NYU Center for Brain Imaging Events

fMRI decoding: what does it reflect, and what can we learn

To encode and make inferences about the world, the brain represents patterns -- visual, auditory, cognitive -- using populations of neurons with diverse and complex forms of selectivity. A coarse-scale method like functional magnetic resonance imaging (fMRI) would, at first glance, appear poorly suited to studying these representations. Consider the case of orientation representation in primary visual cortex (V1). A single fMRI voxel pools responses from many orientation-tuned neurons. Because orientation tuning varies at a fine, columnar spatial-scale, tuning should cancel at the level of fMRI voxels. Surprisingly, results from multivariate decoding analyses imply that voxels in human V1 are weakly but reliably orientation selective. It is widely believed that these small biases arise because of random spatial irregularities in the underlying columnar architecture, and this interpretation, while untested, has been extended to the study of cognitive functions throughout the brain. I will describe a set of experiments that test this hypothesis directly by characterizing the cortical organization of orientation-selective fMRI responses in human V1. We developed an approach for measuring orientation selectivity at multiple spatial scales. Using this technique, we discovered a large-scale map of orientation preference. The existence of this map is both necessary and sufficient for multivariate decoding of orientation, demonstrating that random spatial irregularities do not contribute to decoding. Our results thus imply a parsimonious, but sobering, explanation for why decoding works, and help guide the interpretation of the rapidly growing number of studies based on this technique.

Two distinct mechanisms for negative correlations in rest functional connectivity MRI

Coherent low frequency fluctuations of the BOLD signal in resting state (rest-fcMRI) were shown to contain functional neuronal network information. Resting-state networks (RSN) exhibit positive correlations between the regions that constitute the network, suggesting a functional link between them. However, several RSNs were shown to have an inverse correlation between each other. The underlying physiological mechanisms and the relevance of negative correlations to neurobiology are not clear and are the subject of this study. We compared human and rat rest-fcMRI data, making use of both the similarities (e.g., similar organization: cortical vs. non-cortical structures, inter-hemispheric symmetry etc.) and differences (e.g., different hemodynamic characteristics such as cardiac rates and spatial distances) between them. In addition, the fact that the rats' cortex is relatively unfolded, enables to minimize confounding effects of CSF and large blood vessels on the rest-fcMRI correlations. We show that: (i) Negative correlations observed in rest-fcMRI reflect true physiological traits and are not the mere result of mathematical biases introduced by data analysis. (ii) At least two distinct mechanisms may underlay the appearance of negative correlations, reflecting the actual synchronization between regional neural activities on the one hand and their manifested BOLD signal responses on the other hand. (iii) The variant involvement of CBV in the hemodynamic responses of two different regions may introduce such negative correlations.

Face recognition processes: Insights and perspectives from multiple brain investigation methods

Political Ideology and the Intrinsic Functional Organization of the Brain

Political ideology (i.e., liberalism/conservatism) has long been thought to arise from differences in beliefs about human nature and personal experiences. However, recent research suggests that individual differences in political ideology are associated with more basic cognitive and motivational orientations toward the world (e.g., Oxley et al., 2008), as well as the structure and function of various brain regions implicated in conflict monitoring and emotional processing (Kanai et al., 2011; Amodio et al., 2007).

Resting state fMRI (R-fMRI) has emerged as a mainstream neuroimaging approach providing insight into the functional organization of the brain by characterizing the correlated intrinsic low-frequency fluctuations in the BOLD signal. Our lab and others have shown relationships between intrinsic functional connectivity (iFC) and individual differences in behavior and personality measures (e.g., social competence, Di Martino et al., 2009; empathy, Cox et al., 2011).

I will present preliminary analyses from a study in collaboration with Jay Van Bavel, Ph.D. and John T. Jost, Ph.D. examining whether individual differences in political ideology are associated with the intrinsic functional architecture of the brain, utilizing R-fMRI and iFC methodologies.

Neurobiology of Social Influence on Value: Imaging and Pharmacological Approaches

One of the traits that defines us as individuals and aligns us as groups is the degree to which we are influenced by opinions of other people. This trait can be fairly stable across contexts and vary between people. With a series of studies, I asked: Can social influence on the value of objects be reflected in value-associated neural activity of the brain? Is the tendency to conform a trait that is so stable that it is reflected in neural structure? Is it flexible enough to be modulated by common pharmacological treatments? In this talk, I will talk about how I approached these questions with a variety of methods and provide some intriguing results for discussion.