Volume 4 Supplement 1
Transient process of cortical activity during Necker cube perception: from local clusters to global synchrony
© Shimaoka et al; licensee BioMed Central Ltd. 2010
Published: 3 June 2010
It has been discussed that neural phase-synchrony across distant cortical areas (or global phase-synchrony) was correlated with various aspects of consciousness. The generating process of the synchrony, however, remains largely unknown. As a first step, we investigate transient process of global phase-synchrony, focusing on phase-synchronized clusters. We hypothesize that the phase-synchronized clusters are dynamically organized before global synchrony and clustering patterns depend on perceptual conditions.
In an EEG study, Kitajo reported that phase-synchrony across distant cortical areas was selectively enhanced by top-down attention around 4 Hz in Necker cube perception. Here, we further analyzed the phase-synchronized clusters using hierarchical clustering which sequentially binds up the nearest electrodes based on similarity of phase locking between the cortical signals. First, we classified dominant components of the phase-synchronized clusters over time. We then investigated how the phase-synchronized clusters change with time, focusing on their size and spatial structure.
Phase-locked clusters organized a stable spatial pattern common to the perceptual conditions. In addition, the phase-locked clusters were modulated transiently depending on the perceptual conditions and the time from the perceptual switch. When top-down attention succeeded in switching perception as subjects intended, independent clusters at frontal and occipital areas grew to connect with each other around the time of the perceptual switch. However, the clusters in the occipital and left parietal areas remained divided when top-down attention failed in switching perception. When no primary biases exist, the cluster in the occipital area grew to its maximum at the time of the perceptual switch within the occipital area.
Our study confirmed the existence of stable phase-synchronized clusters. Furthermore, these clusters were transiently connected with each other. The connecting pattern depended on subjects’ internal states. These results suggest that subjects’ attentional states are associated with distinct spatio-temporal patterns of the phase-locked clusters.
It has been discussed that phase-synchrony across distant cortical areas (or global phase-synchrony) was correlated with various aspects of consciousness [1–3]. However, it remains largely unknown what function such synchronous activities play in conscious processing. Recently, some researchers postulate that synchronous activity is not sufficient for consciousness, but assists enhancing a nascent assembly of neurons in its competition with other assemblies [1, 4–6]. To clarify the role of transient global-phase-synchrony, it is crucial to reveal not only synchronous phenomena, but also the dynamical mechanism underlying the phenomena.
As a first step in this clarification, we focus on characterizing the transient process of formation of the global phase-synchrony. One possibility to form a global synchrony is growth of one small “core”. Another possibility could be the merging of coexisting clusters. These processes may depend on perceptual states such as consciousness and attention that accompany the global synchrony. As an example to investigate the transient process towards the global synchrony, we choose to analyze the global phase-synchrony around 4 Hz in Necker cube perception. In this system, phase-synchrony of electroencephalography (EEG) signals across distant cortical areas was selectively enhanced by top-down attention .
In this report, we analyze the spatio-temporal dynamics of phase-synchronized (or phase-locked) clusters using a hierarchical clustering algorithm. Hierarchical clustering binds cortical areas by linking pairs of cortical areas that are close in phase. With this algorithm, we first extract the dominant component of the clusters over time. We subsequently investigate temporal modulation of the clusters, focusing on their size and spatial structure. We thereby characterize the transient process of the attention-enhanced global phase-synchrony.
Sixteen healthy human subjects with normal or corrected-to-normal vision (mean age: 24.7 years; SD: 4.7 years) participated in the study. The study was approved by the RIKEN research ethics committee, and the experiment was undertaken with written informed consent.
Passive view condition (Neutral condition), in which they looked at the cube passively.
Percept A biasing condition (Top down biasing condition 1), in which they tried to maintain percept A as long as possible.
Percept B biasing condition (Top down biasing condition 2), in which they tried to maintain percept B as long as possible.
Self-paced key pressing condition (Control condition), in which the cube was not presented and subjects pressed the keys at their own paces.
In addition to the above conditions, we set the base line condition (referred to as the Resting state condition) in which subjects looked at the centre cross on the blank screen without pressing keys.
In this paper, we will show the results for the switches from Percept B to A. These perceptual switches are classified into the following 3 cases. In the “desired” switch, they try to perceive A, and their percept switches to A. In the “undesired” switch, their percept switches to A, even though they try to continue perceiving B. In the neutral switch, they have no primary biases before the switch.
Effects of top-down attention assessed with behavioural data
Behavioural effects of top-down attention were assessed with selective modulation of dominance durations. In condition (2), the dominance duration of percept A (desired percept) was significantly longer (Mann-Whitney U test, p < 5.00×10-199) than that of percept B (undesired percept). Likewise in condition (3), the dominance duration of percept B (desired perception) was significantly longer (Mann-Whitney U test, p < 5.00×10-125) than that of percept A (undesired perception). These results were consistent with a previous study . We thereby confirmed that the subjects actually viewed the cube with top-down attention in condition (2) and (3).
Cortical activity was recorded using a 62 ch EEG (Figure 1(b)). Electrooculogram (EOG) was simultaneously recorded at the position 1 cm from the outer canthi of both eyes and above and below the left eye. EEG and EOG were amplified with a gain of 500 (Neuroscan, EL Paso, TX), band-pass filtered between 0.1 and 100 Hz, and digitalized at 500 Hz. Scalp voltages were referenced to a linked earlobe reference. Electrode impedances were kept 5 kΩ. One EEG epoch was defined as a period between 1.5 s before and 1.5 s after a key press that reported a perceptual switch. Epochs with artefacts caused by blinks or eye movements or amplifier saturation were detected using an amplitude criterion (±100 μV) and excluded from further analysis. EEG epochs were extracted in which perception of the rivalling views persisted for 1 s or longer both before and after the perceptual switching. A total of 398 epochs for each perceptual condition were obtained from 16 subjects. To reduce the effect of volume conduction, EEG scalp current density  was calculated from the raw signals. For a visualization of topographic maps, EEGLAB Matlab toolbox was used .
Detection of phase-locking
where N (= 398 for each condition) is the number of total epochs and ϕ(t, n) represents the instantaneous phase at time t in the n-th epoch. PLV ranges from 0 (random phase difference or no phase locking) to 1 (constant phase difference or perfect phase-locking).
Relative change of phase-locking
Since PLVz measures the difference from baseline PLV, it is less sensitive to synchrony arising via volume conduction from sources.
Classification of phase-locked clusters
The stable component of phase-locked clusters throughout the test time
To evaluate the significance of phase-locking as a whole pattern of the cortical network, we calculated the Euclid distance between the surrogate matrix and the corresponding original PLV matrix. The resulting distances were: 0.1039±0.7858×10-3 (desired), 0.1010±0.8643×10-3 (undesired), 0.1027±0.7363×10-3 (neutral), 0.0995±0.7852×10-3 (self-paced key press), 0.0936±0.7561×10-3 (rest) (mean ± 3 SD), all of which were clearly larger than 0. These results indicate that the intrinsic phase-locking pattern is significant compared to the background fluctuation.
In Figure 3 (right), we showed the topographical maps of the phase-locked clusters obtained with the clustering threshold of 0.48. Note that different colours in these figures indicate different clusters but the identity of the colours does not have any meaning. A common feature across the conditions is that the clusters are divided into occipital, parietal and frontal areas. A similar clustering pattern was observed for the range of the threshold from 0.44 to 0.52 (Figure 3(a), right). This indicates that the clustering pattern is not sensitive to the clustering threshold. We also confirmed that the clustering pattern was stable over time. These observations suggest that this pattern is stable across time and conditions. The clustering pattern is comparable to structural connectivity obtained from the diffusion spectrum imaging (Figure 6a in ). Although the clustering algorithm adopted in this study is not exactly the same as ours, some features are common to the modules of structural connectivity and the clusters of phase-locking (see Discussion for details).
Transient change of phase-locked cluster
In the previous section, we observed that phase-locked clusters have a common stable structure across the conditions. In this section, we investigate the transient change of phase-locking from the stable structure. For this purpose, we normalize PLV relative to the time-averaged PLV in the resting state condition (called PLVz. See Methods for details.) By applying the hierarchical clustering to the PLVz, we can extract the transient component of the phase-locked activity that runs on the condition-independent clusters observed in the previous section.
To check whether these changes in cluster size were due to the evoked activity by the key press, we compared the cluster size between a perceptual condition (desired, undesired, neutral) and the self-paced key press condition by applying a permutation test to the time series data of maximal cluster size. We first set time windows of 100 ms and did a random partition to the combined data set of the two conditions within the window (we confirmed that the result of the test was not sensitive to the window size). On this random partition, the average of the maximum cluster size was calculated as a test static. We repeated these procedures for 50000 times and obtained a histogram of the test static. From the histogram and the test static that was actually observed, a p-value was obtained. The total significance level (p < 0.05) was controlled by the number of time windows (i.e. 15 windows). The significant periods (p < 0.05 in total, p < 0.003 for each time window) are displayed in Figure 5 as horizontal lines. Within these periods, it is considered that the changes in the maximum cluster size were not due to the key press, but to perceptual changes. We alternatively performed the permutation test between a perceptual condition and the self-paced key press condition, confirming that the null hypothesis was rejected for even wider periods.
Integrating the temporal and spatial characteristics described above leads to a dynamical nature toward global synchrony, especially in the desired switch. Small clusters that coexist at the beginning gradually merge one after another to organize a large cluster around the key press time. It can be postulated that there would be several possible ways to achieve global phase synchrony. One possibility is growth of only one small core cluster. Another possibility is merging of coexisting clusters. The dynamics observed in the desired switch are more consistent with the latter view than the former.
Previous studies have suggested that frontal-occipital synchrony network is enhanced when top-down selective attention is involved in perceptual switch. Our result on transient modulation of phase-locked clusters is consistent with this notion. In addition, our result suggests that processes towards synchrony are also different depending on desired, undesired and neutral switch. In the top-down attentional conditions (desired and undesired switch), phase-synchrony in occipital and frontal areas was independently enhanced before the perceptual switch. In contrast, in the neutral switch, only occipital areas were enhanced. Subsequently, occipital and frontal areas were transiently connected more strongly when top-down attention succeeded in switching perception (desired switch) than when top-down attention did not succeed (undesired switch).
The relationship between functional network and anatomical structure has been investigated in detail [14–17] while transient changes of this relationship have not been yet clarified. However, brain functions can be adaptive faster than anatomical structures so that the non-stationarity or metastability of functional networks should be critical in the processing of perception, as Bullmore and Sporns pointed out in . Our results suggest that stable functional clusters (Figure 3, right) are transiently connected more strongly for the desired switch than for the other conditions. More theoretical and modelling studies are needed to reveal the dynamical mechanism of the transient change of phase-locking, especially the dynamical clustering.
The dynamical clustering observed in this report may be related to functional clustering observed in globally coupled nonlinear systems [19–26]. In these systems, when multiple identical elements with nonlinear dynamics interact through a mean field or through a random network, the elements differentiate into some clusters. In each cluster, elements oscillate synchronously although elements in different clusters oscillate with different amplitude, frequency, or phase. The idea of this clustering has been extended to cell differentiation [27, 28] and pattern dynamics in gas-discharge systems  as well as in experiments of electro-chemical oscillations [30–32]. One feature that may be useful in the context of neuroscience is its flexibility in making clusters.
In investigating the stable component of the phase-locked clusters around the time of perceptual switch, we found that the clusters were divided into occipital, parietal, and frontal areas (Figure 3, right). This clustering pattern is comparable to the modules of structural connectivity reported in . In this study, structural connectivity was estimated with the use of the diffusion spectrum imaging in which nerve fiber architecture is mapped from the 3D spectra of tissue water diffusion with magnetic resonance. From the resulting cortical connection matrix, structural modules were identified employing spectral graph partitioning . Hagmann et al. reported that optimal modularity was achieved with six clusters. In particular, four contralaterally matched modules were localized to the frontal and temporo-parietal areas of a single hemisphere. The two remaining modules comprised regions of the bilateral medial cortex, one centred on the posterior cingulate cortex and another centred on the precuneus and pericalcarine cortex. We also applied the same clustering algorithm  to our data and confirmed that the resulting phase-locked clusters were closer to the structural connectivity reported in  than the cluster obtained from the hierarchical clustering (Figure 3, right). Nevertheless, these clusters had similar structure to the structural connectivity in that the clusters were divided into occipital, parietal and frontal areas. This result is thereby consistent with the previous knowledge obtained from modelling [34, 35] and physiological studies [13–16], which reported that functional clusters coincide with the topological community structures. Thus it is natural to expect that via cortico-cortical synaptic connections , phase-locking around 4Hz is mediated.
Other than the hierarchical clustering that we used, there have been some clustering algorithms, for example, based on network modularity [14, 33] and based on an information-theoretic perspective . In this report, however, we chose a rather simple algorithm, hierarchical clustering, because we were interested in the degree of phase-locking rather than network structure. In addition, our method is suitable for observing clusters that have the same degree of phase-locking over time. In contrast, the criteria for clustering are not fixed in the other methods.
In this report, we characterized the transient process towards attention-enhanced global synchrony as dynamical clustering. It would be interesting to investigate whether the dynamical clustering is observed in the other global synchrony phenomena that are selectively enhanced by perceptual states such as consciousness. We hope our findings will help to reveal the unique process of global synchrony and unveil the dynamical mechanism that constitutes conscious processing.
Through hierarchical clustering, we investigated functional phase-locked clusters during Necker cube perception. The spatial structure of the clusters was stable across time and conserved across the perceptual conditions. We also observed that these stable clusters were transiently connected around the time of perceptual switch. This transient modulation was stronger in the desired switch than in the other conditions. These observations suggest the characteristic dynamics towards the attention-modulated global synchrony in which small clusters that coexisted at the beginning gradually merge one after another to organize a large cluster around the time of perceptual switch.
List of abbreviations
phase locking value
We thank J. Ito for valuable discussion and critical reading of the manuscript, and F. Kleberg for careful reading of the manuscript.
This article has been published as part of Nonlinear Biomedical Physics Volume 4 Supplement 1, 2010: Consciousness and its Measures: Joint Workshop for COST Actions Neuromath and Consciousness. The full contents of the supplement are available online at http://www.nonlinearbiomedphys.com/supplements/4/S1.
- Engel A, Singer W: Temporal binding and the neural correlates of sensory awareness. Trends Cogn Sci. 2001, 5: 16-25. 10.1016/S1364-6613(00)01568-0.View ArticleGoogle Scholar
- Rodriguez E, George N, Lachaux J, Martinerie J, Renault B, Varela F: Perception's shadow: long-distance synchronization of human brain activity. Nature. 1999, 397: 430-433. 10.1038/17120.View ArticleADSGoogle Scholar
- Varela F, Lachaux J, Rodriguez E, Martinerie J: The brainweb: phase synchronization and large-scale integration. Nat Rev Neurosci. 2001, 2: 229-239. 10.1038/35067550.View ArticleGoogle Scholar
- Melloni L, Molina C, Pena M, Torres D, Singer W, Rodriguez E: Synchronization of neural activity across cortical areas correlates with conscious perception. Journal of Neuroscience. 2007, 27: 2858-2865. 10.1523/JNEUROSCI.4623-06.2007.View ArticleGoogle Scholar
- Maia T, Cleeremans A: Consciousness: converging insights from connectionist modeling and neuroscience. Trends Cogn Sci. 2005, 9: 397-404. 10.1016/j.tics.2005.06.016.View ArticleGoogle Scholar
- Crick F, Koch C: A framework for consciousness. Nat Neurosci. 2003, 6: 119-126. 10.1038/nn0203-119.View ArticleGoogle Scholar
- Kitajo K, Ozaki T, Yamaguchi Y: Large-scale neural synchrony associated with top-down selective attentional modulation of Necker cube perception. Proceedings of the 23rd Annual Meeting of the International Society for Psychophysics: 20-23 October 2007; Tokyo, Japan. Edited by: Mori S, Miyaoka T, Wong W. 2007, The International Society for Psychophysics, 345-350.Google Scholar
- Meng M, Tong F: Can attention selectively bias bistable perception? Differences between binocular rivalry and ambiguous figures. J Vis. 2004, 4: 539-551. 10.1167/4.7.2.View ArticleGoogle Scholar
- Kayser J, Tenke C: Principal components analysis of Laplacian waveforms as a generic method for identifying ERP generator patterns: I. Evaluation with auditory oddball tasks. Clin Neurophysiol. 2006, 117: 348-368. 10.1016/j.clinph.2005.08.034.View ArticleGoogle Scholar
- Delorme A, Makeig S: EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods. 2004, 134: 9-21. 10.1016/j.jneumeth.2003.10.009.View ArticleGoogle Scholar
- Lachaux J, Rodriguez E, Martinerie J, Varela F: Measuring phase synchrony in brain signals. Hum Brain Mapp. 1999, 8: 194-208. 10.1002/(SICI)1097-0193(1999)8:4<194::AID-HBM4>3.0.CO;2-C.View ArticleGoogle Scholar
- Doesburg S, Kitajo K, Ward L: Increased gamma-band synchrony precedes switching of conscious perceptual objects in binocular rivalry. Neuroreport. 2005, 16: 1139-1142. 10.1097/00001756-200508010-00001.View ArticleGoogle Scholar
- Hagmann P, Cammoun L, Gigandet X, Meuli R, Honey C, Wedeen V, Sporns O: Mapping the structural core of human cerebral cortex. PLoS Biol. 2008, 6: e159-10.1371/journal.pbio.0060159.View ArticleGoogle Scholar
- Hilgetag C, Burns G, O'Neill M, Scannell J, Young M: Anatomical connectivity defines the organization of clusters of cortical areas in the macaque monkey and the cat. Philos Trans R Soc Lond B Biol Sci. 2000, 355: 91-110. 10.1098/rstb.2000.0551.View ArticleGoogle Scholar
- Hilgetag C, Kaiser M: Clustered organization of cortical connectivity. Neuroinformatics. 2004, 2: 353-360. 10.1385/NI:2:3:353.View ArticleGoogle Scholar
- Tononi G, McIntosh A, Russell D, Edelman G: Functional clustering: identifying strongly interactive brain regions in neuroimaging data. Neuroimage. 1998, 7: 133-149. 10.1006/nimg.1997.0313.View ArticleGoogle Scholar
- Honey C, Sporns O, Cammoun L, Gigandet X, Thiran J, Meuli R, Hagmann P: Predicting human resting-state functional connectivity from structural connectivity. Proc Natl Acad Sci U S A. 2009, 106: 2035-2040. 10.1073/pnas.0811168106.View ArticleADSGoogle Scholar
- Bullmore E, Sporns O: Complex brain networks: graph theoretical analysis of structural and functional systems. Nat Rev Neurosci. 2009, 10: 186-198. 10.1038/nrn2575.View ArticleGoogle Scholar
- Kaneko K: Chaotic but regular posi-nega switch among coded attractors by cluster-size variation. Physical Review Letters. 1989, 63: 219-223. 10.1103/PhysRevLett.63.219.MathSciNetView ArticleADSGoogle Scholar
- Kaneko K: Clustering, coding, switching, hierarchical ordering, and control in a network of chaotic elements. Physica D: Nonlinear Phenomena. 1990, 41: 137-172. 10.1016/0167-2789(90)90119-A.MathSciNetView ArticleADSGoogle Scholar
- Kaneko K: Globally coupled circle maps. Physica D: Nonlinear Phenomena. 1991, 54: 5-19. 10.1016/0167-2789(91)90103-G.MathSciNetView ArticleADSGoogle Scholar
- Kaneko K: Mean field fluctuation of a network of chaotic elements: Remaining fluctuation and correlation in the large size limit. Physica D: Nonlinear Phenomena. 1992, 55: 368-384. 10.1016/0167-2789(92)90066-V.View ArticleADSGoogle Scholar
- Golomb D, Hansel D, Shraiman B, Sompolinsky H: Clustering in globally coupled phase oscillators. Physical Review A. 1992, 45: 3516-3530. 10.1103/PhysRevA.45.3516.View ArticleADSGoogle Scholar
- Okuda K: Variety and generality of clustering in globally coupled oscillators. Physica D: Nonlinear Phenomena. 1993, 63: 424-436. 10.1016/0167-2789(93)90121-G.View ArticleADSGoogle Scholar
- Zanette D, Mikhailov A: Dynamical clustering in large populations of Rossler oscillators under the action of noise. Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics. 2000, 62: R7571-7574.Google Scholar
- Zanette D, Mikhailov A: Dynamical systems with time-dependent coupling: clustering and critical behaviour. Physica D: Nonlinear Phenomena. 2004, 194: 203-218. 10.1016/j.physd.2004.03.002.View ArticleADSGoogle Scholar
- Kaneko K, Yomo T: Cell division, differentiation, and dynamic clustering. Physica D: Nonlinear Phenomena. 1994, 75: 89-102. 10.1016/0167-2789(94)90277-1.View ArticleADSGoogle Scholar
- Furusawa C, Kaneko K: Emergence of multicellular organisms with dynamic differentiation and spatial pattern. Artificial Life. 1998, 4: 79-93. 10.1162/106454698568459.View ArticleGoogle Scholar
- Sugawara T, Kaneko K: Pattern Dynamics in a Two-Dimensional Gas Discharge System. Prog Theor Phys Suppl. 2006, 161: 344-347. 10.1143/PTPS.161.344.View ArticleADSGoogle Scholar
- Wang W, Kiss I, Hudson J: Experiments on arrays of globally coupled chaotic electrochemical oscillators: Synchronization and clustering. Chaos. 2000, 10: 248-256. 10.1063/1.166470.View ArticleADSGoogle Scholar
- Kiss I, Zhai Y, Hudson J: Collective dynamics of chaotic chemical oscillators and the law of large numbers. Phys Rev Lett. 2002, 88: 238301-10.1103/PhysRevLett.88.238301.View ArticleADSGoogle Scholar
- Kiss I, Zhai Y, Hudson J: Emerging coherence in a population of chemical oscillators. Science. 2002, 296: 1676-1678. 10.1126/science.1070757.View ArticleADSGoogle Scholar
- Newman M: Modularity and community structure in networks. Proc Natl Acad Sci U S A. 2006, 103: 8577-8582. 10.1073/pnas.0601602103.View ArticleADSGoogle Scholar
- Zemanova L, Zhou C, Kurths J: Structural and functional clusters of complex brain networks. Physica D: Nonlinear Phenomena. 2006, 224: 202-212. 10.1016/j.physd.2006.09.008.View ArticleADSGoogle Scholar
- Zhou C, Zemanová L, Zamora G, Hilgetag C, Kurths J: Hierarchical organization unveiled by functional connectivity in complex brain networks. Phys Rev Lett. 2006, 97: 238103-10.1103/PhysRevLett.97.238103.View ArticleADSGoogle Scholar
- Felleman D, Van Essen D: Distributed hierarchical processing in the primate cerebral cortex. Cereb Cortex. 1991, 1: 1-47. 10.1093/cercor/1.1.1-a.View ArticleGoogle Scholar
- Slonim N, Atwal G, Tkacik G, Bialek W: Information-based clustering. Proc Natl Acad Sci U S A. 2005, 102: 18297-18302. 10.1073/pnas.0507432102.MathSciNetView ArticleADSGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.