Cognitive chimera states in human brain networks

Introduction

• Brain activity exhibits transient partial synchrony across subsets of regions, while the principles governing its timing and specificity remain poorly understood.The passage connects partial synchrony in cognitive subnetworks with integrating sensory input and stored knowledge while segregating task-irrelevant information.Source: Introduction · p. 3
• A chimera state is a dynamical pattern in which coherent and incoherent oscillator populations coexist.The framework was later demonstrated in opto-electronic and chemical oscillators.Source: Introduction · p. 3
• Prior large-scale chimera studies used animal brain networks whose connectivities were modified for simplicity, leaving a gap in applications to cognitive brain activity.The authors identify the computational complexity of whole-brain modeling as part of this gap.Source: Introduction · p. 3
• The framework combines in silico experiments with predefined cognitive systems and personalized brain network models to study synchronization after regional stimulation.The analysis asks whether patterns depend on the stimulated region and whether participant structural variability reduces their consistency.Source: Introduction · pp. 3–4
• All nine cognitive systems produced chimera states, amid diverse synchronization patterns and region-specific and subject-specific structural variability.The authors suggest this reflects partial synchrony’s foundational role in large-scale brain function.Source: Introduction · p. 4

Classification of brain states using cognitive systems.

• Cognitive-system classifications separate into coherent, chimera, and metastable states and agree with traditional synchronization measures across tested thresholds.Global synchrony decreases from coherent to chimera to metastable states; global synchrony correlates positively with stimulated-region degree (r = 0.81), while chimera-index correlates negatively (r = −0.57).Source: Classification of brain states using… · p. 6

Variable brain states emerge from stimulation of different brain regions.

• Coherent states tend to originate from central brain regions, metastable states from peripheral regions, and chimera states from regions between them.Chimera-producing regions are relatively distributed throughout the brain.Source: Variable brain states emerge from… · p. 6, Figure 3 · p. 8
• Coherent states are associated mainly with medial default mode and subcortical stimulation, while metastable states are associated with auditory, cingulo-opercular, frontoparietal, and ventral temporal association systems.Chimera states emerge across all nine systems.Source: Variable brain states emerge from… · pp. 6, 10
• All nine cognitive systems give rise to chimera states, which occur more often than coherent or metastable states after regional stimulation.Chimera states include diverse patterns of coexisting synchrony and desynchrony.Source: Variable brain states emerge from… · p. 10

Structural variability influences observed dynamic states.

• Metastability dominates cingulo-opercular (45%), frontoparietal (50%), and auditory (50%) stimulation, though these systems also produce coherent or nearly coherent patterns.The cingulo-opercular and frontoparietal systems show similar synchronization patterns.Source: Structural variability influences… · p. 10
• Ventral temporal association stimulation produces metastability 16% of the time and 10 prevalent patterns, matching the motor and somatosensory system’s pattern diversity.Both systems also produce multiple chimera patterns.Source: Structural variability influences… · p. 10
• Attention (36%), default mode (33%), and subcortical (43%) stimulation most often produce coherent states.These systems show less dynamical diversity, with stimulation largely producing high-synchrony chimera states as well.Source: Structural variability influences… · p. 12

Dissociation of subject-specific and region-specific variability

• Tenacity analysis identifies four cognitive-system groups by subject and region variability, separating individually variable from stable systems and cognitively diverse, flexible, and consistent systems.Subject variability is highest in frontoparietal, cingulo-opercular, ventral temporal association, and motor and somatosensory systems; regionally, auditory is diverse, while attention, subcortical, and visual are consistent.Source: Dissociation of subject-specific and… · pp. 12, 14

Discussion

• Chimera states were the most pervasive outcome of regional stimulation, alongside coherent and metastable states, with patterns varying by region and participant.The authors relate this diversity to region-specific and subject-specific structural connectivity.Source: Discussion · p. 14

Prevalence of chimera states across cognitive systems enables segregation and integration in brain dynamics

• All nine cognitive systems produced chimera states after stimulation, consistent with partial synchrony across subsets of systems.The authors connect this balance of integration and segregation with adaptive cognition.Source: Prevalence of chimera states across… · p. 15

Chimera framework reveals subject-specific and region-specific variability in brain connectivity through the analyses of emergent dynamical states

• Medial default mode and subcortical stimulation tended to produce coherent states; these systems contain regional hubs and showed high consistency across 30 individuals.The authors associate coherent states with bridging distributed regions and global communication.Source: Chimera framework reveals… · p. 15
• Metastable states were favored by four sparsely connected systems; auditory, cingulo-opercular, and frontoparietal systems were both least likely to synchronize and most likely to produce metastability.The authors suggest metastability may let cognitive-control systems operate separately, supporting parallel processing.Source: Chimera framework reveals… · pp. 15–16
• Cingulo-opercular, frontoparietal, ventral temporal association, and motor systems showed between-participant variation in synchronization patterns.Frontoparietal and cingulo-opercular systems had the strongest individual variability; ventral temporal association and motor/somatosensory systems had many prevalent patterns without one dominant pattern.Source: Chimera framework reveals… · p. 16

Methodological Considerations and Future Directions

• Because the model captures functional relationships induced through structural connections, its observed dynamics are constrained by network anatomy and may differ from activity measured across functional modalities.The authors also note that modeled metastability may reflect synchronization at a smaller population level, motivating finer spatial resolution in future models.Source: Methodological Considerations and… · p. 16

Conclusion

• The framework links large-scale chimera states to functional segregation and integration, and captures system differences that are stable across people as well as those reflecting individual training and expertise.The authors present it as a way to study how dynamical states may relate to variability in cognitive performance.Source: Conclusion · p. 16

Human diffusion spectrum imaging data acquisition and pre-processing.

• Diffusion MRI analysis used 30 individual participant scans from neurologically healthy participants recruited in Pittsburgh and Aberdeen, Maryland, whose ages ranged from 21 to 45 years at scanning (mean 31).The sample included 20 male and 10 female participants.Source: Human diffusion spectrum imaging data… · p. 17
• The DSI acquisition sampled 257 directions on a 3T scanner, and data were reconstructed with QSDR at 2 mm output resolution.The acquisition used 2.4 mm isotropic voxels.Source: Human diffusion spectrum imaging data… · p. 17

Construction of individual structural brain networks.

• Individual structural networks were estimated by repeating whole-brain tractography 100 times per participant while varying selected tracking parameters.An edge was included in an iteration's binary matrix when at least 5% of streamlines connected a region pair; edge frequency across iterations estimated connection probability.Source: Construction of individual structural… · p. 17
• Connectivity used cortical Desikan-Killiany and subcortical Harvard-Oxford regions, with connection strengths normalized by summed regional brain volumes.These weighted matrices served as individual structural brain representations.Source: Construction of individual structural… · p. 17

• Regional dynamics are represented by Wilson-Cowan oscillators tracking the active fractions of excitatory and inhibitory neurons.The model gives their temporal dynamics using activation functions for the excitatory and inhibitory states.Source: Data-driven network model of brain… · pp. 17–18
• Subject-specific coupling A_ij encodes structural connection strength, while c5 and c6 set global excitatory and inhibitory coupling and P_i(t) supplies external stimulation.The passage specifies c6 = c5/4.Source: Data-driven network model of brain… · p. 18
• Communication delays depend on interregional distance with a 10 m/s transmission velocity; simulations add noise and use second-order Runge-Kutta integration.The model specifies noise with σ = 10^-5, step size 0.1, and initial excitatory and inhibitory values of 0.1.Source: Data-driven network model of brain… · p. 18

• For each individual, c5 was set just below the transition to the excited state, then constant input P_i = 1.15 stimulated one region.The resulting dynamics spread through the network connectivity of the stimulated node.Source: Targeted stimulation · p. 18

Cognitive systems.

• Brain regions were assigned to nine cognitive systems, including attention, frontoparietal, motor and somatosensory, and visual systems.The assignment is similar to Muldoon et al. [36].Source: Cognitive systems · p. 18

Calculation of synchronization within cognitively-informed framework.

• The standard order parameter estimates synchronization after targeted regional stimulation from oscillator phases, averaged over 1 s of simulated activity.The network contains 76 oscillators.Source: Calculation of synchronization within… · pp. 18–19
• For each pair of cognitive systems, an order parameter over their combined oscillators yields synchronization matrices used to identify the dynamical cognitive state after stimulation.The oscillator counts for the two systems determine the size of their combined population.Source: Calculation of synchronization within… · p. 19

• The chimera-index measures normalized average variation in the order parameter within cognitive systems over time.Source: Chimera-index · p. 19
• Its normalization factor CMax = 5/36 represents the maximum order-parameter variability for an ideal state with half the network synchronized and half desynchronized.Source: Chimera-index · p. 19
• The instantaneous system-level quantity averages synchronization across all cognitive systems at a given time.Source: Chimera-index · p. 19

• Two systems are classified as synchronized when their pairwise order parameter reaches the threshold, set to 0.8 for the main-text results.Source: Robust detection of the patterns of… · p. 19
• Community detection groups synchronized systems together, while coherent states form one community, metastable states form nine, and chimera states show varied system distributions.Source: Robust detection of the patterns of… · p. 19
• The method uses generalized Louvain modularity maximization, varying resolution from 0.8 to 0.95 and running consensus to determine community structure.Source: Robust detection of the patterns of… · p. 20

• Each pattern records whether a cognitive system belongs to the synchronized population or remains desynchronized.Source: Pattern tenacity · p. 20
• Tenacity measures whether a system has the same synchrony or desynchrony state in pairs of patterns.Source: Pattern tenacity · p. 20
• Subject tenacity compares patterns across 30 subjects for a given brain region, whereas region tenacity compares system nodes within one subject.The number of nodes, and thus patterns, varies between cognitive systems for region tenacity.Source: Pattern tenacity · p. 20

Clustering of cognitive systems using tenacity features.

• The k-means clustering sought high within-cluster similarity and between-cluster dissimilarity, assessed with silhouette analysis.Source: Clustering of cognitive systems using… · p. 20
• The analysis tested cluster counts of k = 3, 4, 5, and 6.Source: Clustering of cognitive systems using… · p. 20
• The authors identified k = 4 as the optimized clustering and show its silhouette plot in Fig. S7.Source: Clustering of cognitive systems using… · p. 20

Rendering of brain images.

• BrainNet Viewer was used for spatial mapping onto brain images.Source: Rendering of brain images · p. 20