Lynn Lohnas
lyj at mail dot med dot upenn dot edu
____________________ Research Interests ____________________ Education ___________________ Lab Experience ____________________ Publications | I am interested in using computational modeling to study human cognition and behavior. I am also interested in determining the most universal set of rules that govern memory formation, storage and retrieval. My current thesis work combines these two interests by using modeling work to motivate and understand the interaction between episodic and semantic memories. _____________________________________________________________________________________________________________________ PhD Candidate, Neuroscience University of Pennsylvania, 2007-present BA, Neuroscience and Mathematics with Honors Brandeis University, 2003-2007 _____________________________________________________________________________________________________________________ Computational Memory Laboratory, PhD Candidate University of Pennsylvania, 2008-present Over the course of our lives, we form memories into episodes and associate them into a spatiotemporal context. Such episodic memories are autobiographical in nature, such as "Where did I park my car today?" Retrieving the content of an episodic memory can also facilitate retrieval of its associated contexts. In contrast, semantic memory refers to long-standing knowledge that lacks this autobiographical component. For instance, if you answer the questions "Who was the first U.S. president?" and "How many wheels does a car have?" it is rare that you can remember the original context(s) in which you learned this information. My thesis work attempts to characterize properties of longer-term and repeated episodic memories in order to elucidate how episodic memories give rise to semantic memories. Conversely, I am also interested in how semantic memory informs the creation and retrieval of episodic memories. I am investigating the properties of and interaction between episodic and semantic memories in the free-recall paradigm. In this task, human participants are given a list of items (typically words), and then asked to recall as many items as they can from that list in any order. This procedure is repeated for multiple lists in one session. Although participants are instructed to recall items only from the most recently presented list, they occasionally make errors: they may recall words from previously presented lists, or words that were not presented in any experimental list. Typically these errors are disregarded from analyses of free recall. However, studying mistakes of the memory system reveals what it is trying to optimize; the patterns of intrusions reflect influences of episodic and semantic memory systems. To this end, I have extended a neural network model of memory. This model formalizes cognitive correlates of the formation, storage, and retrieval of episodic and semantic memories, and makes predictions of behavioral performance. I am investigating the model's ability to account for results across variants of the free recall task. These findings reveal that participants can use the contexts associated with items, rather than any specific list-tagging mechanism, to target items of particular lists. I am also examining the role of semantic and episodic memory in the beneficial effects of repeating items within and across lists. Another project extends previous work that examined the role of episodic organization to motivate recall order, accuracy and timing. It is a well-known finding that participants are more likely, and more quickly transition, to items that were presented nearby on the list to the just-recalled item. However, these analyses do not address the degree to which previously recalled items may influence recall; participants may use an amalgam of previous items, combined into a compound cue. I have examined influences of compound cueing in free recall, on both recall probability and inter-response time. These results can be explained by a class of context-based models that assume context is used to cue recall, but not by models that assume item representations dominate the recall cue. Kelz Laboratory, Rotation Student University of Pennsylvania, 2008 I investigated different analysis techniques to distinguish intracranial electroencephalographic signals during various arousal states (REM, non-REM, awake, anesthetized). Brainard Laboratory, Rotation Student University of Pennsylvania, 2007 I developed a Bayesian model to examine how the visual system resolves the ambiguity inherent in the retinal image. This model interpreted standard visual stimuli successfully yet was susceptible to visual illusions, ensuring that the model could portray both the accuracies and imperfections of the human visual system. Cognition and Memory Laboratory, Research Assistant Brandeis University, 2004-2007 I analyzed and scored data acquired from behavioral and fMRI experiments from younger and older participants. These experiments investigated the learning process in adaptation to time- or frequency-compressed aural language. I also developed software for EEG analyses, as well as a graphical user interface to automate data processing of behavioral data. When I was a senior in college, I worked in the laboratory enrolled in an independent neuroscience project. I designed and ran a study to determine how participants are biased by previously presented information in anticipation of ambiguous novel stimuli. Laboratory for Computational Cognitive Neuroscience, Research Assistant University of California at Santa Barbara, 2004 I developed a computational model of reward uncertainty, assessing different mathematical approaches to best integrate information from different areas of the model. This work helped to clarify the representations of and interactions between different brain structures implicated in reward processing. Computational Memory Laboratory, Research Assistant Brandeis University, 2003-2004 As a freshman in college, I began working in the lab scoring data from list memory experiments. Although I had originally intended to major in mathematics, I soon learned of the fascinating and illuminating applications of mathematics to neuroscience. On slow days, my supervising graduate student would read and discuss with me articles from mathematical psychology. I learned how to create basic analysis scripts of behavioral data as well. My positive experiences in this lab inspired me to take neuroscience courses. Consequently, my enjoyment of those courses as well as my subsequent labwork led me to major in neuroscience as well as mathematics. _____________________________________________________________________________________________________________________ Lohnas, L.J., Polyn, S.M., Kahana, M.J. (2011). Contextual variability in free recall. Journal of Memory and Language, 64(3), 249-255. Lohnas, L.J., Kahana, M.J. (submitted) The word frequency effect in episodic memory. _____________________________________________________________________________________________________________________ |