Catálogo de publicaciones - libros

The neuropil is a complicated 3D tangle of neural and glial processes. Recent advances in microconnectomics has made it possible to reconstruct neural circuits from serial-section electron microscopy at the micron scale. Electron microscopy allows even higher resolution reconstructions on the nanometer scale. Nanoconnectomic reconstructions approaching molecular resolution allow us to explore the topology of extracellular space and the precision with which synapses are modified by patterns of neural activity.

The diversity and the specialized connectivity and function of inhibitory cortical neurons have been the focus of intense research for many decades (Fishell and Rudy, Ann Rev Neurosci 34:535–567, 2011). Until recently, technical limitations have restricted the power of experiments that could be conducted in vivo. Nevertheless, in vitro studies identified dozens of distinct cortical inhibitory neuron types, each with unique chemical properties, intrinsic firing properties and connection specificity. And at the same time, post-mortem studies from human patients have demonstrated defects of inhibitory circuit markers in diseases such as schizophrenia (Curley and Lewis, J Physiol 590:715–724, 2012; Stan and Lewis, Curr Pharm Biotech 13:1557–1562, 2012; Lewis, Curr Opin Neurobiol 26:22–26, 2014). Together, these observations have led to the hypothesis that distinct types of inhibitory neurons play distinct functional roles in the dynamic regulation of brain states and in the context-dependent extraction of sensory information, cognitive function, and behavioral output—functions thought to be disrupted in disorders such as schizophrenia and autism.

Despite the wealth of evidence in support of this hypothesis, tools have only recently emerged to allow detailed studies of neural circuit mechanisms underlying in vivo dynamics and to implicate specific inhibitory cell types and connections in specific functions (Luo et al., Neuron 57:634–660, 2008). Now, rather than broadly surveying inhibitory neuron properties and connections in vitro, studies have begun to focus more deeply on the in vivo contributions of those inhibitory cell types that are genetically accessible and can therefore be interrogated with modern genetic tools for manipulating and monitoring activity of specific cell types.

Mouse lines that express Cre-recombinase selectively in three major, non-overlapping groups of inhibitory cortical neurons—Parvalbumin-expressing (PV), somatostatin-expressing (SST), and vasoactive intestinal peptide-expressing (VIP; Lee et al., J Neurosci 30:16796–16808, 2010; Xu et al. J Comp Neurol 518:389–404, 2010; Rudy et al., J Comp Neurol 518:389–404, 2011; Taniguchi et al., J Comp Neurol 518:389–404, 2011)—have allowed detailed studies of the connectivity and in vivo functional roles of these cell groups. Such studies have implicated PV inhibitory neurons in gain control (Atallah et al., Neuron 73:159–170, 2012; Lee et al., Nature 488:379–383, 2012; Nienborg et al., J Neurosci 33:11145–11154, 2013), SST interneurons in the suppression of lateral and feedback (top-down) interactions (Adesnik and Scanziani, Nature 464:1155–1160, 2010; Nienborg et al., J Neurosci 33:11145–11154, 2013), and VIP interneurons in the dynamic regulation of SST cells under the control of brain state-dependent neuromodulators (Kawaguchi, J Neurophysiol 78:1743–1747, 1997; Alitto and Dan, Front Syst Neurosci 6:79, 2012; Lee et al., Nat Neurosci 16:1662–1670, 2013; Pi et al., Nature 503:521–524, 2013; Polack et al., Nat Neurosci 16:1331–1339, 2013; Fu et al., Cell 156:1139–1152, 2014; Stryker, Cold Spring Harbor Symp Quant Biol 79:1–9, 2014; Zhang et al., Science 345:660–665, 2014).

Recent advances in neuroscience have enabled increasingly detailed insight into the activity and structure of brain circuitry. In previous work, we have developed and applied methods for precisely controlling the activity of specific cells and projections within neural systems during behavior (optogenetics). Here I review distinct complementary technological approaches for observing natural activity patterns in these cells and projections during behavior (fiber photometry) and for obtaining anatomical insights into the wiring and molecular phenotype of these circuit elements within the intact mammalian brain (CLARITY-optimized lightsheet microscopy). Together these approaches may help further advance understanding of the circuit dynamics and wiring patterns that underlie adaptive and maladaptive behavior.

New techniques for identifying cell types, tracing their synaptic partners, imaging and manipulating their activity in behaving organisms have made mice a widely used model for linking brain circuits to behavior. Most behaviors are tied to vision: identifying objects, guiding movements of body parts, navigating through the environment, and even social interactions. Reason enough to focus on the mouse visual cortex. To find our way around in the occipital cortex, we needed a map. We took a classic approach and traced in the same animal the outputs from multiple retinotopic sites of primary visual cortex (V1) and compared the relative location of projections in the extrastriate cortex. We found nine extrastriate maps and showed by single unit recordings that each of the connectional maps contained visually responsive neurons whose receptive fields were mapped in orderly fashion and completely covered the visual field. Remarkably, a tiny region of one sixth of a dime contained a two- to three-times larger number of areas than the highly developed somatosensory and auditory cortices. By tracing the connections, we found that each of the ten visual areas projected to 25–35 cortical targets and interconnected virtually all of the areas reciprocally with one another. Although the binary graph density of the connection matrix was nearly complete, the connection strengths between areas within the ventral and dorsal cortex differed, indicating that the information from V1 flowed into distinct but interconnected streams. Unit recordings and calcium imaging studies showed that the ventral and dorsal streams processed different spatiotemporal information, which aligned with known properties of streams in primates. Analyses of the laminar patterns of interareal projections showed that areas were organized at multiple levels, suggesting that each stream represented a processing hierarchy.

Recent connectomic tract tracing reveals that, contrary to what was previously thought, the cortical inter-areal network has high density. This finding leads to a necessary revision of the relevance of some of the graph theoretical notions, such as the small-world property, hubs and rich-clubs that have been claimed to characterize the inter-areal cortical network. Weight and projection distance relationships of inter-areal connections inferred from consistent tract tracing data have recently led to the definition of a novel network model, the exponential distance rule (EDR) model, that predicts many observed local and global features of the cortex. The EDR model is a spatially embedded network whose properties are determined by the physical constraints on wiring and geometry, in sharp contrast with the purely topological graph models used heretofore in the description of the cortex. We speculate that, when diving down to finer levels of the embedded cortical network, similar, physically constrained descriptions of connectivity may prove to be equally important for understanding cortical function.

Major efforts are underway to provide highly detailed descriptions of static anatomical brain connectivity in rodents, even down to the level of individual synapses. To fully understand brain functioning and to bridge the gap between rodents and humans, however, I argue in this chapter that effective connectivity studies in nonhuman primates are equally critical. The primate community should embrace the novel, high-precision genetic-based toolkits developed in invertebrates and rodents to study how activity in one brain region influences that in connected brain regions. These methods will allow us to measure true functional weights of anatomical connections during highly varying cognitive and perceptual demands in primates. Why monkeys, and why effective connectivity in addition to anatomical connectivity? First, the nonhuman primate is critically important to understand the functioning of the human brain since important brain regions, such as the granular prefrontal cortex carrying higher cognitive functions, are lacking in rodents as opposed to primates. Second, a pure anatomical description of connections at different scales may be useful to models of brain functioning, however, it has little value to perception and behavior emerging from dynamic neuronal activity in distributed brain networks. Hence, tools that allow us to measure these dynamics at large scale and to causally interfere with the system at high temporal and spatial resolution are required to increase our understanding of changes in information processing at different stages within a functional network. In this chapter I will review past and emerging methods to study effective connectivity (mainly) in nonhuman primates, in other words how activity within a given brain area influences processing in anatomically connected brain regions. I will also argue that high-resolution whole brain imaging in monkeys may be invaluable to guide reversible perturbations and massive neurophysiological recordings simultaneously within multiple nodes of functional networks.

To decipher brain function, it is vital to know how the brain is wired. This entails elucidation of brain circuits at multiple scales, including microscopic, mesoscopic, and macroscopic levels. Here, we review recent progress in mapping the macroscopic brain circuits and functional organization of the cerebral cortex in primates—humans and macaque monkeys, in particular. There are many similarities across species in terms of overall patterns of cortical gray matter myelination as well as functional areas that are presumed to be homologous. However, there are also many important species differences, including cortical convolutions that are much more complex and more variable in humans than in monkeys. Our ability to analyze structure and function has benefited from improved methods for intersubject registration that cope with this individual variability. To characterize long-distance connectivity, powerful but indirect methods are now available, including resting-state functional connectivity and diffusion imaging coupled with probabilistic tractography. We illustrate how connectivity inferred from diffusion imaging and tractography can be evaluated in relation to ‘ground truth’ based on anatomical tracers in the macaque. Interspecies registration between human and macaque cortex based on presumed interspecies homologies demonstrates an impressive degree of areal expansion in regions associated with higher cognitive function.

Nervous systems are networks of neurons and brain regions that are structurally interconnected and dynamically linked in complex patterns. As mapping and recording techniques become increasingly capable of capturing neural structure and activity across widely distributed circuits and systems, there is a growing need for new analysis tools and modeling approaches to make sense of these rich “big data” sets. Modern network science offers a way forward. Both structural and functional brain data sets can be rendered in the form of complex networks and thus become amenable for network modeling and analysis, which can be carried out across scales, from the micro-scale of individual neurons to the macro-scale of whole-brain recordings. In this article, I sketch an overview of structural and functional brain network studies ranging from cells to systems. My emphasis will be on common themes in mapping network attributes across scales. In addition to highlighting important advances, I will outline some major challenges that need to be overcome to achieve a more complete understanding of connectome networks.

Hemispheric specialization (HS), or hemispheric dominance, is a nineteenth century concept that relates to the fact that a given hemisphere is the pilot of a given function such as, for example, the left hemisphere is dominant for language and for right-handedness. HS is grounded in both intra-hemispheric white matter connections, supported by associative bundles, and inter-hemispheric connections between cortical areas located in mirrored positions (homotopic), through the corpus callosum (CC) fiber tracts. Imaging investigations have measured anatomical and/or functional asymmetry, assessing HS at the voxelwise, regional, or hemispheric level. Comparison of these simple measures obtained with functional imaging during language tasks with results from the Wada test has validated that asymmetries do size up HS and pave the way for the investigation of HS in healthy humans. Anatomical asymmetries explain only a fraction of functional variability in lateralization, likely because structural and functional asymmetries develop at different periods of life. Anatomical asymmetries appear as early as the 26th week of gestation; at birth they are identical to those of adults. In contrast, functional neuroimaging investigations have revealed that inter-hemispheric connectivity appears at birth and is leftward asymmetrical in auditory areas, whereas in high-order language areas, this inter-hemispheric connectivity slowly shifts during development to a predominant intra-hemispheric connectivity in the adult. The precise timing and neural basis of this shift are still unknown, but it has been nevertheless shown that the connectivity is not yet in place at the age of seven and that it parallels an increase in leftward asymmetry during language tasks. Abnormal development of this asymmetry is observed in severe mental illnesses that exhibit language symptoms, such as schizophrenia and autism. In addition, after a dominant hemisphere lesion, good language capacities are associated with the recovery of a leftward asymmetry during language tasks. However, neuroimaging studies have shown that HS variability for language, up to rightward dominance, exists in healthy individuals and is partly explained by both behavioral (handedness) and anatomical (i.e., brain volume, size of the left planum temporale) factors, with these factors possibly interacting with one another. Knowledge of the setting up of language HS is still fractional and very little is known about right hemisphere dominance and complementary specialization of the two hemispheres. Considering the complexity of the question, progress will come from the acquisition and analysis of databases developed to answer those questions, such as the BIL&GIN, which includes a sample of 450 healthy volunteers balanced for handedness and gender. Each participant has been characterized for cognitive abilities, anatomy, resting state connectivity and activated networks during motor, language and visuospatial tasks.

Here we give an overview of a worldwide effort, called the ENIGMA Consortium (), which unites scientists worldwide to determine how variants in our genetic code influence the brain, and how 12 major diseases affect the brain worldwide. At the time of writing, ENIGMA involves over 500 scientists from 185 institutions worldwide, working together on around 30 projects to discover factors that may help or harm the brain. By pooling genome-wide genomic data and brain imaging from over 33,000 people, ENIGMA has been able to identify single-nucleotide differences in the genome that are associated with differences in human brain structure and function. Given the broad interest in brain connectivity and the factors that affect it, we outline some tactics adopted by ENIGMA to discover specific genes that affect the brain; then we describe how ENIGMA is extending these methods to discover genetic influences on brain connectivity.