skip to content

Wagner, Anthony David

Overview
Works: 23 works in 24 publications in 1 language and 31 library holdings
Roles: Thesis advisor, Author
Publication Timeline
Key
Publications about Anthony David Wagner
Publications by Anthony David Wagner
Most widely held works by Anthony David Wagner
Prediction and novelty in the human medial temporal lobe by Janice Chen( file )
1 edition published in 2011 in English and held by 2 libraries worldwide
Prediction is at the core of memory. A memory system stores information about the past in service of preparation for the future, and thus the act of memory retrieval may be viewed as an act of prediction about upcoming events. Converging evidence from animal and compuational models suggests that within the hippocampus, stored memories are compared to current sensory input in order to compute novelty -- when expectation deviates from actual outcome. Specifically, hippocampal subfield CA1 is thought to support this computation of mismatch between past and present. Detection of novelty in turn is hypothesized to modulate encoding processes, providing a mechanism for gating the entry of information into long-term memory. Using high-resolution functional MRI, I examined hippocampal subfield (CA1, CA23/DG[dentate gyrus], subiculum) and medial temporal lobe (MTL: entorhinal, perirhinal, parahippocampal) cortical activation during associative retrieval and associative mismatches in humans. In Experiment 1, subjects performed cued image retrieval and made explicit comparisons of memory to matching or mismatching decision probes. Activity in multiple hippocampal and MTL cortical subregions tracked associative retrieval success, whereas activity in CA1 and perirhinal cortex tracked the presence of associative mismatches. In Experiment 2, subjects viewed sequences of images while performing an incidental task (1-back target detection). In CA1, CA23/DG, and perirhinal cortex, activation was greater when image sequences were presented in rearranged order (mismatch with memory) compared to repeated order (match with memory). In CA1 only, this mismatch enhancement was significantly modulated by prediction strength: the mismatch enhancement was greater when predictions were stronger. In a separate behavioral experiment, recognition memory was found to be better for images that had appeared in rearranged-order (mismatch) than repeated-order (match) sequences, supporting the notion that mismatch detection leads to encoding upregulation. In Experiment 3, subjects viewed sequences of images containing either temporal predicitive information only or combined temporal and spatial predictive information. When novel images were embedded within previously viewed sequences, activation in CA1 was significantly related to subsequent memory for items which had violated predictions (remembered> forgotten), with this enhancement modulated by the amount of predictive information. Together, these data are consistent with the hypothesis that CA1 acts as a comparator, detecting when memory for the past and sensory input in the present diverge. More broadly, the current studies reveal the dynamics of the human hippocampus and MTL cortices during the acts of prediction, novelty detection, and memory encoding -- a continuous cycle of events that enable preparation for the future based on past experience
Visual cortical circuitry for building word representations by Andreas Maximilian Rauschecker( file )
1 edition published in 2011 in English and held by 2 libraries worldwide
Reading is the remarkable human ability to decode the sounds and meaning of language from an intricate combination of strokes of ink. This perceptual capacity is tolerant to changes in size, font, and other visual features of text. Transformations of the neural representation between many neural modules, including primary visual cortex (V1), other retinotopically organized areas, and language regions, are necessary for successful reading. An important focus of previous work has been on characterizing word representations in ventral occipito-temporal cortex, in particular in a left-hemisphere region known as the visual word form area (VWFA). To better understand the transformations occurring between V1 and the VWFA, we measured (using functional magnetic resonance imaging, fMRI) and perturbed (using transcranial magnetic stimulation) neural responses in several visual areas while subjects read words defined by atypical visual features (Chapter 2). We show that VWFA responses are invariant to the visual features that define word stimuli, and we show how flexible neural circuitry accounts for these abstract representations. While these studies contribute to understanding the VWFA's inputs, the inner organization of the VWFA remains unexplored. We describe experiments that use fMRI and electrocorticography (ECoG) in the human brain to show that the VWFA is sensitive to visual field position, and that together with a homologous right-hemisphere region, its inner organization encompasses a retinotopic map (Chapter 3). Information about abstract word forms is in turn transferred from the VWFA to language areas of the brain. This transfer occurs via large white-matter bundles that can be measured with diffusion imaging (DTI). We compare data from a severely dyslexic individual to a group of control subjects and show that the individual's deficits are due to a missing arcuate fasciculus, one of the major pathways important for reading and language (Chapter 4). This set of experiments ties together several fields of neuroscientific inquiry, including early visual processing, complex visual object representations, and language. Ultimately, if we are to understand how spots of light are transformed into sounds and meaning, we must unravel the smaller transformations that occur within all these components of the reading circuitry
Functional heterogeneity in parietal cortex during episodic memory retrieval by John Benjamin Hutchinson( file )
1 edition published in 2011 in English and held by 2 libraries worldwide
The memory for specific events and instances in time--episodic memory--is a central aspect of human cognition. It has long been thought that episodic memory is supported by a distributed network of regions in the brain, including frontal cortex and the medial temporal lobes. A growing body of neuroimaging evidence in the last several years has additionally suggested posterior parietal cortex (PPC) involvement during episodic retrieval. Several accounts have argued that the region's involvement reflects the engagement of two specific attentional processes, goal-driven and reflexive attention, distinctly residing in dorsal and ventral regions of PPC, respectively. Strict interpretations of such accounts, however, have recently been called into question, as attention- and memory-related processes might not occupy overlapping subregions of PPC. The research described here systematically assesses the anatomical and functional relationship between attention- and memory-related operations in PPC in a series of three experiments. The first experiment revealed four distinct patterns of activity across left lateral PPC--two thought to relate to mnemonic processes and two thought to relate to attention engaged during retrieval. The second experiment directly tested whether a region thought to index attention during retrieval from the first experiment was recruited in an independent attention task, with results suggesting that attention-sensitive regions are indeed engaged in a systematic manner during retrieval. The final experiment further probed the interaction between memory, attention, and decision-making processes engaged during retrieval, and replicated and extended the findings from the first experiment. Collectively, the findings from the research provide a more detailed depiction of how anatomically and functionally separable regions of PPC support distinct mechanisms of memory and attention
Resting networks in primary insomnia by Michael Cunyuan Chen( file )
1 edition published in 2012 in English and held by 2 libraries worldwide
Insomnia is a prevalent and costly disorder of sleep-related distress, yet little is known about its etiology. To better understand the neurobiology underlying insomnia, we examined resting state and directed sleep brain activity in insomniacs and healthy controls using simultaneous blood oxygen level-dependent (BOLD) signal functional magnetic resonance imaging (fMRI) and electroencephalography (EEG). Using dual regression analysis of BOLD signal brain networks derived from independent component analysis, we found increased coactivation of the insula in salience networks in insomniacs compared to healthy controls. This increase was accompanied by altered EEG power in insomniacs compared to healthy controls, as well as altered BOLD connectivity signatures of EEG frequency bands. These results suggest that aberrant connectivity of the insula and salience networks contribute to neural dysfunction in insomnia
The role of the frontal eye field in gating and maintaining object signals in short-term memory by Kelsey Lynne Clark( file )
1 edition published in 2012 in English and held by 2 libraries worldwide
Spatial attention is known to gate entry into short-term memory, and some evidence suggests that spatial signals may also play a role in binding features or protecting object representations during memory maintenance. To examine a potential role for spatial signals in maintaining object short-term memory, the activity of neurons in the Frontal Eye Field (FEF) of macaque monkeys was recorded during an object-based delayed match-to-sample (DMS) task. In this task monkeys were trained to remember an object identity over a brief delay, irrespective of the locations of the sample or target presentation. FEF neurons exhibited visual, delay, and target period activity, including selectivity for sample location and target location. Delay period activity represented the sample location throughout the delay, despite the irrelevance of spatial information for successful task completion. Furthermore, neurons continued to encode sample position in a variant of the task in which the matching stimulus never appeared in their response field. FEF neurons also exhibited target-position-dependent anticipatory activity immediately prior to target onset, suggesting that the monkeys can predict target position within blocks. These results show that FEF neurons maintain spatial information during short-term memory, even when that information is irrelevant for task performance. Despite the robust delay period activity we observed in FEF during the DMS task, we found little further evidence to support the theory that this activity contributes to object memory maintenance. Noise correlations were present between pairs of simultaneously recorded FEF and IT neurons during the sample and early delay periods, but did not persist into the second half of the delay period, despite the continued elevation of firing rates in both regions throughout the delay. The most direct method of assessing the contribution of the FEF delay period activity observed during the DMS task to object memory was the pharmacological elimination of that activity and evaluation of the impact on task performance. Inactivation of FEF with muscimol produced spatially localized deficits on the memory guided saccade task, but did not selectively impair object memory performance for sample stimuli appearing in the mnemonic scotoma
Incentive processing in the aging brain : individual differences in value-based learning and decision making across the adult life span by Gregory Russell Samanez Larkin( file )
1 edition published in 2010 in English and held by 2 libraries worldwide
As the proportion of older adults continues to grow rapidly here in the U.S. and across the globe, aging adults may be required to make increasingly more independent health-related and financial decisions. Thus, it is increasingly imperative to better understand the impact of age-related psychological changes on decision making. Although a growing body of research has linked age-related deficits in attention, memory, and cognitive control to changes in medial temporal and lateral prefrontal cortical function, remarkably little research has investigated the influence of aging on valuation and associated mesolimbic function in the striatum and medial prefrontal cortex. Likewise, theoretical accounts link age-related declines in a number of basic cognitive abilities to dopamine function, but research has largely neglected age differences in value-based learning and decision making which also rely on the dopamine system. Recent findings reveal age-related declines in the structure of striatal and medial frontal circuits, however it was not previously clear whether these structural declines contribute to functional deficits in incentive processing. Thus, the seven experiments presented here explored potential age differences across a range of value-based tasks from basic anticipatory and consummatory responses to reward cues (Experiments 1--2) to probabilistic value-based learning (Experiments 2--5) to investment decision making (Experiments 6--7). The studies focus on both age-related and non-age-related individual differences in learning and decision making across the adult life span. Overall, three sets of key findings emerge. The first set of experiments on anticipatory affect reveal evidence for an age-related asymmetry in the anticipation of monetary gains and losses, such that older adults appear less sensitive to the prospect of financial loss than younger adults. In a subset of adults, this anticipatory affective bias contributes to loss avoidance learning impairments through the sensitivity of the anterior insula. Thus, although a relative lack of anxiety about potential loss may contribute to increased well-being, this asymmetry may put individuals with blunted loss anticipation at risk for certain types of financial mistakes. In fact, we show that individuals who perform poorly on the laboratory-based loss avoidance learning task accrue more financial debt in the real world. The second set of experiments focus on age differences in value-based learning and reveal that although older adults show intact neural representation of the actual value of reward outcomes, there is an age-related decline in the neural representation of prediction error at outcome in the striatum and medial prefrontal cortex. Age differences in learning are magnified when choice set size is increased, but when the number of trials is extended older adults reach the same performance criterion as younger adults. The third set of experiments focus on age differences in risky financial decision making and reveal that older adults make more suboptimal choices than younger adults when choosing risky assets. Neuroimaging analyses reveal that the representation of expected value in the nucleus accumbens and medial prefrontal cortex is correlated with optimal investment decisions, and that the age-related increase in risky investment mistakes is mediated by a novel neural measure of variability in nucleus accumbens activity. The presentation of value information through visual decision aids improves investment choices in both younger and older adults. These findings are consistent with the notion that mesolimbic circuits play a critical role in optimal choice, and imply that providing simplified information about expected value may improve financial risk taking across the adult life span. Across the experiments, the findings suggest that both age-related affective biases and probabilistic learning impairments can influence decision making both in the laboratory and in the real world through insular and mesolimbic brain regions. Importantly, age-related impairments are reduced under supportive task conditions (designed to target the brain systems identified using neuroimaging). Together, the set of experiments presented here suggests that understanding how the brain processes value information may eventually inform the design of more targeted and effective behavioral interventions for investors of all ages
Visual cortical representations of bottom-up salience and perceptual continuity by Brittany Edna Burrows( file )
1 edition published in 2011 in English and held by 2 libraries worldwide
The amount of information available to the retina far exceeds the processing capacity of the visual system. Visual attention provides a means by which we can make use of the information available with optimal efficiency, allowing us to selectively process objects that are most likely to help us achieve our behavioral goals while filtering out others. Decades of research have revealed that visual representations of objects aligned with the locus of attention are enhanced, resulting in perceptual benefits. The attentional modulation of visual cortical responses, however, has been almost exclusively characterized within the context of volitionally deployed, top-down attention. Although the behavioral correlates of involuntarily deployed, bottom-up attention are well known, the underlying neural mechanism and its relationship to top-down attention are not. In Chapters 2 and 3 of this thesis, I describe experiments that investigate the neural correlates and mechanisms of bottom-up attention. In order to examine the neural correlates of bottom-up attention, we measured responses of single neurons in area V4 to "popout" stimuli, which differ uniformly from surrounding items and draw attention in a bottom-up manner, and compared them to responses to "conjunction" stimuli, which are composed of a combination of surrounding features and require top-down selection. We found that V4 responses are modulated by popout stimuli. This selectivity was more robust when larger numbers of surrounding items and multiple features were included in the display, and it was absent when only a few items were presented immediately outside the receptive field (RF). We next examined the relationship between bottom-up and top-down attention by measuring responses to these same stimuli while top-down attention was directed to locations distant from the RF. We found that popout selectivity was eliminated, indicating that the salience of popout stimuli is not sufficient to drive selection by V4 neurons under these conditions. These results demonstrate that neurons in feature-selective cortex are influenced by bottom-up attention, but that this influence is limited by top-down attention. Whether the source of the signal initiating attentional deployment is top-down or bottom-up, our gaze is typically aligned with the locus of our attention, making use of the high density of photoreceptors at the fovea of the retina. As we move our eyes around a visual scene to briefly fixate objects for further processing, we do not perceive the explosive onset of motion between each fixation that is created by these eye movements, nor do we perceive a series of individual snapshots from each fixation. Instead, we have a seamless visual experience of our surroundings. This perception of continuity across eye movements is crucial to efficient behavior, and yet the neural mechanisms that support a continuous visual experience are not well understood. In Chapter 4 of this thesis, I describe an experiment that investigates the neural representations of perceptual continuity during saccadic eye movements. In order to examine whether V4 neurons are involved in creating a continuous perceptual experience, we measured the responses of V4 neurons to probe stimuli presented during the preparation of saccadic eye movements to targets distant from the RF. We observed that the responses of area V4 neurons to probe stimuli depended on the similarity between the neuronal preference and the target of a saccade. Specifically, the suppression of V4 responses normally present during the preparation of saccades to targets outside of a neuron's RF was absent when the features of the target were congruent with the preference of the recorded neuron. This target-dependent modulation resulted in a substantial effect of the target match/non-match manipulation on neuronal selectivity. The results are consistent with a computational model in which saccade preparation automatically results in a global target-dependent modulation of visual cortical signals. We suggest that this modulation may contribute to the continuity of perception across saccadic eye movements. The results of this thesis demonstrate that V4 neurons are selective for stimuli that draw bottom-up attention. Surprisingly, this selectivity depends on the availability of top-down attentional resources, suggesting that these two distinct varieties of visual attention rely, at least in part, on a common mechanism. Additionally, this thesis provides evidence that V4 neurons contribute to a seamless perceptual experience by maintaining important visual information transaccadically
Prefrontal cortex and recognition memory fMRI evidence for context-dependent retrieval processes by Anthony David Wagner( Book )
2 editions published in 1997 in English and held by 2 libraries worldwide
Neural mechanisms of learning and memory in the human medial temporal lobe by Karen Fossum LaRocque( file )
1 edition published in 2016 in English and held by 1 library worldwide
The medial temporal lobe is critical for enabling memory for the events of one's personal past. However, whether medial temporal lobe subregions perform shared or qualitatively distinct computations to support these memories is a topic of debate. One prominent theoretical perspective proposes that complementary behavioral expressions of memory for past events are enabled via the creation of separated versus overlapping neural representations for events that share similar elements, and that these representations are formed by the hippocampus and medial temporal lobe cortex, respectively. A wealth of empirical data has examined this proposal by asking whether these complementary behavioral expressions of memory are differentially linked to integrity and / or functional activation of the hippocampus and medial temporal lobe cortex. Here, we examined complementary medial temporal lobe computations through a different lens, and asked whether the degree of neural overlap in the hippocampus and medial temporal lobe cortex elicited by stimuli during encoding is differentially linked to the ability to recognize those stimuli in the future. We addressed this question with two experiments using converging methodologies. In Experiment 1 we used high-resolution functional magnetic resonance imaging and pattern similarity analyses to quantify neural overlap across the spatial topography in medial temporal lobe subregions during encoding. We found that the relationship between pattern similarity during encoding and subsequent memory dissociated across the hippocampus and medial temporal lobe cortex: later memory was linked to greater across-item pattern distinctiveness in the hippocampus, but to greater across-item pattern similarity in medial temporal lobe cortex. Additionally, by comparing neural patterns elicited by individual stimuli regardless of later memory for these stimuli, we found that perirhinal cortex and parahippocampal cortex exhibited differential content sensitivity for multiple stimulus categories, whereas the hippocampus failed to demonstrate such content sensitivity. In Experiment 2 we used intracranial electroencephalography to quantify neural overlap across the temporal evolution of population-level neural activity in medial temporal lobe subregions during encoding. We found that later memory was linked to greater across-item pattern distinctiveness in the hippocampus but not in medial temporal lobe cortex, and that the strength of these relationships was modulated by the demands on behavioral expressions of memory during retrieval. These data provide novel evidence for complementary learning mechanisms across the hippocampus and medial temporal lobe cortex, and suggest that these mechanisms operate across both spatial and temporal dimensions of neural codes
The binding problem and the perception of multiple stimuli by Cynthia Marie Henderson( file )
1 edition published in 2014 in English and held by 1 library worldwide
A long-standing assertion from the binding problem is that human vision is limited to identifying only a single stimulus at a time. We test the veracity of this claim in a series of experiments and simulations. We focus in particular on illusory conjunctions (ICs), a phenomenon that has provided support for single-object capacity limits. We find that studies on ICs have conflated to distinct phenomena which we term proximal and distal ICs. Proximal ICs occur between neighboring, peripheral stimuli, and we present evidence that they may be driven by shared or similar mechanisms to crowding. Unlike proximal ICs, distal ICs can occur between distant stimuli but seem to require brief stimulus durations and a competing task. We investigate claims that the competing task drives ICs either through mnemonic errors or through the manipulation of spatial attention, as FIT, a single-object capacity theory, would predict. We find support for neither claim. Instead, attention to and encoding of the competing task stimuli appears to cast a 'shadow' over nearby regions such that IC-susceptible stimuli appearing within those regions become more vulnerable to interference even from distant distractors. In neither phenomenon do we find support for a general inability to process multiple objects in parallel. We further tested claims of a single-object capacity limit with neural network simulations trained to identify abstract representations of stimuli. We found that a range of networks could identify two stimuli in parallel with high accuracy. Despite arguments by von der Malsburg (1999), these networks did not require a mechanism for oscillations and used only weighted summation with a sigmoid filter. Using these networks, we were able to simulate errors similar to crowding and proximal ICs as well as deficits in the identification of multiple stimuli after parietal lesions, a key piece of evidence supporting FIT. These empirical and modeling results demonstrate that it is possible to solve the binding problem without limiting perception to a single stimulus at a time
The neural mechanisms of decision processes underlying memory retrieval by Alan M Gordon( file )
1 edition published in 2014 in English and held by 1 library worldwide
Episodic memory retrieval can be construed as a set of decision processes in which retrieved event information guides a categorical response. Successful retrieval requires the establishment of cortical representations of event features during stimulus encoding, and the cortical reinstatement of these representations at retrieval. Additionally, retrieval requires neural mechanisms, often thought to be localized in lateral frontal and parietal cortex, to transform these representations into a decision variable that guides action selection. Finally, mnemonic decisions ultimately require the implementation of a motor response. The research described in this dissertation examines several aspects of the neural mechanisms that support mnemonic decision making, with three experiments. The first experiment examines how the strength of categorical representations during initial encoding and during retrieval influences behavioral outcomes of the mnemonic decision. The second experiment employs matched perceptual and memory tasks, to assess commonalities between the neural substrates of mnemonic and perceptual decision making. The final experiment examines retrieval acts made with oculomotor and manual responses, to assess whether lateral parietal activity associated with memory retrieval is specific to manual responses or general across response modalities. Collectively, these studies provide an account for how neural representations, decision processes, and response functions work together to support memory retrieval
Functional and structural diversity within the human dopamine system by Kelly Hennigan( file )
1 edition published in 2015 in English and held by 1 library worldwide
The mesencephalic dopamine system is essential for a number of basic neural processes, including reward learning, working memory, and motor control. Maladaptive dopamine functioning has been implicated in a number of psychiatric disorders ranging from addiction to impulse disorders and schizophrenia. Understanding the normal function of the dopamine system and its involvement in cognition and behavior is therefore critical for future advances in promoting mental health. Despite the diversity of functions it subserves, a predominant view has been that dopamine neurons are functionally homogeneous and uniformly broadcast reward-related signals as a population to target structures throughout the brain. The work presented here challenges this view by characterizing functional and structural heterogeneity within the dopamine system in humans. In particular, I show that the substantia nigra (SN) and ventral tegmental area (VTA) are activated by aversive events, which is explicitly at odds with the prevailing idea that dopamine neurons homogeneously encode reward prediction error. These regions exhibit differential patterns of functional connectivity during aversive processing, which suggests that they are components of distinct functional networks. I also identify structural diversity between the SN/VTA and striatum with diffusion-weighted imaging (DWI) and tractography methods. I show that diffusivity in the mesolimbic pathway, but not other nigrostrial pathways, is related to trait impulsiveness, which suggests that these pathways subserve distinct functionality. Collectively, this work suggests that there is functional heterogeneity within the dopamine system, and these functional differences are correlated with anatomically distinct white matter pathways
Vision and revision : cue-triggered perceptual reorganization of two-tone images in US preschoolers and adults by Jennifer Marie Yoon( file )
1 edition published in 2012 in English and held by 1 library worldwide
Two-tones are a class of detail-poor ambiguous images that are puzzles for the mind's eye. The images first confuse, then surprise and delight with a flash of perceptual "insight" when the puzzle is solved and the image transforms in the observer's mind into an easily recognizable figure. Such "aha!" moments are achieved without transforming the original two-tone in any way, but instead by simply providing a clue which allows the observer to "solve" the two-tone, for example by verbally labeling the image, or by showing the corresponding photo from which the two-tone was derived. More precisely, we can call this process cue-triggered perceptual reorganization. Drawing on data contrasting effortless performance in adults with striking failures in preschool-aged children and adults from a remote culture, I argue that this phenomenon may be a unique and important case study in visual enculturation. This framing is in contrast to culture-invariant processes such as high-level visual expertise and visual cortical maturation. Instead, I emphasize the core component process of representation revision, arguably a fundamental component of social cognition broadly as well as in specific material symbolic cultural pursuits such as reading. At least three component processes of representational extraction, alignment, and selection are required for successful cue-triggered perceptual reorganization. Cognitive manipulations that simplify these component processes improve preschoolers' performance, consistent with the argument that cue-triggered perceptual reorganization is a problem of representation revision, which itself may be entrained by a material symbolic culture. The data supporting the claims above are laid out in three chapters. Chapter 1 includes five studies demonstrating the robustness of children's difficulties in cue-triggered perceptual reorganization and culminates in the development of an experimental procedure and analysis method for quantifying this effect. Chapter 2 includes two experiments testing children's recognition across a continuous series of images that increase in difficulty from original full grayscale to two-tone in order to see if we can more precisely identify the point at which recognition breaks down, and instead find evidence that this recognition threshold is not absolute, but depends in part on children's mental representation of the photo cue. Chapter 3 includes four different manipulations of children's understanding of the relationship between the photo cue and two-tone image, revealing that cognitive interventions in the same age group and without stimulus simplification can improve recognition performance. Together these studies support the existence of a striking failure of cue-triggered perceptual reorganization in children that can best be understood as a failure of visual representation revision, a cognitive process already achievable by the preschool years, but not yet fully entrained by US culture
Learning position-tolerant object representations by David Allan Remus( file )
1 edition published in 2011 in English and held by 1 library worldwide
Human observers have a remarkable ability to recognize and discriminate objects despite a high degree of variation in their visual appearance in different contexts. The mechanisms underlying this tolerance to object transformation are largely unknown and it is unclear if learning can influence sensitivity to object transformations. In these studies, I examine the role of learning on our ability to visually discriminate objects across multiple retinal positions. The findings from this research program indicate that prior experience will influence an observer's ability to discriminate objects and that the quality of prior experience is of utmost importance to the development of position-tolerant object representations
Investigations of factors that affect unsupervised learning of 3D object representations by Moqian Tian( file )
1 edition published in 2016 in English and held by 1 library worldwide
Humans have an amazing ability to learn to recognize objects across transformations that present very different retinal stimuli, such as changes in size, illumination, and rotations in space. Such identity-preserving image transformations (DiCarlo, Zoccolan, & Rust, 2012) put extraordinary pressure on our visual system because the computations needed to assign vastly different 2D images of an object to the same identity are non-trivial. However, both behavioral (Biederman & Cooper, 1991a, 1991b; Fiser & Biederman, 1995; Potter, 1976; Thorpe, Fize, & Marlot, 1996) and neural (Hung, Kreiman, Poggio, & DiCarlo, 2005) evidence suggest that the visual system solves this problem accurately and rapidly. While rotations in the image plane preserve the visible features, rotations in-depth may reveal new features of an opaque object and thus present the most difficult transformation for the visual system to resolve, because the resulting 2D image from an in-depth rotation may be unrecoverable from the original image. Thus, understanding how people achieve viewpoint invariance, or the ability to recognize objects from different views and rotations, is key to understanding the visual object recognition system. There is a general consensus that learning is an important component for developing viewpoint invariant object recognition (Logothetis and Pauls, 1992; Tarr and Pinker, 1989). Many studies show that learning can occur in an unsupervised way just from viewing example images of new objects (Edelman and Bulthoff, 1992; Tarr and Pinker, 1989). Two major theories regarding how the visual system achieves viewpoint invariance -- 3D-based theories (Biederman, 1987) and view-based theories (Ullman and Basri, 1989) -- recognize the importance of learning in achieving viewpoint invariant object recognition. However, they differ in what information is used during learning and what representation is consequently built. For example, view-based theories consider spatial and temporal continuities as necessary glue for linking multiple views of an object during unsupervised learning, but 3D-based theories consider feature information to be more important. They also differ on whether the object representation that is built after learning is 3D based or view based. To address these gaps in the published literature, I examined two core questions: What kind of spatial and temporal information in the visual input during unsupervised learning is critical for achieving viewpoint invariant recognition? And what kind of object representation is generated during the learning process? In Chapter 1, I will present a theoretical overview of the issues. Section 1 reviews theories and computational models of viewpoint invariant recognition, with a focus on the debate between 3D-based theories and view-based theories; Section 2 reviews psychophysical and neural evidence supporting each theory; and Section 3 discusses the predictions of the learning mechanisms of each of the competing theories. Chapter 2 presents results from a series of experiments that investigated the spatio-temporal information in the visual input during unsupervised learning that is key for learning the 3D structure of novel objects. Chapter 3 presents data from a series of experiments that examine how the format of the visual information during unsupervised learning affects learning the 3D structure of novel objects. Finally, in Chapter 4, I will discuss the theoretical implications of the findings presented in Chapters 2 & 3, and propose a new framework based on these results
Two mechanisms of human contingency learning by Daniel Alexander Sternberg( file )
1 edition published in 2011 in English and held by 1 library worldwide
Understanding how we learn the relationships between events in the world, and how we learn to respond appropriately to those events has been an important focus of research in psychology and the cognitive sciences for many years. While many different theories have been proposed, two broad classes of accounts have been particularly popular. Findings from studies of animal conditioning and tasks requiring humans learners to make very fast responses have led to proposals that learning in these tasks is based on a process of gradual adjustments to pathways linking representations of cue events to representations of their outcomes and appropriate responses to those outcomes. However, many of the same findings can also be explained as a process of making explicit inferences about the likely causal relationships between events. Findings from human contingency learning tasks have been argued to support the idea that learners in these tasks rely on an inference-based reasoning process. This has led some to doubt that both types of processes necessarily exist. The experiments presented in this thesis set out to look for evidence that both processes exist, and that they are preferentially involved in different kinds of learning situations. In the presented experiments, participants saw displays containing single or paired objects and learned which displays were usually followed by the appearance of a dot shortly afterward. Some participants predicted whether the dot would appear and then saw the outcome, while others were required to respond very quickly if the dot appeared shortly after the objects. For prediction participants, instructions that guided them to infer which objects had the power to cause the dot outcome determined whether contingencies associated with one object affected predictions about its pair mate. For fast-paced responding participants, contingencies associated with one object affected responses to the pair-mate, even when more neutral instructions were provided. These results challenge single-mechanism accounts and support the proposal that the mechanisms underlying performance in the two tasks are distinct. The remainder of the thesis focuses on the development of computational models of the different kinds of processes thought to underlie responding in these tasks
Multiple systems support the neural representation of value over time by Kacey Anne Ballard( file )
1 edition published in 2011 in English and held by 1 library worldwide
Time, when prolonged to engender delays, has a dramatic impact on the way we value events and experiences in the world. Long delays make future rewards less attractive and make memories more difficult to retain. Thus, in the following series of studies, I investigate how time can be used as a manipulation to study the underlying neural processes that support valuation, learning, and decision making. Experiment 1 investigated how delays impact neural representations of value during intertemporal choice, where subjects chose between smaller, immediate rewards and larger, delayed rewards. Using FMRI, we dissociated neural representations of subjective value for delayed rewards according to reward magnitude and delay time. We found that mesolimbic dopamine regions correlated with the magnitude of future rewards, while dissociable activation in lateral cortical regions correlated with the delay of future rewards. The findings suggest that considering rewards far into the future recruits deliberative cognitive control processes in addition to the reward valuation processes of the mesolimbic system. Even when the timing of outcomes is not made explicit, the experience of delays during learning or decision making may affect how values are calculated and updated. Thus, Experiment 2 explored how the introduction of delays into an instrumental reward learning task affects which neural structures and learning strategies are recruited to support successful learning. We found that both mesolimbic (MPFC and NAcc) and declarative (MTL and lateral PFC) systems were recruited to support prediction error and value signals for learning over delay periods. Additionally, we provided evidence that value signals in the MTL were dependent on attention and working memory processes, as a secondary distraction task disrupted value signals in the MTL but not in the MPFC. The results demonstrate that reinforcement learning over delay periods requires shared activation between mesolimbic dopamine regions and declarative memory systems. Finally, Experiment 3 examined more directly how experiencing delays during learning diminishes explicit assessments of value. Participants learned the values associated with different visual cues by playing a repeated gambles game, experiencing reward outcomes that varied in reward magnitude, probability, and delay until delivery. We found that cues associated with longer delays were perceived as less valuable than cues with shorter delays, as indicated by subjects' explicit recall of reward probabilities. Delays also biased subjects' choices towards more risk-averse preferences. These finding suggest that the experience of delays diminishes assessments of value and biases choice behavior, even when people are not explicitly aware of delay times. Taken together, the present studies provide evidence that delays influence the neural processes underlying learning, valuation, and choice, and they suggest new applications for leveraging the timing of events to shift learning strategies and bias decision making heuristics
Overgeneral autobiographical memory in major depressive disorder is associated with abnormal neural activity during retrieval by James Eric Sorenson( file )
1 edition published in 2015 in English and held by 1 library worldwide
Major Depressive Disorder (MDD) is a psychiatric mood disorder that is associated with abnormal cognitive processes, including overgeneral memory (OGM) for autobiographical events. When asked to produce memories from their own lives, depressed individuals produce fewer specific memories than do their nondepressed peers. The neural basis of OGM in MDD, however, is not yet clear. Specifically, we do not know if depressed individuals exhibit normal memory-related neural activation during encoding; we also do not know if individuals with MDD show neural signatures that have been found to be related to retrieval of content, i.e. cortical reinstatement. To address these questions, depressed and nondepressed adults completed a standard measure of autobiographical memory specificity and, inside a Magnetic Resonance Imaging (MRI) scanner, a paired-associates learning task. At the behavioral level, depressed individuals showed evidence of decreased specificity of autobiographical memory relative to healthy controls, but did not exhibit impairment in associative learning. At the neural level, depressed and nondepressed participants did not differ during encoding. During retrieval, a predicted difference between depressed and nondepressed participants in neural activity in the hippocampus was also not obtained. Importantly, however, compared to their nondepressed peers, depressed individuals showed less activity in the right prefrontal cortex during the successful retrieval of images. In addition, and also counter to hypotheses, there was robust evidence of cortical reinstatement of image category information in both depressed and nondepressed individuals. Together, these findings suggest that OGM in MDD is not related to an inability to reactivate visual information or to an inability to form or encode associative memories. Instead, it seems that if OGM in MDD is related to some systematic aberrant neural function, this abnormality is a subtle difference between depressed and nondepressed individuals in retrieval processes that is mediated by the prefrontal cortex
Myelination of the brain in major depressive disorder : an in vivo magnetic resonance imaging study by Matthew Daniel Sacchet( file )
1 edition published in 2016 in English and held by 1 library worldwide
Major depressive disorder (MDD) is a debilitating psychiatric condition and a leading contributor to the global burden of disease. Characterizing MDD-related abnormalities in neurobiological processes will inform more comprehensive etiological frameworks of MDD that will facilitate the development of more targeted approaches to the prevention and identification of, and intervention for, this disorder. In this context, one promising biological target is myelin, a specialized biological tissue and fundamental facilitator of neuronal communication. Myelin ensheaths axons and facilitates saltatory conduction of electrical signaling in the nervous system. Postmortem studies of brains of depressed individuals, and non-human animal, genetic, and neuroimaging studies suggest that abnormalities in myelin are associated with MDD. Growing evidence suggests that neural activity and myelin influence each other to support an effective nervous system, and that stress-related neuroinflammation may result in the degradation of myelin in MDD. Brain regions implicated in this research, and in MDD more generally, include the nucleus accumbens (NAcc) and the dorsolateral prefrontal cortex (DLPFC), core regions involved in reward and cognitive control processes, respectively. Recent developments in quantitative magnetic resonance imaging (qMRI) allow for improved assessment of myelin content at the whole brain level, in vivo, in humans through the measure of R1. In this study we used qMRI to measure R1 to examine whether the brains and, in particular, the NAcc and DLPFC, of individuals diagnosed with MDD are characterized by reductions in myelin content compared to individuals without a history of psychiatric disorder (i.e., healthy controls [CTLs]). We found that the MDD group had lower levels of myelin than did the CTL group at the whole brain level and in the NAcc. Furthermore, myelin content of the DLPFC was reduced in MDD participants who had experienced a greater number of depressive episodes compared to both MDD participants who had experienced fewer depressive episodes and participants in the CTL group. Taken together, these results offer new evidence that MDD is characterized by reduced myelin content of the brain and in the NAcc in particular, and that the chronicity of MDD is associated with reduced myelin in the DLPFC. While further research is needed to elucidate the role of myelin in influencing affective, cognitive, behavioral, and clinical aspects of MDD, the current study provides important evidence that a fundamental property of brain composition, myelin, is altered in this disorder
Interactions between selective attention and working memory in emotion regulation : an EEG investigation by Ravi Thiruchselvam( file )
1 edition published in 2014 in English and held by 1 library worldwide
Selective attention and working memory (WM) are central to our ability to internally represent and act on the world, and a growing body of research suggests that the two interact in several ways. These interactions hold critical implications for emotion regulation, as they may underlie distinct forms of cognitive control that enable individuals to alter emotion. In this dissertation, I examine two types of selective attention-working memory interactions that may be important for emotion regulation: the gating of affective content into WM, and the biasing of affective content active within WM. Study 1 examined the immediate and delayed emotional consequences of controlling the gating of affective content into WM. Participants were presented with affective (and neutral) images during two phases. In an initial regulation phase, participants attempted to restrict the access of image representations into WM by loading WM with unrelated content. In a subsequent re-exposure phase, participants simply attended to these images. Results showed that, during the regulatory episode, loading WM while exposed to images rapidly and powerfully attenuated a robust electrocortical index of emotional response, the late positive potential (LPP). Upon re-exposure, however, images with a WM-load history paradoxically elicited heightened LPP responses (compared to images with a simple-viewing history). This pattern of findings diverged significantly from those obtained for another major form of emotion regulation -- cognitive reappraisal -- that permits encoding affective inputs into WM. Together, these results suggest that blocking the access of affective events into WM can quickly dampen emotion during the regulatory episode, a feature that may have unintended consequences upon passive re-exposure. Study 2 examined the emotional consequences of biasing affective content active within WM. Participants were cued to attend to either an arousing or neutral aspect of an affective image representation maintained within WM. Results showed that, relative to focusing on an arousing portion of a negative image representation within WM, focusing on a neutral portion reduced both LPP responses and self-reported negative emotion. These data suggest that deploying attention within affective representations active within WM can successfully alter emotion. Study 3 sought to replicate the findings obtained in Study 2, using a novel system to record electrocortical responses. Even under different recording conditions, results showed strong preservation of the original findings. Specifically, attending to a neutral (versus arousing) aspect of an unpleasant image representation within WM reduced both LPP responses and self-reported negative emotion. Study 4 examined whether the ability to bias affective content within WM via attention is impaired in a particular form of psychopathology -- generalized anxiety disorder (GAD) -- that may be characterized by deficits in the control of affective WM representations. Participants diagnosed with GAD and never-disordered healthy control (HC) participants were cued to attend to either an arousing or neutral aspect of an affective image representation active within WM. Contrary to predictions, results showed a failure to replicate the original findings in HC participants. Although the GAD group showed an inability to dampen LPP responses by shifting attention to neutral (versus arousing) aspects of affective WM content, the pattern of results is inconclusive due to a lack of expected findings in the HC group. Future studies will need to clarify the causes underlying the unexpected pattern of results amongst healthy participants
 
moreShow More Titles
fewerShow Fewer Titles
Languages
English (21)
Close Window

Please sign in to WorldCat 

Don't have an account? You can easily create a free account.