Maurizio Corbetta - US grants

The long term objective of this application is to understand how behavioral goals and expectations about visual events are represented and used within the human brain. Event-related fMRI methods are developed to distinguish the neural signals representing behavioral goals and expectations from the signals reflecting their effects on perceptual analysis. These methods are then used to study the dependence of these signals on the content of the expectation (e.g. an expectation about the motion of an object or its color), on how the expectation is generated (e.g. from verbal or non- verbal commands, or from instructions presented in different sensory modalities), and on how it is used during visual perception (e.g. to detect a stimulus or to categorize it). This information will advance our theoretical understanding of visual perception and awareness. These studies will also help to elucidate the pathophysiology of clinical brain disorders, that involve attentional and visual perceptual deficits, e.g. unilateral neglect, attentional deficit disorders, or dyslexia. A finer understanding of attentional mechanisms will also facilitate the development of neurobiologically- driven rehabilitative strategies.

The long term goal of this project is to understand the neural basis of functional recovery after focal brain injury. Neuroimaging (functional magnetic resonance imaging, fMRI, and positron emission tomography, PET) and methods will be used to study the recovery of aphasia after focal damage to the left frontal lobe. In particular, we will assess the relative contribution to recovery of tissue surrounding the infarction, and/or of homologous regions in the contralateral hemisphere. These findings will have implications for the design of neurobiologically- oriented strategies of rehabilitation, and the evaluation of therapies aimed at salvaging brain tissue around the core of a stroke.

DESCRIPTION (provided by applicant): Disorders of attention and action are often produced by brain injury. Unilateral spatial neglect is the most common neurobehavioral syndrome following focal brain injury to the human right hemisphere. Neglect spontaneously recovers to some extent in most patients, and therefore represents an ideal model for studying mechanisms of neurological recovery. The long-term goal of this proposal is to understand the neural and cognitive mechanisms of spatial neglect, its recovery, and how they relate to normal attentional mechanisms. Our overarching hypothesis is that neglect reflects the malfunction of attentional systems that are localized in posterior parietal and frontal cortex. We ask three basic questions. First, what processing deficits most contribute to the syndrome of neglect, and are these deficits anatomically dissociable? For example, what is the relative importance of deficits related to spatial attention and perception, directional motor actions, and arousal, and do they map to different brain regions? Second, which of these processing deficit(s) is/are more closely associated with the recovery of neglect? Third, what are the neural correlates of neglect recovery? Does recovery depend on the degree to which activation recovers in the lesioned and/or non-lesioned hemisphere, or does it depend on the balance of activity across the hemispheres? To answer these questions we carry out a prospective longitudinal behavioral, anatomical and functional magnetic resonance imaging (MRI) study in a large group of neglect patients with single unilateral hemispheric strokes, and several psychophysical experiments in patients with lesions in core areas of the attention system.

DESCRIPTION (provided by applicant): The relationship between the behavioral deficits and anatomical damage produced by a stroke is only partial, since physiological dsyfunction can be measured in brain regions far removed from the lesion. These non-local physiological deficits can be assessed using functional connectivity magnetic resonance imaging (fcMRl), which measures the temporal correlation between brain regions in the blood-oxygenation-level-dependent (BOLD) signal. Studies of fcMRl in healthy adults have identified distributed brain networks that underlie different behavioral functions, such as attention, motor control, and language. Our previous work on stroke patients with spatial neglect has shown that fcMRl deficits within the attention network correlate with a patient's deficit in the neglected field. This grant proposes that fcMRl can be used more broadly to understand the deficits produced by stroke across behavioral domains. We will measure fcMRl in a large, heterogeneous sample of stroke patients, and test several hypotheses of how strokes produce dysfunction in brain networks and how this dysfunction correlates with behavioral deficits. We predict that decreases in inter-hemispheric connectivity in motor and attention networks, which are bilaterally organized, will correlate with corresponding behavioral deficits and that these correlations will show functional specificity. For example, connectivity within an arm-defined motor network will predict upper-extremity function better than lower-extremity function. We determine the correlation between connectivity and behavior in an asymmetrically organized network, language, and compare the importance of inter-hemispheric vs intra-hemispheric connectivity. We examine how the functioning of multiple brain networks interact to produce a single behavioral deficit, and hypothesize that the connectivity of some networks, such as attention, correlates with behavior across domains. We measure fcMRl longtitudinally to determine if connectivity recovers in different networks at different rates and whether that recovery correlates with behavioral recovery. Finally, we study how changes in connectivity produced by a stroke depend on the type of tracts that are structurally damaged. PUBLIC HEALTH RELEVANCE: The goal of this proposal is to understand the effects of a stroke on the physiology of the brain and how they relate to the behavioral deficits produced by the stroke. We plan to measure these physiological effects in stroke patients using a novel method that will allow us to simultaneously assess the functioning of many different brain networks that are thought to underlie our ability to use language, attend in the environment, perform motor acts, and so forth. We will measure the ability of stroke patients to carry out these activities at different time points following their stroke, allowing us to understand how recovery of the patient's behavioral functions possibly relate to recovery of their physiological function.

DESCRIPTION (provided by applicant): Neuroimaging studies have provided a wealth of information on the cortical and subcortical regions of the human brain active during cognitive tasks. Recent studies have shown that regions that are co-activated during tasks maintain, even at rest, a high level of inter-regional correlation, or 'functional connectivity'. However, these interregional correlations are measured over long time scales (e.g. minutes). In contrast, little is known about the temporal dynamics and interactions of these brain regions over the short timescales (e.g. sec or ms) that are typical of most tasks. The analysis of temporal dynamics and interactions is strongly limited by the low temporal resolution of neuroimaging methods (functional magnetic resonance imaging, fMRI; Positron emission tomography, PET), and the low spatial resolution of methods for recording extracranial electro-magnetic activity (electroencephalography, EEG, magnetoencephalography, MEG). To solve these fundamental limitations, we have combined fMRI measures of blood-oxgyenation-level- dependent (BOLD) signals with electrocorticographic (ECoG) signals recorded from invasively monitored human subjects. We have developed methods to co-register functional networks localized with fMRI with intracranial electrodes that record surface cortical local field potentials (LFP). We have applied these novel methods to study the dynamics and interactions of cortical networks involved in spatial attention. Our preliminary results indicate that cortical networks observed with fMRI during a spatial attention task show multiple coherence modulations with ECoG. Maintenance of spatial attention correlates with sustained phase synchronization in the delta band (1-3 Hz) across multiple occipital, parietal, and frontal task-relevant regions, while shifts of spatial attention are associated with transient increases of phase synchronization in the theta band (3-7 Hz). We propose a series of experiments in which we first localize with fMRI cortical regions/networks specialized for spatial attention and then study their dynamics and interaction with ECoG on well- characterized cognitive tasks. Our first specific aim is to determine the role of delta/theta band phase synchronization in linking cortical regions during voluntary orienting of spatial attention. Our second specific aim studies how delta/theta band phase synchronization is affected by the temporal structure of a task. Our third specific aims focuses on the interaction between two different attention networks (dorsal, DAN; ventral, VAN) during stimulus-driven re-orienting.

? DESCRIPTION (provided by applicant): The long-term goal of our research program is to develop new treatments for improving behavioral deficits post-stroke based on a formal theory of brain function and organization. However, the immediate goal of this competitive renewal is to identify patterns of disrupted structural and functional connectivity that are related to stroke-induced deficits, determine how they change during recovery, and see if they predict outcome. Computational theories have proposed that lesions to the brain will have different effects based on the underlying network architecture. Mechanisms controlling recovery of function may similarly depend on network architecture. Specific Aim 1 measures changes in whole brain functional connectivity (FC) after damage to peripheral networks that are mainly connected within the network vs. central networks that communicate more broadly with other brain networks. We measure separately the effect of cortical vs. white matter pathway damage. We predict that peripheral lesions (either to gray or white matter) will produce FC changes that are mainly limited to the affected network, while central lesions (either gray or white matter) will produce multi- network FC changes. Peripheral networks are connected to input/output pathways and mediate predominantly sensory or motor functions. Central networks are connected through association pathways and mediate predominantly attention and memory functions. Specific Aim 2 uses a machine learning ridge regression method to test the idea that sensory-motor deficits and their recovery are more dependent on structural damage and anatomical disconnection, whereas cognitive impairments are better captured by multi-network functional connectivity dysfunction. An important clinical goal of the project is to determine whether advanced neuroimaging techniques, such as structural imaging, DTI, and fMRI can provide useful clinical information on an individual patient's recovery and outcome, over and above that provided by measures of acute behavior, demographic variables, and amount of treatment. In Specific Aim 3 we enter neuroimaging data into a classifier to predict outcome in individual patients, as well as to predict the effect of rehabilitation interventions. Finally, in Specific Aim 4 we move from correlation and prediction to mechanism. Our basic hypothesis is that changes in functional connectivity after focal injury reflect alterations in brain dynamics, specifically a decrease in the variability of inter-regional phase differences. This hypothesis is tested empirically by phase measurements, but also computationally, using a biophysically based model that can simulate the FC changes observed in patients. Importantly, the model also estimates the information capacity of the damaged brain, which we will correlate with the patient's empirically measured degree of modularity and profile of behavioral deficits.

SUMMARY A fundamental advance in Neuroimaging in the last 10 years has been the development of methods to derive regional parcellations of the cerebral cortex, or `parcellated connectomes', from structural and/or functional MRI data. Within the Conte Center, the Neuroimaging core will focus on providing single subject and group connectomes of the brain in humans and macaques. Specifically, connectomes will allows for: (1) the localization of recording sites in relation to independently localized human cortical regions and networks; (2) standardization of recording site localization across centers, which will in turn provide a way to compare results across experiments; (3) definition of a macaque monkey cortical parcellation atlas for the comparison, sharing, and meta-analysis of physiological studies within our consortium, but also the wider non-human primates physiological community. The Neuroinformatics core will facilitate data sharing across the Conte Center sites and the neuroscience community through an open source web interface.

Visual perception is limited by two fundamental rhythms, a 7-10 Hz theta/alpha rhythm that describes fluctuations in psychophysical performance and a 3-5 Hz rhythm related to saccadic eye movements. The structure of a task may impose or entrain an additional rhythm, such as when a target stimulus follows a cue at a fixed interval. The goal of this project is to identify in humans the neural correlates of these rhythms and determine their relationship to intrinsic rhythms of spontaneous activity. Neural activity is measured using electrocorticography (ECoG) in patients undergoing surgery for epilepsy. In Expt. 1, localizers are conducted to identify electrodes that respond to saccadic eye movements, foveal stimuli, and/or show spatially selective responses to peripheral stimuli. Functional magnetic resonance imaging is used to identify the large-scale brain network associated with each electrode. In Expt. 2 subjects are cued to detect a target under conditions of temporal uncertainty. In the one-location condition, the target only appears at the cued location. In the two- location condition, the target appears equi-probably at one of two locations. Consistent with previous studies indicating a fixed sampling rhythm, performance should fluctuate in the 1-location condition at twice the frequency as the fluctuations in each location of the 2-location condition. We then identify the neural correlate(s) of this rhythm in electrodes identified by the localizer. These correlates may be associated with local field potentials, modulations of band-limited power, or phase-amplitude relationships that couple low frequencies to high frequencies. In addition, we determine whether these correlates can be identified in intrinsic activity measured at rest. Expt. 3 compares the 1-location and 2-location conditions when the interval between the cue and target is fixed, corresponding to a task-imposed rhythm and temporal certainty. The question is whether the neural correlates of the task-imposed rhythm are independent of the intrinsic rhythms measured in Expt. 2. Finally, Expt. 4 compares the neural rhythms that are generated when subjects process foveal stimuli during a sequence of saccades as compared to the same foveal stimuli when fixation is maintained. We test the hypothesis that saccades produce a phase reset that aligns the maximal excitability phase of internal rhythms with incoming sensory signals. This hypothesis predicts that high gamma sensory evoked responses and behavioral performance should be facilitated by saccades. We also determine the relationship between the neural correlates of the saccadic rhythm, the 7-10 Hz sampling rhythm, and spontaneous rhythms measured at rest. We will interact closely with Project 2, which uses 2 of these tasks in monkey intracortical recordings, and with Projects 3 and 4 that study parallel auditory tasks in humans and monkeys, respectively. Along with Project 3, we will supply data to dynamical network modeling studies (Project 5), and use their findings to refine and/or modify our paradigms as work progresses. Integration of findings across these studies will build more robust models of brain mechanisms operating in Active Sensing.