Abstract

It has been hypothesized that neural synchrony underlies perceptual coherence. The hypothesis of loss of central perceptual coherence has been proposed to be at the origin of abnormal cognition in autism spectrum disorders and Williams syndrome, a neurodevelopmental disorder linked with autism, and a clearcut model for impaired central coherence. We took advantage of this model of impaired holistic processing to test the hypothesis that loss of neural synchrony plays a separable role in visual integration using EEG and a set of experimental tasks requiring coherent integration of local elements leading to 3-D face perception. A profound reorganization of brain activity was identified. Neural synchrony was reduced across stimulus conditions, and this was associated with increased amplitude modulation at 25–45 Hz. This combination of a dramatic loss of synchrony despite increased oscillatory activity is strong evidence that synchrony underlies central coherence. This is the first time, to our knowledge, that dissociation between amplitude and synchrony is reported in a human model of impaired perceptual coherence, suggesting that loss of phase coherence is more directly related to disruption of holistic perception.

INTRODUCTION

The central coherence hypothesis is a well-known model for cognitive dysfunction in autism spectrum disorders (ASD; Brock, Brown, Boucher, & Rippon, 2002) and basically refers to the impairment in a wide range of tasks requiring coherent integration of local and global object parts (Bernardino, Mouga, Almeida, et al., 2012). A more evident pattern of visual processing characterized by deficits in central coherence and in particular in local–global grouping was found in other clinical model, Williams syndrome (WS). WS is a rare genetic neurodevelopmental disorder with a prevalence of 1 in 7500 to 20,000 live births (Strømme, Bjørnstad, & Ramstad, 2002), which shares an intriguing link with autism (Klein-Tasman, Phillips, Lord, Mervis, & Gallo, 2009). Interestingly, the WS visual phenotype resembles the one found in ASD, although these neurodevelopmental disorders show distinct features in other domains such as in social interaction. WS patients are often described as showing social disinhibition (Capitão, Sampaio, Férnandez, et al., 2011; Capitão, Sampaio, Sampaio, et al., 2011; Dodd, Porter, Peters, & Rapee, 2010), increased approachability (Jawaid et al., 2010), and a particular interest in faces (Riby et al., 2011). Thus, experimental tasks using face stimuli (as in this study) would benefit from the addition of motivational aspects. Face processing in WS was described as being comparable to that of typically developing individuals albeit driven by different mechanisms including a bias for local processing (Karmiloff-Smith et al., 2004; Deruelle, Mancini, Livet, Cassé-Perrot, & de Schonen, 1999). Evidence from studies exploring visual processing of faces and objects in these patients has suggested WS as a model of low (perceptual) coherence (Bernardino, Mouga, Almeida, et al., 2012; Deruelle et al., 1999). WS thereby provides a suitable model to address the hypothesis of dissociation between amplitude modulation and synchrony (Singer, 2009). Although no synchrony studies have been reported in this population, there are already previous reports of neural synchronization disruption in autism (Sun et al., 2012; Dinstein et al., 2011). Adding to this, WS is a well-established genetic neurodevelopmental (Korenberg et al., 2000) model of impaired spatial perception associated with abnormal visual integration and coherence (Bellugi, Lichtenberger, Jones, Lai, & St George, 2000). 2-D and 3-D motion coherence thresholds are impaired (Mendes et al., 2005) as participants fail to integrate local and global information for object perception to a much larger extent than ASD participants (Bernardino, Mouga, Castelo-Branco, & van Asselen, 2012). A distinctive low-level visual phenotype was also observed in WS including decreased retinal thickness (Castelo-Branco et al., 2007), reduced visual acuity (Atkinson et al., 2001), reduced depth perception (van der Geest et al., 2005), and impaired eye movements (van der Geest et al., 2004). Although these features may contribute to altered visual perception in WS, they do not predict high-level deficits in the integration of motion, 3-D information, and face discrimination (Castelo-Branco et al., 2007). WS does therefore represent a relevant model to test for a functional role of synchronization in perceptual integration. Previous studies demonstrated impairments in 3-D structure-from-motion (SFM) perception in WS (Bernardino, Castelhano, Farivar, Silva, & Castelo-Branco, 2013; Mendes et al., 2005) or in response to static faces both in autism and WS (Grice et al., 2001), which inspired the combination of all these paradigms in the current study. Perception of SFM stimuli requires dorsal-ventral stream integration of local features into a coherent whole to achieve holistic perception (Graewe, De Weerd, Farivar, & Castelo-Branco, 2012). This paradigm may therefore help elucidate why gamma oscillatory patterns (variation of oscillatory activity over time) are differentially distributed for different subbands in a task- and/or disease-dependent manner (Castelhano, Duarte, Wibral, Rodriguez, & Castelo-Branco, 2014; Bernardino et al., 2013) and provide further insights into the hypothesis that synchronization predicts perception (Hipp, Engel, & Siegel, 2011).

Moreover, the role of temporal coding in perceptual coherence, defined as hierarchical grouping of local elements, remains controversial (Palva, Palva, & Kaila, 2005; Engel, Fries, & Singer, 2001). It is unclear whether oscillation amplitude is relevant for encoding global stimulus properties or if neural synchrony plays instead a pivotal role in gestalt formation (Gray, 1999). The fact that synchrony and amplitude modulation do often covary has jeopardized their separation in terms of a mechanistic explanation (Palva & Palva, 2012; Singer, 2009). Gamma-band activity has been related to holistic integration and proposed to represent a neural correlate of perceptual coherence, object representation, and binding (Güntekin & Başar, 2014). Accordingly, phase synchronization was shown to correlate with object perception, feature integration (Gray, 1999; Lachaux, Rodriguez, Martinerie, & Varela, 1999; Rodriguez et al., 1999), and regional brain coupling (Varela, Lachaux, Rodriguez, & Martinerie, 2001). Neural coherence, defined as increased spatiotemporal phase-locking, would therefore reflect perceptual coherence (Hipp et al., 2011). The observed gamma frequency ranges in the previous mentioned studies and also during invasive recordings do however vary and seem to have distinct genetic and neurochemical determinants (Muthukumaraswamy, Edden, Jones, Swettenham, & Singh, 2009) and different functional meaning (Groppe et al., 2013; Tsuchiya et al., 2008; Crone, Sinai, & Korzeniewska, 2006). These make it hard to understand the functional significance of given modulations at particular frequency ranges. In fact, recent findings using multimodal approaches have shown distinct subbands within the gamma-band that have very distinct neurophysiological substrates and functional correlates (Castelhano et al., (2014); Sedley & Cunningham, 2013; Martinovic & Busch, 2011; Pockett, Bold, & Freeman, 2009). An important way to disentangle their relevance to perceptual integration would be to explicitly look at synchronization patterns at each band, as more direct indicators of neural coherence, which we hypothesize to subserve perceptual integration and coherence.

Additional evidence for a role of oscillatory patterning in perceptual organization comes from studies in clinical populations (Başar, Başar-Eroğlu, Güntekin, & Yener, 2013) with fragmented perception such as autism (Uhlhaas & Singer, 2012; Brock et al., 2002) and schizophrenia, which have been associated with disrupted gamma-band activity (Uhlhaas & Singer, 2010; Spencer et al., 2003).

Here we tested whether synchronization and amplitude modulation of gamma patterns can be segregated in WS, a model of impaired central coherence (Bernardino, Mouga, Almeida, et al., 2012), as motivated by the hypothesis that neural coherence more directly relates to perceptual coherence. We designed a set of experimental tasks requiring feature integration at different complexity levels (static and motion stimuli as well as 2-D and 3-D face stimuli) and compared changes in synchrony and concurrent changes in signal amplitude (Hipp et al., 2011) between groups. We found that neural synchrony and the amplitude of oscillatory patterns dissociate under impaired coherence, with loss of coherence at perception-related frequency bands, thereby providing a novel clue for their functional significance in health and disease.

METHODS

Participants

Participants were 17 individuals. Nine WS participants (four male and five female) were aged between 15 and 37 years (mean ± SE = 21.44 ± 2.30). The WS participants were recruited from a database used in previous studies (Bernardino et al., 2013; Castelo-Branco et al., 2007; Mendes et al., 2005). All patients were diagnosed based on clinical and genetic examinations confirmed by FISH analysis, which demonstrated the hemyzigous Elastin deletion. Additional genetic analysis sequenced the breakpoint regions and revealed the same deletion size (1.55 Mb) in all WS participants. All WS patients underwent a complete ophthalmologic examination performed by an experienced ophthalmologist, including best-corrected visual acuity (Snellen optotypes) complete oculomotor examination, stereopsis evaluation using the Randot test, slit lamp examination of anterior chamber structures, and fundus examination. No abnormalities that could affect vision were identified.

The eight control participants (four male and four female) were aged between 15 and 34 years (mean ± SE = 21.89 ± 2.40). Groups were matched on chronological age (Mann–Whitney U test, p > .05), sex, and handedness. Two participants who demonstrated left-hand dominance participated in each group. All participants had normal or corrected-to-normal vision and were naive regarding the purpose of the study.

Written informed consent was obtained from all participants or their parents. The study was conducted according to the Declaration of Helsinki and was approved by the local ethics committees of the Faculty of Medicine of the University of Coimbra.

Task Setup and Procedure

Participants were seated in a comfortable chair and were positioned at a distance of 120 cm from the computer screen in a darkened, acoustically and electrically shielded room. The stimuli (size: 13° horizontally and 10° vertically) were presented using Presentation software (v14.9, Neurobehavioral Systems, Victoria, Australia) in the center of a CRT monitor screen (1024 × 768 with a refresh rate of 60 Hz). The experimental paradigm included four different stimulus conditions: Static Dots, Static Neutral Face, SFM-defined neutral face, and Random Dots (Figure 1). The emphasis on dynamic stimuli was based on the fact that motion integration requires spatiotemporal correlation in the sense of Reichardt coincidence detectors (Reichardt, 1987). The static dot stimulus consisted of one frame of the SFM video (Figure 1A). In the static face condition, participants viewed a mask composed of a large set of dots drawn to be similar to the SFM stimulus appearance (Figure 1B). In the SFM condition, the stimuli consisted of videos of SFM-defined faces with 980 msec of duration. In this type of stimulus, a large set of coherent dots is moving, rotating in a cycle from −22.5° to 22.5° centered at the frontal plane. This enables the formation of 3-D shape percepts (3-D face). Importantly, shape extraction can only be achieved when dots are moving, because otherwise the shape remains invisible (Figure 1C; Bernardino et al., 2013; Graewe et al., 2012). In the random dot condition, a set of white dots was randomly moving on a black background (Figure 1D). The four experimental conditions were randomly presented during 980 msec and were followed by a fixation period in which participants were required to press a button if the presented stimulus was a face and another button if the presented stimulus was not a face. The response was reported using a Cedrus response box (RB-530 model) and was required only after the stimulus offset to avoid contamination by motor activity. RT was unlimited, and the next trial was only presented after the participants' response. A 300-msec fixation period separated the RT from the test trial presentation (Figure 1E). The experimental task was composed of 100 test trials of each task condition resulting in a total of 400 test trials, which were divided in four blocks. The behavioral task was conducted also to ensure that participants were able to maintain attention on the visual stimuli throughout the acquisition session. Twelve learning trials were administered, and the practice phase was repeated whenever necessary to ensure full understanding of the task and motor coordination.

Figure 1. 

Examples of stimulus conditions and task timeline. (A–D) Stimuli used in the task. Static (A, B) and motion (C, D) stimuli. Note that integration of dots is required for face perception. (E) Timeline of the task. Stimuli (100 trials per condition divided in four runs) were randomly presented for 980 msec, and participants had to report at the end of the trial if a face was present.

Figure 1. 

Examples of stimulus conditions and task timeline. (A–D) Stimuli used in the task. Static (A, B) and motion (C, D) stimuli. Note that integration of dots is required for face perception. (E) Timeline of the task. Stimuli (100 trials per condition divided in four runs) were randomly presented for 980 msec, and participants had to report at the end of the trial if a face was present.

Electrophysiological Recordings

EEG signals were recorded using a SynAmps 2 amplifier (64 channels Quick-Cap, Compumedics Neuroscan, Charlotte, NC) and digitized at a 2-kHz sampling rate. The EOG was monitored via electrodes positioned at the standard positions (vertical, VEOG and horizontal, HEOG) to correct artifacts due to blinking and eye movements. No notch filters were used during EEG recording (50 Hz notch was applied offline), and impedances of each electrode were kept below 10 kΩ (electrodes that did not fulfill this criterion, no more than 4 of 64 channels were marked as “bad”). The reference was an electrode placed in the quick-cap close to CZ (the same for all participants) as supplied by the cap manufacturer. Data were re-referenced offline to the average reference.

ERP Data Processing and Analysis

The data were processed offline using the Scan 4.5 Edit Software (Neuroscan). Bad channels were removed and interpolated (as implemented in EEGLAB) with spherical interpolation so that every participant has the same number of channels. A digitally low- and high-pass filter (at 30 Hz, 24 dB/oct and 1 Hz, 12 dB/oct, respectively) was applied. Data were then segmented into epochs starting 100 msec before stimulus onset to 500 msec after stimulus onset. The baseline time interval was set to the time before stimulus onset (−100 to 0 msec). An automatic artifact rejection was applied, and epochs with amplitudes exceeding ±75 μV were rejected (>80% of the trials were accepted for further processing). The epochs for each stimulus type were averaged together, and ERPs were computed. We defined four different clusters of electrodes for further analysis (Bernardino et al., 2013; Graewe et al., 2012), namely occipital, temporal, parieto-occipital, and parietal. The occipital cluster included the O1, OZ, and O2 electrodes, the temporal cluster was constituted by the TP7, CP5, P7, P5, PO7, TP8, CP6, P6, P8, and PO8 electrodes, the posterior-occipital cluster comprised the PO5, PO3, POZ, PO4, and PO6 electrodes, and, finally, the parietal cluster included the CP3, CP1, P3, P1, CP2, CP4, P2, and P4 electrodes.

Time–Frequency and Phase Synchrony Analysis

After filtering the data (1–100 Hz), we performed signal correction of eye movement-related artifacts as described previously (Keren, Yuval-Greenberg, & Deouell, 2010), using independent component analysis based on all electrodes (including four EOG channels) as implemented in EEGLAB (version 10.2.5.6). Artifact components were identified based on their correlation with the EOG electrodes and on the scalp topography (increased activity distribution around the orbits) and removed from the data. Moreover, other artifact components (e.g., muscular artifacts) were also identified with independent component analysis, automatic epoch rejection, and visual inspection and removed from the data. Epochs with artifacts exceeding ±75 μV were rejected. On average 72.31 ± 12.17% for controls and 79.48 ± 7.65% for WS of the epochs remained for further analysis. The number of accepted epochs did not significantly differ between groups (p > .05). Data were segmented into epochs locked to the stimulus onset (from −0.5 to 1.5 sec) for all conditions.

Time–frequency and phase synchrony analysis were performed as in Uhlhaas et al. (2006) and are described elsewhere (Castelhano, Rebola, Leitão, Rodriguez, & Castelo-Branco, 2013; Uhlhaas et al., 2006; Lachaux et al., 1999; Rodriguez et al., 1999). Time–frequency analysis was carried out in Matlab (The MathWorks, Natick, MA) across distinct defined electrode clusters (parietal, temporal, parieto-occipital, and occipital; see above). Signals were time–frequency-analyzed using the pseudo Wigner–Ville transformation. For every time window and frequency bin (frequency resolution of 1 Hz/frequency bin), the amplitude and phase were computed (10–90 Hz in steps of 1 Hz) particularly in the high-beta/gamma frequency range and in the time period of the epochs described above. In this way we focused on the modulations of activity in the gamma-band range. We analyzed the so-called “induced” gamma activity, which is not phase-locked to stimulus onset (by computing time–frequency transforms of single epochs and averaging them across all trials). From the phase values, we calculated the phase-locking value that measures the intertrial variability of the phase difference (Lachaux et al., 1999). Because we were interested in long-range coordination of neural activity, the phase difference was calculated between all electrode pairs and averaged over trials across a wide frequency range (Uhlhaas et al., 2006; Lachaux et al., 1999; Rodriguez et al., 1999).

Time–frequency and phase were then normalized by subtracting the baseline average (prestimuli) and dividing by the baseline standard deviation on a frequency-by-frequency basis (Lachaux et al., 1999). The spatiotemporal distributions of coherence (higher phase locking across channels are drawn in a 2-D head map) were calculated for distinct frequency bands and five consecutive time windows of 200-msec length locked to the stimulus onset. Graphs of synchrony lines, representing enhanced phase locking, are shown per group and stimulus condition for two gamma frequency sub-bands (low frequency: LF, 25–45 Hz; high frequency: HF, 60–90 Hz) across time. Synchrony lines are drawn only if the synchrony value is above a two-tailed probability value of p = .01 (Rodriguez et al., 1999).

Statistical Analysis

Statistical comparisons were performed to compare ERPs and time–frequency results between groups and conditions with the alpha level set to .05. Standard statistical comparisons of the ERPs were performed as they are implemented in EEGLAB (including Holms correction for multiple comparisons, for Conditions × Groups) to assess significant differences between groups. For the time–frequency and synchrony statistical comparison between groups, Mann–Whitney U tests and Wilcoxon rank sum test were carried out. Statistics were obtained per time, and frequency points using the Wilcoxon test and p values were corrected for multiple comparisons using a false discovery rate (FDR) approach. p values are shown for each time–frequency point in a colormap scale in the correspondent panel of each figure. Additionally, bootstrap statistics with 1000 bootstrap samples were used to assess significance comparing the number of synchrony lines between groups. All statistical analyses were performed with the IBM (Armank, NY) SPSS Statistics 19.0 software package and Matlab.

RESULTS

Behavioral and ERP Data Analysis

In the present task, participants were required to discriminate between face stimuli and nonface stimuli. Both WS and control participants were performance-matched when discriminating face stimuli from nonface stimuli (Mann–Whitney tests, p > .05; error rates per condition: static dots, 11.63 ± 4.93%, 3.29 ± 1.52%; static faces, 9.63 ± 4.42%, 1.43 ± 0.95%; SFM, 17.12 ± 12.90%, 1.93 ± 1.20%; random dots, 11.75 ± 5.56%, 7.43 ± 6.44%; mean ± SE for WS and controls, respectively). Accordingly, results indicated that all the participants were successfully performing the tasks, which were adjusted for difficulty to ensure matching. This is relevant because it allows the study of putative compensatory neural mechanisms underlying visual motion integration in WS when performance is matched across groups.

Figure 2 illustrates the ERP data (grand-averaged waveforms for WS and typical controls) for each stimuli condition in the parieto-occipital cluster. For control participants, we found the expected early positive P100 visual component, followed by the expected face-related negative N170 component. We confirmed that WS participants showed lower P100 amplitude (particularly in the moving task conditions) than control participants, an earlier N150 component as well as a novel positive peak—P200 (Figure 2).

Figure 2. 

ERP plots are represented for each condition (columns) for the parieto-occipital cluster of electrodes. Controls (black line) and WS (gray dashed line) differed mostly around N170 and P200 peaks (see text for details). Gray bars represent the time windows of significant differences (p < .05 with Holms correction for multiple comparisons).

Figure 2. 

ERP plots are represented for each condition (columns) for the parieto-occipital cluster of electrodes. Controls (black line) and WS (gray dashed line) differed mostly around N170 and P200 peaks (see text for details). Gray bars represent the time windows of significant differences (p < .05 with Holms correction for multiple comparisons).

We performed statistical analyses for all clusters (occipital, occipito-temporal, occipito-parietal, and parietal regions) and task conditions (static dots, static face, SFM, and random dots). Significant differences between the two groups (p < .05) were mostly identified just around the emergence of the N170 standard face component (120–180 msec time window) and the P200 component (220–280 msec time window). These differences were observed for all task conditions (static dots, p < .05; static faces, p < .038; SFM, p < .048; random dots, p < .043; see gray bars in Figure 2).

Time–Frequency Analysis: Amplitude and Synchrony Differences across Distinct Gamma-band Ranges for the High versus Low Perceptual Coherence Groups

We focused on induced gamma oscillations known to be important for holistic processing, in particular in the low 25–45 Hz frequency band (Kaiser & Lutzenberger, 2003). We found a dual gamma activity pattern at low- and high-gamma frequency ranges in both groups and dependent on the type of stimulus for the different stimulus conditions in the parieto-occipital cluster (Figure 3). Details on statistical analyses concerning the group comparisons for all conditions are reported in the respective figures. p values are shown for each time–frequency point in a colormap scale in the correspondent figure panel. The number of channels differs across clusters, but the same results also hold true even when one considers individual channels.

Figure 3. 

Time–frequency analysis results per experimental condition shows distinct oscillatory patterns across groups. (A) Normalized (to prestimulus baseline) gamma-band activity patterns for the four conditions for the control group. (B) Time–frequency analysis for the WS group. Color codes indicate normalized scores. (C) Statistical maps for the between-group comparison (Ctr vs. WS, Wilcoxon test with FDR-corrected p values shown per time and frequency points). The blue line represents the stimulus onset, and the gray background highlights the low-frequency range. The analysis depicted here corresponds to the parieto-occipital cluster of electrodes. Stimulus-driven 60–90 Hz gamma oscillations dominate in the control group, whereas in WS modulations have larger amplitudes near 25–45 Hz.

Figure 3. 

Time–frequency analysis results per experimental condition shows distinct oscillatory patterns across groups. (A) Normalized (to prestimulus baseline) gamma-band activity patterns for the four conditions for the control group. (B) Time–frequency analysis for the WS group. Color codes indicate normalized scores. (C) Statistical maps for the between-group comparison (Ctr vs. WS, Wilcoxon test with FDR-corrected p values shown per time and frequency points). The blue line represents the stimulus onset, and the gray background highlights the low-frequency range. The analysis depicted here corresponds to the parieto-occipital cluster of electrodes. Stimulus-driven 60–90 Hz gamma oscillations dominate in the control group, whereas in WS modulations have larger amplitudes near 25–45 Hz.

Reorganization of gamma-band activity patterns was found in the impaired central coherence clinical group, both concerning the 25–45 Hz (believed to be involved in gestalt formation) and 60–90 Hz bands. Regarding static face stimuli, time–frequency results confirmed (Grice et al., 2001) that in the 60–90 Hz gamma-frequency range WS patients show decreased activity in comparison to controls (see Figure 3C for detailed FDR corrected p values).

Regarding the conditions containing moving stimuli (SFM and random dots), between-group comparisons revealed differences in both conditions for both the 60–90 and 25–45 Hz ranges. SFM results replicate our previous finding (Bernardino et al., 2013, using a much more restricted set of stimuli than in the current study and not assessing phase coherence) that control participants show increased activity for the 60–90 Hz band whereas WS participants show significantly stronger activation at the lower gamma-band range (25–45 Hz; Figure 3). The same pattern of results was also observed for the random dot condition for similar subbands.

Loss of Synchrony in the Low Perceptual Coherence Group

We evaluated the phase synchrony, a putative measure of binding between distinct brain regions (Varela et al., 2001). We computed synchronization between all electrode pairs for the different stimulus conditions (see Figure 4A and B). Phase synchrony was significantly different between groups for all conditions, except for static faces, in distinct frequency subbands (Figure 4C). Synchrony plots show a decrease of synchronization in the lower-frequency range in the WS group (Figure 4C) in spite of the larger amplitude modulation.

Figure 4. 

Phase synchrony results for the four conditions. Synchrony was calculated for all possible combination of pairs of electrodes and is locked to the beginning of the stimuli (blue line) for the control (A) and WS (B) groups. Color scale depicts normalized units. Baseline normalization was performed for the time window preceding the stimulus onset. (C) Representation of the time–frequency points with significant differences between groups (color scale shows the range of p values, and the gray background highlights the low-frequency range). (D) Synchrony spectrum (within group averages over time of the results in A and B) separated by stimuli conditions. For sake of clarity, standard deviations were omitted in these plots. Note that controls (red line) show higher synchronization for the lower frequencies. Note that for the static faces condition, this difference is attenuated. (E) Group-averaged synchrony spectrum pooled over conditions is shown. Red and green bars at the bottom mark the frequencies where groups differed (p < .02, FDR-corrected): Controls showed higher synchrony at 10–22 and 30–43 Hz (with large effect size, see text for details on statistics) and WS (green line) with increased synchrony for the 62–79 Hz frequency band.

Figure 4. 

Phase synchrony results for the four conditions. Synchrony was calculated for all possible combination of pairs of electrodes and is locked to the beginning of the stimuli (blue line) for the control (A) and WS (B) groups. Color scale depicts normalized units. Baseline normalization was performed for the time window preceding the stimulus onset. (C) Representation of the time–frequency points with significant differences between groups (color scale shows the range of p values, and the gray background highlights the low-frequency range). (D) Synchrony spectrum (within group averages over time of the results in A and B) separated by stimuli conditions. For sake of clarity, standard deviations were omitted in these plots. Note that controls (red line) show higher synchronization for the lower frequencies. Note that for the static faces condition, this difference is attenuated. (E) Group-averaged synchrony spectrum pooled over conditions is shown. Red and green bars at the bottom mark the frequencies where groups differed (p < .02, FDR-corrected): Controls showed higher synchrony at 10–22 and 30–43 Hz (with large effect size, see text for details on statistics) and WS (green line) with increased synchrony for the 62–79 Hz frequency band.

This finding is further detailed in Figure 4D, showing the synchrony spectrum average over time in the interval after stimulus onset for individual stimuli for both groups. Most importantly, spectral comparisons between groups show that controls have higher synchrony for the low-frequency range (25–45 Hz) whereas WS participants have slightly increased synchrony for the higher gamma-band (60–90 Hz).

Regarding the detailed spectral analysis by stimulus type, controls have also increased synchrony in the low range for the static dot condition (2.1 < z < 2.9, .003 < p < .036), for the SFM condition (1.73 < z < 2.02, .043 < p < .046), as well as for the random dot condition (1.92 < z < 2.50, .011 < p < .027).

WS participants showed increased synchrony for the higher band only for the static dot condition (.012 < p < .036, 2.1 < z < 2.52) as well as for the random dot condition (2.02 < z < 2.41, .016 < p < .043). Moreover, this finding was spatially restricted (see analysis of synchrony lines below), and differences did not reach statistical significance for the static face condition, in line with the idea that lower band (25–45 Hz) oscillations that have been classically related to feature binding were more relevant.

We also performed a pooled analysis across conditions and found that control participants have higher synchronization also at 10–22 Hz (p < .017 FDR-corrected, 2.52 < z < 3.09) and 30–45 Hz (p < .02 FDR-corrected, 4.50 < z < 6.36) when compared to WS. This analysis further confirmed that WS patients exhibit increased synchronization only at the higher frequency band (p < .015 FDR-corrected, 2.59 < z < 3.31) in comparison to controls.

Because synchronization near the 40-Hz frequency is believed to reflect coherent perception (Tallon-Baudry, 2003), we analyzed topography maps of synchrony across multiple gamma subbands. The spatiotemporal distribution of these subbands revealed distinct patterns of synchronization for different frequencies and time intervals. Figure 5 shows the synchrony patterns between electrodes pairs. Only pairs are considered whose synchrony value reaches a two-tailed probability value of p = .01 (SFM condition; similar results were obtained for the other conditions, data not shown).

Figure 5. 

Phase synchrony changes in the LF (25–45 Hz) and HF (60–90 Hz) bands plotted across space and time. Circles indicate electrode positions, with anterior sites at top and posterior sites at bottom. Results are shown for the SFM condition over six time windows spanning the entire duration of the stimuli (−200–0 msec was defined as baseline). Red lines mark increased synchrony between electrode sites for the control group, and green lines mark higher synchrony for the WS group (lines are plotted with a significance threshold of p < .01). Note the increased number of lines (representing increased synchrony) for the LF band in the control group.

Figure 5. 

Phase synchrony changes in the LF (25–45 Hz) and HF (60–90 Hz) bands plotted across space and time. Circles indicate electrode positions, with anterior sites at top and posterior sites at bottom. Results are shown for the SFM condition over six time windows spanning the entire duration of the stimuli (−200–0 msec was defined as baseline). Red lines mark increased synchrony between electrode sites for the control group, and green lines mark higher synchrony for the WS group (lines are plotted with a significance threshold of p < .01). Note the increased number of lines (representing increased synchrony) for the LF band in the control group.

These topographic maps show that synchronization pairs for the low gamma frequency band (LF, 25–45 Hz) cover a large brain network over an extended time period in the control group. This is in contrast with the pattern exhibited by the WS group that shows substantially less synchrony in accordance with the reduced ability to form coherent percepts. Bootstrap statistical analysis including all stimuli combined revealed that this reduction of synchrony is significant for the first two time windows after stimulus onset for the LF (25–45 Hz) band (0–200 msec: mean difference = 16.82, p < .027, bias = 0.32, SE = 7.53; 200–400 msec: mean difference = 27.32, p < .020, bias = 0.34, SE = 10.04). Interestingly, in the control group, synchronization is highly attenuated for higher frequencies (HF, 60–90 Hz) for all stimulus conditions. However, between-group differences in the degree of synchrony for this band were not significant (p > .11), further supporting the notion that it is the 25–45 Hz band that critically differentiates the two groups. The topography of synchronization patterns shows the electrodes that were consistently participating in these coherent patterns. These were located at the parieto-occipital and temporal clusters, which lie over posterior dorsal stream areas and temporal ventral regions, respectively. These electrodes synchronized mainly with the frontocentral electrodes over decision-related brain regions (Castelhano et al., 2013; Rebola, Castelhano, Ferreira, & Castelo-Branco, 2012). Moreover, we have assessed the effect of age and found that it does not influence our neurophysiological results.

These results demonstrate that synchronization is altered in WS, particularly in the lower-frequency band (25–45 Hz) in particular for dynamic stimuli, which inherently require more spatiotemporal integration and phase/coincidence detection. Imaginary coherence analysis (Nolte et al., 2004) further confirmed our findings. This helps ruling out volume conduction biases. A particularly important observation was that synchrony was remarkably increased for the frequency range in which the amplitude was reduced.

DISCUSSION

In this study, we took advantage of a well-characterized model of impaired central coherence to examine the functional role of neural synchrony (Başar & Güntekin, 2013; Brock et al., 2002). WS patients show markedly disrupted visual coherence and holistic visual perception (Bernardino, Mouga, Almeida, et al., 2012; Meyer-Lindenberg et al., 2004). If synchrony is supporting holistic perception, we predicted that it should be diminished (Tallon-Baudry, 2003) in this model, in tasks involving visual integration. In previous studies, gamma-band activity and synchrony covaried in association with perceptual integration (e.g., Grützner et al., 2010; Martinovic, Gruber, Hantsch, & Müller, 2008; Kaiser & Lutzenberger, 2003; Keil, Müller, Ray, Gruber, & Elbert, 1999), rendering it difficult to separate amplitude modulations from changes in synchrony. Our hypothesis was that synchrony and oscillatory power do not necessarily covary upon loss of perceptual coherence and that the former is likely lost at bands related to perceptual decision.

Patients showed increased amplitude but reduced synchrony compared with controls in the LF gamma range, particularly during processing of SFM stimuli requiring the integration of information from the visual dorsal and ventral streams. This is consistent with the notion that, even if synchrony remains constant, changes in the components' amplitudes might still change or vice versa (Hipp et al., 2011). The pattern of synchrony loss as opposed to amplitude increases was seen regardless of similar behavioural performance across groups. In fact, as in our previous study using SFM stimulus (Bernardino, Rebola, Farivar, Silva, & Castelo-Branco, 2014; Bernardino et al., 2013), we ensured that WS participants were able to perceive the presented visual stimulus at matched performance levels to test whether the neural strategies (reflecting compensatory mechanisms) are distinct. Compensatory mechanisms were found in previous studies using SFM stimulus (Bernardino et al., 2013, 2014), but also for faces, a stimulus type in which WS patients exhibit normal or near-normal behavioural performance. We hypothesize that these compensatory mechanisms contribute to the distinct electrophysiological pattern found in WS. In other words, compensatory brain activity patterns can be identified irrespective of behavioral differences between groups.

Here, we show for the first time that a model of fragmented coherence in perception is associated with abnormal phase-synchrony even when gamma-band activity increases in the same time–frequency interval. This allowed us to avoid the confound of synchrony with amplitude covariance.

Interestingly, the task condition possibly requiring the highest level of integration (SFM coherence) was the one associated with the most dramatic loss of synchrony. This suggests a dependence on visual complexity and integration requirements and is in line with the proposal that perception requires mechanisms for integration and feature binding that rely on the neuronal synchrony (Hipp et al., 2011; Engel et al., 2001; Varela et al., 2001). Moreover, effects were more prominent for all dynamic conditions. This also suggests that this effect was more closely related to motion processing rather than gestalt perception per se, although it is worth pointing out that motion perception implies path integration and analysis of dot motion coherence. Accordingly, motion perception requires spatiotemporal correlation and is often impaired in syndromes associated with a reduction of perceptual coherence. Finally, influential theoretical models postulate correlation-based motion detection (Reichardt correlation or coincidence detection models; Reichardt, 1987).

Previous studies in WS did not address synchrony and used a much more restricted set of paradigms (Bernardino et al., 2013; Grice et al., 2001). It is known that coherent visual motion induces high-frequency gamma oscillatory activity (Müller, Junghöfer, Elbert, & Rochstroh, 1997) in parieto-occipital regions of the visual dorsal stream (Gruber, Müller, Keil, & Elbert, 1999). On the other hand, a recent report showed that the gamma-band does not index visual form-motion integration but the beta band does, suggesting that multiple bands with different functional significance have to be considered (Aissani, Martinerie, Yahia-Cherif, Paradis, & Lorenceau, 2014). However, this was an MEG study, and the frequency ranges reported (55–85 Hz) are distinct from the significant ones reported here (25–45 Hz), for which we showed a clear role in perceptual decision in a previous article (Castelhano et al., 2014). Sun et al. (2012) found reductions of both gamma amplitude (up to 150 Hz) and phase-locking in ASD. They did not observe the dissociation reported here, which is in line with the notion that power variations may show increases or decreases that may or may not covary. In our study, we could show that they are dissociable and that low-frequency gamma activity reflects integration of dynamic information based on gamma synchrony irrespective of changes of amplitude. This provides an intriguing link to dorsal stream functions, which have been implicated in the analysis of motion and spatial relationships between objects (Ungerleider & Haxby, 1994; Mishkin, Ungerleider, & Macko, 1983). Our findings extend the notion that gamma-band activity serves an important role in a variety of processes (in particular feature integration) reflecting distinct mechanisms (Sedley & Cunningham, 2013; Tallon-Baudry, 2003) and that different activity patterns for particular gamma subbands may have different roles and sources within the gamma frequency range (see, e.g., Castelhano et al., 2014; Bernardino et al., 2013; Sedley & Cunningham, 2013; Herrmann & Kaiser, 2011; Gruber, Martinovic, & Muller, 2008). It is worth pointing out that other effects might have been missed due to the low-sample size, which may have reduced our statistical power. The decision of dividing EEG channels into clusters is also a limitation of this study.

The finding that controls and WS have distinct synchrony patterns is relevant to akin conditions such as ASD, as well as schizophrenia and other neuropsychiatric conditions with fragmented perception (Uhlhaas & Singer, 2010; Klein-Tasman et al., 2009; Spencer et al., 2003; Grice et al., 2001). The identified signature of a dramatic loss of synchrony as opposed to concomitantly increased gamma activity provides evidence that it is a marker for central coherence, irrespective of performance levels. WS is a neurodevelopmental disorder, and in spite of the inclusion of both adolescents and adults, our correlation analysis showed that age does not change the pattern of these results.

Furthermore, given that WS has a genetic basis, which is implicated in cortical circuit specification with a direct impact on the phenotype (Castelo-Branco et al., 2007), it is relevant to consider the molecular mechanisms underlying abnormal oscillatory patterning (Bernardino et al., 2013). It is known that high-frequency oscillatory activity in the cortex results from interactions between GABAergic inhibitory interneurons (Chen et al., 2014; Muthukumaraswamy et al., 2009) and is impaired in some neuropsychiatric disorders such as schizophrenia, epilepsy, and ADHD (Herrmann & Demiralp, 2005). Moreover, it has been shown that GABA concentration is positively correlated with stimulus-induced gamma-oscillations during a visual task (Muthukumaraswamy et al., 2009). Hence, abnormalities in synchronous neural activity could be disrupted in WS due to a dysfunction of inhibitory interneuron networks. Thus, future studies should explore mechanisms underlying abnormal patterning of gamma-band activity in WS, in particular with respect to neural mechanisms of synchronization. Given the evidence that GABA levels are altered in other pathologies with visuospatial impairment associated with dorsal stream dysfunction, such as neurofibromatosis type I (Violante et al., 2013) and autism (Coghlan et al., 2012), this theoretical thread is worth pursuing. Further studies focusing on the connectivity of distributed cortical networks, interhemispheric interactions, and neurochemical phenotypes may also help elucidate this issue.

Our experimental results support the notion that gamma-band synchronization deficits are associated with loss of central coherence and suggest a new neural correlate for impaired cognition in neuropsychiatric disorders (Spencer et al., 2003). This abnormality may be at the basis of some of the manifestations of disorders known to have central coherence deficits. In summary, our data indicate that neural synchrony in the lower gamma frequency range is more important in the generation of perceptual coherence than increases in oscillatory power.

Acknowledgments

We are grateful to the participants and their families for their participation and support. We thank Dr. Eduardo Silva for helping in participant recruitment and ophthalmological assessment. This research was supported by Bial Foundation grants 132 and133, CENTRO-07-ST24-FEDER-00205 From molecules to man, FP7-HEALTH-2013-INNOVATION-1-602186-BRAINTRAIN, PTDC/SAU-ORG/118380/2010, COMPETE, FCT-UID/NEU/04539/2013 to M. C. B. (PI) and individual scholarship SFRH/BD/65341/2009 to J. C.

Reprint requests should be sent to Miguel Castelo-Branco, IBILI-Faculty of Medicine, University of Coimbra, Azinhaga de Santa Comba, 3000-548 Coimbra, Portugal, or via e-mail: mcbranco@fmed.uc.pt.

REFERENCES

Aissani
,
C.
,
Martinerie
,
J.
,
Yahia-Cherif
,
L.
,
Paradis
,
A.-L.
, &
Lorenceau
,
J.
(
2014
).
Beta, but not gamma, band oscillations index visual form-motion integration
.
PloS One
,
9
,
e95541
.
Atkinson
,
J.
,
Anker
,
S.
,
Braddick
,
O.
,
Nokes
,
L.
,
Mason
,
A.
, &
Braddick
,
F.
(
2001
).
Visual and visuospatial development in young children with Williams syndrome
.
Developmental Medicine & Child Neurology
,
43
,
330
337
.
Başar
,
E.
,
Başar-Eroğlu
,
C.
,
Güntekin
,
B.
, &
Yener
,
G. G.
(
2013
).
Brain's alpha, beta, gamma, delta, and theta oscillations in neuropsychiatric diseases: Proposal for biomarker strategies
.
Supplements to Clinical Neurophysiology
,
62
,
19
54
.
Başar
,
E.
, &
Güntekin
,
B.
(
2013
).
Review of delta, theta, alpha, beta, and gamma response oscillations in neuropsychiatric disorders
.
Supplements to Clinical Neurophysiology
,
62
,
303
341
.
Bellugi
,
U.
,
Lichtenberger
,
L.
,
Jones
,
W.
,
Lai
,
Z.
, &
St George
,
M.
(
2000
).
I. The neurocognitive profile of Williams syndrome: A complex pattern of strengths and weaknesses
.
Journal of Cognitive Neuroscience
,
12(Suppl. 1)
,
7
29
.
Bernardino
,
I.
,
Castelhano
,
J.
,
Farivar
,
R.
,
Silva
,
E. D.
, &
Castelo-Branco
,
M.
(
2013
).
Neural correlates of visual integration in Williams syndrome: Gamma oscillation patterns in a model of impaired coherence
.
Neuropsychologia
,
51
,
1287
1295
.
Bernardino
,
I.
,
Mouga
,
S.
,
Almeida
,
J.
,
van Asselen
,
M.
,
Oliveira
,
G.
, &
Castelo-Branco
,
M.
(
2012
).
A direct comparison of local-global integration in autism and other developmental disorders: Implications for the central coherence hypothesis
.
PloS One
,
7
,
e39351
.
Bernardino
,
I.
,
Mouga
,
S.
,
Castelo-Branco
,
M.
, &
van Asselen
,
M.
(
2012
).
Egocentric and allocentric spatial representations in Williams syndrome
.
Journal of the International Neuropsychological Society
,
19
,
54
62
.
Bernardino
,
I.
,
Rebola
,
J.
,
Farivar
,
R.
,
Silva
,
E.
, &
Castelo-Branco
,
M.
(
2014
).
Functional reorganization of the visual dorsal stream as probed by 3-D visual coherence in Williams syndrome
.
Journal of Cognitive Neuroscience
,
26
,
2624
2636
.
Brock
,
J.
,
Brown
,
C. C.
,
Boucher
,
J.
, &
Rippon
,
G.
(
2002
).
The temporal binding deficit hypothesis of autism
.
Development and Psychopathology
,
14
,
209
224
.
Capitão
,
L.
,
Sampaio
,
A.
,
Férnandez
,
M.
,
Sousa
,
N.
,
Pinheiro
,
A.
, &
Gonçalves
,
Ó. F.
(
2011
).
Williams syndrome hypersociability: A neuropsychological study of the amygdala and prefrontal cortex hypotheses
.
Research in Developmental Disabilities
,
32
,
1169
1179
.
Capitão
,
L.
,
Sampaio
,
A.
,
Sampaio
,
C.
,
Vasconcelos
,
C.
,
Férnandez
,
M.
,
Garayzábal
,
E.
, et al
(
2011
).
MRI amygdala volume in Williams syndrome
.
Research in Developmental Disabilities
,
32
,
2767
2772
.
Castelhano
,
J.
,
Duarte
,
I. C.
,
Wibral
,
M.
,
Rodriguez
,
E.
, &
Castelo-Branco
,
M.
(
2014
).
The dual facet of gamma oscillations: Separate visual and decision making circuits as revealed by simultaneous EEG/fMRI
.
Human Brain Mapping
,
35
,
5219
5235
.
Castelhano
,
J.
,
Rebola
,
J.
,
Leitão
,
B.
,
Rodriguez
,
E.
, &
Castelo-Branco
,
M.
(
2013
).
To perceive or not perceive: The role of gamma-band activity in signaling object percepts
.
PLoS One
,
8
,
e66363
.
Castelo-Branco
,
M.
,
Mendes
,
M.
,
Sebastião
,
A. R.
,
Reis
,
A.
,
Soares
,
M.
,
Saraiva
,
J.
, et al
(
2007
).
Visual phenotype in Williams-Beuren syndrome challenges magnocellular theories explaining human neurodevelopmental visual cortical disorders
.
Journal of Clinical Investigation
,
117
,
3720
3729
.
Chen
,
C.-M. A.
,
Stanford
,
A. D.
,
Mao
,
X.
,
Abi-Dargham
,
A.
,
Shungu
,
D. C.
,
Lisanby
,
S. H.
, et al
(
2014
).
GABA level, gamma oscillation, and working memory performance in schizophrenia
.
Neuroimage: Clinical
,
4
,
531
539
.
Coghlan
,
S.
,
Horder
,
J.
,
Inkster
,
B.
,
Mendez
,
M. A.
,
Murphy
,
D. G.
, &
Nutt
,
D. J.
(
2012
).
GABA system dysfunction in autism and related disorders: From synapse to symptoms
.
Neuroscience and Biobehavioral Reviews
,
36
,
2044
2055
.
Crone
,
N.
,
Sinai
,
A.
, &
Korzeniewska
,
A.
(
2006
).
High-frequency gamma oscillations and human brain mapping with electrocorticography
.
Progress in Brain Research
,
159
,
275
295
.
Deruelle
,
C.
,
Mancini
,
J.
,
Livet
,
M. O.
,
Cassé-Perrot
,
C.
, &
de Schonen
,
S.
(
1999
).
Configural and local processing of faces in children with Williams syndrome
.
Brain Cognition
,
41
,
276
298
.
Dinstein
,
I.
,
Pierce
,
K.
,
Eyler
,
L.
,
Solso
,
S.
,
Malach
,
R.
,
Behrmann
,
M.
, et al
(
2011
).
Disrupted neural synchronization in toddlers with autism
.
Neuron
,
70
,
1218
1225
.
Dodd
,
H. F.
,
Porter
,
M. A.
,
Peters
,
G. L.
, &
Rapee
,
R. M.
(
2010
).
Social approach in pre-school children with Williams syndrome: The role of the face
.
Journal of Intellectual Disability Research
,
54
,
194
203
.
Engel
,
A. K.
,
Fries
,
P.
, &
Singer
,
W.
(
2001
).
Dynamic predictions: Oscillations and synchrony in top–down processing
.
Nature Reviews Neuroscience
,
2
,
704
716
.
Graewe
,
B.
,
De Weerd
,
P.
,
Farivar
,
R.
, &
Castelo-Branco
,
M.
(
2012
).
Stimulus dependency of object-evoked responses in human visual cortex: An inverse problem for category specificity
.
PloS One
,
7
,
e30727
.
Gray
,
C.
(
1999
).
The temporal correlation hypothesis of visual feature integration: Still alive and well
.
Neuron
,
24
,
31
47
.
Grice
,
S. J.
,
Spratling
,
M. W.
,
Karmiloff-Smith
,
A.
,
Halit
,
H.
,
Csibra
,
G.
,
de Haan
,
M.
, et al
(
2001
).
Disordered visual processing and oscillatory brain activity in autism and Williams syndrome
.
NeuroReport
,
12
,
2697
2700
.
Groppe
,
D. M.
,
Bickel
,
S.
,
Keller
,
C. J.
,
Jain
,
S. K.
,
Hwang
,
S. T.
,
Harden
,
C.
, et al
(
2013
).
Dominant frequencies of resting human brain activity as measured by the electrocorticogram
.
Neuroimage
,
79
,
223
233
.
Gruber
,
T.
,
Martinovic
,
J.
, &
Muller
,
M.
(
2008
).
It's all in your eyes? Induced gamma band responses in the human EEG. Comment on Yuval-Greenberg, 2008
.
Neuron Online Comments
,
301
302
.
Gruber
,
T.
,
Müller
,
M. M.
,
Keil
,
A.
, &
Elbert
,
T.
(
1999
).
Selective visual-spatial attention alters induced gamma band responses in the human EEG
.
Clinical Neurophysiology
,
110
,
2074
2085
.
Grützner
,
C.
,
Uhlhaas
,
P. J.
,
Genc
,
E.
,
Kohler
,
A.
,
Singer
,
W.
, &
Wibral
,
M.
(
2010
).
Neuroelectromagnetic correlates of perceptual closure processes
.
Journal of Neuroscience
,
30
,
8342
8352
.
Güntekin
,
B.
, &
Başar
,
E.
(
2014
).
A review of brain oscillations in perception of faces and emotional pictures
.
Neuropsychologia
,
58
,
33
51
.
Herrmann
,
C. S.
, &
Demiralp
,
T.
(
2005
).
Human EEG gamma oscillations in neuropsychiatric disorders
.
Clinical Neurophysiology
,
116
,
2719
2733
.
Herrmann
,
C. S.
, &
Kaiser
,
J.
(
2011
).
EEG γ-band responses reflect human behavior: An overview
.
International Journal of Psychophysiology
,
79
,
1
2
.
Hipp
,
J. F.
,
Engel
,
A. K.
, &
Siegel
,
M.
(
2011
).
Oscillatory synchronization in large-scale cortical networks predicts perception
.
Neuron
,
69
,
387
396
.
Jawaid
,
A.
,
Riby
,
D. M.
,
Egridere
,
S.
,
Schmolck
,
H.
,
Kass
,
J. S.
, &
Schulz
,
P. E.
(
2010
).
Approachability in Williams syndrome
.
Neuropsychologia
,
48
,
1521
1523
.
Kaiser
,
J.
, &
Lutzenberger
,
W.
(
2003
).
Induced gamma-band activity and human brain function
.
Neuroscientist
,
9
,
475
484
.
Karmiloff-Smith
,
A.
,
Thomas
,
M.
,
Annaz
,
D.
,
Humphreys
,
K.
,
Ewing
,
S.
,
Brace
,
N.
, et al
(
2004
).
Exploring the Williams syndrome face-processing debate: The importance of building developmental trajectories
.
Journal of Child Psychology and Psychiatry
,
45
,
1258
1274
.
Keil
,
A.
,
Müller
,
M. M.
,
Ray
,
W. J.
,
Gruber
,
T.
, &
Elbert
,
T.
(
1999
).
Human gamma band activity and perception of a gestalt
.
The Journal of Neuroscience
,
19
,
7152
7161
.
Keren
,
A. S.
,
Yuval-Greenberg
,
S.
, &
Deouell
,
L. Y.
(
2010
).
Saccadic spike potentials in gamma-band EEG: Characterization, detection and suppression
.
Neuroimage
,
49
,
2248
2263
.
Klein-Tasman
,
B. P.
,
Phillips
,
K. D.
,
Lord
,
C.
,
Mervis
,
C. B.
, &
Gallo
,
F. J.
(
2009
).
Overlap with the autism spectrum in young children with Williams syndrome
.
Journal of Developmental and Behavioral Pediatrics
,
30
,
289
299
.
Korenberg
,
J. R.
,
Chen
,
X. N.
,
Hirota
,
H.
,
Lai
,
Z.
,
Bellugi
,
U.
,
Burian
,
D.
, et al
(
2000
).
VI. Genome structure and cognitive map of Williams syndrome
.
Journal of Cognitive Neuroscience
,
12(Suppl. 1)
,
89
107
.
Lachaux
,
J. P.
,
Rodriguez
,
E.
,
Martinerie
,
J.
, &
Varela
,
F. J.
(
1999
).
Measuring phase synchrony in brain signals
.
Human Brain Mapping
,
8
,
194
208
.
Martinovic
,
J.
, &
Busch
,
N. A.
(
2011
).
High frequency oscillations as a correlate of visual perception
.
International Journal of Psychophysiology
,
79
,
32
38
.
Martinovic
,
J.
,
Gruber
,
T.
,
Hantsch
,
A.
, &
Müller
,
M. M.
(
2008
).
Induced gamma-band activity is related to the time point of object identification
.
Brain Research
,
1198
,
93
106
.
Mendes
,
M.
,
Silva
,
F.
,
Simões
,
L.
,
Jorge
,
M.
,
Saraiva
,
J.
, &
Castelo-Branco
,
M.
(
2005
).
Visual magnocellular and structure from motion perceptual deficits in a neurodevelopmental model of dorsal stream function
.
Brain Research: Cognitive Brain Research
,
25
,
788
798
.
Meyer-Lindenberg
,
A.
,
Kohn
,
P.
,
Mervis
,
C. B.
,
Kippenhan
,
J. S.
,
Olsen
,
R. K.
,
Morris
,
C. A.
, et al
(
2004
).
Neural basis of genetically determined visuospatial construction deficit in Williams syndrome
.
Neuron
,
43
,
623
631
.
Mishkin
,
M.
,
Ungerleider
,
L.
, &
Macko
,
K.
(
1983
).
Object vision and spatial vision: Two cortical pathways
.
Trends in Neurosciences
,
6
,
414
417
.
Müller
,
M. M.
,
Junghöfer
,
M.
,
Elbert
,
T.
, &
Rochstroh
,
B.
(
1997
).
Visually induced gamma-band responses to coherent and incoherent motion: A replication study
.
NeuroReport
,
8
,
2575
2579
.
Muthukumaraswamy
,
S. D.
,
Edden
,
R. A. E.
,
Jones
,
D. K.
,
Swettenham
,
J. B.
, &
Singh
,
K. D.
(
2009
).
Resting GABA concentration predicts peak gamma frequency and fMRI amplitude in response to visual stimulation in humans
.
Proceedings of the National Academy of Sciences, U.S.A.
,
106
,
8356
8361
.
Nolte
,
G.
,
Bai
,
O.
,
Wheaton
,
L.
,
Mari
,
Z.
,
Vorbach
,
S.
, &
Hallett
,
M.
(
2004
).
Identifying true brain interaction from EEG data using the imaginary part of coherency
.
Clinical Neurophysiology
,
115
,
2292
2307
.
Palva
,
J. M.
,
Palva
,
S.
, &
Kaila
,
K.
(
2005
).
Phase synchrony among neuronal oscillations in the human cortex
.
Journal of Neuroscience
,
25
,
3962
3972
.
Palva
,
S.
, &
Palva
,
J. M.
(
2012
).
Discovering oscillatory interaction networks with M/EEG: Challenges and breakthroughs
.
Trends in Cognitive Sciences
,
16
,
219
230
.
Pockett
,
S.
,
Bold
,
G. E. J.
, &
Freeman
,
W. J.
(
2009
).
EEG synchrony during a perceptual-cognitive task: Widespread phase synchrony at all frequencies
.
Clinical Neurophysiology
,
120
,
695
708
.
Rebola
,
J.
,
Castelhano
,
J.
,
Ferreira
,
C.
, &
Castelo-Branco
,
M.
(
2012
).
Functional parcellation of the operculo-insular cortex in perceptual decision making: An fMRI study
.
Neuropsychologia
,
50
,
3693
3701
.
Reichardt
,
W.
(
1987
).
Evaluation of optical motion information by movement detectors
.
Journal of Comparative Physiology A: Sensory, Neural, and Behavioral Physiology
,
161
,
533
547
.
Riby
,
D. M.
,
Jones
,
N.
,
Brown
,
P. H.
,
Robinson
,
L. J.
,
Langton
,
S. R. H.
,
Bruce
,
V.
, et al
(
2011
).
Attention to faces in Williams syndrome
.
Journal of Autism and Developmental Disorders
,
41
,
1228
1239
.
Rodriguez
,
E.
,
George
,
N.
,
Lachaux
,
J. P.
,
Martinerie
,
J.
,
Renault
,
B.
, &
Varela
,
F. J.
(
1999
).
Perception's shadow: Long-distance synchronization of human brain activity
.
Nature
,
397
,
430
433
.
Sedley
,
W.
, &
Cunningham
,
M. O.
(
2013
).
Do cortical gamma oscillations promote or suppress perception? An under-asked question with an over-assumed answer
.
Frontiers in Human Neuroscience
,
7
,
595
.
Singer
,
W.
(
2009
).
Distributed processing and temporal codes in neuronal networks
.
Cognitive Neurodynamics
,
3
,
189
196
.
Spencer
,
K. M.
,
Nestor
,
P. G.
,
Niznikiewicz
,
M. A.
,
Salisbury
,
D. F.
,
Shenton
,
M. E.
, &
McCarley
,
R. W.
(
2003
).
Abnormal neural synchrony in schizophrenia
.
Journal of Neuroscience
,
23
,
7407
7411
.
Strømme
,
P.
,
Bjørnstad
,
P. G.
, &
Ramstad
,
K.
(
2002
).
Prevalence estimation of Williams syndrome
.
Journal of Child Neurology
,
17
,
269
271
.
Sun
,
L.
,
Grützner
,
C.
,
Bölte
,
S.
,
Wibral
,
M.
,
Tozman
,
T.
,
Schlitt
,
S.
, et al
(
2012
).
Impaired gamma-band activity during perceptual organization in adults with autism spectrum disorders: Evidence for dysfunctional network activity in frontal-posterior cortices
.
Journal of Neuroscience
,
32
,
9563
9573
.
Tallon-Baudry
,
C.
(
2003
).
Oscillatory synchrony and human visual cognition
.
Journal of Physiology
,
97
,
355
363
.
Tsuchiya
,
N.
,
Kawasaki
,
H.
,
Oya
,
H.
,
Howard
,
M. A.
,
Adolphs
,
R.
, &
Howard
,
M. A.
, III
(
2008
).
Decoding face information in time, frequency and space from direct intracranial recordings of the human brain
.
PloS One
,
3
,
e3892
.
Uhlhaas
,
P. J.
,
Linden
,
D. E. J.
,
Singer
,
W.
,
Haenschel
,
C.
,
Lindner
,
M.
,
Maurer
,
K.
, et al
(
2006
).
Dysfunctional long-range coordination of neural activity during Gestalt perception in schizophrenia
.
Journal of Neuroscience
,
26
,
8168
8175
.
Uhlhaas
,
P. J.
, &
Singer
,
W.
(
2010
).
Abnormal neural oscillations and synchrony in schizophrenia
.
Nature Reviews Neuroscience
,
11
,
100
113
.
Uhlhaas
,
P. J. J.
, &
Singer
,
W.
(
2012
).
Neuronal dynamics and neuropsychiatric disorders: Toward a translational paradigm for dysfunctional large-scale networks
.
Neuron
,
75
,
963
980
.
Ungerleider
,
L. G.
, &
Haxby
,
J. V.
(
1994
).
“What” and “where” in the human brain
.
Current Opinion in Neurobiology
,
4
,
157
165
.
van der Geest
,
J. N.
,
Lagers-van Haselen
,
G. C.
,
van Hagen
,
J. M.
,
Brenner
,
E.
,
Govaerts
,
L. C. P.
,
de Coo
,
I. F. M.
, et al
(
2005
).
Visual depth processing in Williams–Beuren syndrome
.
Experimental Brain Research
,
166
,
200
209
.
van der Geest
,
J. N.
,
Lagers-van Haselen
,
G. C.
,
van Hagen
,
J. M.
,
Govaerts
,
L. C. P.
,
de Coo
,
I. F. M.
,
de Zeeuw
,
C. I.
, et al
(
2004
).
Saccade dysmetria in Williams–Beuren syndrome
.
Neuropsychologia
,
42
,
569
576
.
Varela
,
F.
,
Lachaux
,
J.
,
Rodriguez
,
E.
, &
Martinerie
,
J.
(
2001
).
The brainweb: Phase large-scale integration
.
Nature Reviews Neuroscience
,
2
,
229
239
.
Violante
,
I. R.
,
Ribeiro
,
M. J.
,
Edden
,
R. A. E.
,
Guimarães
,
P.
,
Bernardino
,
I.
,
Rebola
,
J.
, et al
(
2013
).
GABA deficit in the visual cortex of patients with neurofibromatosis type 1: Genotype-phenotype correlations and functional impact
.
Brain
,
136
,
918
925
.