Neuropsychologia 48 (2010) 1644–1651 Contents lists available at ScienceDirect Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia Coherent motion processing in autism spectrum disorder (ASD): An fMRI study Sarah Brieber a,b,c , Beate Herpertz-Dahlmann b , Gereon R. Fink c,d , Inge Kamp-Becker e , Helmut Remschmidt e , Kerstin Konrad a,∗ a Child Neuropsychology Section, Department of Child and Adolescent Psychiatry and Psychotherapy, University Hospital of the RWTH Aachen, D–52074 Aachen, Germany Department of Child and Adolescent Psychiatry and Psychotherapy, University Hospital of the RWTH Aachen, Germany c Cognitive Neurology Section, Institute of Neuroscience and Medicine (INM-3), Research Center Juelich, Germany d Department of Neurology, University Hospital Cologne, Germany e Department of Child and Adolescent Psychiatry, Phillips University Marburg, Germany b a r t i c l e i n f o Article history: Received 30 September 2009 Received in revised form 2 February 2010 Accepted 5 February 2010 Available online 12 February 2010 Keywords: Autism Functional MRI Visual Motion perception a b s t r a c t A deficit in global motion processing caused by a specific dysfunction of the visual dorsal pathway has been suggested to underlie perceptual abnormalities in subjects with autism spectrum disorders (ASD). However, the neural mechanisms associated with abnormal motion processing in ASD remain poorly understood. We investigated brain responses related to the detection of coherent and random motion in 15 male subjects with ASD and 15 age- and IQ-matched healthy controls (aged 13–19 years) using eventrelated functional magnetic resonance imaging (fMRI). Behaviorally, no significant group differences were observed between subjects with ASD and controls. Neurally, subjects with ASD showed increased brain activation in the left primary visual cortex across all conditions compared with controls. A significant interaction effect between group and condition was observed in the right superior parietal cortex resulting from increased neural activity in the coherent compared with the random motion conditions only in the control group. In addition, neural activity in area V5 was not differentially modulated by specific motion conditions in subjects with ASD. Functional connectivity analyses revealed positive correlations between the primary visual cortex and area V5 within both hemispheres, but no significant between-group differences in functional connectivity patterns along the dorsal stream. The data suggest that motion processing in ASD results in deviant activations in both the lower and higher processing stages of the dorsal pathway. This might reflect differences in the perception of visual stimuli in ASD, which possibly result in impaired integration of motion signals. © 2010 Elsevier Ltd. All rights reserved. 1. Introduction Autism spectrum disorders (ASD) are neurodevelopmental disorders characterized by impairments in social communication and interaction as well as repetitive, stereotyped behaviors. An additional non-social feature of autism is sensory abnormality (Rogers & Ozonoff, 2005). Heightened sensitivity to small differences between stimuli, increased attention to fragments or surface features of objects, and a relative failure to extract the information context have been described frequently in autistic individuals (Dakin & Frith, 2005; Hill & Frith, 2003). This information processing style, characterized by favoring local over global feature aspects, is specific to autism (Bertone & Faubert, 2006) and could explain both impaired and superior perceptual processing in ASD (Happé & Frith, 2006). ∗ Corresponding author. Tel.: +49 241 8088768; fax: +49 241 8082544. E-mail address: kkonrad@ukaachen.de (K. Konrad). 0028-3932/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropsychologia.2010.02.007 Recently, there has been a growing interest in low-level perceptual processing in ASD especially in the visual domain, but the neural basis of these abnormalities is still unknown. Furthermore, different theories suggest that abnormal visual processing in ASD is characterized by an integration deficit at an early stage of visual processing and/or a dysfunction of the visual pathways (Bertone, Mottron, Jelenic, & Faubert, 2005; Davis, Bockbrader, Murphy, Hetrick, & O’Donnell, 2006; Happé & Frith, 2006). The visual system beyond the primary visual cortex segregates into at least two streams, the dorsal and ventral pathways, each of which processes and transmits the information about objects and events in different ways and for different purposes (Goodale & Milner, 1992; Livingstone & Hubel, 1987). The ventral stream projects from the primary visual cortex to area V4 to the inferotemporal cortex. It is mainly responsible for the perceptual identification of objects, and it receives input from the parvocellular system for high spatial frequencies. The dorsal pathway on the other hand projects from the striate cortex to the middle temporal area (MT/V5) to the posterior parietal cortex. It receives input mainly from the magnocellular layers of the lateral geniculate S. Brieber et al. / Neuropsychologia 48 (2010) 1644–1651 nucleus (LGN), which are sensitive to low spatial frequencies, and it plays an important role in the localization of visual stimuli and the control of object-related actions towards an object (Goodale & Milner, 1992; Milner & Goodale, 2008). Notably, the cells in the dorsal stream, especially in MT/V5, are highly responsive to motion stimuli (Culham, He, Dukelow, & Verstraten, 2001; Tootell et al., 1995). A classical psychophysical task for investigating the function of this pathway is coherent motion detection (Braddick et al., 2001; Newsome & Pare, 1988), in which the perception threshold of an observer is measured by determining the lowest proportion of dots that is needed to correctly identify the direction of coherent motion in a stimulus patch (Dakin & Frith, 2005). Functional imaging studies have shown that in healthy subjects, coherent motion stimuli compared with static or random motion stimuli, activate the specialized motion-sensitive area MT/V5 along with other occipital and parietal areas (Braddick et al., 2001; Culham et al., 2001). In autism, several behavioral studies have reported increased motion coherence thresholds or motion-processing deficits, although this pattern could not consistently be replicated across all studies (Del Viva, Igliozzi, Tancredi, & Brizzolara, 2006; Milne et al., 2006; Vandenbroucke, Scholte, van Engeland, Lamme, & Kemner, 2008), in particular no group differences were found if IQ differences were controlled (Koldewyn, Whitney, & Rivera, 2009). At the neural level, a disturbed dorsal pathway for motion processing, despite an intact ventral pathway for form processing, has been suggested to reflect a dorsal stream vulnerability in neurodevelopmental disorders (Braddick, Atkinson, & Wattam-Bell, 2003). Furthermore, Milne et al. (2002) interpreted abnormally high motion coherence thresholds in relation to the local processing bias in ASD by postulating that both are a consequence of low levels of activity in the low spatial frequency channels of the magnocellular pathway, which possibly lead to abnormal development of the parietal lobe. Contrary to this hypothesis, other researchers have reported intact lower level dorsal stream functioning (processed in the primary visual cortex) and suggested that second-order or coherent motion-processing deficits in autism stem from a higher order deficit in the integration of ‘complex’ information at the global level instead of a motion-processing deficit per se (Bertone, Mottron, Jelenic, & Faubert, 2003; Pellicano, Gibson, Maybery, Durkin, & Badcock, 2005). Hence, the motion-sensitive area V5 might be responsible for increased motion thresholds in ASD. This region is crucial for motion processing in the brain, and at this stage, local directional signals are combined to form a global percept, which involves additional cooperative mechanisms in the cortex (Braddick et al., 2001). Furthermore, neural responses in V5 have shown a linear dependence on increasing coherence levels (Rees, Friston, & Koch, 2000). Recently, it has also been suggested that perceptual deficits in ASD are not only due to dysfunctions within a specific pathway or region but are rather related to reduced connectivity between distant brain regions that might lead to a reduced top–down modulation of lower level sensory processing (Just, Cherkassky, Keller, Kana, & Minshew, 2007; Just, Cherkassky, Keller, & Minshew, 2004; Koshino et al., 2005; Villalobos, Mizuno, Dahl, Kemmotsu, & Muller, 2005), but non-replications are puzzling (for review: Muller, 2008). Thus, the exact neural mechanism underlying deficits in motion processing have yet to be elucidated. In particular, it remains unclear at which stage of information processing deficits occurs. In the present functional magnetic resonance imaging (fMRI) study, we used a coherent motion detection task that reliably activates the dorsal visual stream to investigate the neural mechanisms underlying abnormal motion processing in ASD in comparison with a healthy control group. This kind of task exemplifies early neurointegrative processing and is mediated mainly by area MT/V5, which is a specialized motion-sensitive region in extrastriate cortex (Braddick et al., 2001). On the basis of behavioral coherent motion 1645 Table 1 Sample characteristic for both groups. Age FSIQ VIQ PIQ ASD CON T score p 16.42 (±2.28) 108.53 (±15.34) 114.07 (±16.69) 101.2 (±12.43) 15.35 (±1.80) 115.07 (±11.62) 115.07 (±12.79) 112.13 (±9.45) 1.421 −1.315 −.184 −2.712 .166 .199 .855 .011 studies in ASD, we hypothesized that the neural activity of the motion-sensitive region MT/V5 is abnormal in subjects with ASD compared with controls. Since previous neuroimaging studies in autism also reported increased activation in early visuo-perceptual processing areas (Manjaly et al., 2007), but reduced neural activity in “higher order” cortical regions (Mottron, Dawson, Soulieres, Hubert, & Burack, 2006), we conducted a whole-brain imaging analysis to examine the extent to which higher order or early perceptual differences are involved in abnormal visual processing in ASD. Moreover, the functional connectivity among activated areas of the dorsal motion pathway were investigated to further test the hypothesis that deficits in the integration of complex information in ASD are associated with abnormal coupling between these brain regions. 2. Methods and materials 2.1. Subjects Fifteen male adolescents with ASD and 15 male healthy controls, all aged between 13 and 19 years with a full-scale intelligence quotient (FSIQ) above 80, were included in the present study. There were no significant differences between groups in the FSIQ, the verbal IQ (VIQ) or age, but the performance IQ (PIQ) differed significantly between the groups (Table 1). One participant in the autism group and two in the control group were left-handed; all other subjects were right-handed. An experienced clinician reviewed the clinical diagnosis according to the standard criteria of ICD-10 (World Health Organization, 1993) and DSM-IV (American Psychiatric Association, 1994), and all subjects underwent an extensive psychiatric and neurological examination. Their psychiatric classification was then determined on the basis of a semi-structured diagnostic interview (K-SADS; Kaufman et al., 1997; German translation: Delmo, Weiffenbach, Gabriel, Stadler, & Poustka, 2001). In the autistic group, the expression of the autistic symptoms was further assessed by the German version of the Autism Diagnostic Observation Scale (ADOS-G; Lord et al., 2000) and a semi-structured autism specific parent interview (ADI-R; Le Couteur et al., 1989), both conducted by a trained examiner (IKB) (Table 2). The ASD group consisted of thirteen participants with Asperger syndrome and two participants with the diagnosis of high-functioning autism (HFA). All children and their parents or caregivers gave their written consent after having been informed about the details and the purpose of this study. The study was approved by the Ethics Committee of the University Hospital Aachen. In addition, the subjects (14 subjects with ASD (one subject was excluded due to technical problems) and 15 controls) participated in a short experimental paradigm outside the scanner using the stimuli of the fMRI task to make sure that all subjects were able to perform the fMRI task. The stimuli (see next paragraph) were presented on a screen with a resolution of 800 × 600 pixels. The subjects were asked to sit still while performing the task and were seated 60 cm in front of the screen. The patch of dots subtended ca. 9◦ × 12◦ . The size of each pixel/dot was 0.05◦ and the speed of moving dots was 2.5◦ /s. The participants had to decide for each stimulus if the dots were moving coherently to the left or the right direction. The percentage of coherent moving dots was systematically reduced with a starting point of 50% coherence and then 10% enhanced if more than four errors were made or the three percent coherence level were reached. The percentage of correct trials of the resulting coherence levels around the assumed threshold were used for the analysis. Thresholds were defined at the signal-to-noise ratio needed to perform correctly at 75% of the trials and was obtained by fitting a logistic curve to each individual participant’s data. Table 2 Diagnostic data for ASD individuals. ADI communication ADI social interaction ADI stereotype behavior ADOS communication ADOS social interaction ADOS stereotype behavior Mean SD Min Max 14.47 18.73 6.47 4.27 8.00 1.47 4.78 5.68 3.11 2.19 2.27 1.30 8.00 11.00 3.00 .00 4.00 .00 25.00 31.00 14.00 7.00 12.00 4.00 1646 S. Brieber et al. / Neuropsychologia 48 (2010) 1644–1651 2.2. fMRI task Participants completed a coherent motion task in which they had to decide if the dots were moving coherently or randomly by pressing the corresponding button as fast as possible. Three coherence levels with 40%, 60% or 80% of the dots moving to the left or the right were used, while the remaining dots were moving randomly every frame in one of eight possible directions. These coherence levels were chosen in order to exclude behavioral differences between the groups. Accordingly, these levels are in line with previous studies suggesting that motion detection thresholds in ASD subjects are typically below these levels (Milne et al., 2002; Spencer et al., 2000). Thus, any between-group differences in neural activations are not confounded by differential capacity to perform the task (Price & Friston, 1999). The stimuli comprised a small rectangular patch containing 300 randomly arranged white dots on a black background. Each dot had a limited lifetime of six animation frames after which it disappeared and reappeared at a random location within the stimulus patch. This procedure ensures that the subjects were not able to track single dots. The fMRI task consisted of 180 trials in pseudo-randomized order with 90 trials of randomly moving dots and 90 trials of coherently moving dots for three different coherence levels (40%, 60%, and 80%). Stimuli were presented for 1000 ms with an interstimulus interval of 1200 ms. Blank trials were included, resulting in variable stimulus-onset-asynchronies. After half of the trials participants were given a break of 60 s. Stimuli were presented via a MRI-compatible goggle system using Presentation software. The patch of dots subtended 6.4◦ × 7.5◦ . The size of each pixel/dot was 0.03◦ and the speed of moving dots was 1.8◦ /s. During scanning, reaction times and error rates were recorded. Subjects were instructed to respond as quickly and as accurately as possible. A short training session was performed before scanning to familiarize the subjects with the stimuli and task requirements. 2.3. Data acquisition For each subject, 230 functional whole-brain images were acquired with a Siemens scanner (1.5 T) and echo planar imaging (EPI) with the following parameters: TR (repetition time) = 3000 ms, TE (echo time) = 60 ms, thirty 4-mm-thick axial slices with a 0.4-mm gap, matrix size = 64 × 64, voxel size = 3.1 mm × 3.1 mm × 4.0 mm, field of view (FOV) = 200 mm, flip angle = 90◦ . Images were analyzed using SPM5 (http://www.fil.ion.ucl.ac.uk/spm) and MATLAB 7 (The Mathworks Inc., Natick, USA) as follows: the EPI images were corrected for head movement between scans by an affine registration. Then, the images were initially realigned to the first image of the time-series and subsequently re-realigned to the mean of all images after the first step. The mean EPI image for each subject was computed and spatially normalized to the MNI single subject template using the “unified segmentation” function in SPM5. The ensuing deformation was subsequently applied to the individual EPI volumes as well as to the T1 scan, which was coregistered to the mean of the realigned EPIs beforehand. All images were hereby transformed into standard stereotaxic space and resampled at 3 mm × 3 mm × 3 mm voxel size. The normalized images were spatially smoothed using an 8-mm FWHM Gaussian kernel to meet the statistical requirements of the General Linear Model and to compensate for residual macroanatomical variations across subjects. During scanning, eye data were recorded using IVIEW and evaluated by determining the mean eye-motion variance for each condition in every subject using MATLAB7. The mean variance was compared by analysis of variance (ANOVA) using SPSS 14. After acquisition of the functional scan, high-resolution T1-weighed anatomical brain scans were collected using a rapid acquisition gradient-echo (MPRAGE) pulse sequence (TR = 2200 ms, TE = 3.93 ms, flip angle = 15◦ ; FOV = 256 mm; matrix = 180 × 256; 160 slices, slice thickness 1 mm, inter-slice gap = 0.5 mm). (missed responses/errors) and the time and dispersion derivatives for all effects. Errors were excluded since it has been demonstrated previously that even a small number of errors might alter activation maps (Murphy & Garavan, 2004). The six head movement parameters were included as confounds. For each subject, the contrasts for coherent and random motion (vs. implicit baseline) were computed and were then fed to a second-level group analysis using ANCOVA models with the performance IQ as a covariate. Activations are reported at a p < .05 level of significance, family-wise error (FWE) corrected for whole brain. Post hoc t-tests for the percent signal change values were used and plotted to further evaluate possible differences between groups. Based on the hypothesis of a linear increase of cortical activity with respect to the degree of motion coherence, we calculated in an additional analysis a parametric modulation of functional images (i.e., coherence level: 40% coherent motion as 1, 60% coherent motion as 2, and 80% coherent motion as 3). Functional activations were anatomically localized by version 14 of the SPM anatomy toolbox (Eickhoff et al., 2005; http://www.fz-juelich.de/ime/spm anatomy toolbox) using a maximum probability map (MPM). In addition the localization of the activations was confirmed by superimposing the activations on a structural mean group image and using a standard anatomical atlas. Region-of-interest (ROI)-based analyses for area 17 and human area MT/V5 based on the probability maps included in the SPM anatomy toolbox (Amunts, Malikovic, Mohlberg, Schormann, & Zilles, 2000; Eickhoff, Heim, Zilles, & Amunts, 2006; Malikovic et al., 2007) were performed for each contrast. 3.2. Analysis of functional connectivity 3. Data analysis Functional connectivity was computed (separately for each participant) as a correlation between the average time course of all the activated voxels in each member of a pair of ROIs. Four ROIs [area 17 and human area MT/V5 bilaterally based on the probability maps included in the SPM anatomy toolbox] were defined to encompass the main clusters of activation in the group activation map for both groups (conjunction) of the average effect of condition (coherent and random motion vs. implicit baseline). If there were three or more activated voxels in the ROI for each participant, a mean time course was computed across all activated voxels in each ROI. A correlation coefficient was then calculated between the time courses of pairs of ROIs. Fisher’s z transformation was applied to the correlation coefficients [z(r) = (0.5) ln [(1 + r)/(1 − r)]], in order to calculate between-group differences in functional connectivity patterns [Z = z(r1) − z(r2)/ [1/(n1 − 3)] + [1/(n2 − 3)]]. 3.1. Imaging data 4. Results The data were analyzed using a general linear model as implemented in SPM5. Each experimental condition was modeled using an event-related reference vector convolved with a canonical hemodynamic response function and its first-order temporal and dispersion derivative. Low-frequency signal drifts were filtered using a cutoff period of 128 s. Parameter estimates were subsequently calculated for each voxel using weighted least squares to provide maximum likelihood estimators based on the temporal autocorrelation of the data in order to get identical and independently distributed error terms/ordinary least square estimation. No global scaling was applied. At the first level, 3 event types were defined and blank trials (null events) were used as an implicit baseline condition. The event-types consisted of two effects of interest (random motion and coherent motion) and one effect of no interest 4.1. Behavioral data Fifteen control and 14 ASD subjects participated in the motion detection threshold task outside the scanner. An analysis of covariance with performance IQ as a covariate revealed no significant group differences between the two groups (control: mean: 6.88 ± 1.83; ASD: mean: 7.13 ± 2.44; F(1,26) = 0.349, p = .56). For the behavioral data of the fMRI experiment, a 2 × 2 analysis of covariance (ANCOVA) with “GROUP” (ASD and healthy subjects (CON)) as a between-subject factor, “MOTION” (random and coherent motion) as a within-subject factor and the performance IQ as a covariate was calculated for the percentage of errors and reaction times. The ANCOVA revealed a main effect of “MOTION” for both reaction times (in millisecond) and errors (in percent). Reaction S. Brieber et al. / Neuropsychologia 48 (2010) 1644–1651 1647 Fig. 1. Mean of error rates for both groups for the random motion condition and the three coherent motion levels of the fMRI task. ASD = autism spectrum disorder; CON = healthy controls; RND = random motion; CM = coherent motion times were significantly slower for the random motion task (random: 869.5 ± 127.94; coherent: 766.33 ± 121.27; p = .013), but accuracy was superior (random: 5.79 ± 5.59; coherent: 8.77 ± 7.08; p = .023) for both groups. Furthermore, no main effect of group (RT: F(1,27) = 0.997, p = .327; errors: F(1,27) = 0.757, p = .392) and no interaction (RT: F(1,27) = 1.928, p = .176; errors: F(1,27 = 2.848, p = .103)) could be observed (Fig. 1). Eye movements were successfully recorded in 11 ASD subjects and 10 control subjects. An ANCOVA (group × condition) revealed no main effect of eye-motion variance (F(1,18) = 4.268, p = .054), no main effect of group (F(1,18) = 0.946, p = .344) and no interaction (F(1,18) = 1.335, p = .263). 4.2. fMRI analyses We evaluated the effect of coherent motion compared with random motion (cm > rm) across both groups. The whole-brain analysis revealed significant neural activation in the left postcentral gyrus. In addition, increased neural activity was found in area MT/V5 bilaterally in the ROI analysis (Table 3). Post hoc comparisons revealed that the difference between coherent and random motion in left V5 was only significant for the control group (CON) but not in ASD subjects (CON: p = .001; ASD: p = .113). In addition, subjects with ASD showed a significantly higher activation in the random motion condition in comparison to the CON group (Fig. 2A Table 3 Results from fMRI data analysis. Region Side WB/ROI MNI coordinates x y Z score p z cm > rm MT/V5 MT/V5 Postcentral gyrus L R L ROI ROI WB −42 52 −14 −66 −68 −34 −1 9 61 3.6 2.76 5.14 .003 .041 .013 ASD > CON V1 L ROI −6 −100 11 3.87 .031 Interaction (CON > ASD × cm > rm) SPL/precuneus R WB 16 −54 55 4.86 .044 ASD = autism spectrum disorder; CON = healthy controls; cm = coherent motion; rm = random/incoherent motion; ROI = based on region of interest analysis; WB = based on whole brain analysis; MT/V5 = middle temporal gyrus/human analogue of area V5; V1 = primary visual cortex; SPL = superior parietal lobe. Fig. 2. Brain areas showing significant higher activation in area MT/V5 to coherent motion as compared to the random motion condition for both groups in the ROI analysis. A statistical threshold of p < .001 was used for display purposes (A). Plots of the parameter estimates are shown separately for both groups as a function of trial type (coherent and random motion) for the respective activation maximum (B). ASD = autism spectrum disorder; CON = healthy controls and B). In the right area V5 the coherent vs. random motion contrast showed a trend for significance in the controls (p = .055) and no significant effect for the autistic group (p = .087). A significant main effect of group was found in the left primary visual cortex in the ROI analysis resulting from increased neural activity across conditions in subjects with ASD compared with control subjects (ASD(cm + rnd) > CON(cm + rnd)) (Fig. 3, Table 3). No significant effect emerged in the reverse contrast (CON > ASD) either in the ROI or in the whole-brain analysis. In addition, an interaction effect for group × motion condition was observed in the posterior parietal cortex (superior parietal gyrus/precuneus). Plots of the parameter estimates revealed that only control subjects showed increased neural activity in the coherent compared with the random motion condition (Fig. 4A and B, Table 3). No significant effect was found in the analysis of the parametric modulation of coherence levels (40% coherent motion modeled as 1, 60% coherent motion as 2, 80% coherent motion as 3) either in area V5 (ROI analysis) or in any other brain region in the whole-brain analysis. 1648 S. Brieber et al. / Neuropsychologia 48 (2010) 1644–1651 Fig. 3. Significant higher activations in the primary visual cortex in the ROI analysis in participants with ASD in comparison to healthy controls in both motion conditions. A statistical threshold of p < .001 was used for display purposes. Fig. 5. Positive effect of condition (coherent and random motion vs. implicit baseline) for both groups is overlaid on the probability atlas of Eickhoff et al. (2005) (whole brain, FWE corrected .05). 4.3. Functional connectivity along the dorsal stream Functional connectivity of the brain regions of the dorsal stream was calculated and analyzed for between-group differences. Fourteen subjects per group were included in the analysis. The conjunction of the activation maps of the average effect of condition (coherent and random motion vs. implicit baseline) for both groups revealed significant activations in the bilateral visual cortex, and, in the ROI analysis, a significant effect in bilateral V5 and V1 was specified (Fig. 5). The time course in these four regions was extracted separately for each participant, correlated and z-transformed for group comparisons. Data examining patterns of functional connectivity in the entire sample of 28 participants revealed strongly positive correlations between neural activity in V1 and V5 within both hemispheres as well as between right and left V5. These findings suggest that task performance was associated with strong functional connectivity within the dorsal stream. Between-group differences in functional connectivity between V1 and V5, however, were not evident (left V1–left V5 Z(rASD) = 0.4341; Z(rCON) = 0.3839; z = 0.1177), between right V1–right V5 (Z(rASD) = 0.4134; Z(rCON) = 0.4398; z = −0.0619), and between left V5–right V5 (Z(rASD) = 0.4944; Z(rCON) = 0.5141; z = −0.0462), assuming a significance level of 0.05 and a critical value of 1.96. 5. Discussion Fig. 4. Brain areas showing significant higher activation in superior parietal cortex to coherent motion as compared to the random motion condition for the healthy controls in contrast to the autistic subjects. A statistical threshold of p < .001 was used for display purposes (A). Plots of the parameter estimates are shown separately for both groups as a function of trial type (coherent and random motion) for the respective activation maximum (B). ASD = autism spectrum disorder; CON = healthy controls The present study investigated the neural mechanism of coherent motion perception in participants with autism spectrum disorder. We hypothesized abnormal neural activity in the dorsal stream in autistic subjects. Indeed, our data revealed significant between-group differences in neural activation between healthy control subjects and participants with ASD at different processing stages of the dorsal pathway, including increased neural activity in the primary visual cortex, unmodulated activation in the motion-sensitive area V5 in autistic subjects and an interaction effect between group and motion type in parietal cortex activation. S. Brieber et al. / Neuropsychologia 48 (2010) 1644–1651 Our fMRI analyses revealed increased activity in the left primary visual cortex in the ASD group as compared with the control group, irrespective of stimulus type. Increased activity in visual areas in autism has already been reported in several imaging studies (e.g., Boelte, Hubl, Dierks, Holtmann, & Poustka, 2008; Manjaly et al., 2007; Ring et al., 1999). For example, Manjaly and colleagues studied the neural correlates of a visual search task in autism and suggested that increased V1 activation may be a general phenomenon and may lead to an advantage in local visual processing by boosting features of individual stimuli. This conclusion is supported by a recent behavioral study by Joseph, Keehn, Connolly, Wolfe, and Horowitz (2009). They showed that non-search factors in general and enhanced visual perception in particular account for superior visual search performance in ASD. Previous studies in healthy subjects (Braddick et al., 2001) have shown that the activity in area V1 is increased in random compared with coherent motion conditions, presumably due to the small receptive fields of the primary visual cortex. In the present study, such an activity pattern was not observed either in the control group or the ASD group. However, the overall higher activation in the primary visual cortex in the ASD group compared with the control group may reflect a different processing of the motion stimuli in both conditions, with a stronger focus on local details (i.e., the random motion parts). The stimuli used in our study prevent the possibility of tracking single dots (see Section 2), and therefore, an enhanced local processing or attention to local stimuli—for example, only to one part of the stimulus—would lead to enhanced perception of random motion and thus might increase the activity in V1. However, the subjects with ASD in our study were able to detect coherent motion and hence were processing the stimulus as a whole. Therefore the assumption of pure local processing and/or attention to local details might not be a sufficient explanation for the increased neural activity in V1. Furthermore, Belmonte and Yurgelun-Todd (2003) reported hyperarousal of visual areas in a selective attention task and suggested an impaired signal-to-noise ratio and impaired selectivity of primary visual processing. In line with this hypothesis Vandenbroucke, Scholte, van Engeland, Lamme, and Kemner (2008) proposed a dysfunction of horizontal connections within early visual areas in autism. They suggested that the malfunction of horizontal connections, which play an important role via lateral inhibition, provide an explanation for atypical perception in autism as a more general neurobiological deficit. In the present study, it was observed that dysfunctional lateral inhibition might lead to abnormal enhanced activity in primary visual cortex. Alternatively, it has been suggested that deficits in visual perception in ASD might be due to an abnormal increase of the magnitude of internal noise (proportional to external noise) and not to the sensation of visual stimuli per se (Sanchez-Marin & Padilla-Medina, 2008). Thus, one might speculate that the presence of random motion in both conditions in our study might have lead to an abnormally high amount of internal noise, possibly due to dysfunctional horizontal connections in subjects with ASD. This possibly makes the detection of motion embedded in noise more demanding for subjects with ASD and therefore might be an explanation for their deficits in detecting coherent motion signals. However, the increased activity of the primary visual cortex in response to motion stimuli in subjects with ASD needs to be further elucidated. The motion-sensitive area V5 in the middle temporal/occipital gyrus is known to integrate local components of motion, mainly signaled by V1, into a coherent motion signal. Previous behavioral studies have shown deficits in global motion perception in autism (e.g. Pellicano et al., 2005), and therefore we hypothesized an abnormal activity in area V5. Contradictory to this hypothesis, we found a significant activation of this area bilaterally in both groups in the coherent > random motion contrast and no interaction effect or between-group difference. However, this finding is 1649 in line with a recent study by Pelphrey et al. (2007). The authors investigated the neural activity of dynamic changes in facial displays of emotional states and found a comparable activation of V5 in the autistic and the control group. Nevertheless, in the present study post hoc t-tests on the parameter estimates revealed that the difference between coherent and random motion was only significant for the control group (significant for the left and trend for the right side), whereas the activation in the random motion condition in the left area V5 was significantly higher in the ASD group compared with the controls. Bertone and Faubert (2006) suggested that impaired motion perception in autism may result from diffuse, non-specific neural dysfunction of early neurointegrative mechanisms, which lead to deficits in the perception of complex stimuli. The unmodulated activity in V5 in the present study would support this assumption of disturbed complex motion processing. Furthermore, an increase of activity in area V5 depending on coherence levels, as proposed by Rees et al. (2000), was not observed, either in the control or in the ASD group. Several imaging studies have proposed differences in connectivity in ASD (e.g. Just et al., 2004). Therefore, we further analyzed the functional connectivity between regions of the dorsal pathway that were activated by the task in both groups, namely the primary visual cortex and human V5. The analysis revealed interactions between these areas in ASD and control subjects, and the group comparison showed no difference of correlation coefficients between groups. This result leads to the conclusion that the interaction within the visual cortex is intact in our ASD sample, and we found no evidence that possible differences in motion perception are a result of abnormal forward or feedback connections between visual areas. This is in line with other study by Bertone et al. (2003, 2005). They suggested that the differences in low-level information processing in autism due to atypical connectivity of lateral (within low-level visual areas) rather than feedback (between visual areas) connections. Furthermore Villalobos et al. (2005) investigated functional connectivity along the dorsal pathway in a visuomotor task. They showed, in contrast to the predictions of the global underconnectivity hypothesis (e.g. Just et al., 2004), no differences of occipitoparietal connectivity. Furthermore, a significant interaction effect was found in the superior parietal cortex (extending to the precuneus), resulting from reduced activation in the coherent (vs. random) motion condition in the ASD group. This is in line with other studies that reported parietal dysfunctions in autism during visuo-spatial attention or visual search tasks (Belmonte & Yurgelun-Todd, 2003; Ring et al., 1999). Moreover, a recent study by Koldewyn et al. (2009) found no support of a general dorsal stream deficit in ASD. Instead they speculated that differences in dynamic or ‘spatiotemporal’ attention lead to the visual motion impairments in ASD. Our result of differences in parietal activation might result in attentional differences between groups and therefore support this hypothesis. Interestingly, Bartels, Zeki, and Logothetis (2008) showed that motion perception during natural vision (movies) in healthy subjects enhances activity (besides other regions) in the medial posterior parietal cortex (mPPC). They further demonstrated that the mPPC is especially activated by global and contrast-independent motion and that this area is highly connected with anterior V5. They proposed that the mPPC mainly responds to stimulus flow and heading and might be a candidate region for comparing visual stimulus flow with self-induced motion. This might be an indication of a deficit in higher level motion processing in autism that might lead to abnormal spatial processing as well as gross motor impairment and clumsiness in Asperger’s disorder (Milne et al., 2006; Muller, Kleinhans, Kemmotsu, Pierce, & Courchesne, 2003; Villalobos et al., 2005). In addition, the whole-brain analysis revealed an activation in the left postcentral gyrus that was significantly activated by 1650 S. Brieber et al. / Neuropsychologia 48 (2010) 1644–1651 coherent motion vs. random motion in both groups. This activation is somewhat surprising, but most studies only focused on specific regions in the visual cortex and might have overlooked additional activations. For example, Culham et al. (1998) showed in an early fMRI study that the foci of attentive-tracking of moving targets are not limited to the SPL but also include adjacent parietal regions in the IPS and the postcentral gyrus. As expected, we did not find performance differences in the fMRI task when using coherence levels above the motion detection threshold shown by recent studies (Milne et al., 2002; Spencer et al., 2000). As group differences in performance levels can be considered confounding factors in the experiments, the absence of these effects facilitates meaningful interpretations of observed group differences in neural mechanisms. More surprisingly, we also did not observe any differences between autistic and healthy subjects in our behavioral motion detection task. Although this is in line with some previous studies (Del Viva et al., 2006; Vandenbroucke et al., 2008b), other earlier studies have reported enhanced motion coherence thresholds in autistic participants (Milne et al., 2002; Spencer et al., 2000). Interestingly, it has been suggested recently that only high-functioning autistic subjects, but not subjects with Asperger’s disorder (Spencer and O’Brien, 2006; Tsermentseli, O’Brien, & Spencer, 2008) might manifest decreased sensitivity to complex motion. The ASD group in the present study included 13 participants with Asperger’s disorder and 2 with high-functioning autism, and therefore, the unimpaired performance of our ASD group is reasonable. Thus, it might be that only a subgroup of participants with ASD, possibly subjects with high-functioning autism, shows a deficit in global motion detection. Another possibility is that, with Asperger participants occupying the upper end of the autism spectrum, they might be able to partly compensate for the global motion deficits. Therefore our fMRI results might reflect early abnormal processing in participants with Asperger’s disorders that can be compensated by higher order activity or might lead to differences in the higher level global perception and integration of visual stimuli. However, for the analysis of subgroups of ASD (Asperger’s disorder and HFA) our sample size was too small and an exclusion of the HFA subjects from the autistic sample did not change the pattern of neural abnormalities. Nevertheless, some of the group differences did no longer survive the WB correction thereshold most likely due to the loss of power and future studies should focus on the direct comparison of both autistic subgroups. Another limitation of this study was that we did not find any significant behavioral differences between the control and the ASD group and further studies are needed to investigate brain activity in ASD subjects with actual deficits in coherent motion processing. In conclusion, this is, to our knowledge, the first study that explicitly investigates visual coherent motion perception in autism by functional MRI. To date, several functional and effective connectivity studies have shown underconnectivity between visual areas (striate and extrastriate areas) and higher social and cognitive processing areas in individuals with autism (Bird, Catmur, Silani, Frith, & Frith, 2006; Just, Cherkassky, Keller, & Minshew, 2004; Villalobos, Mizuno, Dahl, Kemmotsu, & Muller, 2005), but behavioral studies have indicated that visual processing per se is also abnormal in autism. Therefore, the present study investigated the neural correlates of coherent motion perception, which exemplifies early neurointegrative processing, to enhance the understanding of visual processing in the dorsal pathway in participants with autism. Our results indicate an abnormal functioning of the dorsal pathway in autism at both the lower and higher processing stages despite an equal behavioral performance. We hypothesize that hyperactivation in the primary visual cortex might lead to a lack of specific activation in the extrastriate cortex (V5) and that an underactivation of the parietal cortex might contribute to deficits in effective integration and differences of the processing of motion signals. The importance of the investigation of the visual system in ASD participants is further highlighted by a recent study from McCleery et al. (2007). They hypothesized that abnormalities in the magnocellular pathway in early infancy might be associated with ASD and investigated sensitivity for luminance contrasts (mediated by the magnocellular pathway) in infants with a high risk for autism. Their data suggested that increased sensitivity might be an endophenotypic marker for autism, which leads to an atypical development of the input regions of the magnocellular pathway, such as the subcortical face-processing circuits and the dorsal visual pathway. The authors suggested that such early abnormalities could result in visual-perceptual abnormalities, which could lead to the known deficits in ASD, such as abnormal social, emotional, communicative and visual behavior and processing. Financial disclosures Dr. Herpertz-Dahlmann is a consultant for Eli Lilly and has received industry research funding from AstraZeneca, Eli Lilly, Novartis, and Janssen Cilag. The other authors declare that no conflicts of interest exist. Acknowledgements We are grateful to all our volunteers and our colleagues at the Institute of Neuroscience and Biophysics, Department of Medicine, Research Center Juelich. Furthermore we wish to thank Andre Knops for his support in programming the experiment and Ralph Weidner for his helpful comments on an earlier version of the manuscript. This study was supported by a grant to K.K. and G.R.F. by the Interdisciplinary Centre for Clinical Research (IZKF N68a). References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: APA. Amunts, K., Malikovic, A., Mohlberg, H., Schormann, T., & Zilles, K. (2000). Brodmann’s areas 17 and 18 brought into stereotaxic space—where and how variable? NeuroImage, 11, 66–84. Bartels, A., Zeki, S., & Logothetis, N. K. (2008). Natural vision reveals regional specialization to local motion and to contrast-invariant, global flow in the human brain. Cerebral Cortex, 18, 705–717. Belmonte, M. K., & Yurgelun-Todd, D. A. (2003). Functional anatomy of impaired selective attention and compensatory processing in autism. Brain Research Cognitive Brain Research, 17, 651–664. Bertone, A., & Faubert, J. (2006). Demonstrations of decreased sensitivity to complex motion information not enough to propose an autism-specific neural etiology. Journal of Autism and Developmental Disorders, 36, 55–64. Bertone, A., Mottron, L., Jelenic, P., & Faubert, J. (2003). Motion perception in autism: A ‘complex’ issue. Journal of Cognitive Neuroscience, 15, 218–225. Bertone, A., Mottron, L., Jelenic, P., & Faubert, J. (2005). Enhanced and diminished visuo-spatial information processing in autism depends on stimulus complexity. Brain, 128, 2430–2441. Bird, G., Catmur, C., Silani, G., Frith, C., & Frith, U. (2006). Attention does not modulate neural responses to social stimuli in autism spectrum disorders. NeuroImage, 31, 1614–1624. Boelte, S., Hubl, D., Dierks, T., Holtmann, M., & Poustka, F. (2008). An fMRI-study of locally oriented perception in autism: Altered early visual processing of the block design test. Journal of Neural Transmission, 115, 545–552. Braddick, O., Atkinson, J., & Wattam-Bell, J. (2003). Normal and anomalous development of visual motion processing: Motion coherence and ‘dorsal-stream vulnerability’. Neuropsychologia, 41, 1769–1784. Braddick, O. J., O’Brien, J. M., Wattam-Bell, J., Atkinson, J., Hartley, T., & Turner, R. (2001). Brain areas sensitive to coherent visual motion. Perception, 30, 61–72. Culham, J. C., Brandt, S. A., Cavanagh, P., Kanwisher, N. G., Dale, A. M., & Tootell, R. B. (1998). Cortical fMRI activation produced by attentive tracking of moving targets. Journal of Neurophysiology, 80, 2657–2670. Culham, J., He, S., Dukelow, S., & Verstraten, F. A. (2001). Visual motion and the human brain: What has neuroimaging told us? Acta Psychologica (Amst.), 107, 69–94. Dakin, S., & Frith, U. (2005). Vagaries of visual perception in autism. Neuron, 48, 497–507. Davis, R. A. O., Bockbrader, M. A., Murphy, R. R., Hetrick, W. P., & O’Donnell, B. F. (2006). Subjective perceptual distortions and visual dysfunction in children with autism. Journal of Autism and Developmental Disorders, 36, 199–210. S. Brieber et al. / Neuropsychologia 48 (2010) 1644–1651 Del Viva, M. M., Igliozzi, R., Tancredi, R., & Brizzolara, D. (2006). Spatial and motion integration in children with autism. Vision Research, 46, 1242–1252. Delmo, C., Weiffenbach, O., Gabriel, M., Stadler, C., & Poustka, F. (2001). Diagnostisches Interview Kiddie-Sads-Present and Lifetime Version (K-SADS-PL). 5. Auflage der deutschen Forschungsversion, erweitert um ICD-10-Diagnostik [5th edition of the German research version with the addition of ICD-10-diagnosis] Frankfurt: Klinik für Psychiatrie und Psychotherapie des Kindes-und Jugendalters, (pp. 1–241). Eickhoff, S. B., Heim, S., Zilles, K., & Amunts, K. (2006). Testing anatomically specified hypotheses in functional imaging using cytoarchitectonic maps. NeuroImage, 32, 570–582. Eickhoff, S. B., Stephan, K. E., Mohlberg, H., Grefkes, C., Fink, G. R., Amunts, K., et al. (2005). A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. NeuroImage, 25, 1325–1335. Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15, 20–25. Happé, F., & Frith, U. (2006). The weak coherence account: Detail-focused cognitive style in autism spectrum disorders. Journal of Autism and Developmental Disorders, 36, 5–25. Hill, E. L., & Frith, U. (2003). Understanding autism: Insights from mind and brain. Philosophical Transactions of the Royal Society B: Biological Sciences, 358, 281–289. Joseph, R. M., Keehn, B., Connolly, C., Wolfe, J. M., & Horowitz, T. S. (2009). Why is visual search superior in autism spectrum disorder? Developmental Science, 12, 1083–1096. Just, M. A., Cherkassky, V. L., Keller, T. A., Kana, R. K., & Minshew, N. J. (2007). Functional and anatomical cortical underconnectivity in autism: evidence from an FMRI study of an executive function task and corpus callosum morphometry. Cerebral Cortex, 17, 951–961. Just, M. A., Cherkassky, V. L., Keller, T. A., & Minshew, N. J. (2004). Cortical activation and synchronization during sentence comprehension in high-functioning autism: Evidence of underconnectivity. Brain, 127, 1811–1821. Kaufman, J., Birmaher, B., Brent, D., Rao, U., Flynn, C., Moreci, P., et al. (1997). Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): Initial reliability and validity data. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 980–988. Koldewyn, K., Whitney, D., & Rivera, S. M. (2009). The psychophysics of visual motion and global form processing in autism. Brain,. Advance Access published November 3, 2009 Koshino, H., Carpenter, P. A., Minshew, N. J., Cherkassky, V. L., Keller, T. A., & Just, M. A. (2005). Functional connectivity in an fMRI working memory task in highfunctioning autism. NeuroImage, 24, 810–821. Le Couteur, A., Rutter, M., Lord, C., Rios, P., Robertson, S., Holdgrafer, M., et al. (1989). Autism diagnostic interview: A standardized investigator-based instrument. Journal of Autism and Developmental Disorders, 19, 363–387. Livingstone, M. S., & Hubel, D. H. (1987). Psychophysical evidence for separate channels for the perception of form, color, movement, and depth. Journal of Neuroscience, 7, 3416–3468. Lord, C., Risi, S., Lambrecht, L., Cook, E. H. J., Leventhal, B. L., DiLavore, P. C., et al. (2000). The autism diagnostic observation schedule-generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30, 205–223. Malikovic, A., Amunts, K., Schleicher, A., Mohlberg, H., Eickhoff, S. B., Wilms, M., et al. (2007). Cytoarchitectonic analysis of the human extrastriate cortex in the region of V5/MT+: A probabilistic, stereotaxic map of area hOc5. Cerebral Cortex, 17, 562–574. Manjaly, Z. M., Bruning, N., Neufang, S., Stephan, K. E., Brieber, S., Marshall, J. C., et al. (2007). Neurophysiological correlates of relatively enhanced local visual search in autistic adolescents. NeuroImage, 35, 283–291. McCleery, J. P., Allman, E., Carver, L. J., & Dobkins, K. R. (2007). Abnormal magnocellular pathway visual processing in infants at risk for autism. Biological Psychiatry, 62, 1007–1014. Milne, E., Swettenham, J., Hansen, P., Campbell, R., Jeffries, H., & Plaisted, K. (2002). High motion coherence thresholds in children with autism. Journal of Child Psychology and Psychiatry, 43, 255–263. 1651 Milne, E., White, S., Campbell, R., Swettenham, J., Hansen, P., & Ramus, F. (2006). Motion and form coherence detection in autistic spectrum disorder: Relationship to motor control and 2:4 digit ratio. Journal of Autism and Developmental Disorders, 36, 225–237. Milner, A. D., & Goodale, M. A. (2008). Two visual systems re-viewed. Neuropsychologia, 46, 774–785. Mottron, L., Dawson, M., Soulieres, I., Hubert, B., & Burack, J. (2006). Enhanced perceptual functioning in autism: An update, and eight principles of autistic perception. Journal of Autism and Developmental Disorders, 36, 27–43. Muller, R. A. (2008). From loci to networks and back again: Anomalies in the study of autism. Annals of the New York Academy of Sciences, 1145, 300–315. Muller, R. A., Kleinhans, N., Kemmotsu, N., Pierce, K., & Courchesne, E. (2003). Abnormal variability and distribution of functional maps in autism: An FMRI study of visuomotor learning. American Journal of Psychiatry, 160, 1847–1862. Murphy, K., & Garavan, H. (2004). Artifactual fMRI group and condition differences driven by performance confounds. NeuroImage, 21, 219–228. Newsome, W. T., & Pare, E. B. (1988). A selective impairment of motion perception following lesions of the middle temporal visual area (MT). Journal of Neuroscience, 8, 2201–2211. Pellicano, E., Gibson, L., Maybery, M., Durkin, K., & Badcock, D. R. (2005). Abnormal global processing along the dorsal visual pathway in autism: A possible mechanism for weak visuospatial coherence? Neuropsychologia, 43, 1044–1053. Pelphrey, K. A., Morris, J. P., McCarthy, G., & Labar, K. S. (2007). Perception of dynamic changes in facial affect and identity in autism. Social Cognitive and Affective Neuroscience, 2, 140–149. Price, C. J., & Friston, K. J. (1999). Scanning patients with tasks they can perform. Human Brain Mapping, 8, 102–108. Rees, G., Friston, K., & Koch, C. (2000). A direct quantitative relationship between the functional properties of human and macaque V5. Nature Neuroscience, 3, 716–723. Ring, H. A., Baron-Cohen, S., Wheelwright, S., Williams, S. C., Brammer, M., Andrew, C., et al. (1999). Cerebral correlates of preserved cognitive skills in autism: A functional MRI study of embedded figures task performance. Brain, 122, 1305–1315. Rogers, S. J., & Ozonoff, S. (2005). Annotation: What do we know about sensory dysfunction in autism? A critical review of the empirical evidence. Journal of Child Psychology and Psychiatry, 46, 1255–1268. Sanchez-Marin, F. J., & Padilla-Medina, J. A. (2008). A psychophysical test of the visual pathway of children with autism. Journal of Autism and Developmental Disorders, 38, 1270–1277. Spencer, J. V., & O’Brien, J. M. D. (2006). Visual form-processing deficits in autism. Perception, 35, 1047–1055. Spencer, J., O’Brien, J., Riggs, K., Braddick, O., Atkinson, J., & Wattam-Bell, J. (2000). Motion processing in autism: Evidence for a dorsal stream deficiency. Neuroreport, 11, 2765–2767. Tootell, R. B., Reppas, J. B., Kwong, K. K., Malach, R., Born, R. T., Brady, T. J., et al. (1995). Functional analysis of human MT and related visual cortical areas using magnetic resonance imaging. Journal of Neuroscience, 15, 3215–3230. Tsermentseli, S., O’Brien, J. M., & Spencer, J. V. (2008). Comparison of form and motion coherence processing in autistic spectrum disorders and dyslexia. Journal of Autism and Developmental Disorders, 38, 1201–1210. Vandenbroucke, M. W. G., Scholte, H. S., van Engeland, H., Lamme, V. A. F., & Kemner, C. (2008a). Coherent versus component motion perception in autism spectrum disorder. Journal of Autism and Developmental Disorders, 38, 941–949. Vandenbroucke, M. W. G., Scholte, H. S., van Engeland, H., Lamme, V. A. F., & Kemner, C. (2008b). A neural substrate for atypical low-level visual processing in autism spectrum disorder. Brain, 131, 1013–1024. Villalobos, M. E., Mizuno, A., Dahl, B. C., Kemmotsu, N., & Muller, R. (2005). Reduced functional connectivity between V1 and inferior frontal cortex associated with visuomotor performance in autism. Neuroimage, 25, 916–925. World Health Organization. (1993). The ICD-10 classification of mental and behavioural disorders: Diagnostic criteria for research. Geneva: WHO.