Thesis: Ph. D., Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2017. / Cataloged from PDF version of thesis. "September 2016." / Includes bibliographical references (pages 115-118). / For the recognition of general objects in natural scenes, the current top-performing computer vision models owe a debt to visual neuroscience. The hierarchical architecture of convolutional networks, and related models such as HMAX, mimics that of the ventral stream of visual cortex. In essence, they apply the model of Hubel and Wiesel recursively, alternating layers of 'simple' cells, which are tuned to certain local features, and 'complex' cells, which pool the outputs of simple cells within a local region. With recent advances in deep learning, for many tasks in vision and speech, emphasis has moved away from so-called 'hand-designed' models and toward big data and high throughput computing, with models learning from millions of labeled examples. Yet CNNs only learn their features - the weights of connections in the network. All other aspects of the network (size, connectivity, response functions, etc.) are unlearned architectural choices made by their designers. Vision has not yet been reduced to a pure learning problem - human insight into the nature of visual problems continues to be important. To design a good vision system, one still has to understand vision. And, as evidenced by performance for many complex visual tasks, natural vision systems still 'understand' vision better than we do; there is still much to be learned from them. Our work is based on the HMAX model, which places greater weight on biological realism. Our goals are threefold: to better understand the ventral stream algorithm, as well as the visual problem it solves, and to improve the performance of artificial vision systems. In this work we take two main approaches. i-theory is an ongoing effort to explain the good performance of hierarchical models in terms of a formal theory of invariance to transformations. We provide a reinterpretation of V1 simple and complex cells in the context of i-theory as computing a high-dimensional, locally translation-invariant signature for the contents of a V1 receptive field. We describe a simple algorithm for learning them which can extend without modification to the learning of higher-order representations for V2 and beyond. The algorithm yields model V1 cells having a good fit to data from several animal species. We also demonstrate that a precondition of i-theory, covariance, can hold in upper layers, even for transformations not anticipated in the training of lower layers. No current hierarchical object recognition model incorporates realistic retinal resolution. Incorporating this detail forces a reevaluation of the role of the ventral stream's feedforward core in the larger task of scene understanding as well as many details of the model itself, particularly with respect to scale. We investigate the optimal shape of the input window used to select a subset of the visual information available in a scene for processing in a single feedforward pass, defined as a region in (x, y, A), the handling of the A dimension within the hierarchy, and the problem of clutter. Our main experimental results are (1) spatial wavelengths too small for the retina to perceive across the entire object do not play a significant role in the no-clutter case, but confer robustness in the presence of clutter, and (2) preservation by the hierarchy of information about the relative scale (distance along A) of feature activations is more important than current models reflect. / by Jim Mutch. / Ph. D.

Thesis: Ph. D. in Neuroscience, Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2016. / Cataloged from PDF version of thesis. Page 114 blank. / Includes bibliographical references (pages 107-113). / While animals tend to prefer immediate rewards to delayed ones [1], delayed gratification is often advantageous [2]. Appropriate choice about future rewards is critical for survival. The dorsal raphe serotonergic neurons have been long implicated in the control of temporal discounting of reward [3] [4], but it is not clear whether their activities in fact direct the decision making process. In this thesis, I designed a cued intertemporal choice task for mice that allows the combination of highly specific genetic manipulations with sophisticated behavioral interrogations. The task utilizes odors to communicate upcoming reward contingencies to the mouse subjects. I found that optogenetically augmenting or silencing the activities of dorsal raphe serotonergic neurons precisely at decision epochs resulted in an increase or a reduction in the choice for the delayed and larger reward, respectively. These manipulations do not alter the subjects' choice in trials involving immediate rewards, suggesting that serotonin might only be important for conditions in which difficult trade-offs are required. I also demonstrated that the nucleus accumbens, a major component of the mesolimbic reward pathway, is a possible downstream target of the aforementioned serotonin action. Taken together, these results show that serotonergic neurons regulate inter-temporal choice behavior bidirectionally, possibly through actions in nucleus accumbens. / by Sangyu Xu. / Ph. D. in Neuroscience

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 158-173). / Neuropsychiatric Disorders are the leading category contributing to disability-adjusted life years (DALYs) in the U.S. according to the World Health Organization. These findings underline the vast burden caused by neuropsychiatric disorders on patients. However, effective treatments do not exist for many of the neuropsychiatric disorders mostly due to lack of understanding of disease pathology. Evidence from whole genome sequencing of psychiatric disorder patients increasingly suggest that Wnt signaling and cortical development - in addition to other perturbations - may underlie the pathophysiology of multiple disorders. Furthermore, besides autism spectrum disorder, contribution of neurodevelopmental dysregulations to disease etiology in late-onset disorder such as schizophrenia are becoming widely accepted. Therefore, a better understanding of cortical development and functions of Wnt signaling could prove critical in determining the cellular and molecular mechanisms underlying the causes of psychiatric disorders. The work presented in this thesis aims to understand the functions of multiple neuropsychiatric disorder risk genes in brain development, and the converging role of Wnt signaling in neurodevelopment. First, we determined ASD risk gene Chd8 to be a positive regulator neural progenitor proliferation in the developing mouse brain through its transcriptional regulation of cell cycle and Wnt signaling genes. Surprisingly, Chd8 exhibits a cell type-specific modulation of Wnt signaling. Furthermore, knockdown of Chd8 in the upper cortical layer neurons caused ASD-related behavioral abnormalities in adult mice, which could be rescued via induction of Wnt signaling. Secondly, we made the novel observation that bipolar disorder risk gene Ank3 (ankyrin-G) plays a crucial role in cortical neurogenesis through regulation of subcellular localization of [beta]-catenin, which is an essential component of Wnt signaling. Finally, the effects of brain-specific deletion of BcI9 on brain development and behavior were characterized using a heterozygous BcI9 deletion transgenic mouse line. Behavioral and brain development defects associated with BcI9 were shown to mimic some of the clinical symptoms observed in patients. Collectively, our results demonstrate a central role for Wnt signaling and cortical development in pathophysiology of neurodevelopmental and neuropsychiatric disorders. / by Omer Durak. / Ph. D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 37-42). / Chronic stress has been linked to variation in gene regulation in the hippocampus (HIP) among other areas. These lead to cytoskeletal and volumetric rearrangements in various nuclei of the central nervous system and are thought to contribute to several stress-sensitive disorders. One such gene that has been shown to be downregulated in HIP in response to stress is somatotropin, colloquially known as growth hormone (GH). These experiments were conducted to develop a novel assay for examination of working memory in rats and explore the nature of stress-induced impairment of hippocampal function and determine whether infusion of a modified herpes simplex virus (HSV) carrying the recombinant rodent growth hormone (GH) would be sufficient to restore normal hippocampal function. After 21 days of chronic immobilization stress (CIS), animals received bilateral infusions into the dorsal HIP of 2[mu]l HSV carrying either GH with green florescent protein (GFP) or GFP only. On the second day following the infusion, the animals received trace conditioning, a HIP-dependent task, with five tone-shock pairings of a 16 second tone followed by a 30 second trace interval terminating with a 1 second 0.85 milliamp footshock. An inter-trial interval of 3 minutes was used to separate the tone-shock pairings. The following day the animals were tested for fear to the context and for fear to the tone in a novel context, measured by amount of time the animal spent freezing. Using this criterion, animals that had undergone stress that received the control vector were less likely to freeze when presented with the tone, indicating an impairment of hippocampal function. Viral-mediated overexpression of GH in the dorsal HIP was able to reverse the CIS-related impairment in hippocampal function. ELISA was used to verify the expression of GH from the infused vector. These experiments may yield future directions of investigation for stress-based disorders. / by Christopher M. Saenz. / S.M.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2004. / Includes bibliographical references (leaves 99-112). / Synaptic plasticity is the rearrangement of neuronal connections that likely underlies learning and memory. It requires the expression of a set of genes essential for the synaptic changes that occur during plasticity, candidate plasticity gene 2 (cpg2) was isolated in a screen for genes that effect synaptic plasticity. In this thesis, I analyze the regulation and function of cpg2 in neurons. I find that cpg2 is a splice-variant of the syne-1 gene that is expressed only in brain regions capable of plasticity and encodes a protein specifically localized to a postsynaptic endocytic zone of excitatory synapses, often in the vicinity of clathrin-coated pits. I further show that, through its C-terminal coiled-coil motifs, CPG2 binds to the actin cytoskeleton and to endophilin B2, a member of a family of proteins involved in membrane trafficking. RNAi-mediated knock-down of CPG2 increased the number of postsynaptic clathrin-coated vesicles, some of which trafficked NMDA receptors, and disrupted the internalization of glutamate receptors. In addition, alterations in its protein levels affected dendritic spine size, supporting a role for CPG2 in regulating membrane trafficking. These data suggest that CPG2 organizes a network of proteins at the postsynaptic endocytic zone critical for glutamate receptor internalization. Due to its unique expression profile and subcellular localization, CPG2 may underlie a novel adaptation of the clathrin-mediated endocytosis pathway that enables the capacity for postsynaptic plasticity in excitatory synapses. / by Jeffrey Richard Cottrell. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 297-205). / This thesis introduces EMRE, an expectation-based model of referring expressions. EMRE is proposed as a model of non-syntactic dependencies - in particular, discourse-level semantic dependencies that bridge sentence gaps. These include but are not limited to anaphora (references to noun phrases in previous sentences) and coherence predicates such as causality, temporal ordering and resemblance -- two domains that have typically been treated as entirely distinct aspects of language. EMRE is a computational-level model, and is agnostic about any particular algorithms, cognitive faculties, or neurological substrates that might be applied to the problem of semantic reference. Instead, it describes reference as a computational problem framed in terms of expectation and inference, and describes a solution to the problem based on rational top-down expectations about the likely targets of referring expressions, and on bottom-up feature-based matching that occurs when a referring expression is encountered. EMRE is used to derive novel empirical predictions about how people will construe particular discourse constructions involving NP anaphora and coherence predicates. These predictions are tested in controlled behavioral experiments, in which participants read and answer questions about short texts. The results of these experiments are shown to be consistent with a model of reference as an expectation-based computational structure with different underlying rules than those governing syntactic processing. / by John Kræmer. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2007. / Includes bibliographical references (leaves 98-102). / This thesis contains three sets of studies designed to explore the identification and function of prosodic features in English. The first set of studies explores the identification of prosodic features using prosodic annotation. We compared inter-rater agreement for two current prosodic annotation schemes, ToBI (Silverman, et al., 1992) and RaP (Dilley & Brown, 2005) which provide guidelines for the identification of English prosodic features. The studies described here survey inter-rater agreement for both novice and expert raters in both systems, and for both spontaneous and read speech. The results indicate high agreement for both systems on binary classification, but only moderate agreement for categories with more than two levels. The second section explores an aspect of the function of prosody in determining the propositional content of a sentence by investigating the relationship between syntactic structure and intonational phrasing. The first study tests and refines a model designed to predict the intonational phrasing of a sentence given the syntactic structure. In further analysis, we demonstrate that specific acoustic cues-word duration and the presence of silence after a word, can give rise to the perception of intonational boundaries. The final set of experiments explores the relationship between prosody and information structure, and how this relationship is realized acoustically. In a series of four experiments, we manipulated the information status of elements of declarative sentences by varying the questions that preceded those sentences. We found that all of the acoustic features we tested-duration, f0, and intensity-were utilized by speakers to indicate the location of an accented element. However, speakers did not consistently indicate differences in information status type (wide focus, new information, contrastive information) with the acoustic features we investigated. / by Mara E. Breen. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 1992. / Includes bibliographical references (leaves 156-164). / by Thomas M. Breuel. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 2007. / Includes bibliographical references (leaves 139-145). / This thesis examines two questions concerning the working memory system underlying language processing: (1) To what extent is the working memory system underlying language processing domain-specific? and (2) What is the relationship between the working memory system and the long-term memory system in language processing? In Chapter 1, I describe ten experiments investigating the extent to which the working memory system underlying linguistic integrations is domain-specific. I argue that the results of these experiments demonstrate that at least some aspects of the working memory system used for linguistic integrations are not domain-specific, being involved in arithmetic, and possibly, musical processing. In Chapter 2, I describe six experiments investigating the relationship between the two retrieval operations that are required when an incoming word is integrated into an evolving structure: the retrieval of the lexical properties of the word from long-term memory and the retrieval of its structural dependents from working memory. I demonstrate that the relative ease or difficulty of retrieving the lexical properties of an incoming word affect the difficulty of retrieving its structural dependents. I therefore argue that the two retrieval operations rely on overlapping pools of resources. / by Evelina G. Fedorenko. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Brain and Cognitive Sciences, 1993. / Includes bibliographical references. / by Steven A. Farber. / Ph.D.