This study uses statistical techniques to determine anesthetic depths of three females undergoing total abdominal hysterectomies. Spectral analysis of the electronencephalogram is employed to define changes in brain wave activity under different levels of anesthesia after administration of diazepam and isoflurane. The multivariate statistical technique of discriminant function analysis is used to determine which frequencies, or linear combinations of frequencies, yield the most information for classification of the electronencephalogram samples into one of the three anesthetic depths (mild sedation, moderate anesthesia, and anesthetic sleep). Spectral analysis of the electronencephalogram showed similar results for all three patients after administration of diazepam (mild sedation), but widely varying results among patients during anesthesia using isoflurane. The combination of spectral analysis and discriminant function analysis showed reliable discrimination among the three anesthetic depths. The ability to discriminate was significantly improved when only two anesthetic depths were used.

Simultaneous electroencephalogram (EEG) and functional magnetic resonance imaging (fMRI) have gained increasing popularity in studying human cognition due to their potential to map the brain dynamics with high spatial and temporal fidelity. Such detailed mapping of the brain is crucial for understanding the neural mechanisms by which humans make perceptual decisions. Despite recent advances in data acquisition and analysis of simultaneous EEG-fMRI, the lack of effective computational tools for optimal fusion of the two modalities remains a major challenge. The goal of this dissertation is to provide a recipe of machine learning methods for fusion of simultaneous EEG-fMRI data. Specifically, we investigate three types of fusion approaches and apply them to study the whole-brain spatiotemporal dynamics during a rapid object recognition task where subjects discriminate face, car, and house images under ambiguity. We first use an asymmetric fusion approach capitalizing on temporal single-trial EEG variability to identify early and late neural subsystems selective to categorical choice of faces versus nonfaces. We find that the degree of interaction in these networks accounts for a substantial fraction of our bias to see faces. Based on a computational modeling of behavioral measures, we further dissociate separate neural correlates of the face decision bias modulated by varying levels of stimulus evidence. Secondly, we develop a state-space model based symmetric fusion approach to integrate EEG and fMRI in a probabilistic generative framework. We use a variational Bayesian method to infer the network connectivity among latent neural states shared by EEG and fMRI. Finally, we use a data-driven symmetric fusion approach to compare representations of the EEG and fMRI against those of a deep convolutional neural network (CNN) in a common similarity space. We show a spatiotemporal hierarchical correspondence in visual processing stages between the human brain and the CNN. Collectively, our results show that the spatiotemporal properties of neural circuits revealed by the analysis of simultaneous EEG-fMRI data can reflect the choice behavior of subjects during rapid perceptual decision making.

The human brain relies on the complex interactions of distinct brain regions to support cognitive processes. The interplay between the hippocampus and neocortical regions plays a key role in the formation, storage, and retrieval of long-term episodic memories. Oscillatory activities during sleep are a fundamental mechanism that binds distributed neuronal networks into functionally coherent ensembles. However, the large-scale hippocampal-neocortical oscillatory mechanisms that support flexible modulation of long-term memory remain poorly understood. Furthermore, alterations to physiologic spatiotemporal patterns that are essential for intact memory function can result in pathophysiology in brain disorders such as focal epilepsy. Investigating how epileptic network activity disrupts connectivity in distributed networks and the organization of oscillatory activity are additional crucial areas that require further research. Our experiments on rodents and human patients with epilepsy have provided valuable insights into these mechanisms. In rodents, we used high-density microelectrode arrays in tandem with hippocampal probes to analyze intracranial electroencephalography (iEEG) from multiple cortical regions and the hippocampus. We identified key hippocampal-cortical oscillatory biomarkers that were differentially coordinated based on the age, strength, and type of memory. We also analyzed iEEG from patients with focal epilepsy and demonstrated how individualized pattern of pathologic-physiologic coupling can disrupt large-scale memory circuits. Our findings may offer new opportunities for therapies aimed at addressing distributed network dysfunction in various neuropsychiatric disorders.

Over 30 years of functional imaging studies have demonstrated that the human brain operates as a complex and interconnected system, with distinct functional networks and long-range coordination of neural activity. Yet, how our brains coordinate our behavior from moment to moment, permitting us to think, talk, and move at the same time, has been almost impossible to decode (Chapter 1). The invasive, long-term, and often multi-regional iEEG monitoring utilized for epilepsy surgery evaluation presents a valuable opportunity for studying brain-wide dynamic neural activity in behaving human subjects. In this study, we analyzed over 93 hours of iEEG recordings along with simultaneously acquired video recordings from 10 patients with drug-resistant focal epilepsy syndromes, who underwent invasive iEEG with broadly distributed bilateral depth electrodes for clinical evaluation. Initially, we explored the dynamic connectivity patterns quantified from band-limited neural activities using metrics from previous literature in a subset of subjects. These metrics can characterize long-range connectivity across brain regions and reveal variations over time. They have shown success in identifying state differences using controlled task presentations and trial-averaged data. However, we found that replicating this success with naturalistic, complex behaviors in our subjects is challenging. Although they demonstrate differences across wake and sleep states, they are less sensitive in differentiating more complicated and subtle state transitions during wakefulness. In addition, patterns identified from individual frequency bands exhibit patient-to-patient differences, making it difficult to generalize results across frequency bands and subjects. (Chapter 2). Inspired by clinical electrocortical stimulation mapping studies, which seek to identify critical brain sites for language and motor function, and the frequency gradient observed from human scalp and intracranial EEG recordings, we developed a new approach to meet the requirements for real-time analysis and frequency band selection. It is worth mentioning that detecting state transitions in naturalistic behavior requires analyzing raw EEG during individual transitions. We refer to this as "real-time analysis," to distinguish it from formal task performance and trial-averaging techniques. Rather than representing data as time-varying signals within specific frequency bands, we incorporated all frequencies (2-55 Hz) into our analysis by calculating the power spectral density (PSD) at each electrode. This analysis confirmed that the human brain’s neural activity PSD is heterogenous, exhibiting a distinct topography with bilateral symmetry, consistent with prior resting-state MEG and iEEG studies. However, investigating the variability of each region’s PSD over time (within a 2-second moving window), we discovered the tendency of individual electrode channel to switch back and forth between 2 distinct power spectral densities (PSDs, 2-55Hz) (Chapter 3). We further recognized that this ‘spectral switching’ occurs synchronously between distant sites, even between regions with differing baseline PSDs, revealing long-range functional networks that could be obscured in the analysis of individual frequency bands. Moreover, the real-time PSD-switching dynamics of specific networks exhibited striking alignment with activities such as conversation, hand movements, and eyes open versus closed, revealing a multi-threaded functional network representation of concurrent naturalistic behaviors. These network-behavior relationships were stable over multiple days but were altered during sleep, suggesting state-dependent plasticity of brain-wide network organization (Chapter 4). Our results provide robust evidence for the presence of multiple synchronous neuronal networks across the human brain. The real-time PSD switching dynamics of these networks provide physiologically interpretable read-outs, demonstrating the parallel engagement of multiple brain regions in a range of concurrent naturalistic behaviors (Chapter 5).

The field of cognitive neuroscience has benefited greatly from multimodal investigations of the human brain, which integrate various tools and neuroimaging data to understand brain functions and guide treatments for brain disorders. In this dissertation, we present a series of studies that illustrate the use of multimodal approaches to investigate brain structure and function, brain connectivity, and neuromodulation effects. Firstly, we propose a novel landmark-guided region-based spatial normalization technique to accurately quantify brain morphology, which can improve the sensitivity and specificity of functional imaging studies. Subsequently, we shift the investigation to the characteristics of functional brain activity due to visual stimulations. Our findings reveal that the task-evoked positive blood-oxygen-level dependent (BOLD) response is accompanied by sustained negative BOLD responses in the visual cortex. These negative BOLD responses are likely generated through subcortical neuromodulatory systems with distributed ascending projections to the cortex. To further explore the cortico-subcortical relationship, we conduct a multimodal investigation that involves simultaneous data acquisition of pupillometry, electroencephalography (EEG), and functional magnetic resonance imaging (fMRI). This investigation aims to examine the connectivity of brain circuits involved in the cognitive processes of salient stimuli. Using pupillary response as a surrogate measure of activity in the locus coeruleus-norepinephrine system, we find that the pupillary response is associated with the reorganization of functional brain networks during salience processing. In addition, we propose a cortico-subcortical integrated network reorganization model with potential implications for understanding attentional processing and network switching. Lastly, we employ a multimodal investigation that involves concurrent transcranial magnetic stimulation (TMS), EEG, and fMRI to explore network perturbations and measurements of the propagation effects. In a preliminary exploration on brain-state dependency of TMS-induced effects, we find that the propagation of left dorsolateral prefrontal cortex TMS to regions in the lateral frontoparietal network might depend on the brain-state, as indexed by the EEG prefrontal alpha phase. Overall, the studies in this dissertation contribute to the understanding of the structural and functional characteristics of brain network systems, and may inform future investigations that use multimodal methodological approaches, such as pupillometry, brain connectivity, and neuromodulation tools. The work presented in this dissertation has potential implications for the development of efficient and personalized treatments for major depressive disorder, attention deficit hyperactivity disorder, and Alzheimer's disease.