Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2004. / Includes bibliographical references. / (cont.) manipulation of intronic elements that enables fish genes to be spliced properly in mammalian cells; (iii) A computational analysis using EST data, genome sequence data, and microarray expression data of tissue- specific alternative splicing is conducted, which distinguishes human brain, testis and liver as having unusually high levels of AS, highlights differences in the types of AS occurring commonly in different tissues, and identifies candidate cis-regulatory elements and trans-factors likely to play important roles in tissue-specific AS in human cells; (iv) The identification of a set of discriminatory sequence features and their integration into a statistical machine-learning algorithm, ACEScan, which distinguishes exons subject to evolutionarily conserved alternative splicing from constitutively spliced or lineage-specifically-spliced exons is described; (v) The genome-wide search for and experimental validation of exon-skipping events using the combination of two silencing cis-elements, UAGG and GGGG. / The regulation of pre-messenger RNA splicing by the spliceosomal machinery via interactions between cis-regulatory elements and splicing trans-factors to generate a specific mRNA i.e. constitutive splicing, or sometimes many distinct mRNA isoforms i.e. alternative splicing, is still a poorly understood process. Progress into illuminating this process is further exacerbated by the variation of splicing in the multitude of tissues and cell types present, as well as the variation of cis and trans elements in different organisms, and the possibility that some alternative splicing events present in expressed sequence tag (EST) databases may constitute biochemical 'noise' or transient evolutionary fluctuations. Several studies, mainly computational in nature, addressing different questions regarding constitutive and alternative splicing are described here, ranging from improved modeling of splicing signals, studying the variation of alternative splicing in various tissues, analyzing evolutionary differences of cis and trans elements of splicing in various vertebrates, and utilizing attributes indicative of alternative splicing events conserved in human and mouse to identify novel alternatively spliced exons. In particular: (i) A general approach for improved modeling of short sequence motifs, based on the Maximum Entropy principle, that incorporates local adjacent and non-adjacent position dependencies is introduced, and applied to understanding splice site signals. The splice site recognition algorithm, MaxENTScan, performs better than previous models that utilize as input similar length sequences; (ii) The first large-scale bioinformatics study is conducted that identifies similarities and differences in candidate cis-regulatory elements and trans-acting splicing / by Gene W. Yeo. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 1994. / Includes bibliographical references. / by Todd Clayton Holmes. / Ph.D.

Thesis: Ph. D. in Neuroscience, Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2016. / Page 127 blank. Cataloged from PDF version of thesis. / Includes bibliographical references. / Loss of function mutations in the X-linked gene encoding for MeCP2 are the underlying genetic cause for Rett Syndrome (RTT), a devastating neurodevelopmental disorder that primarily affects girls. While the function of this transcriptional regulator remains elusive and complex, recent focus has turned to downstream signaling pathways as putative targets for novel therapeutics. The complexity of MeCP2 function is compounded by the heterogeneity of cell types in the brain, with recent evidence implicating glia cells in RTT pathophysiology. The focus of my thesis has been two-fold: exploring signaling mechanisms downstream of MeCP2 and the potential of IGF-1 as a therapeutic for RTT, and examining functional astrocyte sensory processing in healthy and impaired circuits. I present evidence that IGF-1 levels are reduced in mouse models of RTT and systemic treatment with IGF-1 leads to improvements in lifespan, respiratory patterns, and social behaviors. These effects are accompanied by increased synaptic proteins, activation of signaling pathways, and enhanced excitatory transmission, as well as effects on plasticity in visual cortex circuits. Astrocytes, known to contribute to synapse formation and maintenance, have been implicated alongside neurons as contributors to the RTT phenotype. They express the two most abundant glutamate transporters in the brain responsible for the majority of glutamate clearance from synapses. Indeed, lack of MeCP2 in astrocytes leads to a reduction in signaling pathways and aberrant glutamate transporter expression, with strong implications for synaptic and circuit activity. Efficient processing of visual information requires processing salient features while overcoming the inherent variability in neuronal networks. Natural movies evoke reliable responses from pyramidal neurons in visual cortex and my work reveals that discrete microdomain regions of visual cortex astrocytes also exhibit temporally reliable and spatially correlated responses to natural scenes. I show that glutamate transporters, which influence astrocytic Ca 2 signaling and synaptic transmission, regulate the reliability of astrocyte microdomain responses and thus contribute crucially to visual information processing. Finally, I show that in a RTT mouse model, astrocyte microdomains elicited by visual stimuli are reduced in size, consistent with the reduced synaptic transmission and neuronal responses observed in these mice. / by Rodrigo I. Garcia. / Ph. D. in Neuroscience

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2005. / Includes bibliographical references (p. 52-56). / The effects of attending to one or another of an object's attributes on neuronal representations of that object were investigated using extracellular recordings. A female rhesus monkey performed a delayed match to object attribute (DMSA) task, in which she alternately matched object orientations and object colors. In half of the task conditions, only one attribute matched the sample, forcing the animal to apply the current matching rule and ignore the irrelevant-attribute. Multiple simultaneous single-unit extracellular recordings were made in the ventrolateral prefrontal cortex (VLPFC) and anterior cingulate cortex (ACC) while the monkey performed the task. Neuronal selectivity for matching rule, object attributes, attribute relevance, response choice, and congruency were assessed using multi-factor ANOVAs. Attribute-selective responses were common in both cortical areas during the sample and delay periods, but were not significantly modulated by attribute relevance. There were few interactions between color-selective and orientation-selective responses according to the ANOVAs, suggesting that these attributes were represented independently. / (cont.) Significant effects of attribute relevance, response choice, and congruency appeared in both areas after the delay period, when the probe appeared onscreen. VLPFC cells were more active during incongruent and non-match conditions, when responses had to be suppressed. ACC cells were more active during congruent and match conditions, when active response suppression was not required. The results indicate that although prefrontal cortex often shows a bias for relevant information (Rainer et al, 1998), it may not do so if the task requires frequent alternation of attentional sets or active suppression of conflicting responses. The data also indicate that the VLPFC's role in managing attentional 'set' (Banich et al, 2000; Milham et al, 2001) is performed in conjunction with active stimulus comparison and response selection (e.g., Rushworth et al, 1997), not during working memory maintenance. The ACC may facilitate the reactivation of response tendencies that had been actively suppressed, possibly as part of a larger role in managing response conflict (Botvinick et al, 2004). / by Cynthia E. Kiddoo. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references. / People vary significantly across multiple cognitive domains, and that variation may be related to individual differences in the intrinsic functional architecture of the human brain. Such individual differences can be characterized via resting-state measurement of temporal synchrony between brain regions (i.e. functional connectivity). Here, I examine individual differences in working memory and mindfulness in relation to connectivity between the medial prefrontal cortex (MPFC) and dorsolateral prefrontal cortex (DLPFC). The DLPFC is a major component of the executive control network that increased in activation during tasks involving working memory and mental control. Young adults exhibit greater working memory capacity (how much information could be maintained and manipulated in mind) with greater MPFC-DLPFC negative correlation (whereas older adults show reduced working memory capacity and MPFC-DLPFC correlation relative to younger adults). Increased dispositional mindfulness (everyday awareness of the present moment) is associated with greater MPFC-DLPFC positive correlations, at rest. Thus, variation in two advantageous human abilities, working memory capacity and mindfulness, may be related to opposite patterns of internetwork functional connectivity (MPFC-DLPFC). The MPFC is a prominent component of the default-mode network (DMN), which typically decreases in activation during attention-demanding task requiring external focus. Given its relevance to mindfulness tendencies, mind-wandering (MW) episodes engage distributed regions within the DMN. I also investigate mind-wandering frequency using experience sampling with fMRI. MW transiently engages cortical regions associated with executive control and state-level mind wandering is reduced in those with greater trait-level mindfulness, as measured by resting-state functional connectivity. / by Joseph Barrington Keller. / Ph. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 81-93). / Developmental dyslexia is a neurological condition that specifically impairs the development of expert reading ability. Phonological processing deficits -- impaired representation of, or access to, the abstract units of spoken language -- have been implicated as the principal source of reading difficulties in dyslexia, independent of other cognitive factors. However, the source of these phonological impairments remains unknown: What mechanisms preclude development of the robust phonological representations critical for reading development? Experiments with phonological processing in dyslexia typically employ metalinguistic tasks that require explicit knowledge about phonological structure, failing to distinguish between access to representations and the representations themselves. Here I report a series of experiments that elucidate the nature of phonological impairments in dyslexia by examining the implicit processing of phonetic variability. Phonetic variability affects language processing at the interface between perceiving the physical speech signal and mapping it onto stored linguistic representations. This approach is well-suited to interrogate the integrity of phonological processing in dyslexia and to provide insight into how phonological representations may come to be impaired in this disorder. In Experiment 1, individuals with dyslexia demonstrated profoundly reduced ability to learn to use phonetic consistency in talker identification, thus reifying the status of phonological representations themselves as fundamentally impaired in this disorder. In Experiment 2, functional magnetic resonance imaging (fMRI) adaptation revealed reduced neural sensitivity to phonetic consistency during speech perception in individuals with dyslexia, indicating impaired rapid, implicit learning of phonetic-phonological consistency. The neural mechanisms that support such learning may be a specific instance of general brain mechanisms for adapting to stimulus consistency. In Experiment 3, fMRI adaptation further revealed that such exiguous neural plasticity in dyslexia is not limited to speech phonetics; instead, the core mechanisms of rapid adaptation to stimulus consistency appear to be dysfunctional in dyslexia, such that neural adaptation was reduced to all stimuli measured, whether auditory or visual, linguistic or non-linguistic. Deficits in neural adaptation may represent disruption of a core rapid plasticity mechanism for perceptual learning, dysfunction of which would impair the ability to develop the robust perceptual (phonological) representations critical to reading development. / by Tyler K. Perrachione. / Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 1999. / Includes bibliographical references (leaves 27-34). / by William Joseph Turkel. / S.M.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2010. / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Cataloged from student submitted PDF version of thesis. / Includes bibliographical references (p. 130-140). / We present an approach to solving computer vision problems in which the goal is to produce a high-dimensional, pixel-based interpretation of some aspect of the underlying structure of an image. Such tasks have traditionally been categorized as ''low-level vision'' problems, and examples include image denoising, boundary detection, and motion estimation. Our approach is characterized by two main elements, both of which represent a departure from previous work. The first is a focus on convolutional networks, a machine learning strategy that operates directly on an input image with no use of hand-designed features and employs many thousands of free parameters that are learned from data. Previous work in low-level vision has been largely focused on completely hand-designed algorithms or learning methods with a hand-designed feature space. We demonstrate that a learning approach with high model complexity, but zero prior knowledge about any specific image domain, can outperform existing techniques even in the challenging area of natural image processing. We also present results that establish how convolutional networks are closely related to Markov random fields (MRFs), a popular probabilistic approach to image analysis, but can in practice can achieve significantly greater model complexity. The second aspect of our approach is the use of domain specific cost functions and learning algorithms that reflect the structured nature of certain prediction problems in image analysis. / (cont.) In particular, we show how concepts from digital topology can be used in the context of boundary detection to both evaluate and optimize the high-order property of topological accuracy. We demonstrate that these techniques can significantly improve the machine learning approach and outperform state of the art boundary detection and segmentation methods. Throughout our work we maintain a special interest and focus on application of our methods to connectomics, an emerging scientific discipline that seeks high-throughput methods for recovering neural connectivity data from brains. This application requires solving low-level image analysis problems on a tera-voxel or peta-voxel scale, and therefore represents an extremely challenging and exciting arena for the development of computer vision methods. / by Viren Jain. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2009. / "June 2009." / Includes bibliographical references (leaves 127-150). / Modification of synapses by neural activity has been proposed to be the substrate for experience-dependent brain development, learning, and recovery of function after brain damage. In the visual cortex, the strength of cortical synapses can be bidiredionally modified, where in response to a critical level of postsynaptic activation, synapses are strengthened (long-term potentiation; LTP) and below this level, synapses are weakened (long-term depression; LTD). Previous work in visual cortex has suggested that the threshold for synaptic modifications is dependent on the recent history of visual experience, a phenomenon called metaplaticity. Recent mechanistic studies have shown that experience-dependent adjustments of the modification threshold correlate with changes in the subunit composition and function of NMDA-type glutamate receptors (NMDARs). However, causality has not been conclusively established. Here we examined the mechanistic basis of metaplaticity, and specifically how this process is mediated by a switch in NMDAR subunit composition by focusing on the NR2A subunit of the NMDA receptor in visual cortex. We provide evidence for the functional significance of the NR2A subunit in metaplastic changes both in synaptic platicity elicited in vitro and in naturally-occurring platicity in vivo. We also performed a comparison of in vitro methods of inducing plasticity and those which subserve in vivo experience-dependent changes in synaptic strength. These findings represent an important step forward in understanding how plasticity thresholds are regulated in the brain. / by Kathleen K. A. Cho. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2014. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 122-138). / The invariance hypothesis is the claim that properties of the ventral stream can be deduced from 1. a consideration of the computational problem with which it is engaged: invariant object recognition, 2. constraints of the neural "hardare", and 3. the visual environment. We are interested here in a particular instantiation of this idea. A recent general theory of hierarchical networks for invariant recognition [1] describes many modern convolutional networks as special cases, and also implies the existence of a wider class of algorithms, which we are only now beginning to explore. Our version of the invariance hypothesis is the claim that the algorithm implemented by the ventral stream is also in this class. As applied to the brain, the theory follows from a few simple and commonly accepted premises. This thesis contributes several models/studies in which properties of the ventral stream are deduced and explained in the context of the theory. The main contribution here is providing a general framework through which disparate results concerning many parts of the ventral stream, and even different levels of analysis [2], can be bridged and understood. In that sense, it is primarily a Neuroscience contribution. However, the ideas and algorithms it suggests may also have implications for the broader question of how to learn representations capable of supporting intelligence. / by Joel Zaidspiner Leibo. / Ph. D.