The question of how we form memories has fascinated scientists for decades. The hippocampus and surrounding medial-temporal-lobe (MTL) structures are critical for both memory and spatial navigation, yet we do not fully understand the neuronal representations used to support these behaviors. Much research has examined how the MTL neurally represents spatial information, such as with “place cells” that represent an animal’s current location or “head-direction cells” that code for an animal’s current heading. In addition to attending to current spatial locations, navigating to remote destinations is a common part of daily life. In this dissertation I investigate how the human MTL represents the relevant information in a goal-directed spatial-memory task. Specifically, I analyze single-neuron and local field potential (LFP) data from neurosurgical patients with respect to their spatial navigation and memory behavior, with a focus on probing the link between neuronal firing, oscillations, and memory. In Chapter 2, I find that the firing rates of many MTL neurons during navigation significantly change depending on the position of the current spatial target. In addition, I observe neurons whose firing rates during navigation are tuned to specific heading directions in the environment, and others whose activity changes depending on the timing within the trial. By showing that neurons in our task represent remote locations rather than the subject’s own position, my results suggest that the human MTL can represent remote spatial information according to task demands. In Chapter 3, I find that during encoding the left hippocampus exhibits greater low theta power for subsequently recalled items compared to unrecalled items. I also find that high frequency activity and neuronal firing in the hippocampus distinguish between item-filled compared to empty chests. Finally, I find that MTL cells’ firing rates and the differential timing of spikes relative to low frequency oscillations in the LFP distinguish between subsequent recall conditions. These results provide evidence for a distinct processing state during the encoding of successful spatial memory in the human MTL. Overall, in this thesis I show new aspects of the neural code for spatial memories, and how the human MTL supports these representations.

No description available.

Some of the most important discoveries in cognitive neuroscience have come from recent innovations in experimental tools. Computational models that simulate human perception of environmental inputs have revealed the internal processes and features by which those inputs are learned and represented by the brain. We advance this line of work across two separate research studies in which we leveraged these models to both generate experimental task stimuli and make predictions about behavioral and neural responses to those stimuli. Chapter 1 details how nine language models were used to generate controversial sentence pairs for which two of the models disagreed about which sentence is more likely to occur. Human judgments about these sentence pairs were collected and compared to model preferences in order to identify model-specific pitfalls and provide a behavioral performance benchmark for future research. We found that transformer models GPT-2, RoBERTa and ELECTRA were most aligned with human judgments. Chapter 2 utilizes the GloVe model of semantic word vectors to generate a set of schematically structured poems comprising ten different topics whose specific temporal order was learned by a group of participants. The GloVe model was then used to investigate learning-induced changes in the spatial geometry of the representations of the topics across the cortex. A Hidden Markov Model was also used to measure neural event segmentation during poem listening. In both analyses we identified a consistent topography of learning-induced changes in the default mode network, which could be partially explained by the models.

The entorhinal cortex (EC) is central to the brain’s navigation system. Its subregions are conventionally thought to compute dichotomous representations for spatial processing: medial entorhinal cortex (MEC) provides a global spatial map, while lateral entorhinal cortex (LEC) encodes specific sensory details of experience. While local recordings of EC circuits have amassed a vast catalogue of specialized cell types that could support navigational computations in the brain, we have little direct evidence for how these signals are actually transmitted outside of the EC to its primary downstream reader, the hippocampus, which itself is critical for the formation of spatial and episodic memories. Here we exploit in vivo sub-cellular imaging to directly record from EC axon terminals as they locally innervate hippocampal area CA1, while mice performed navigation and spatial learning tasks in virtual reality. We find both distinct and overlapping representations of task, location, and context in both MEC and LEC axons. While MEC transmitted a highly location- and context-specific code, LEC inputs were strongly biased by ongoing navigational goals and reward. Surprisingly, the position of the animal could be accurately decoded from either entorhinal subregion. Our results challenge prevailing dogma on the routing of spatial and non-spatial information from the cortex to the hippocampus, indicating that cortical interactions upstream of the hippocampus are critical for combining these processing streams to support navigation and memory.

This research was undertaken to determine the effect that computer aided design (CAD) had on the spatial visualization abilities of selected gifted adolescents. The following hypotheses was tested: Subjects receiving instruction in CAD will show improvement in spatial visualization ability, as measured by the Revised Minnesota Paper Form Board (RMPFB) test. when compared to the subjects not receiving CAD instruction. The experimental group consisted of 20 students enrolled in the CAD course offered in the 1991 Virginia Governor's School of Technology. The control group consisted of 20 Governor's School students not enrolled in the CAD course. Both groups were pretested using the RMPFB test Form AA to measure entry level spatial visualization. A treatment consisting of three weeks of CAD instruction using CADKEY 3.5 was given to the experimental group. Following the treatment both groups were post tested using the RMPFB test Form BB to determine their existing level of spatial visualization ability. The nonequivalent control group design was used in this study since the experimental group was an intact group and therefore not randomly assigned. ANCOVA statistical analysis was used to determine if there was statistical significance of the post test scores. / Ph. D.

Accounting education research has explained some variation in student performance by aptitude, attitude, and experience variables, as well as gender. The unexplained portion of variance, however, suggests the existence of unidentified variables. This study examines the relationship of visual-spatial aptitude (VSA) to student completion and continuing behaviors and to performance in four accounting courses. VSA is a group of cognitive abilities which facilitate building mental representations and solving problems and which are positively related to performance in mathematics and science courses. This study hypothesizes that high VSA students will complete Accounting Principles I at a higher rate, continue to Principles II at a higher rate, and perform better in Principles I than will low VSA students. Also hypothesized are gender differences and course differences in the relationship between VSA and performance. Subjects are students tracked in accounting courses for three semesters. Independent variables include gender, prior bookkeeping coursework, and major, as well as SAT scores, GPA, and scores on two VSA tests--the MAP Planning Test (MAP) and the Mental Rotations Test (MRT). MAP and MRT measures include the number right, the number wrong, and the percentage right. Dependent variables include student completion and continuing status as well as performance scores in each of four courses. Results indicate that high VSA subjects have a higher completion rate than do low VSA subjects and that completers of Principles I have higher VSA than do droppers. Also, continuers to Principles II have higher VSA than do non continuers. Results indicate relationships between VSA and Principles I scores. Subjects with high MRT percentages score higher on exams and lower on homework/quizzes than do those with low MRT percentages. Subjects with few MRT wrong have higher exam scores than do those with more MRT wrong. MAP is related only to Principles II and Intermediate, while MRT is related only to Principles I, II, and Cost. Relationships of VSA to exams are positive and frequent. Relationships to homework/quizzes are negative and less frequent. Computerized practice set scores are rarely related to VSA. In separate analyses of students taking Principles I, VSA is related to homework and exams for females and only to exams for males. For the smaller sample of students continuing to Principles II, models which contain general aptitude covariates indicate that VSA is only related to female performance and only in Principles I. / Ph. D.

The purpose of this research was to investigate hypothesized relations of visuospatial and logical reasoning skills, and span of short-term memory to achievement in geometry. In addition, major subfactors of visuospatial ability (visualization, speeded rotations, spatial orientation, and disembedding) were assessed to determine which were significant predictors of geometry achievement. Vernon's (1965) model of intelligence and Baddeley's model of working memory provided the theoretical framework for these hypotheses. Subjects (N = 110) were students in seven sophomore level geometry classes in two schools in southwest Virginia. Cognitive measures of speeded rotations, visualization, spatial orientation, disembedding, Gestalt closure, logical reasoning, and short-term memory span were administered. Two measures of geometry achievement were used: The standardized New York Regents Geometry Exam, and z-transformations of the classroom final grade. A model of geometry achievement is proposed and major predictions of the model were supported. within this sample, regression analysis showed the measures of visualization, logical reasoning, and short-term memory predicted achievement on the New York Regents Geometry Exam. Separate regression analyses for each gender revealed visualization predicted geometry achievement for the girls, while logical reasoning and short-term memory span predicted geometry achievement for the boys. Gender differences favoring boys were found on measures of speeded rotations, spatial orientation, and Gestalt closure. Girls had significantly higher scores on the measure of short-term memory span and the classroom measure of geometry achievement. / Master of Science

The fundamental premise for this project was the desire to explore one of my fascinations with "movement" of space and “movement” of structure. This was somewhat a reaction to “static” boxed structures which have pervaded our eastern cities for the last few decades, and moreover, a want to inspire life into a fixed object. The work presented was a process of discovery, the most enlightening being that of "contrast". It became the key to achieve meaning My original intentions combined with changes experienced over time have resulted in a building, but better, in a transformation and another new place from which to begin. / Master of Architecture

Decision Support Systems (DSS) should increase the effectiveness of a decision and the efficiency of the decision making process. The success of DSSs has varied among individual users. One explanation for this variation is that individual’s spatial ability has a moderating effect on performance. Another factor found to impact decision performance is the structure of the task. The purpose of this study is to determine whether spatial ability factors have a moderating effect on decision making performance in a DSS environment under differing task structure. Three of the major factors of spatial ability (spatial scanning, spatial relations, and field independence) and two levels of task structure (moderately complex and complex) are considered in the study. Spatial scanning and field independence were assessed by the Kit Factor-Reference Cognitive Tests, Map Planning Test and Hidden Figures Test, respectively [Ekstrom et al., 1976]. Spatial relations was assessed by the Mental Rotations Test [Vanderberg & Kuse, 1978]. Model formulation and data analysis are two stages of decision making considered in the study. Decision making performance is assessed by time to complete the task, DSS features used, decision confidence, and decision accuracy. Computer experience is treated as a control variable. Fifty Master level students in the School of Business attended three experimental sessions which involved completing several spatial ability tests, participating in a ninety minute lecture on the software package, and completing four practice problems and two experimental cases. The software package employed in the experiment is IFPS, a DSS generator. The results of the study indicate that spatial relations has a moderating effect on decision confidence during the data analysis stage of decision making. Furthermore, there is a significant moderating interaction effect between spatial relations and task complexity when performance is assessed by decision confidence. Spatial relations is found to be more significantly related to performance in the complex case than in the moderately complex case. / Ph. D.

Over the past several years, machine learning has enabled incredible progress on many tasks, such as mastering board games, recognizing objects, conversing in natural language, and generating images or videos. Despite these accomplishments, state-of-the-art techniques in artificial intelligence lack the foundations necessary to flexibly and robustly understand and manipulate their three-dimensional spatial surroundings. For instance, before their second birthday, children learn that objects persist during occlusion, they know how containment works, and they are surprised by novel physics. In contrast, a true notion of object permanence has remained elusive for computer vision, despite its vitality in perceiving and interacting with everyday situations. In this thesis, I will outline my work on enhancing spatial reasoning within dynamic scenes, where I have integrated machine learning, intuitive physics, geometry, and world knowledge to create powerful frameworks that can capture, represent, and generate their complex, cluttered visual environment. Specifically, I will present models to reconstruct 4D scenes, track objects through occlusions, and perform dynamic view synthesis, all from a single camera viewpoint, and often successfully generalizing to real-world settings. These capabilities are pivotal for applications in embodied intelligence (such as robotics and self-driving), content creation and editing, or augmented and mixed reality, where machines need to accurately represent their surroundings and deeply understand how they evolve over time.