Context Change Shapes the Organization of Memory Recall by Lindsay I. Rait A dissertation accepted and approved in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Psychology Dissertation Committee: Dr. Brice A. Kuhl, Chair Dr. Vishnu (Deepu) P. Murty, Core Member Dr. Dagmar (Dasa) Zeithamova, Core Member Dr. James M. Murray, Institutional Representative University of Oregon Spring 2025 2 © 2025 Lindsay I. Rait This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International license. 3 DISSERTATION ABSTRACT Lindsay I. Rait Doctor of Philosophy in Psychology Title: Context Change Shapes the Organization of Memory Recall Our memories of past experiences are strongly linked to rich contextual details—that is, memory for where or when an event took place. This contextual information not only supports memory retrieval but also shapes how memories are organized. Free recall tasks provide a unique window into these organizational processes, revealing that memory organization is often guided by the similarity of contextual features during learning. Contexts can be similar along a number of dimensions, including temporal, source, and motivational context. However, the context in which we form memories is constantly changing, and while context is known to support memory, it remains unclear how such changes influence what is remembered and how memories are organized. Across three studies, this dissertation explores how features of context change influence the organization of free recall memory, using both behavioral measures and brain activity measured with functional magnetic resonance imaging (fMRI). Chapter II presents two behavioral studies using a context switching paradigm to examine how the frequency and relative novelty of context changes interact to influence free recall. I found that recall performance is worse only when rapidly switching to novel contexts. This suggests not only a benefit for switching to a familiar context, but that impairments of switching to novel environments only emerged in the context of rapid versus slower switches. This difference may have been due, in part, to differences in memory organization. Chapter III is an fMRI study in which the rate of external context change was manipulated during encoding. I tested whether this affected memory organization, measured using 4 temporal clustering—the tendency to recall items in the order they were studied. I also examined whether these behavioral patterns were mirrored in the brain by looking at gradually changing activity in the hippocampus. A higher rate of context change was associated with both less temporal clustering and lower hippocampal autocorrelation, a measure of the stability of hippocampal activity patterns over time. Moreover, participants who exhibited greater hippocampal autocorrelation during encoding also exhibited stronger temporal clustering during recall, establishing a link between hippocampal autocorrelation and temporal organization of memory. Chapter IV used a between-subjects free recall paradigm to test whether agency— having control over one’s choices—can act as a context to organize memories. I found that participants with agency showed reduced temporal clustering compared to yoked participants who had no choice. Instead, participants with agency organized their memories more around the meaningful connections they constructed through their choices. Collectively, these findings provide novel insight into how contextual factors shape the organization of memory. This dissertation includes previously published and unpublished co-authored material. 5 ACKNOWLEDGMENTS I feel incredibly fortunate to have had the opportunity to work with and learn from so many remarkable individuals during my time in graduate school. It truly takes a village to raise a scientist, and I am profoundly grateful to mine. First and foremost, I would like to thank Dr. Brice Kuhl. I have learned so much from his thoughtful feedback, and I am especially thankful for the support he offered at pivotal moments. I cannot adequately express how grateful I am for his role in seeing me through to the finish line. I am also deeply grateful to Dr. Deepu Murty. His guidance has been a constant presence throughout my time in graduate school, even before he moved to Oregon. His mentorship has been invaluable in helping me navigate both the scientific journey and its most challenging moments. I will always appreciate his encouragement, humor, and his belief in me. I am incredibly fortunate to have had the opportunity to work with Dr. Sarah DuBrow, whose passion for science and unwavering support profoundly shaped my graduate school experience. She gave me the space to grow as both a scientist and a person, and her mentorship has left a lasting impact that I will always cherish. This dissertation would not have been possible without her. I would also like to thank Dr. Ben Hutchinson for his constructive feedback and for ensuring the lab environment was always welcoming—and well-stocked with snacks. Additionally, I would like to thank other members of my dissertation committee, Dr. Dasa Zeithamova and Dr. James Murray for their time, attention, and thoughtful insight throughout this process. Furthermore, I would like to thank all of the friends that I made along the way. I am so grateful to the many members of the DuBrow, Kuhl, Murty, and Hutchinson Labs, past and present. I am so appreciative of their support, advice, friendship, and laughter over the years. I am so lucky to have been a member of such a supportive environment filled with so many amazing people. 6 My biggest and most heartfelt thank you goes to my friends and family for their support throughout this process. I would like to thank my parents, Lauren and Jeff, for being my biggest cheerleaders and standing by me in everything I do. I am so grateful for their endless love and encouragement, and for always taking the time to listen—no matter how many times I interrupted Jeopardy! to share the details of my day. To my sisters, Rebecca and Megan, for filling my life with laughter, for always offering the world’s best advice, and for being my constant sources of strength and joy. I am also deeply thankful to the Silversteins—Lisa, Kevin, Marlee, and Steve —for such a warm welcome into the family and for always brightening my day. Finally, I would like to thank my husband, Michael. Growing as a researcher has gone hand in hand with growing our life together, and I am so grateful to have had him by my side throughout all of graduate school. His confidence in me, unwavering support, and the joy he brings to everyday life made all the difference. 7 To my parents—your unconditional love and support have meant the world to me. 8 TABLE OF CONTENTS Chapter Page I. INTRODUCTION .................................................................................................... 14 Context change shapes free recall .......................................................................... 17 Context change influences overall recall performance .................................... 17 Context change and memory for boundary information .................................. 19 Context change restructures memory organization .......................................... 20 Motivational states as a context for memory organization .............................. 21 The hippocampus and context memory ................................................................. 22 Hippocampal involvement in context processing ............................................ 22 Hippocampal drift ............................................................................................ 23 Overview of the present work ................................................................................ 24 References Cited .................................................................................................... 26 II. CONTEXTUAL FAMILIARITY RESCUES THE COST OF SWITCHING ....... 34 Introduction ............................................................................................................ 34 Experiment 1 .......................................................................................................... 37 Methods............................................................................................................ 37 Participants ................................................................................................. 37 Stimuli .................................................................................................. 37 Procedure .................................................................................................. 39 Design .................................................................................................. 40 Data analysis .............................................................................................. 41 Results .............................................................................................................. 43 9 Encoding performance ............................................................................... 43 Context switching and immediate recall performance ............................... 43 Context switching and recall organization ................................................. 44 Experiment 1: Discussion ................................................................................ 45 Experiment 2 .......................................................................................................... 46 Methods............................................................................................................ 46 Participants ................................................................................................. 46 Stimuli .................................................................................................. 47 Design and procedure ................................................................................ 47 Data analysis .............................................................................................. 48 Results .............................................................................................................. 50 Encoding performance ............................................................................... 50 Context switching and immediate recall performance ............................... 51 Context switching and recall organization ................................................. 52 Temporal clustering ............................................................................. 52 Recall transitions by context ................................................................ 53 Experiment 2: Discussion ................................................................................ 54 General Discussion ................................................................................................ 55 References Cited .................................................................................................... 58 10 Chapter Page III. HIPPOCAMPAL DRIFT RATE REFLECTS THE TEMPORAL ORGANIZATION OF MEMORIES ..................................................................... 62 Introduction ............................................................................................................ 62 Methods.................................................................................................................. 64 Participants ....................................................................................................... 64 Stimuli .............................................................................................................. 64 Experimental procedure ................................................................................... 65 Design .............................................................................................................. 68 Analysis of behavioral data .............................................................................. 69 fMRI data acquisition ...................................................................................... 71 Anatomical data preprocessing ........................................................................ 72 Functional data preprocessing .......................................................................... 73 Region of interest (ROI) definition .................................................................. 74 Autocorrelation analysis .................................................................................. 76 Correlation between temporal clustering and hippocampal autocorrelation ... 77 Statistical tests .................................................................................................. 77 Results .................................................................................................................... 78 Behavioral measures of recall .......................................................................... 78 fMRI measures of autocorrelation ................................................................... 83 Relationship between hippocampal autocorrelation and temporal clustering in recall ........................................................................................... 89 Discussion .............................................................................................................. 92 References Cited .................................................................................................... 97 11 IV. AGENCY ALTERS MEMORY ORGANIZATION DURING FREE RECALL... ............................................................................................................. 104 Introduction ............................................................................................................ 104 Methods.................................................................................................................. 106 Participants ....................................................................................................... 106 Stimuli .............................................................................................................. 107 Experimental procedure ................................................................................... 108 Data Analysis ................................................................................................... 110 Temporal clustering analysis ............................................................... 112 Decision clustering analysis ................................................................. 113 Results .................................................................................................................... 114 Discussion .............................................................................................................. 118 References Cited .................................................................................................... 121 V. GENERAL DISCUSSION. .................................................................................... 125 Features of context change shape recall performance and organization ................ 125 Memory organization relates to context representations in the hippocampus ....... 128 Conclusion ............................................................................................................. 132 References Cited .................................................................................................... 132 APPENDICES ............................................................................................................. 137 A. CHAPTER II SUPPLEMENTARY MATERIAL ............................................ 137 B. CHAPTER IV SUPPLEMENTARY MATERIAL .......................................... 144 12 LIST OF FIGURES Figure Page 1.1.Schematic depicting internal context representations based on the rate of external context changes ........................................................................................ 19 2.1. Task design for experiments 1 & 2 ....................................................................... 38 2.2. Experiment 1 results ............................................................................................. 44 2.3. Experiment 2 immediate recall performance ........................................................ 52 2.4. Context switching and recall organization ............................................................ 53 3.1. Encoding procedure and design ............................................................................ 67 3.2. Memory recall by condition and boundaries ........................................................ 79 3.3. Temporal clustering during memory recall ........................................................... 81 3.4. Temporal autocorrelation for hippocampal subregions ........................................ 84 3.5. Hippocampal autocorrelation as a function of switch rate condition ................... 87 3.6. Relationship between hippocampal autocorrelation (during memory encoding) and temporal clustering (during subsequent recall) .............................................. 91 4.1. Experimental design .............................................................................................. 109 4.2. Temporal clustering by group ............................................................................... 115 4.3. Decision clustering ................................................................................................ 117 S2.1. Final recall for experiments 1 and 2 ................................................................... 138 S2.2. Proportion of local transitions by context for High and Low Switch rates for experiment 2 .................................................................................................. 139 13 LIST OF TABLES Table Page 3.1. Autocorrelation effects in non-hippocampal ROIs ............................................... 89 S2.1. Multilevel logistic regression model examining the interaction between switch rate, switch type, and contextual familiarity on percent recall for immediate recall .................................................................................................................... 137 S4.1. Mixed-effects linear regression model examining the effect of Group (Choice vs. Fixed) and average number of words recalled per list (Mean Recall) on temporal clustering ......................................................................................... 144 S4.2. Mixed-effects linear regression model examining the effect of Group (Choice vs. Fixed) and average number of words recalled per list (Mean Recall) on the probability of making stay transitions ...................................................... 144 14 CHAPTER I INTRODUCTION Memories are not stored in isolation; they are embedded within the context in which they were formed. What we remember is deeply shaped by the places we visit, how long we were there, and the events unfolding around us. These contextual details help anchor our memories, allowing us to mentally re-experience past events—such as vividly recalling the setting of a meaningful conversation or the atmosphere of a significant life event. The ability to retrieve specific experiences, complete with the contextual details that define them, is known as episodic memory (Tulving, 2002). Context not only supports memory retrieval but also shapes how memories are structured and organized over time. One way to examine how context influences memory is through free recall, a task in which individuals remember studied information without external cues. This method allows researchers to investigate not only how much information is remembered but also the order in which it is recalled—that is, how memories are organized. Studies of free recall have demonstrated that the order in which items are recalled from memory is influenced by the similarity of the contexts in which they were encoded (Howard & Kahana, 2002; Polyn et al., 2009a; Sederberg et al., 2008). This is evidenced by recall transitions—where it is more likely that individuals will recall two items in immediate succession that share a similar context. In this way, the organization of items in free recall is thought to provide a window into the structure of natural memories and the underlying contextual representations. However, our environment is constantly changing. As we move through different locations, engage with different people, or shift between tasks, the context in which memories are formed evolves. These features of our environment may change abruptly or shift gradually over time 15 (DuBrow et al., 2017; Polyn et al., 2009a). These fluctuations in context play a crucial role in shaping both how much we remember and how our memories are organized. For example, imagine attending a conference where you listen to multiple talks throughout the day. If you were asked to recall which talks you went to, you might recall the talks in the order they happened, remembering first what you heard in the morning before lunch and then what you heard in the afternoon. Alternatively, your memory could be structured by topic—you might group together all the talks related to episodic memory, even if they were spread throughout the day, and separately recall those about decision-making. Prior research has shown that sudden context changes—such as changes in perceptual attributes (Heusser et al., 2018), task set (Polyn et al., 2009b), and stimulus class (DuBrow & Davachi, 2013, 2016)—can shape how information is remembered. However, it remains unclear how different features of context change influence the organization of free recall memory. One key aspect of context change that may influence memory organization is its switch rate. Slower changes in context have been shown to provide structure for memory (Polyn et al., 2009b), but the effects of more rapid context changes remain unknown. These effects may also depend on an individual’s familiarity with the environment—while distinct or novel events can serve as meaningful boundaries that support memory, too much novelty or frequent disruptions may impair recall. Beyond external context shifts, internal factors also shape how memories are organized during free recall (Wang et al., 2023). For instance, having agency over a decision influences subsequent temporal memory judgments (Houser et al., 2022), yet little is known about whether actively making a choice between contexts affects memory organization. Clarifying how these factors interact is essential for uncovering the mechanisms underlying memory organization. 16 While the link between context processing and recall has been well studied across a diverse range of behavioral studies, a growing number of studies additionally provide insight into the underlying neural mechanisms at play. Specifically, the hippocampus is thought to be a key structure involved in both context processing and our ability to recall the past (Baldassano et al., 2017; Ben-Yakov & Dudai, 2011; DuBrow & Davachi, 2016). Famously, the case study of Patient H.M., who had damage to his hippocampus and other nearby structures, revealed that he was unable to form or subsequently recall any new episodic memories following his injury (Scoville & Milner, 1957). More recently, a number of studies across modalities and species have also implicated the hippocampus in various aspects of context processing (see Davachi, 2006; Eichenbaum et al., 2007, 2012; Maren et al., 2013; Ranganath, 2010a, 2010b; Ross & Easton, 2021; Rudy, 2009; Smith & Mizumori, 2006 for reviews). Taken together, these lines of literature strongly suggest that the hippocampus plays a fundamental role in linking contextual information to how and what we recall. The aim of this dissertation is to better understand how changes in context during encoding influence free recall memory. The work in this dissertation describes a series of experiments that use novel behavioral paradigms to examine how memories are shaped by contextual information, alongside advanced functional magnetic resonance imaging (fMRI) techniques. Specifically, this dissertation investigates: 1) how qualitative features of context change influence memory structure and organization, 2) whether hippocampal context representations are linked to temporal memory organization during free recall, and 3) whether having agency serves as a context for structuring memories. The remaining sections of this chapter will review relevant background before moving on to detailed descriptions of each study in the subsequent chapters. 17 Context change shapes free recall Context change influences overall recall performance Context plays a critical role in both remembering and forgetting, with changes in context shaping how much information is later recalled. For instance, research using both the segmentation of continuous events (Flores et al., 2017; Schwan et al., 2000) and shifts in experimental context (Pettijohn et al., 2016; Smith et al., 1978) found that memory performance is often enhanced when an experience includes a meaningful context change. This overall recall benefit is especially pronounced when the task requires participants to actively attend to the meaningful change points in an event (Boltz, 1992; Gold et al., 2017; Schwan et al., 2000; Schwan & Garsoffky, 2004). However, while context changes can enhance memory for new information, they may also come at a cost. Findings from context-dependent memory literature indicate that recall is often impaired when the context at study does not match the context at retrieval (Godden & Baddeley, 1975; Sahakyan & Kelley, 2002; Shin et al., 2021; Smith et al., 1978; Smith & Vela, 2001; Unsworth et al., 2012). Taken together, these findings suggest that the effects of context changes on memory depend on specific features of the change itself. It remains unclear what determines whether a change in context enhances or disrupts recall memory. One key feature that may shape the effects of context on memory is the rate of context change. In the real world, different situations are often accompanied by dramatically different numbers of contexts. For instance, a postal worker can spend their whole day delivering packages to multiple different houses or alternatively, to just a few large apartment building mailrooms. Studies suggest that learning across multiple contexts can reduce the impairment typically seen when memory is tested in a novel context (Smith, 1982). For example, recall performance improves when word lists are studied in multiple spatial locations (two, three, or four) rather than 18 in a single setting (Smith, 1982; Smith, 1984; Smith et al., 1978; Smith & Rothkopf, 1984). Similarly, one study found that in narratives with event shifts, stories with two shifts enhanced overall recall performance compared with those with only one shift (Pettijohn et al., 2016). However, it remains unclear whether increasing the rate of context changes continues to benefit recall or if there is a point at which too many shifts become disruptive. According to contextual variability theory, items will be more easily recalled if they are tagged with more varied contexts, as each distinct context provides additional paths to retrieve the item (Lohnas et al., 2011). Even if an item is only encountered once, our internal context is thought to drift slowly, so items may be tagged with the current as well as the previous external context. As depicted in Figure 1.1, when slowly alternating between two different tasks the internal context that tags individual items is highly differentiated (i.e., tends to be either red or blue). When trying to remember those items, recall may be limited by only having a single contextual cue for retrieval. By contrast, when more rapidly switching, it is possible that part of the prior context not only lingers into the next context, but also overlaps, creating a “blended” contextual representation, such that individual items are tagged with both contextual cues (depicted in purple; Polyn et al., 2012). This could be helpful for memory in that items now have more contextual cues for retrieval (Lohnas et al., 2011; Siegel & Kahana, 2014). However, it could also be harmful for memory due to heightened interference and competition (Anderson, 2003), or decreased novelty typically associated with a context switch (Polyn et al., 2012). The effects of the rate of context change on memory performance may also interact with other factors, such as the individual’s familiarity with the contexts. 19 Figure 1.1 Schematic depicting internal context representations based on the rate of external context changes (slow vs. rapid). Colors represent different tasks or categories. Outline represents external context and gradient bars represent internal context. Context change and memory for boundary information Context changes can also prioritize the encoding of information at the change point, or event boundary. In one task-switching study, participants were more likely to recall the boundary item first in the recall sequence, and there was a recall advantage for items studied immediately after the context shift (Polyn et al., 2009b). A similar boundary enhancement was observed using naturalistic navigation (Jeunehomme & D’Argembeau, 2020). These findings are consistent with the idea that event boundaries trigger increased attention to salient, new information, which in turn is associated with better memory for information encountered at the boundary (Clewett et al., 2019; Zacks & Swallow, 2007). However, some studies found no consistent boost in free recall for boundary items (Heusser et al., 2018; Pettijohn et al., 2016). Additionally, one study even reported BabyDog Lily Lake Internal context Internal context BabyDog Lily Lake Slow Context Switching Rapid Context Switching 20 both a benefit and a lack of a clear advantage in recalling boundary items within the same paradigm. In this movie viewing study, explicitly cueing boundaries through editing led to higher recall for those segments compared with unedited clips (which had the same boundaries), suggesting that directing attention to boundaries may be essential for memory benefits (Gold et al., 2017). Determining the specific conditions under which recall memory for boundary information is enhanced is an important avenue for future research. Context change restructures memory organization According to computational models of free recall, memory search is thought to be guided by internally maintained representations of context. This context representation contains information related to temporal context, source information, and semantic characteristics (Polyn et al., 2009a). An item’s temporal context consists of recently studied items, other environmental cues, as well as the individuals’ internal thoughts and emotions. This temporal context representation has been thought to drift slowly over time (Howard & Kahana, 2002; Polyn et al., 2009a). When two temporally adjacent items are studied within the same slowly drifting context, an indirect association forms between them, linking them over time (DuBrow et al., 2017). During recall, the current context serves as a retrieval cue, activating memories of items encoded in similar contextual states. With each recalled item, the context updates, influencing the order of subsequently recalled items. This explains the well-established phenomenon of temporal clustering, where individuals tend to recall items in the order in which they were originally studied (Kahana, 1996; Sederberg et al., 2010). Temporal clustering has been previously thought to be very durable, surviving the insertion of a distractor task before every item (Howard & Kahana, 1999). While prior studies have since shown that temporal clustering can be influenced by 21 manipulating features of the stimuli themselves (Hong et al., 2024; Manning et al., 2023), future work is needed to determine how changing the external context influences temporal clustering. Source context reflects the content of the information present at encoding beyond the gradually drifting temporal representation. This includes features such as the type of stimulus used or the task participants are instructed to perform (DuBrow & Davachi, 2013; Murdock & Walker, 1969; Polyn et al., 2009a, 2009b). A sudden shift in the external environment, task, or internally maintained goals can lead to a disruption or reorganization of the source context representation. Although free recall is inherently unconstrained–allowing participants to recall items in any order– substantial evidence suggests that salient environmental changes impose structure on memory by isolating items studied before and after a context change. Specifically, items encountered within the same context are more likely to be recalled together, a phenomenon known as source clustering (Hintzman et al., 1972; Murdock & Walker, 1969; Polyn et al., 2009a, 2009b, 2012; Smith, 1982). Moreover, research indicates that participants are more likely to transition between successive recalls within the same context rather than across contexts (Chan et al., 2017; Heusser et al., 2018; Lohnas et al., 2023; Polyn et al., 2009b). This suggest that while context changes can facilitate the organization of memories by grouping related information together (e.g., recalling items studied with the same task together), they can also make it more difficult to retrieve information across contexts. How temporal and source context interact when contexts change is still unclear. Motivational states as a context for memory organization The studies discussed so far have focused on how memories are organized for neutral information. However, many studies have explored motivational state as a source context (Rouhani et al., 2020; Talmi et al., 2021; Wang et al., 2023). While much of this work examines how 22 motivation influences memory performance, several studies have also shown that motivational states structure memory organization. For instance, one study had participants study positive, neutral, and negative words in a free recall task. Participants were significantly more likely to recall successive words of the same valence (e.g., positive to positive) than words of different valence (e.g., positive to negative; Long et al., 2015). This effect has been extended to reward motivation, where participants learned lists of words that were associated with high or low rewards (Horwath et al., 2023). Similarly, participants were more likely to transition between words of the same reward category, but this effect was specific to highly rewarded items, suggesting that reward served as an organizational category for valuable items (Horwath et al., 2023). In contrast, a study examining motivation through threat vs. instruction-based incentives found that threat impaired the organization of memory around motivation value compared to instructed motivation (Horwath et al., 2024). Notably, in the both studies, temporal clustering did not differ across motivation conditions (Horwath et al., 2023, 2024). Together these findings suggest that recalling items in temporal order is not always adaptive. Instead, motivational states may serve as a context to organize memories, though this organization appears disrupted with threat. However, it remains unknown whether these effects persist when individuals have agency over their choices. The hippocampus and context memory Hippocampal involvement in context processing Episodic memory involves forming associations between individual events and the context in which they occurred. The hippocampus is thought to play a critical role in this process, as evidenced by research in humans (Chun & Phelps, 1999; Herz et al., 2023; Long et al., 2017; Miller et al., 2013) and rodents (Corcoran & Maren, 2001; Komorowski et al., 2009, 2013; 23 McKenzie et al., 2014). For instance, in a visual search task, amnesic patients with hippocampal damage showed normal improvements in general search speed. However, unlike healthy individuals, they did not exhibit any context-dependent learning—where repeated exposure to the same displays would typically help locate the target more quickly over time (Chun & Phelps, 1999). Additionally, contextual factors have also been shown to directly influence hippocampal activity (Baldassano et al., 2017; Dimsdale-Zucker et al., 2018; DuBrow & Davachi, 2014, 2016; Geva-Sagiv et al., 2023). In one line of work, researchers found that the offset of movie clips was related to increased hippocampal activity, which correlated with later memory (Ben-Yakov et al., 2013; Ben-Yakov & Dudai, 2011). Further research is needed to examine how changes in context influence activity in the hippocampus. Hippocampal drift Much of our internal temporal context representations drift more slowly than our changing external environment (DuBrow et al., 2017; Polyn et al., 2009a). This slowly drifting temporal context representation has been thought to be reflected in gradually changing patterns of activity in the hippocampus. For instance, evidence in both humans (Folkerts et al., 2018) and rodents (Manns et al., 2007) suggests that hippocampal activity patterns become more dissimilar as the temporal distance between experiences increases. Additionally, hippocampal time cells gradually change their activity patterns over seconds during moments of empty periods in a sequence (Eichenbaum, 2014; MacDonald et al., 2011; Umbach et al., 2020), reinforcing the idea that the hippocampus maintains a continuous representation of temporal context. If the hippocampus is continuously tracking changes in context, how might this be reflected in behavior? 24 Several lines of evidence support the idea that this hippocampal drift is relevant to episodic memory. For example, the degree of drift in the hippocampus during the encoding of a sequence of stimuli is predictive of subsequent temporal memory judgments for those stimuli (DuBrow & Davachi, 2014; Ezzyat & Davachi, 2014; Jenkins & Ranganath, 2016; Manns et al., 2007). More recently, researchers have used autocorrelation analyses to quantify how hippocampal activity changes slowly over time (Bouffard et al., 2023; Brunec et al., 2018; Coughlan et al., 2023). These researchers found that hippocampal drift is modulated by navigation behavior (Bouffard et al., 2023; Brunec et al., 2018), such that as navigation difficulty increases, the amount of drift decreases, as reflected by increased autocorrelation. Yet, there is a surprising lack of evidence directly linking hippocampal drift rate to behavior, specifically in free recall. Overview of the present work The goals of this dissertation are to: 1) examine how specific features of context change shape recall performance and organization and 2) test whether memory organization relates to drifting context representations in the hippocampus. In three empirical studies, I will show that manipulations of external (Chapters II and III) and internal context (Chapter IV) influence the rate of internal contextual drift specifically in the hippocampus (Chapter III only), and influence how free recall memory is organized. Importantly, these chapters will demonstrate that different features of context change will lead to different ways to organize memory. Together, these studies add to a growing literature characterizing the effects of context on episodic memory and further our understanding of the role of the hippocampus in free recall memory. In Chapter II, I test how the interaction between the rate of context change and an individual’s familiarity with the context influences free recall memory. I found and replicated that 25 free recall memory was only impaired when switching quickly between novel contexts, not when switching quickly between familiar contexts. In fact, when the context was familiar, memory was just as good when the context was changing rapidly vs. when it was not changing at all. Lastly, using clustering analyses, we observed that the order in which participants recalled the items differed when the context switched rapidly depending on whether the contexts were familiar or novel. In Chapter III, I test whether there is a direct link between temporal clustering in recall and drifting contextual representations in the hippocampus during encoding. While context switch rate had no effect on the total number of words recalled, I found that it significantly influenced the degree of temporal clustering. Specifically, a higher context switch rate was associated with less temporal clustering. Using autocorrelation analyses to measure hippocampal drift (Bouffard et al., 2023; Brunec et al., 2018), we found that this pattern of data was mirrored by autocorrelation in the hippocampus: autocorrelation significantly decreased when switch rate increased. Most importantly, I found that hippocampal autocorrelation during encoding was positively correlated with temporal clustering during free recall. In Chapter IV, I test whether imbuing participants with agency over a decision can act as a context to organize memories. I found that participants with agency showed reduced temporal clustering compared to participants who had no control over their choices, suggesting that recalling items in strict temporal order may not always be the most adaptive strategy. Instead, I found that having agency over a choice led participants to organize their memories more around the relationships they actively constructed through their choices. Finally, Chapter V will summarize the presented empirical findings and provide a general discussion that ties them to the broader literature. 26 References Cited Anderson, M. C. (2003). Rethinking interference theory: Executive control and the mechanisms of forgetting. Journal of Memory and Language, 49(4), 415–445. https://doi.org/10.1016/j.jml.2003.08.006 Baldassano, C., Chen, J., Zadbood, A., Pillow, J. W., Hasson, U., & Norman, K. A. (2017). Discovering Event Structure in Continuous Narrative Perception and Memory. Neuron, 95(3), 709-721.e5. https://doi.org/10.1016/j.neuron.2017.06.041 Ben-Yakov, A., & Dudai, Y. (2011). Constructing Realistic Engrams: Poststimulus Activity of Hippocampus and Dorsal Striatum Predicts Subsequent Episodic Memory. Journal of Neuroscience, 31(24), 9032–9042. https://doi.org/10.1523/JNEUROSCI.0702-11.2011 Ben-Yakov, A., Eshel, N., & Dudai, Y. (2013). Hippocampal immediate poststimulus activity in the encoding of consecutive naturalistic episodes. Journal of Experimental Psychology: General, 142(4), 1255–1263. https://doi.org/10.1037/a0033558 Boltz, M. (1992). Temporal accent structure and the remembering of filmed narratives. Journal of Experimental Psychology: Human Perception and Performance, 18(1), 90–105. https://doi.org/10.1037/0096-1523.18.1.90 Bouffard, N. R., Golestani, A., Brunec, I. K., Bellana, B., Park, J. Y., Barense, M. D., & Moscovitch, M. (2023). Single voxel autocorrelation uncovers gradients of temporal dynamics in the hippocampus and entorhinal cortex during rest and navigation. Cerebral Cortex, 33(6), 3265–3283. https://doi.org/10.1093/cercor/bhac480 Brunec, I. K., Bellana, B., Ozubko, J. D., Man, V., Robin, J., Liu, Z.-X., Grady, C., Rosenbaum, R. S., Winocur, G., Barense, M. D., & Moscovitch, M. (2018). Multiple Scales of Representation along the Hippocampal Anteroposterior Axis in Humans. Current Biology, 28(13), 2129-2135.e6. https://doi.org/10.1016/j.cub.2018.05.016 Chan, S. C. Y., Applegate, M. C., Morton, N. W., Polyn, S. M., & Norman, K. A. (2017). Lingering representations of stimuli influence recall organization. Neuropsychologia, 97, 72–82. https://doi.org/10.1016/j.neuropsychologia.2017.01.029 Chun, M. M., & Phelps, E. A. (1999). Memory deficits for implicit contextual information in amnesic subjects with hippocampal damage. Nature Neuroscience, 2(9), 844–847. https://doi.org/10.1038/12222 Clewett, D., DuBrow, S., & Davachi, L. (2019). Transcending time in the brain: How event memories are constructed from experience. Hippocampus, 29(3), 162–183. https://doi.org/10.1002/hipo.23074 Corcoran, K. A., & Maren, S. (2001). Hippocampal Inactivation Disrupts Contextual Retrieval of 27 Fear Memory after Extinction. The Journal of Neuroscience, 21(5), 1720–1726. https://doi.org/10.1523/JNEUROSCI.21-05-01720.2001 Coughlan, G., Bouffard, N. R., Golestani, A., Thakral, P. P., Schacter, D. L., Grady, C., & Moscovitch, M. (2023). Transcranial magnetic stimulation to the angular gyrus modulates the temporal dynamics of the hippocampus and entorhinal cortex. Cerebral Cortex, 33(6), 3255–3264. https://doi.org/10.1093/cercor/bhac273 Davachi, L. (2006). Item, context and relational episodic encoding in humans. Current Opinion in Neurobiology, 16(6), 693–700. https://doi.org/10.1016/j.conb.2006.10.012 Dimsdale-Zucker, H. R., Ritchey, M., Ekstrom, A. D., Yonelinas, A. P., & Ranganath, C. (2018). CA1 and CA3 differentially support spontaneous retrieval of episodic contexts within human hippocampal subfields. Nature Communications, 9(1), 294. https://doi.org/10.1038/s41467-017-02752-1 DuBrow, S., & Davachi, L. (2013). The influence of context boundaries on memory for the sequential order of events. Journal of Experimental Psychology. General, 142(4), 1277– 1286. https://doi.org/10.1037/a0034024 DuBrow, S., & Davachi, L. (2014). Temporal Memory Is Shaped by Encoding Stability and Intervening Item Reactivation. Journal of Neuroscience, 34(42), 13998–14005. https://doi.org/10.1523/JNEUROSCI.2535-14.2014 DuBrow, S., & Davachi, L. (2016). Temporal binding within and across events. Neurobiology of Learning and Memory, 134, 107–114. https://doi.org/10.1016/j.nlm.2016.07.011 DuBrow, S., Rouhani, N., Niv, Y., & Norman, K. A. (2017). Does mental context drift or shift? Current Opinion in Behavioral Sciences, 17, 141–146. https://doi.org/10.1016/j.cobeha.2017.08.003 Eichenbaum, H. (2014). Time cells in the hippocampus: A new dimension for mapping memories. Nature Reviews Neuroscience, 15(11), Article 11. https://doi.org/10.1038/nrn3827 Eichenbaum, H., Sauvage, M., Fortin, N., Komorowski, R., & Lipton, P. (2012). Towards a functional organization of episodic memory in the medial temporal lobe. Neuroscience and Biobehavioral Reviews, 36(7), 1597–1608. https://doi.org/10.1016/j.neubiorev.2011.07.006 Eichenbaum, H., Yonelinas, A. P., & Ranganath, C. (2007). The Medial Temporal Lobe and Recognition Memory. Annual Review of Neuroscience, 30(1), 123–152. https://doi.org/10.1146/annurev.neuro.30.051606.094328 Ezzyat, Y., & Davachi, L. (2014). Similarity Breeds Proximity: Pattern Similarity 28 within and across Contexts Is Related to Later Mnemonic Judgments of Temporal Proximity—ScienceDirect. https://www.sciencedirect.com/science/article/pii/S0896627314000737 Flores, S., Bailey, H. R., Eisenberg, M. L., & Zacks, J. M. (2017). Event segmentation improves event memory up to one month later. Journal of Experimental Psychology: Learning, Memory, and Cognition, 43(8), 1183–1202. https://doi.org/10.1037/xlm0000367 Folkerts, S., Rutishauser, U., & Howard, M. W. (2018). Human Episodic Memory Retrieval Is Accompanied by a Neural Contiguity Effect. Journal of Neuroscience, 38(17), 4200– 4211. https://doi.org/10.1523/JNEUROSCI.2312-17.2018 Geva-Sagiv, M., Dimsdale-Zucker, H. R., Williams, A. B., & Ranganath, C. (2023). Proximity to boundaries reveals spatial context representation in human hippocampal CA1. Neuropsychologia, 189, 108656. https://doi.org/10.1016/j.neuropsychologia.2023.108656 Godden, D. R., & Baddeley, A. D. (1975). Context-Dependent Memory in Two Natural Environments: On Land and Underwater. British Journal of Psychology, 66(3), 325–331. https://doi.org/10.1111/j.2044-8295.1975.tb01468.x Gold, D. A., Zacks, J. M., & Flores, S. (2017). Effects of cues to event segmentation on subsequent memory. Cognitive Research: Principles and Implications, 2(1), 1. https://doi.org/10.1186/s41235-016-0043-2 Herz, N., Bukala, B. R., Kragel, J. E., & Kahana, M. J. (2023). Hippocampal activity predicts contextual misattribution of false memories. Proceedings of the National Academy of Sciences, 120(40), e2305292120. https://doi.org/10.1073/pnas.2305292120 Heusser, A. C., Ezzyat, Y., Shiff, I., & Davachi, L. (2018). Perceptual boundaries cause mnemonic trade-offs between local boundary processing and across-trial associative binding. Journal of Experimental Psychology: Learning, Memory, and Cognition, 44(7), 1075. https://doi.org/10.1037/xlm0000503 Hintzman, D., Block, R. A., & Inskeep, N. R. (1972). Memory for mode of input. Journal of Verbal Learning and Verbal Behavior, 11(6), 741-749. https://doi.org/10.1016/S0022- 5371(72)80008-2 Hong, M. K., Gunn, J. B., Fazio, L. K., & Polyn, S. M. (2024). The modulation and elimination of temporal organization in free recall. Journal of Experimental Psychology: Learning, Memory, and Cognition, 50(7), 1035–1068. https://doi.org/10.1037/xlm0001311 Horwath, E. A., Katerman, B. S., Biju, M., DuBrow, S., & Murty, V. P. (2024). Threat Impairs the Organization of Memory Around Motivational Context. Journal of Cognitive Neuroscience, 1–10. https://doi.org/10.1162/jocn_a_02242 Horwath, E. A., Rouhani, N., DuBrow, S., & Murty, V. P. (2023). Value restructures the 29 organization of free recall. Cognition, 231, 105315. https://doi.org/10.1016/j.cognition.2022.105315 Houser, T. M., Tompary, A., & Murty, V. P. (2022). Agency enhances temporal order memory in an interactive exploration game. Psychonomic Bulletin & Review, 29(6), 2219–2228. https://doi.org/10.3758/s13423-022-02152-7 Howard, M. W., & Kahana, M. J. (1999). Contextual variability and serial position effects in free recall. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25(4), 923–941. https://doi.org/10.1037/0278-7393.25.4.923 Howard, M. W., & Kahana, M. J. (2002). A Distributed Representation of Temporal Context. Journal of Mathematical Psychology, 46(3), 269–299. https://doi.org/10.1006/jmps.2001.1388 Jenkins, L. J., & Ranganath, C. (2016). Distinct neural mechanisms for remembering when an event occurred. Hippocampus, 26(5), 554–559. https://doi.org/10.1002/hipo.22571 Jeunehomme, O., & D’Argembeau, A. (2020). Event segmentation and the temporal compression of experience in episodic memory. Psychological Research, 84. https://doi.org/10.1007/s00426-018-1047-y Kahana, M. J. (1996). Associative retrieval processes in free recall. Memory & Cognition, 24(1), 103–109. https://doi.org/10.3758/BF03197276 Komorowski, R. W., Garcia, C. G., Wilson, A., Hattori, S., Howard, M. W., & Eichenbaum, H. (2013). Ventral Hippocampal Neurons Are Shaped by Experience to Represent Behaviorally Relevant Contexts. Journal of Neuroscience, 33(18), 8079–8087. https://doi.org/10.1523/JNEUROSCI.5458-12.2013 Komorowski, R. W., Manns, J. R., & Eichenbaum, H. (2009). Robust Conjunctive Item-Place Coding by Hippocampal Neurons Parallels Learning What Happens Where. Journal of Neuroscience, 29(31), 9918–9929. https://doi.org/10.1523/JNEUROSCI.1378-09.2009 Lohnas, L. J., Healey, M. K., & Davachi, L. (2023). Neural Temporal Context Reinstatement of Event Structure During Memory Recall. Journal of Experimental Psychology: General, 152(7), 1840–1872. https://doi.org/10.1037/xge0001354 Lohnas, L. J., Polyn, S. M., & Kahana, M. J. (2011). Contextual variability in free recall. Journal of Memory and Language, 64(3), 249–255. https://doi.org/10.1016/j.jml.2010.11.003 Long, N. M., Danoff, M. S., & Kahana, M. J. (2015). Recall dynamics reveal the retrieval of emotional context. Psychonomic Bulletin & Review, 22(5), 1328–1333. https://doi.org/10.3758/s13423-014-0791-2 Long, N. M., Sperling, M. R., Worrell, G. A., Davis, K. A., Gross, R. E., Lega, B. C., Jobst, B. 30 C., Sheth, S. A., Zaghloul, K., Stein, J. M., & Kahana, M. J. (2017). Contextually Mediated Spontaneous Retrieval Is Specific to the Hippocampus. Current Biology, 27(7), 1074–1079. https://doi.org/10.1016/j.cub.2017.02.054 MacDonald, C. J., Lepage, K. Q., Eden, U. T., & Eichenbaum, H. (2011). Hippocampal “Time Cells” Bridge the Gap in Memory for Discontiguous Events. Neuron, 71(4), 737–749. https://doi.org/10.1016/j.neuron.2011.07.012 Manning, J. R., Whitaker, E. C., Fitzpatrick, P. C., Lee, M. R., Frantz, A. M., Bollinger, B. J., Romanova, D., Field, C. E., & C, H. A. (2023). Feature and order manipulations in a free recall task affect memory for current and future lists. arXivorg. Manns, J. R., Howard, M. W., & Eichenbaum, H. (2007). Gradual Changes in Hippocampal Activity Support Remembering the Order of Events. Neuron, 56(3), 530–540. https://doi.org/10.1016/j.neuron.2007.08.017 Maren, S., Phan, K. L., & Liberzon, I. (2013). The contextual brain: Implications for fear conditioning, extinction and psychopathology. Nature Reviews Neuroscience, 14(6), 417– 428. https://doi.org/10.1038/nrn3492 McKenzie, S., Frank, A. J., Kinsky, N. R., Porter, B., Rivière, P. D., & Eichenbaum, H. (2014). Hippocampal Representation of Related and Opposing Memories Develop within Distinct, Hierarchically Organized Neural Schemas. Neuron, 83(1), 202–215. https://doi.org/10.1016/j.neuron.2014.05.019 Miller, J. F., Neufang, M., Solway, A., Brandt, A., Trippel, M., Mader, I., Hefft, S., Merkow, M., Polyn, S. M., Jacobs, J., Kahana, M. J., & Schulze-Bonhage, A. (2013). Neural Activity in Human Hippocampal Formation Reveals the Spatial Context of Retrieved Memories. Science, 342(6162), 1111–1114. https://doi.org/10.1126/science.1244056 Murdock, B. B., & Walker, K. D. (1969). Modality effects in free recall. Journal of Verbal Learning and Verbal Behavior, 8(5), 665–676. https://doi.org/10.1016/S0022- 5371(69)80120-9 Pettijohn, K. A., Thompson, A. N., Tamplin, A. K., Krawietz, S. A., & Radvansky, G. A. (2016). Event boundaries and memory improvement. Cognition, 148, 136–144. https://doi.org/10.1016/j.cognition.2015.12.013 Polyn, S. M., Kragel, J. E., Morton, N. W., McCluey, J. D., & Cohen, Z. D. (2012). The neural dynamics of task context in free recall. Neuropsychologia, 50(4), 447–457. https://doi.org/10.1016/j.neuropsychologia.2011.08.025 Polyn, S. M., Norman, K. A., & Kahana, M. J. (2009a). A context maintenance and retrieval model of organizational processes in free recall. Psychological Review, 116(1), 129–156. https://doi.org/10.1037/a0014420 31 Polyn, S. M., Norman, K. A., & Kahana, M. J. (2009b). Task context and organization in free recall. Neuropsychologia, 47(11), 2158–2163. https://doi.org/10.1016/j.neuropsychologia.2009.02.013 Ranganath, C. (2010a). A unified framework for the functional organization of the medial temporal lobes and the phenomenology of episodic memory. Hippocampus, 20(11), 1263–1290. https://doi.org/10.1002/hipo.20852 Ranganath, C. (2010b). Binding Items and Contexts: The Cognitive Neuroscience of Episodic Memory. Current Directions in Psychological Science, 19(3), 131–137. https://doi.org/10.1177/0963721410368805 Ross, T. W., & Easton, A. (2021). The Hippocampal Horizon: Constructing and Segmenting Experience for Episodic Memory. Neuroscience and Biobehavioral Reviews, 132. https://doi.org/10.1016/j.neubiorev.2021.11.038 Rouhani, N., Norman, K. A., Niv, Y., & Bornstein, A. M. (2020). Reward prediction errors create event boundaries in memory. Cognition, 203, 104269. https://doi.org/10.1016/j.cognition.2020.104269 Rudy, J. W. (2009). Context representations, context functions, and the parahippocampal– hippocampal system. Learning & Memory, 16(10), 573–585. https://doi.org/10.1101/lm.1494409 Sahakyan, L., & Kelley, C. M. (2002). A contextual change account of the directed forgetting effect. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28(6), 1064. https://doi.org/10.1037/0278-7393.28.6.1064 Schwan, S., & Garsoffky, B. (2004). The cognitive representation of filmic event summaries. Applied Cognitive Psychology, 18(1), 37–55. https://doi.org/10.1002/acp.940 Schwan, S., Garsoffky, B., & Hesse, F. W. (2000). Do film cuts facilitate the perceptual and cognitive organization of activitiy sequences? Memory & Cognition, 28(2), 214–223. https://doi.org/10.3758/BF03213801 Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery, and Psychiatry, 20(1), 11-21. https://doi.org/10.1136/jnnp.20.1.11 Sederberg, P. B., Howard, M. W., & Kahana, M. J. (2008). A context-based theory of recency and contiguity in free recall. Psychological Review, 115(4), 893. https://doi.org/10.1037/a0013396 Sederberg, P., F. Miller, J., W. Howard, M., & J. Kahana, M. (2010). The temporal contiguity effect predicts episodic memory performance. Memory & Cognition, 38(6), 689–699. https://doi.org/10.3758/MC.38.6.689 32 Shin, Y. S., Masís-Obando, R., Keshavarzian, N., Dáve, R., & Norman, K. A. (2021). Context- dependent memory effects in two immersive virtual reality environments: On Mars and underwater. Psychonomic Bulletin & Review, 28(2), 574–582. https://doi.org/10.3758/s13423-020-01835-3 Siegel, L. L., & Kahana, M. J. (2014). A retrieved context account of spacing and repetition effects in free recall. Journal of Experimental Psychology: Learning, Memory, and Cognition, 40(3), 755–764. https://doi.org/10.1037/a0035585 Smith, D. M., & Mizumori, S. J. Y. (2006). Hippocampal place cells, context, and episodic memory. Hippocampus, 16(9), 716–729. https://doi.org/10.1002/hipo.20208 Smith, S. (1982). Enhancement of recall using multiple environmental contexts during learning. Memory & Cognition, 10(5), 405–412. Smith, S. M. (1984). A comparison of two techniques for reducing context-dependent forgetting. Memory & Cognition, 12(5), 477–482. https://doi.org/10.3758/BF03198309 Smith, S. M., Glenberg, A., & Bjork, R. A. (1978). Environmental context and human memory. Memory & Cognition, 6(4), 342–353. https://doi.org/10.3758/BF03197465 Smith, S. M., & Rothkopf, E. Z. (1984). Contextual Enrichment and Distribution of Practice in the Classroom. Cognition and Instruction, 1(3), 341–358. https://doi.org/10.1207/s1532690xci0103_4 Smith, S. M., & Vela, E. (2001). Environmental context-dependent memory: A review and meta- analysis. Psychonomic Bulletin & Review, 8(2), 203–220. https://doi.org/10.3758/BF03196157 Talmi, D., Kavaliauskaite, D., & Daw, N. D. (2021). In for a penny, in for a pound: Examining motivated memory through the lens of retrieved context models. Learning & Memory, 28(12), 445–456. https://doi.org/10.1101/lm.053470.121 Tulving, E. (2002). Episodic Memory: From Mind to Brain. Annual Review of Psychology, 53(1), 1–25. https://doi.org/10.1146/annurev.psych.53.100901.135114 Umbach, G., Kantak, P., Jacobs, J., Kahana, M., Pfeiffer, B. E., Sperling, M., & Lega, B. (2020). Time cells in the human hippocampus and entorhinal cortex support episodic memory. Proceedings of the National Academy of Sciences, 117(45), 28463–28474. https://doi.org/10.1073/pnas.2013250117 Unsworth, N., Spillers, G. J., & Brewer, G. A. (2012). Dynamics of context-dependent recall: An examination of internal and external context change. Journal of Memory and Language, 66(1), 1–16. https://doi.org/10.1016/j.jml.2011.05.001 33 Wang, Y. C., Adcock, R. A., & Egner, T. (2023). Toward an integrative account of internal and external determinants of event segmentation. Psychonomic Bulletin & Review, 31(2), 484-506. https://doi.org/10.3758/s13423-023-02375-2 Zacks, J. M., & Swallow, K. M. (2007). Event Segmentation. Current Directions in Psychological Science, 16(2), 80–84. https://doi.org/10.1111/j.1467-8721.2007.00480.x 34 CHAPTER II CONTEXTUAL FAMILIARITY RESCUES THE COST OF SWITCHING From Rait, L. I., Murty, V. P., & DuBrow, S. (2024). Contextual familiarity rescues the cost of switching. Psychonomic Bulletin & Review, 31(3), 1103-1113. Introduction Our environment is ever-changing, filled with switches in context varying in how often switches occur as well as prior exposure to the contexts (i.e., relative novelty). These types of changes profoundly influence how we form and structure memories. For instance, imagine watching a movie in the living room while preparing dinner in the kitchen. If asked to recall what happened in the movie, your later memory of a specific scene may become associated with memories of preparing dinner, supporting the recall of those memories. Instead, you could run errands in a new shopping mall. You may have difficulty recalling all the items purchased since you were rapidly switching between each new store. Prior research showed how memory performance is influenced by several different qualia of context, including perceptual attributes (Heusser et al., 2018), task set (Polyn et al., 2009b), and stimulus class (DuBrow & Davachi, 2013, 2016). Questions remain about how the qualitative features of context change influence memory structure and organization. There are two particular features of contextual change known to influence memory but have yet to be studied in the context of free recall: rate of context changes and relative novelty of the context to which one is switching. Typically, when an item is presented during study, it is stamped into a continuously drifting context representation. The detection of a sufficiently novel representation, however, can cause a sudden shift in context (Polyn et al., 2009a). When slowly 35 alternating between different contexts, the internal context that tags individual items is highly differentiated, which has been shown to provide structure for participants’ free recall (Heusser et al., 2018; Polyn et al., 2009b). However, when rapidly switching between two repeated contexts, part of the prior context may not only linger into the new context at transitions, but also overlap, creating a “blended” contextual representation. Little is known about whether a “blended” internal context representation would be beneficial or detrimental for free recall memory. One possibility is that recall accuracy would benefit from rapid contextual switching as items will be more easily recalled if tagged with more varied contexts, providing more contextual retrieval cues (Lohnas et al., 2011; Siegel & Kahana, 2014). However, recall could also be impaired due to interference from overlapping memories that share similar internal context representations. When memories share features, retrieving those memories may be more difficult compared to memories that do not share features due to heightened competition during retrieval (Anderson, 2003). The current set of experiments was designed to arbitrate between these two hypotheses. The factors that contribute to a “blended” context may not solely rely on the frequency of switching, but could also be modulated by the learner’s familiarity with the contexts. Accumulating evidence suggests that salient events, like encountering a novel scene (Zacks & Swallow, 2007) or oddball items (Ranganath & Rainer, 2003; Von Restoroff, 1933), can separate the overlap amongst two contexts and boost memory performance. However, increasing the amount of novelty or frequency of novel events leads to worse memory performance (Radvansky et al., 2011; Reggev et al., 2018; Shepherdson, 2021) and less access to information immediately following the shift (Dux & Marois, 2009). These findings lend us to expect that recall memory 36 would be impaired specifically when individuals are rapidly switching to novel vs. repeated events due to less access to retrieval cues. One mechanism by which switch rate may influence recall performance is by inducing competition during memory search between items learned with contexts that are switching vs. not switching. With a limited time to recall, the most memorable items will “win”. When an item “wins” this competition, the context representation is updated and the next item recalled will likely have been encoded with a similar context (Lohnas et al., 2015; Lohnas & Kahana, 2014; Polyn et al., 2009a). A related, yet parallel question will investigate whether the presence of competition during memory search (competitive vs. pure lists) differentially influences memory performance when switching contexts at different rates. This present study uses a context switching paradigm in which we independently manipulate switching rate and relative novelty of the contexts. We sought to characterize how these qualitatively different features of context change influence how well items are remembered and organized. We predict that memory will be differentially affected by switch rate with exposure to novel contexts, such that rapidly switching to novel contexts will be more harmful for memory compared to slower switching as participants may have less access to retrieval cues. We also predict that returning to a repeated context may independently benefit memory as a “blended” context could provide more contextual retrieval cues. We then determine the extent to which items are organized by the order in which they were encoded or the type of context with which they were originally presented. 37 Experiment 1 Methods Participants One hundred and ten participants from the University of Oregon completed this experiment online for course credit. One participant was excluded for chance-level performance on the encoding task, 19 participants were excluded for failing to provide audio usable for verbal recall, and six participants were excluded for writing down words as indicated on a post-experiment questionnaire. The final sample size for analysis was 84 participants (65 female, mean age 19.46 +/- 3.17 SD). Participants were randomly assigned to one of two contextual familiarity groups (repeated = 41; novel = 43). Sample size was determined based on standards in the literature of free recall (for example Polyn et al., 2009a) and doubled given that this study has a between- subjects manipulation. Consent was obtained in a manner approved by University of Oregon’s Institutional Review Board. Stimuli In brief, encoding consisted of alternating presentations of word and scene stimuli. Scene stimuli consisted of 46 unique scene contexts, where half depicted an indoor scene and half depicted an outdoor scene (Chang et al., 2019). A scene context is defined as the image immediately preceding a particular study item (word). We randomized the presentation of scene contexts appearing in each condition across participants. Cue words were 240 two-syllable nouns presented in capitalized letters (e.g., “GIRAFFE”). Nouns were based on object image labels from the Bank of Standardized Stimuli (Brodeur et al., 2014). Words were randomly assigned to scenes 38 and conditions uniquely for each participant. Stimuli were presented using Inquisit 6 [computer software]. (2020). Retrieved from https://www.millisecond.com. Figure 2.1. Task design for experiments 1 & 2. a) Procedure. Each trial began with the encoding phase, which consisted of alternating presentations of word and scene stimuli. Participants were instructed to respond as to whether the item depicted by the word would fit in the scene. After a 10s distractor task, participants verbally recalled as many items as possible from the list that they could. After all eight lists were completed, participants completed a final recall. b) Within Subjects Switch Rate Manipulation. In Experiment 1, participants learned lists with a high or low contextual switch rate. These were competitive lists with two switch types per list (switch vs. no-switch). In Experiment 2, participants learned pure lists at just one switch rate (no vs. low vs. high). Participants in Experiment 1 learned lists of 24 words, whereas participants in Experiment 2 learned lists of 16 words. c) Between Subjects Contextual Familiarity Manipulation. We manipulated the level of contextual familiarity with the scene contexts (repeated vs. novel). Distractor Free Recall (Immediate, Final) 3 + 4 +1 = 8 TRUEFALSE 10s Say the words that you remember outloud Exp 2: Immediate Min 60s Max 120s Final Min 180s Max 300s Encoding Repeated Context Novel Context PIZZA CORAL APPLE LILY CANDY CELLO LION RIVER Repeated Novel time a cb 1s Cue 500ms ISI 2.5s Word 1s ITI PIZZA+ + Task: Does the item fit in the scene (y/n)? Experiment 1: Competitive Lists of 24 Words Experiment 2: Pure Lists of 16 Words Low Rate No-Switch Rate Low-Switch Rate High-Switch Rate High Rate Switch Items No-Switch Items Switch Items No-Switch Items time https://www.millisecond.com/ 39 Procedure After a brief practice, the experiment consisted of eight lists, with each list consisting of three sequential phases: encoding, distractor, and recall (Figure 2.1a). On each trial, participants viewed a scene context for 1,000ms. The scene disappeared for 500ms and was then followed by a word presented in the center of the screen for 2,500ms. Alternating presentations of scene and word stimuli were chosen as it has previously been shown as an effective way to manipulate temporal context (Chan et al., 2017; Manning et al., 2016). During the word presentation, participants were instructed to respond as to whether the item depicted by the word would fit in the previous scene (yes/no). This is a subjective judgment as to whether the participant could picture a given item in the scene previously presented, which did not contain the item. The word remained on the screen for 2,500ms regardless of a button press to equate encoding time. Trials were separated by a 1,000ms intertrial interval (ITI) which consisted of a blank screen. Each list in the encoding phase included a total of 24 words. Immediately following each encoding phase, participants completed a math distractor task to reduce rehearsal. Participants were presented with math equations in the form of A + B + C = D, where the values of A, B, and C were set to single digit integers (Howard & Kahana, 1999). Participants were instructed to indicate whether the statement was true or false with a key press. The distractor phase lasted 10s in total, but the number of equations completed was variable depending on speed of completion. After the distraction period, participants were given up to three minutes to verbally recall as many items as possible from the list that they could, without any explicit instructions about the order of the of recall. A written cue indicated the start of the recall period, and participants’ microphones were turned on for recording. Participants could move onto the next list whenever 40 they felt that they recalled as many words as they could remember. After all eight lists were completed, participants moved onto the final recall portion of the experiment. Participants were instructed to verbally recall as many words as they could from the entire experiment for up to three minutes. Design The encoding task contained two main conditions of interest. For the first condition of interest, we manipulated the switch rate between scenes to generate two switch rates within subjects: low rate and high rate (Figure 2.1b). For the low rate, the scene contexts changed after every four items. The high rate was the most rapid switch rate, where the scene context switched after every two items. Each list had two different switch types within the list: switch (high or low) and no-switch (Figure 2.1b). The no-switch rate (baseline) is the slowest switch rate, where all of the items were studied with the same scene (no context switches). The rationale was to create competitive lists and to include a baseline, no-switch rate for comparison within each list. Therefore, given that a list was composed of 24 unique items, 12 are presented at the no-switch rate and 12 are presented at the switch rate (either the low-switch or high-switch rate). List order and switch type order was randomized for each participant. The second condition of interest is the level of contextual familiarity that participants have with the scene contexts. We manipulated contextual familiarity (repeated vs. novel) as a between- subjects variable (Figure 2.1c). Within a given list in the repeated context condition, participants switched back and forth between the same two scene contexts (see Figure 2.1b for example list). In other words, when there was a change in the scene context, it would be a repeat of a scene that had already been seen in the list previously. However, in the novel context condition, each time 41 there was a switch in scene context, participants would see a novel scene that had not been seen before in the experiment. In both groups, every list contained new scenes and included both switch and no-switch items. Data analysis Statistical analyses were conducted in R 3.6.3 (R Core Team (2020); https://www.R- project.org/). For the encoding task, given the subjectivity of responses, accuracy was calculated based on normative responses. We determined whether each response matched the modal response for when each word was presented with each scene. Mean responses were calculated for each level of contextual familiarity and compared using Welch’s Two Sample t-tests, as there are unequal sample sizes. Effect sizes (e.g., Cohen’s d) were calculated using the lsr package in R. Using the lme4 package in R, Generalized Linear Mixed-Effects Models were used to determine whether switch rate (low vs. high), switch type (no-switch vs. switch) and contextual familiarity (repeated vs. novel) predicted the percent of words that participants recalled, with subject and word identity as random effects. Specifically, we ran a model that assessed the relationship between percent of words recalled and the interaction between switch rate, switch type, and contextual familiarity. In a parallel question, we were interested in how the nature of competition between switch rates during memory search influences memory performance. Therefore, we ran an additional model that further unpacked the two-way interaction between switch rate and switch type. Given that our main predictions were about the high-switch rate, we ran a third model that specifically looked at the two-way interaction between high-switch rate and contextual familiarity (repeated vs. novel) to better interpret this interaction. These analyses were run for both immediate and final recall data (see Supplementary Figure 2.1 for final recall results). https://www.r-project.org/ https://www.r-project.org/ 42 Additionally, we ran a final model to determine whether item position in the event predicted the percent of words that participants recalled (see Supplementary Analysis 1). All models additionally controlled for list number and list half (whether the word appeared in the first 12 or last 12 items in the list) as fixed effects. To determine the extent to which participants tend to successively recall nearby items, we calculated a temporal clustering score for each participant (Polyn et al., 2009a). For each recall transition, we determined the temporal distance (in absolute lag) between the serial position of the just-recalled word and the set of not-yet-recalled words. The temporal clustering score is calculated as the proportion of possible lags greater than the observed lag. A score of 1 indicates high temporal clustering, meaning that participants made the shortest transitions possible. A score of 0.5 indicates chance-level temporal clustering, meaning that transitions were just as likely to be to a neighboring or remote item. For this analysis, each participant received two temporal clustering scores: one for high rate lists (which includes both the switch and no-switch items) and one for low rate lists (which includes both the switch and no-switch items). Temporal clustering scores were computed using publicly available MATLAB (The MathWorks, Natick, MA) scripts from the Behavioral Toolbox (Version 1.01) from the Computational Memory Lab (http://memory.psych.upenn.edu/Behavioral_toolbox). Mean temporal clustering scores were calculated and compared across levels of contextual familiarity (repeated vs. novel) using t-tests. Verbal recall responses were digitally recorded and annotated offline using Penn Total Recall (http://memory.psych.upenn.edu/TotalRecall). Four undergraduate research assistants, who were blind to which words were randomly assigned to which switch rate and level of contextual familiarity (repeated vs. novel), annotated the verbal responses. A recall was classified as valid if http://memory.psych.upenn.edu/Behavioral_toolbox 43 the item recalled came from the current list. Items from previous lists, words not in the wordpool, or other vocalizations (e.g, “umm”) were not included in analysis. Results Encoding performance Overall accuracy was 87.71%. Accuracy and response time (RT) did not differ for repeated and novel context switches (Accuracy: repeated- M = 87.72%, SE = 0.96, novel- M = 87.71%, SE = 0.9, t(81.42) = 0.01, p = .99, d = .002; RT: F(1, 82) = 1.42, p = .24, d = .11). Context switching and immediate recall performance Participants recalled 22.37% of total words. Free recall accuracy was greater for repeated (24.45%) vs. novel (20.39%) switches, t(81.99) = 2.18, p = .03, d = .48, suggesting that participants’ memory was better for words from lists containing repeated scenes. We next examined whether switch rate, switch type, and contextual familiarity influenced recall performance. There was a significant three-way interaction, highlighting that the relationship between switch rate (high vs. low) and switch type (no-switch vs. switch) differed depending on whether scenes were repeating or novel, z = -2.48, p = .01, partial R2 < .001 (see Supplementary Table 2.1 for full model). Next, we unpacked the two-way interaction between switch rate and switch type. With exposure to novel contexts, recall performance was reduced when switching contexts at a high rate, compared to not switching, z = 3.26, p = .001, partial R2 = .003 (Figure 2.2a). There was no detriment to memory for low-switch, compared to no-switch, items with novel context switching, z = .30, p = .76, partial R2 = .001. This resulted in a reliable high x low rate interaction, which 44 highlights that performance is disrupted only when switching to novel contexts at a high rate, z = 2.05, p = .04, partial R2 < .001. Interestingly, there was a boost in recall performance for high- switch, over no-switch, items when switching back to repeated contexts, z = 2.83, p = .005, partial R2 = .001 (Figure 2.2a). Given that performance differences were specific to rapid switching, a final analysis was conducted to directly test the interaction between high-switch rate and contextual familiarity. This resulted in a reliable high rate x context familiarity interaction, z = 4.21, p < .001, partial R2 = .002 (Figure 2.2a). This means that switching back to repeated contexts at a rapid rate may rescue the cost associated with rapidly switching to novel contexts. Figure 2.2. Experiment 1 results. a) Immediate Recall Performance. b) Temporal Clustering. Error bars reflect within subject standard error. ** p < .01, *** p <.001. Context switching and recall organization We next investigated how switch rate and contextual familiarity influenced temporal organization. Results showed that both switch rates across both levels of familiarity showed 0.50 0.55 0.60 0.65 0.70 Repeated Novel Contextual Familiarity Te m po ra l C lu st er in g Sc or e *** Low High Low High 0% 10% 20% Switch Rate Pe rc en t R ec al le d NovelRepeated ** ** a b Switch Items No-Switch Items Switch Items No-Switch Items 45 significant binding of items to their temporal context, as measured by greater than chance-level temporal clustering (ps < .001). However, when individuals were switching to novel vs. repeated contexts, there was a greater reliance on temporal information (i.e., higher temporal clustering), t(157.6) = 4.99, p < .001, d = .77 (Figure 2.2b). Experiment 1: Discussion The results suggest that costs to memory only occur when individuals are rapidly switching to novel contexts. In fact, participants’ memory performance was improved when rapidly switching to repeated contexts over not switching. This suggests not only a benefit for switching to repeated contexts, but that impairments of switching to novel environments only emerged in the context of rapid versus slower switches. Experiment 1 used competitive lists (containing both switch and no-switch items) to test how competition between switch and no-switch items during memory search influenced memory performance. Results demonstrated that this competition shaped memory performance. First, performance for no-switch items differed depending on switch rate (high vs. low), given that no- switch items were competing with different switching rates in memory. Additionally, memory differences between switch and no-switch items were only observed in high-switch rate lists. For Experiment 2, we will investigate whether competition between switch rates was necessary to observe memory differences. We additionally found that switching to novel contexts was more likely to increase reliance on temporal information. This may seem counterintuitive as recall performance is disrupted at a high switch rate to novel contexts, supporting the idea that participants may be making more remote transitions when switching to repeated contexts. These findings support the framework that 46 when scenes are repeatedly encountered, a “blended” context representation is created that spans the entire list, supporting long-distance transitions. We quantitatively test this in Experiment 2 by further investigating the types of recall transitions made by participants. Experiment 2 Experiment 2 aimed to replicate the above findings with increased sample size and recall performance. The goals for Experiment 2 were: 1) replicate the recall performance findings that the negative effect of switching on memory recall was rescued (at least as good as not switching) when switching to a repeated context and 2) replicate and expand on the recall organization results to further investigate how participants structure their memory. To expand, we investigated the types of recall transitions participants made. Methods Participants One hundred ninety-two native English speakers were recruited from Prolific. In order to increase power for Experiment 2, we doubled the sample size from Experiment 1. Participants were compensated an initial $6.50 and could receive an additional bonus payment of up to $6.00 for good performance on the encoding and recall portions of the experiment. Nine participants were excluded for chance-level performance on the encoding task, six participants were excluded for failing to provide audio usable for verbal recall, and ten participants were excluded for writing down words as indicated on a post-experiment questionnaire. The final sample size for analysis was 167 participants (90 female, mean age 35.74 +/- 12.98 SD). Participants were randomly 47 assigned to one of two context switching groups (repeated = 83; novel = 84). Consent was obtained in a manner approved by University of Oregon’s Institutional Review Board. Stimuli The stimuli used for Experiment 2 were the same as those used in Experiment 1. However, in Experiment 2, we only used 30 of the scene contexts as the list length was shortened (see below). Design & procedure The procedure for Experiment 2 was identical to that of Experiment 1, except for the following changes aimed at improving participants’ verbal recall performance. The first set of changes were made to the encoding portion of the experiment. First, the list length was shortened to only contain 16 items per list. The rationale for this change was to help improve recall performance. Second, given that the lists were shorter, the switch rate variable changed such that there are now three distinct switch rates (no-switch, low-switch, and high-switch) presented in their own list to optimize the number of switch items per condition. Thus, participants learned pure lists where there was no competition between switch rates during memory search as each rate was learned and tested separately. This allowed for investigating whether competition is necessary for observing differences in recall performance between switch rates as no competition is present in Experiment 2. Additionally, this eliminated the switch type variable. Participants saw two lists of each switch rate for a total of six lists. One additional change was made in the instructions to improve the clarity of the encoding task and create a more even distribution of yes/no responses. Participants were completing the same encoding task as Experiment 1, but were now instructed to make a yes/no judgment as to whether they could find the item in the scene. 48 We also changed the recall portion of the experiment. In Experiment 1, participants could move onto the next list whenever they felt that they recalled as many words as they could remember. However, in Experiment 2, a minimum time was added to the immediate and final recall. The instructions were changed to encourage participants to continue to search their memory until at least the minimum time was up. After the minimum time was up, participants could move onto the next list, or continue to search their memory until the maximum time has passed. For immediate recall, participants had up to two minutes to recall, but would not be allowed to continue until after one minute. For final recall, participants had up to five minutes to recall, but would not be allowed to continue until after three minutes. This change was aimed at increasing recall performance by preventing participants from recalling just a few words and moving on and rather encouraging them to really search their memory. Data analysis Data analysis was identical to Experiment 1 with the following changes and additions. First, given that Experiment 2 used pure lists, the switch type variable is eliminated from analyses. Therefore, Generalized Linear Mixed-Effects Models were used to determine whether switch rate (no-switch vs. low-switch vs. high-switch) and contextual familiarity (repeated vs. novel) predicted the percent of words that participants recalled, with subject and word identity as random effects. Specifically, we ran a model that assessed the relationship between percent of words recalled and the interaction between switch rate and context familiarity. As in Experiment 1, these analyses were run for both immediate and final recall data (see Supplementary Figure 2.1 for final recall results), and all models additionally controlled for list number and list half as fixed effects. 49 Additionally, we sought to determine the extent to which participants successively recalled items shown with the same scene, or source context. During the encoding task, items (represented by words) were paired with the scene context immediately preceding. We were interested in the question: When a participant makes a local transition during recall, how often is it to the same or a neighboring context? There are three types of local transitions. Same context transitions were when participants made their next recall to an item paired with the same context as the context of the just-recalled item. Backward context transitions were when participants made their next recall to an item paired with the context immediately preceding (backwards) the context of the just- recalled item. Lastly, forward context transitions were when participants made their next recall to the item paired with context immediately following (forwards) the context of the just-recalled item. Only local transitions were analyzed as this is a fair comparison between switching to repeated vs. novel contexts (see Supplementary Analysis 2 for further investigation into local and remote transitions). Inclusion of remote transitions would allow participants switching to repeated contexts to transition between items paired with the same scene throughout in the list, which was not possible when switching to novel contexts. Therefore, inclusion of only local transitions allowed for an analysis of the same types of transitions for lists with repeated and novel contexts. For this analysis, we calculated the conditional response probabilities by local context type, similar to (Polyn et al., 2009a). For each participant, we tallied the number of recall transitions that were between items studied with the same image, the previous image (backwards transition), and the following image (forwards transition). We then conducted two different analyses: 1) To account for differences in the number of transitions each participant made, we divided each context type by each participant’s total number of recall transitions. This gave the proportion of local transitions for each participant that was then averaged across all participants in each group (see 50 Supplementary Figure 2.2). 2) Separately, to account for the fact that there were a different number of opportunities to transition to each context type (same, forward, or backward), the number of recalls from each type was divided by the total number of recall transitions possible for that type. This gave the probability of local transitions for each participant that was then averaged across all participants per type. Here, we were interested in how organization by source context differed between repeated and novel context switches, within each switch rate. Therefore, we used the lme4 package in R to run Linear Mixed-Effects Models to make such comparisons. The high-switch and low-switch rates had a different number of items between each transition (2 vs. 4 items), so therefore it would not make sense to compare these two rates, as the number of potential switch transitions differs across rates. Thus, we ran two separate models, one for the low-switch rate comparing repeated vs. novel context switches and one for the high-switch rate again comparing repeated vs. novel context switches. Additionally, this analysis was unable to be run in Experiment 1 because within a given list, recalls included items from both a no-switch and a switching (low-switch or high- switch) rate. Therefore, transitions between items are not matched. Results Encoding performance Overall accuracy was 85.3%. Accuracy and RT did not differ between repeated and novel switches (Accuracy: repeated- M = 84.61%, SE = 0.94, novel- M = 85.99%, SE = 0.64, t(158.27) = 1.13, p = .26, d = .17; RT: F(1, 165) = 1.47, p = .23, d = .09). 51 Context switching and immediate recall performance Participants recalled 50.47% of total words. There were no differences in the percent of words recalled between repeated (51.15%) and novel (49.79%) switches, t(164.97) = .51, p = .61, d = .08. This demonstrates that design changes made in Experiment 2 were successful in raising recall performance and equating overall accuracy across levels of contextual familiarity. We next tested for interactive effects of switch rate (no-switch vs. low-switch vs. high- switch) and contextual familiarity (repeated vs. novel). Replicating Experiment 1, rapidly switching to novel, z = -2.64, p = .008, partial R2 = .001, but not repeated, contexts reduced memory performance and resulted in a reliable interaction, z = 2.19, p = .03, partial R2 = .001 (Figure 2.3). There were no performance differences for low-switch between repeated and novel contexts, z = .07, p = .94, partial R2 <.001. As expected, recall performance when not switching was similar between novel and repeated contexts, z = -.003, p = .99, partial R2 = .001, since there was no competition between rates in memory search. Thus, we directly replicated that memory recall is hindered only when switching to novel contexts at a high rate. 52 Figure 2.3. Experiment 2 immediate recall performance. Error bars reflect across subject standard error. ns p > .05, ** p < .01. Context switching and recall organization Temporal clustering. Results showed that all switch rates across both levels of contextual familiarity exhibited greater than chance-level temporal clustering (ps < .001). Replicating results from Experiment 1, there was greater reliance on temporal information (i.e., higher temporal clustering) when switching to novel, compared to repeated, contexts, t(330.14) = 3.07, p = .002, d = .34 (Figure 2.4a). No-Switch Low High No-Switch Low High 0% 10% 20% 30% 40% 50% Switch Rate Pe rc en t R ec al le d NovelRepeated **ns 53 Figure 2.4. Context switching and recall organization. a) Temporal Clustering. b) Probability of Local Recall Transitions by Context for High-Switch (left) and Low-Switch (right) Rates. Here we calculated the probability of making a local transition compared to the number of possible transitions per type. The low-switch and high-switch results are unable to be directly compared due to differences in the number of items between each transition (2 vs. 4 items). The main comparison of interest is between repeated and novel context switches within switch rates. Error bars reflect across subject standard error. ** p < .01, *** p <.001. Recall transitions by context. One reason that performance differences were specific to switching at a high rate may be differences in how participants organized their memory. Therefore, this next analysis examined the proportion of recall transitions made to the same or neighboring contexts (forwards or backwards). At the high-switch rate, participants transitioned significantly more to items in the same context compared to neighboring contexts when switching to repeated, (Forwards: t(164.0) = 3.03, p = .002, partial R2 = .029, Backwards: t(164.00) = 5.33, p < .001, partial R2 = .085) and novel (Forwards: t(166.0) = 7.87, p < .001, partial R2 = .177, Backwards: t(166.0) = 9.17, p < .001, partial R2 = .226) contexts. Although all participants made most of their recall transitions to the same context, there was a significant interaction such that there was less of a difference in the probability of making same vs. forward, t(330.0) = 3.503, p <.001, partial R2 = .020, or same vs. backward, t(330.0) = 2.83, p = .004, partial R2 = .013, context transitions when ba Repeated Novel Contextual Familiarity 0.50 0.55 0.60 0.65 0.70 Te m po ra l C lu st er in g Sc or e ** Transition Type Backward Context Same Context Forward Context Backward Context Same Context Forward Context 0 0.04 0.08 0.12 0.16 Lo ca l T ra ns iti on P ro ba bi lit y ***** NovelRepeated High Switch Transition Type Backward Context Same Context Forward Context Backward Context Same Context Forward Context NovelRepeated 0.04 0.08 0.12 0.16 Lo ca l T ra ns iti on P ro ba bi lit y 0 *** *** *** ********* Low Switch 54 participants were switching to repeated, compared to novel, contexts (Figure 2.4b). This demonstrates a notable difference in how participants switching between repeated vs. novel contexts organize their recalls when switching at a high rate. For the low-switch rate, participants transitioned significantly more to items in the same context compared to neighboring contexts in repeated, (Forwards: t(246) = 6.85, p < .001, partial R2 = .163, Backwards: t(246) = 7.13, p < .001, partial R2 = .174), and novel contexts, (Forwards: t(249) = 9.68, p < .001 partial R2 = .277, Backwards: t(249) = 9.90, p < .001, partial R2 = .287; Figure 2.4b). There were no significant differences between repeated and novel switches. Experiment 2: Discussion Replicating Experiment 1, results demonstrated that memory was hindered only when switching to novel contexts at a high rate, highlighting a unique function of rapid switching on memory performance. Additionally, although there was no direct memory benefit, memory for items during repeated context switches was just as good as memory with no context switch