Wavelet-based Gaussian-mixture hidden Markov model for the detection of multistage seizure dynamics: a proof-of-concept study

Behavioral Studies Using Large-Scale Brain Networks – Methods and Validations

Mapping human behaviors to brain activity has become a key focus in modern cognitive neuroscience. As methods such as functional MRI (fMRI) advance cognitive scientists show an increasing interest in investigating neural activity in terms of functional connectivity and brain networks, rather than activation in a single brain region. Due to the noisy nature of neural activity, determining how behaviors are associated with specific neural signals is not well-established. Previous research has suggested graph theory techniques as a solution. Graph theory provides an opportunity to interpret human behaviors in terms of the topological organization of brain network architecture. Graph theory-based approaches, however, only scratch the surface of what neural connections relate to human behavior. Recently, the development of data-driven methods, e.g., machine learning and deep learning approaches, provide a new perspective to study the relationship between brain networks and human behaviors across the whole brain, expanding upon past literatures. In this review, we sought to revisit these data-driven approaches to facilitate our understanding of neural mechanisms and build models of human behaviors. We start with the popular graph theory approach and then discuss other data-driven approaches such as connectome-based predictive modeling, multivariate pattern analysis, network dynamic modeling, and deep learning techniques that quantify meaningful networks and connectivity related to cognition and behaviors. Importantly, for each topic, we discuss the pros and cons of the methods in addition to providing examples using our own data for each technique to describe how these methods can be applied to real-world neuroimaging data.

Dynamics underlying interictal to ictal transition in temporal lobe epilepsy: insights from a neural mass model

We propose an approach that combines a neural mass model and clinical intracranial electroencephalographic (iEEG) recordings to explore the potential pathophysiological mechanisms (at the neuronal population level) of ictogenesis. Thirty iEEG recordings from 10 temporal lobe epilepsy (TLE) patients around seizure onset were investigated. Physiologically meaningful parameters [average excitatory (A_e ), slow (B), and fast (G) inhibitory synaptic gain] were identified during interictal to ictal transition. Four ratios (A_e /G, A_e /B, A_e /(B + G), and B/G) were derived from these parameters, and their evolution over time was analyzed. The excitation/inhibition ratio increased around seizure onset and decreased before seizure offset, indicating the impairment and re-emergence of excitation/inhibition balance around seizure onset and before seizure offset, respectively. Moreover, the slow inhibition may have an earlier effect on excitation/inhibition imbalance. We confirm the decrease in excitation/inhibition ratio upon seizure termination in human temporal lobe epilepsy, as revealed by optogenetic approaches both in vivo in animal models and in vitro. The increase in excitation/inhibition ratio around seizure occurrence could be an indicator to detect seizures.

On the Role of Situational Stressors in the Disruption of Global Neural Network Stability during Problem Solving

When individuals are placed in stressful situations, they are likely to exhibit deficits in cognitive capacity over and above situational demands. Despite this, individuals may still persevere and ultimately succeed in these situations. Little is known, however, about neural network properties that instantiate success or failure in both neutral and stressful situations, particularly with respect to regions integral for problem-solving processes that are necessary for optimal performance on more complex tasks. In this study, we outline how hidden Markov modeling based on multivoxel pattern analysis can be used to quantify unique brain states underlying complex network interactions that yield either successful or unsuccessful problem solving in more neutral or stressful situations. We provide evidence that brain network stability and states underlying synchronous interactions in regions integral for problem-solving processes are key predictors of whether individuals succeed or fail in stressful situations. Findings also suggested that individuals utilize discriminate neural patterns in successfully solving problems in stressful or neutral situations. Findings overall highlight how hidden Markov modeling can provide myriad possibilities for quantifying and better understanding the role of global network interactions in the problem-solving process and how the said interactions predict success or failure in different contexts.

Reduced information transmission in the internal segment of the globus pallidus of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced rhesus monkey models of Parkinson's disease

Rhesus monkey models of Parkinson's disease were induced by injection of N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine. Neural firings were recorded using microelectrodes placed in the internal segment of the globus pallidus. The wavelets and power spectra show gradual power reduction during the disease process along with increased firing rates in the Parkinson's disease state. Singular values of coefficients decreased considerably during tremor-related activity as well as in the Parkinson's disease state compared with normal signals, revealing that higher-frequency components weaken when Parkinson's disease occurs. We speculate that the death of neurons could be reflected by irregular frequency spike trains, and that wavelet packet decomposition can effectively detect the degradation of neurons and the loss of information transmission in the neural circuitry.

Kernel duration and modulation gain in a coupled oscillator model and their implications on the progression of seizures

The coupled oscillator model has previously been used for the simulation of neuronal activities in in vitro rat hippocampal slice seizure data and the evaluation of seizure suppression algorithms. Each model unit can be described as either an oscillator which can generate action potential spike trains without inputs, or a threshold-based unit. With the change of only one parameter, each unit can either be an oscillator or a threshold-based spiking unit. This would eliminate the need of a new set of equations for each type of unit. Previous analysis has suggested that long kernel duration and imbalance of inhibitory feedback can cause the system to intermittently transition into and out of ictal activities. The state transitions of seizure-like events were investigated here; specifically, how the system excitability may change when the system underwent transitions in the preictal and postictal processes. Analysis showed that the area of the excitation kernel is positively correlated with the mean firing rate of ictal activity. The kernel duration is also correlated to the amount of ictal activity. The transition into ictal involved the escape from the saddle point foci in the state space trajectory identified using Newton's method.