studies were ambiguous as to whether differences in performance
level affected brain activity in this area. Here we could show that this result was independent of
performance level. Additionally, we were able to show that BM related brain activity in the right
FG was sensitive to performance, since, only poorly performing children showed activity in this
area. Our results are in line with findings that reveal maturation occurring well into childhood
and even adolescence [14, 15]. The former study showed that neural activity within dorsal
regions of the occipital and parietal lobule was not mature at age 6 [14]. We also agree with
results indicating that ventral regions are engaged more in 6-year-old children during effortful
high-order visual perception [14]. Our findings provide additional support for the hypothesis that
functional development of the dorsal stream is not yet complete at age 6. They further agree
with behavioural reports that children do not reach adult levels in tasks depending on dorsal
stream functioning before the age of 7-8 years. [20].
-9-
Conclusion
Taken together, these findings show that in preschool and school age children BM perception
engages ventral areas only when performance is poor, whereas the classical network for BM
processing in the dorsal stream and pSTS is recruited later during development, independently
of performance. Therefore, we conclude that the functional segregation into dorsal and ventral
stream is still immature at the age of 6.
- 10 -
Acknowledgment
The Swiss National foundation (Project No. B0-109983) and EMDO foundation (Switzerland)
supported this research. We would like to thank O. Braddick and J. Atkinson for inspirational
scientific exchange. We are, above all, grateful to the numerous volunteers and their families.
11
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Blake R, Turner LM, Smoski MJ, Pozdol SL and Stone WL. Visual
recognition of biological motion is impaired in children with autism. Psychol
Sci, 2003; 14(2): 151-7.
Pavlova M, Staudt M, Sokolov A, Birbaumer N and Krageloh-Mann I.
Perception and production of biological movement in patients with early
periventricular brain lesions. Brain, 2003; 126(Pt 3): 692-701.
Pavlova M, Krageloh-Mann I, Sokolov A and Birbaumer N. Recognition of
point-light biological motion displays by young children. Perception, 2001;
30(8): 925-33.
Freire A, Lewis TL, Maurer D and Blake R. The development of sensitivity
to biological motion in noise. Perception, 2006; 35(5): 647-57.
Beauchamp MS, Lee KE, Haxby JV and Martin A. FMRI responses to video
and point-light displays of moving humans and manipulable objects. J
Cogn Neurosci, 2003; 15(7): 991-1001.
Grossman E, Donnelly M, Price R, Pickens D, Morgan V, Neighbor G, et al.
Brain areas involved in perception of biological motion. J Cogn Neurosci,
2000; 12(5): 711-20.
Vaina LM, Solomon J, Chowdhury S, Sinha P and Belliveau JW. Functional
neuroanatomy of biological motion perception in humans. Proc Natl Acad
Sci U S A, 2001; 98(20): 11656-61.
Ungerleider LG and Mishkin M. Two cortical visual systems. In Ingle DJ,
Goodale MA and Mansfield RJW (eds). Analysis of visual behaviour.
Cambridge: MIT Press; 1982. pp. 549-586.
Giese MA. A Neural Model for Biological Movement Recognition: A
Neurophysiologically Plausible Theory. In Vaina LM, Breadsley SA and
Rushton S (eds). Optic Flow and Beyond (Synthese Library). Norwell, MA,
USA: Kluwer Academic Publishers; 2004. pp. 443-470.
Carter EJ and Pelphrey KA. School-age children exhibit domain-sepcific
responses to biological motion. Social Neuroscience, 2006; 1(3-4): 396411.
Mosconi MW, Mack PB, McCarthy G and Pelphrey KA. Taking an
"intentional stance" on eye-gaze shifts: a functional neuroimaging study of
social perception in children. Neuroimage, 2005; 27(1): 247-52.
Gathers AD, Bhatt R, Corbly CR, Farley AB and Joseph JE. Developmental
shifts in cortical loci for face and object recognition. Neuroreport, 2004;
15(10): 1549-53.
Passarotti AM, Smith J, DeLano M and Huang J. Developmental differences
in the neural bases of the face inversion effect show progressive tuning of
face-selective regions to the upright orientation. Neuroimage, 2007;
34(4): 1708-22.
Klaver P, Lichtensteiger J, Bucher K, Dietrich T, Loenneker T and Martin E.
Dorsal stream development in motion and structure-from-motion
perception. Neuroimage, 2008; 39(4): 1815-23.
Bucher K, Dietrich T, Marcar VL, Brem S, Halder P, Boujraf S, et al.
Maturation of luminance- and motion-defined form perception beyond
adolescence: a combined ERP and fMRI study. Neuroimage, 2006; 31(4):
1625-36.
Gunn A, Cory E, Atkinson J, Braddick O, Wattam-Bell J, Guzzetta A, et al.
Dorsal and ventral stream sensitivity in normal development and
hemiplegia. Neuroreport, 2002; 13(6): 843-7.
12
17.
18.
19.
20.
Vanrie J and Verfaillie K. Perception of biological motion: a stimulus set of
human point-light actions. Behav Res Methods Instrum Comput, 2004;
36(4): 625-9.
Kang HC, Burgund ED, Lugar HM, Petersen SE and Schlaggar BL.
Comparison of functional activation foci in children and adults using a
common stereotactic space. Neuroimage, 2003; 19(1): 16-28.
Gilmore RO and Johnson MH. Learning what is where: Oculomotor
contributions to the development of spatial cognition. In Simion F and
Butterworth G (eds). The Development of Sensory, Motor and Cognitive
Capacities in Early Infancy: From Sensation to Cognition. Hove, East
Sussex, UK: Psychology Press; 1998. pp. 25-47.
Giaschi D and Regan D. Development of motion-defined figure-ground
segregation in preschool and older children, using a letter-identification
task. Optom Vis Sci, 1997; 74(9): 761-7.
13
Legend
Figure 1. Activation maps for the contrast biological motion (BM_all) >
scrambled motion (SM_all) are thresholded at a voxelwise uncorrected P < 0.001
(two-tailed) and a spatial extent of five contiguous voxels. The maps are shown
superimposed onto selected coronal/sagital/axial slices of a canonical T1
template provided by SPM2. The anterior-posterior level is based on MNI
coordinates.
Table
Table1 Cerebral region with corresponding Brodman area (BA), Z-values and MNI
coordinates for peaks. All significant voxels are reported at a threshold of p<0.001
uncorrected. L, left; R, right; pSTS, posterior superior temporal sulcus; hMT, human
middle temporal lobe; MNI, Montreal Neurological Institute; k = cluster size
Region
k
Side
MNI coordinates
(x, y, z)
SPM
{Z}
Cingulate
Gyrus
pSTS
156
R
3
3.82
BA
Adults
BM_all > SM_all
Supramarginal
Gyrus
Cuneus
-45
39
31
19
R
63
-45
18
3.44
22
21
R
45
-42
30
3.42
40
14
R
24
-93
-3
3.39
18
Children
BM_all > SM_all
Fusiform
Gyrus
hMT
27
R
39
-54
-15
3.78
37
30
L
-48
-69
6
3.42
37
pSTS
15
R
66
-27
17
3.49
42
Nucleus
caudatus
Fusiform
Gyrus
15
R
12
-6
24
3.58
-
9
R
42
-51
-12
3.34
37
Adults > Children
BM_all > SM_all
Children > Adults
BM_all > SM_all
Figure
Click here to download high resolution image