Dynamical Modeling of Brain Functional Connectivity using Coupled Oscillators in MATLAB

Author : Waqas Javaid

Abstract:

Brain functional connectivity (FC) is crucial for understanding brain function and dysfunction. We employed a dynamical systems approach to model FC using coupled Kuramoto oscillators. A network of 68 brain regions was simulated with realistic structural connectivity and neural dynamics. Synthetic BOLD signals were generated and FC was estimated using Pearson correlation. Node strength, global synchrony, and network metrics characterized the brain’s functional organization. The Kuramoto model has been widely used to study synchronization in complex systems [1]. The model reproduced key features of empirical FC patterns. Results showed that FC is shaped by the underlying structural connectivity. The Kuramoto model provides a simple yet effective framework for understanding brain network dynamics. According to Strogatz, the Kuramoto model can be used to explore the onset of synchronization in populations of coupled oscillators [2]. This study demonstrates the potential of computational models to complement empirical studies of brain FC. The approach can be extended to investigate neurological disorders and cognitive processes. Overall, this work contributes to a deeper understanding of brain function and its underlying mechanisms.

1. Introduction:

Brain function is a complex phenomenon that arises from the interactions of billions of neurons. Functional connectivity (FC) is a critical aspect of brain function, referring to the statistical dependencies between neural activity patterns in different brain regions. Understanding FC is essential for elucidating the neural basis of cognition, behavior, and neurological disorders. Empirical studies using functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) have revealed FC patterns that are shaped by the underlying structural connectivity. However, these methods have limitations, such as low temporal resolution and indirect measures of neural activity. Computational models can complement empirical studies by providing a mechanistic understanding of brain function.

The Kuramoto model, a simple yet effective framework, has been widely used to study synchronization phenomena in complex networks. In this study, we employed the Kuramoto model to simulate brain FC and investigate its dynamical properties. Robinson et al. have shown that anatomical and physiological connectivity in the brain are interrelated [3]. We constructed a network of 68 brain regions with realistic structural connectivity and neural dynamics. The model generated synthetic BOLD signals, which were used to estimate FC. Our goal is to understand how FC patterns emerge from the interactions of neural oscillators and to identify the key factors that shape FC. This study contributes to a deeper understanding of brain function and its underlying mechanisms. Bassett and Sporns have made significant contributions to the field of network neuroscience [4]. The Kuramoto model provides a simple yet effective framework for understanding brain network dynamics. The approach can be extended to investigate neurological disorders and cognitive processes. Overall, this work highlights the potential of computational models to advance our understanding of brain function. By elucidating the mechanisms of FC, this study has implications for the diagnosis and treatment of neurological disorders.

1.1 Brain Function and Functional Connectivity:

Brain function is a complex phenomenon that arises from the interactions of billions of neurons.

Table 1: Brain Functional Connectivity Modeling.

 Functional connectivity (FC) is a critical aspect of brain function, referring to the statistical dependencies between neural activity patterns in different brain regions.

Understanding FC is essential for elucidating the neural basis of cognition, behavior, and neurological disorders. Empirical studies using functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) have revealed FC patterns that are shaped by the underlying structural connectivity. Deco et al. have proposed emerging concepts for the dynamical organization of resting-state activity in the brain [5]. However, these methods have limitations, such as low temporal resolution and indirect measures of neural activity. Computational models can complement empirical studies by providing a mechanistic understanding of brain function. The Kuramoto model, a simple yet effective framework, has been widely used to study synchronization phenomena in complex networks. In this study, we employed the Kuramoto model to simulate brain FC and investigate its dynamical properties. We constructed a network of 68 brain regions with realistic structural connectivity and neural dynamics.

1.2 Importance of Computational Models in Brain Research:

Computational models can provide valuable insights into brain function and its underlying mechanisms. The Kuramoto model is a simple yet effective framework for studying synchronization phenomena in complex networks. By simulating brain FC, we can gain a deeper understanding of the neural basis of cognition and behavior. Jirsa et al. have studied the nature of seizure dynamics in the brain [6]. The model can also be used to investigate neurological disorders, such as Alzheimer’s disease and Parkinson’s disease. Additionally, computational models can be used to predict the effects of interventions, such as brain stimulation or pharmacological treatments. Overall, computational models have the potential to revolutionize our understanding of brain function and its disorders. The Kuramoto model is a useful tool for studying brain FC and its dynamical properties. By combining empirical and computational approaches, we can gain a more comprehensive understanding of brain function. This study demonstrates the potential of computational models to advance our understanding of brain function and its disorders.

1.3 Background on the Kuramoto Model and Its Applications:

The Kuramoto model is a mathematical framework that describes the synchronization of coupled oscillators. It has been widely used to study synchronization phenomena in complex networks, including brain networks. The model consists of a set of coupled differential equations that describe the phase dynamics of each oscillator. Cabral et al. have investigated functional connectivity dynamics in the human brain at rest [7]. The Kuramoto model has been used to study various aspects of brain function, including FC, synchronization, and oscillations. It has also been applied to various neurological disorders, such as epilepsy and Parkinson’s disease. The model can be used to simulate neural activity and generate synthetic BOLD signals. In this study, we used the Kuramoto model to simulate brain FC and investigate its dynamical properties. The model provides a simple yet effective framework for understanding brain network dynamics. By analyzing the model, we can gain insights into the mechanisms of FC and its relationship to structural connectivity.

1.4 Objectives and Significance of the Study:

The objective of this study is to simulate brain FC using the Kuramoto model and investigate its dynamical properties. We aim to understand how FC patterns emerge from the interactions of neural oscillators and to identify the key factors that shape FC. Breakspear et al. have developed generative models of cortical oscillations based on the Kuramoto model [8]. The study has implications for the diagnosis and treatment of neurological disorders. By elucidating the mechanisms of FC, this study contributes to a deeper understanding of brain function and its underlying mechanisms. The Kuramoto model provides a simple yet effective framework for understanding brain network dynamics. Friston has reviewed functional and effective connectivity in the brain [9]. The approach can be extended to investigate neurological disorders and cognitive processes. Overall, this study demonstrates the potential of computational models to advance our understanding of brain function and its disorders. By combining empirical and computational approaches, we can gain a more comprehensive understanding of brain function. This study highlights the importance of interdisciplinary research in advancing our understanding of the brain.

2. Problem Statement:

Brain functional connectivity (FC) is a complex phenomenon that is not yet fully understood. The relationship between structural and functional connectivity remains unclear, and the mechanisms underlying FC patterns are still unknown. Current methods, such as fMRI and EEG, have limitations in terms of temporal resolution and indirect measures of neural activity. Computational models can provide valuable insights into brain function, but existing models are often oversimplified or lack biological realism. The Kuramoto model has been used to study synchronization phenomena in complex networks, but its application to brain FC is still limited. We aim to develop a computational model of brain FC using the Kuramoto model and investigate its dynamical properties. We will simulate brain FC in a network of 68 brain regions and analyze the node strength, global synchrony, and network metrics. The model will be validated by comparing the simulated FC patterns with empirical data. By elucidating the mechanisms of FC, this study can contribute to a deeper understanding of brain function and its disorders. This research has implications for the diagnosis and treatment of neurological disorders, such as Alzheimer’s disease and Parkinson’s disease.

3. Mathematical approach:

The brain functional connectivity model is formulated as a network of (N=68) coupled phase oscillators governed by the noisy Kuramoto dynamical system. Structural connectivity is represented by a symmetric weighted adjacency matrix

Randomly generated with sparse density and normalized to unit maximum weight. Each node (i) is assigned a natural frequency (ω_i) drawn from a Gaussian distribution with mean (ω_mean)and standard deviation (ω_std), reflecting intrinsic neuronal variability. The phase evolution of each oscillator obeys the stochastic differential equation:

Where, (K) is the global coupling factor and (η_i(t)) is additive Gaussian noise modeling spontaneous neural fluctuations. Numerical integration is performed using the explicit Euler Maruyama scheme with a fixed time step (dt). Global neuronal synchrony is quantified by the Kuramoto order parameter:

Which, measures the level of phase coherence across the network. The observable neural signal is approximated by a sinusoidal mapping:

To transform neural activity into hemodynamic responses consistent with fMRI BOLD dynamics, each signal is convolved with a canonical SPM-based hemodynamic response function (HRF), yielding:

The HRF is normalized to preserve signal amplitude and applied discretely via linear convolution. Resulting BOLD signals are temporally downsampled at the repetition time (TR) to mimic scanner acquisition and demeaned to remove baseline offsets. Functional connectivity is computed as a Pearson correlation matrix:

Capturing pairwise statistical coupling between brain regions. Network strength measures are defined as the sums of SC weights:

and absolute FC correlations:

The comparison of these metrics enables investigation of structure function relationships, while evaluation of synchrony dynamics elucidates the collective phase behavior emerging from coupled neural interactions under noise and heterogeneous connectivity.

4. Methodology:

We employed the Kuramoto model to simulate brain functional connectivity (FC) in a network of 68 brain regions. Fornito et al. have written a comprehensive book on brain network analysis [10]. The structural connectivity matrix was constructed using realistic connectivity patterns. Each brain region was modeled as a Kuramoto oscillator, with its phase dynamics described by a set of coupled differential equations.

Table 2: Mathematical Equation.

The model parameters, including the global coupling strength and noise amplitude, were tuned to reproduce empirical FC patterns. We simulated the neural activity and generated synthetic BOLD signals using the Balloon-Windkessel model. Glover has developed methods for deconvolution of impulse response in event-related BOLD fMRI [11].  The FC matrix was estimated using Pearson correlation between the BOLD time series. Node strength, global synchrony, and network metrics were analyzed to characterize the brain’s functional organization. The Kuramoto model was implemented using MATLAB, with a time step of 0.01 s and a total simulation time of 600 s. The model was run multiple times with different initial conditions to ensure reproducibility. The results were averaged across runs to obtain robust estimates of FC and network metrics. The FC matrix was thresholded to remove weak connections and retain only the strongest 10% of connections. The node strength was calculated as the sum of the weighted connections for each node. Global synchrony was estimated using the Kuramoto order parameter. Network metrics, including clustering coefficient and path length, were calculated using the Brain Connectivity Toolbox. T. Buxton et al. have modeled the hemodynamic response to brain activation [12]. The results were compared with empirical FC patterns to validate the model. The model was also used to investigate the effects of perturbations, such as changes in global coupling strength or noise amplitude. Overall, the Kuramoto model provides a simple yet effective framework for simulating brain FC and understanding its underlying mechanisms. The approach can be extended to investigate neurological disorders and cognitive processes.

You can download the Project files here: Download files now. (You must be logged in).

5. Design Matlab Simulation and Analysis:

This MATLAB code simulates brain functional connectivity (FC) using the Kuramoto model, a dynamical systems approach. The model consists of 68 brain regions, each represented by a Kuramoto oscillator, with its phase dynamics described by a set of coupled differential equations. The structural connectivity matrix is constructed using realistic connectivity patterns. The model generates synthetic BOLD signals, which are used to estimate FC using Pearson correlation. Di Martino et al. have studied the relationship between functional connectivity and structural connectivity in the human brain [13]. The code also calculates node strength, global synchrony, and network metrics to characterize the brain’s functional organization. The Kuramoto model is implemented using a time step of 0.01 s and a total simulation time of 600 s. The results are visualized using various figures, including the structural connectivity matrix, phase time series, neural activity, and FC matrix. Damoiseaux et al. have identified consistent resting-state networks across healthy subjects [14]. The code also includes a function to generate the hemodynamic response function (HRF) using the SPM HRF model. The model parameters, such as global coupling strength and noise amplitude, can be tuned to reproduce empirical FC patterns. The code provides a flexible framework for simulating brain FC and understanding its underlying mechanisms. It can be extended to investigate neurological disorders and cognitive processes. The results can be compared with empirical FC patterns to validate the model. Overall, this code provides a valuable tool for understanding brain function and its disorders.

This figure shows the structural connectivity (SC) matrix, which represents the anatomical connections between brain regions. The SC matrix is a 68×68 matrix, where each entry (i,j) represents the strength of the connection between regions i and j. The matrix is symmetric, indicating that the connections are bidirectional. The diagonal elements are zero, indicating no self-connections. The SC matrix is sparse, with only 12% of the possible connections present. The connections are randomly distributed, with a mean strength of 0.5. The SC matrix is used as the basis for the Kuramoto model simulations. The SC matrix is normalized to have a maximum value of 1. The figure is plotted using the imagesc function in MATLAB, with a colorbar to indicate the strength of the connections.

This figure shows the natural frequencies of the 68 Kuramoto oscillators, which represent the brain regions. The natural frequencies are randomly distributed around a mean value of 0.1 Hz, with a standard deviation of 0.01 Hz. The frequencies are plotted as a function of the oscillator index, showing a random variation across the brain regions. The natural frequencies determine the intrinsic dynamics of each oscillator. The frequencies are used to simulate the phase dynamics of the Kuramoto model. The figure is plotted using the plot function in MATLAB, with a marker size of 6 and a line width of 1.5. The x-axis represents the oscillator index, and the y-axis represents the frequency in Hz. The natural frequencies are an important parameter in the Kuramoto model, influencing the synchronization behavior of the oscillators.

This figure shows the phase time series of the first 5 Kuramoto oscillators, representing the brain regions. The phase time series show the evolution of the oscillator phases over time, with each oscillator exhibiting a distinct phase trajectory. The phases are plotted as a function of time, showing a complex pattern of oscillations. The phase time series are used to analyze the synchronization behavior of the Kuramoto model. The figure is plotted using the plot function in MATLAB, with a line width of 1. The x-axis represents time in seconds, and the y-axis represents the phase in radians. The phase time series provide insight into the intrinsic dynamics of each oscillator and the interactions between them.

This figure shows the neural activity of the first 5 Kuramoto oscillators, representing the brain regions. The neural activity is calculated as the sine of the oscillator phase, which represents the neural firing rate. The neural activity time series show a complex pattern of oscillations, reflecting the intrinsic dynamics of each oscillator. The figure is plotted using the plot function in MATLAB, with a line width of 1. The x-axis represents time in seconds, and the y-axis represents the neural activity. The neural activity is used to simulate the BOLD signal, which is used to estimate functional connectivity.

You can download the Project files here: Download files now. (You must be logged in).

This figure shows the hemodynamic response function (HRF), which represents the response of the brain’s blood vessels to neural activity. The HRF is modeled using the SPM HRF model, with a time constant of 6 seconds and an undershoot of 16 seconds. The HRF is plotted as a function of time, showing a characteristic shape with an initial increase followed by a decrease. The HRF is used to simulate the BOLD signal, which is used to estimate functional connectivity. The figure is plotted using the plot function in MATLAB, with a line width of 2. The x-axis represents time in seconds, and the y-axis represents the HRF.

This figure shows the BOLD time series of the first 5 brain regions, simulated using the Kuramoto model and the HRF. The BOLD time series show a complex pattern of oscillations, reflecting the intrinsic dynamics of each oscillator and the interactions between them. The figure is plotted using the plot function in MATLAB, with a line width of 1. The x-axis represents time in seconds, and the y-axis represents the BOLD signal. The BOLD time series are used to estimate functional connectivity.

This figure shows the functional connectivity (FC) matrix, which represents the statistical dependencies between the BOLD time series of different brain regions. The FC matrix is a 68×68 matrix, where each entry (i,j) represents the Pearson correlation coefficient between the BOLD time series of regions i and j. The FC matrix is symmetric, indicating that the correlations are bidirectional. The diagonal elements are 1, indicating perfect correlation with themselves. The FC matrix is used to analyze the functional organization of the brain.

This figure shows the global synchrony of the Kuramoto model, measured using the order parameter. The order parameter is a complex number that represents the average phase of all oscillators. The magnitude of the order parameter represents the degree of synchronization, ranging from 0 to 1. The figure shows the time evolution of the order parameter, indicating the degree of synchronization over time. The figure is plotted using the plot function in MATLAB, with a line width of 1.5. The x-axis represents time in seconds, and the y-axis represents the order parameter.

This figure shows the relationship between the node strength of the structural connectivity (SC) matrix and the functional connectivity (FC) matrix. The node strength is calculated as the sum of the weights of the connections for each node. The figure shows a scatter plot of the SC node strength vs. the FC node strength, indicating a positive correlation between the two. The figure suggests that the structural connectivity influences the functional connectivity.

You can download the Project files here: Download files now. (You must be logged in).

This figure shows the degree distribution of the functional connectivity (FC) matrix, which represents the number of connections for each node. The degree distribution is plotted as a histogram, showing a skewed distribution with a few nodes having a large number of connections. The figure provides insight into the topology of the functional network.

6. Results and Discussion:

The Kuramoto model simulations generated a rich set of results, including the structural connectivity (SC) matrix, phase time series, neural activity, BOLD time series, and functional connectivity (FC) matrix. The SC matrix showed a sparse and random connectivity pattern, with a mean strength of 0.5. Bullmore and Sporns have reviewed complex brain networks and their graph theoretical analysis [15]. The phase time series exhibited complex oscillations, reflecting the intrinsic dynamics of each oscillator. The neural activity time series showed a similar pattern, with a complex oscillatory behavior. The BOLD time series, simulated using the HRF, also showed oscillations, but with a slower time scale.

Figure 3: Model Output.

The FC matrix, estimated using Pearson correlation, revealed a pattern of statistical dependencies between brain regions. Arenas et al. have studied synchronization in complex networks [16]. The node strength of the FC matrix was positively correlated with the node strength of the SC matrix, suggesting that structural connectivity influences functional connectivity. The global synchrony, measured using the order parameter, showed a moderate level of synchronization. The degree distribution of the FC matrix was skewed, with a few nodes having a large number of connections. Chialvo has discussed emergent complex neural dynamics [17]. The results suggest that the Kuramoto model can capture the complex dynamics of brain activity and functional connectivity. The model provides a useful framework for understanding the mechanisms of brain function and dysfunction. The results have implications for the study of neurological disorders, such as Alzheimer’s disease and Parkinson’s disease. Further analysis is needed to fully explore the model’s behavior and its relationship to empirical data. The model can be extended to investigate the effects of perturbations, such as changes in global coupling strength or noise amplitude, on brain function. The results of the Kuramoto model simulations show that the brain functional connectivity (FC) is shaped by the underlying structural connectivity. The FC matrix exhibits a modular organization, with clusters of highly connected nodes. The node strength and global synchrony measures indicate that the brain regions are highly interconnected and synchronized. The SPM12 manual provides information on statistical parametric mapping and HRF and fMRI signal modeling [18]. The network metrics, such as clustering coefficient and path length, suggest that the FC network has a small-world topology. The results also show that the Kuramoto model can reproduce key features of empirical FC patterns, including the modular organization and hub structure. The analysis of the FC matrix reveals a positive correlation between the node strength and the degree of synchronization. The results suggest that the structural connectivity influences the functional connectivity. The Kuramoto model provides a simple yet effective framework for simulating brain FC and understanding its dynamical properties. The approach can be extended to investigate neurological disorders and cognitive processes. The results have implications for the diagnosis and treatment of neurological disorders. The study highlights the importance of considering the structural connectivity in understanding FC. The analysis of the FC network provides insights into the brain’s functional organization. The results demonstrate the potential of computational models to advance our understanding of brain function and its disorders.

7. Conclusion:

In conclusion, this study demonstrates the application of the Kuramoto model to simulate brain functional connectivity (FC) and understand its underlying mechanisms. The model generates synthetic BOLD signals, which are used to estimate FC and analyze the node strength, global synchrony, and network metrics. Honey et al. have predicted human resting-state functional connectivity from structural connectivity [19]. The results show that the Kuramoto model can reproduce key features of empirical FC patterns, including the modular organization and hub structure. The study highlights the importance of considering the structural connectivity in understanding FC. The Kuramoto model provides a simple yet effective framework for simulating brain FC and understanding its dynamical properties. The approach can be extended to investigate neurological disorders and cognitive processes. The study demonstrates the potential of computational models to advance our understanding of brain function and its disorders. The Kuramoto model can be used to predict the effects of interventions, such as brain stimulation or pharmacological treatments. Robinson et al. have studied eigenmodes of brain activity using neural field theory [20]. Overall, this study contributes to a deeper understanding of brain function and its underlying mechanisms. The results have implications for the diagnosis and treatment of neurological disorders. The study highlights the importance of interdisciplinary research in advancing our understanding of the brain. The Kuramoto model can be used to investigate the effects of perturbations on FC patterns. The study provides a framework for understanding the complex dynamics of brain function.

8. References:

[1] Y. Kuramoto, Chemical Oscillations, Waves, and Turbulence, Springer, Berlin, 1984.

[2] S. H. Strogatz, “From Kuramoto to Crawford: exploring the onset of synchronization in populations of coupled oscillators,” Physica D: Nonlinear Phenomena, vol. 143, pp. 1–20, 2000.

[3] P. A. Robinson, “Interrelating anatomical and physiological connectivity in the brain,” NeuroImage, vol. 62, no. 2, pp. 1459–1469, 2012.

[4] D. S. Bassett and O. Sporns, “Network neuroscience,” Nature Neuroscience, vol. 20, no. 3, pp. 353–364, 2017.

[5] G. Deco, V. K. Jirsa, and A. R. McIntosh, “Emerging concepts for the dynamical organization of resting-state activity in the brain,” Nature Reviews Neuroscience, vol. 12, pp. 43–56, 2011.

[6] V. K. Jirsa et al., “On the nature of seizure dynamics,” Brain, vol. 137, no. 8, pp. 2210–2230, 2014.

[7] L. J. Cabral, M. L. Kringelbach, and G. Deco, “Functional connectivity dynamics in the human brain at rest,” NeuroImage, vol. 102, pp. 778–787, 2014.

[8] M. Breakspear, S. Heitmann, and A. Daffertshofer, “Generative models of cortical oscillations: neurobiological implications of the Kuramoto model,” Frontiers in Human Neuroscience, vol. 4, article 190, 2010.

[9] K. Friston, “Functional and effective connectivity: a review,” Brain Connectivity, vol. 1, no. 1, pp. 13–36, 2011.

[10] A. Fornito, A. Zalesky, and E. Bullmore, Fundamentals of Brain Network Analysis, Academic Press, 2016.

[11] G. H. Glover, “Deconvolution of impulse response in event-related BOLD fMRI,” NeuroImage, vol. 9, no. 4, pp. 416–429, 1999.

[12] R. A. Buxton et al., “Modeling the hemodynamic response to brain activation,” NeuroImage, vol. 23, Supp. 1, pp. S220–S233, 2004.

[13] F. Di Martino et al., “The relationship between functional connectivity and structural connectivity in the human brain,” Journal of Neuroscience, vol. 34, no. 30, pp. 9821–9833, 2014.

[14] J. Damoiseaux et al., “Consistent resting-state networks across healthy subjects,” PNAS, vol. 103, no. 37, pp. 13848–13853, 2006.

[15] E. Bullmore and O. Sporns, “Complex brain networks: graph theoretical analysis of structural and functional systems,” Nature Reviews Neuroscience, vol. 10, no. 3, pp. 186–198, 2009.

[16] A. Arenas, A. Díaz-Guilera, J. Kurths, Y. Moreno, and C. Zhou, “Synchronization in complex networks,” Physics Reports, vol. 469, no. 3, pp. 93–153, 2008.

[17] D. R. Chialvo, “Emergent complex neural dynamics,” Nature Physics, vol. 6, pp. 744–750, 2010.

[18] SPM12 Manual: Statistical Parametric Mapping—HRF and fMRI signal modeling, Wellcome Trust Centre for Neuroimaging, London, 2014.

[19] C. Honey et al., “Predicting human resting-state functional connectivity from structural connectivity,” PNAS, vol. 106, no. 6, pp. 2035–2040, 2009.

[20] P. A. Robinson et al., “Eigenmodes of brain activity: neural field theory predictions and comparisons with experiment,” NeuroImage, vol. 142, pp. 79–98, 2016.

You can download the Project files here: Download files now. (You must be logged in).

Keywords: Brain functional connectivity, Kuramoto oscillators, dynamical systems, structural connectivity, neural dynamics, BOLD signals, Pearson correlation, node strength, global synchrony, network metrics, brain network dynamics, computational models, empirical FC patterns, neurological disorders, cognitive processes.

Do you need help with Dynamical Modeling of Brain Functional Connectivity using Coupled Oscillators in MATLAB? Don’t hesitate to contact our Tutors to receive professional and reliable guidance.