Mutations in the SHANK3 gene have been identified as a risk factor for autism spectrum disorder (ASD), with known effects on synaptic structure and function. Recent work has started mapping SHANK3-related changes in non-human primate models, but the cell-type-specific molecular composition of synapses–especially local mRNA populations at postsynaptic sites–remains poorly understood. A long-term goal of this project is to develop an AAV-based, single-synaptosome RNA sequencing (SSS) pipeline that can profile synaptic mRNAs from defined neuron types. As a first step toward that goal, this thesis focuses on building and validating compact synaptic-targeting tools that can later be used in AAV constructs. I designed two fluorescent reporter constructs, pABP-18 and pABP-53, by fusing TagRFP to short N-terminal motifs derived from the palmitoylation domain of the postsynaptic protein pABP-L. These constructs were cloned into a pCAGGs backbone, expressed in primary mouse neurons, and imaged together with the postsynaptic marker PSD-95. To test whether these motifs are able to drive synaptic localization on their own, I quantified PSD-95 enrichment at pABP-positive puncta compared to nearby dendritic background using Fiji. A non-targeted pCAGGs-RFP control showed no enrichment (~1.0x), while pABP-53 produced moderate enrichment (1.3-2.4x). Surprisingly, the minimal pABP-18 construct showed the strongest effect, with PSD-95 levels ~3.5-4.2x higher at pABP puncta than in surrounding dendrite. These results show that short, palmitoylated pABP fragments–especially pABP-18–are sufficient on their own to drive robust postsynaptic targeting of a fluorescent reporter. The AAV implementation, in vivo validation, and SSS pipeline remain in progress. In the next stage of the project, these validated motifs will be placed into enhancer-driven AAV constructs to test whether they can drive cell-type-specific synaptic labeling in vivo. Because AAVs have size limits, having a compact sequence such as pABP-18 that reliably targets synapses is a major advantage. The goal is to use these AAVs to label postsynaptic terminals from defined neuronal populations and eventually isolate those synapses for single-synaptosome RNA sequencing, with the long-term goal of applying this approach to SHANK3 mutant circuits.

Alzheimer’s Disease (AD) is a neurodegenerative disorder characterized by progressive cognitive decline. Current therapeutics for AD only temporarily alleviate the symptoms of cognitive impairment, and broadly impact multiple brain regions without targeting localized cell populations that specifically drive AD pathology. Recent research suggests that the entorhinal cortex may be one of the first brain regions impacted by AD pathology. Layer 2 (L2) neurons within the entorhinal cortex are particularly vulnerable to neurodegeneration, as this layer exhibits the initial and most severe cellular degeneration of any other brain region in AD patients. This identifies these neurons as a reasonable target for studying AD pathology. This is a proof of principle study with the goal of identifying and validating glutamatergic L2/3 cell-type specific entorhinal cortex candidate enhancers. The study aims to address a knowledge gap in current AD research as there is currently a primary focus on Aβ accumulation and tau hyperphosphorylation as causal drivers of the disease pathology, even though dysfunction of the entorhinal cortex is one of the first signs of AD. Cell-type specific targeting of neurons in this region will provide a better understanding of the role of neural circuits containing L2 neurons, which are known to be particularly vulnerable to neurodegeneration. Successful identification and validation of a cell type-specific L2 entorhinal enhancer may lead to therapeutics aimed at preserving cognitive function using targeted adeno-associated virus (AAV) delivery.

Early life adversity (ELA) is a major risk factor for the development of psychiatric disorders, including major depressive disorder (MDD), as it often leads to persistent dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis. However, the mechanisms linking ELA to HPA axis dysregulation are not entirely understood. This thesis proposes a model in which ELA induces long-term HPA axis dysregulation primarily through reductions of the neurosteroid allopregnanolone (ALLO) in the ventral hippocampus. Decreased ALLO weakens GABAergic inhibition of corticotropin-releasing hormone (CRH) neurons, promoting increased CRH release from the hypothalamic paraventricular nucleus (PVN) and, subsequently, increased glucocorticoid secretion from the adrenal glands. This leads to increases in the intensity, density, and C4S sulfation of perineuronal nets (PNNs) surrounding parvalbumin-containing (PV+) interneurons in the ventral hippocampus, stabilizing these neurons and altering their firing patterns. This further disrupts the inhibition of CRH neurons, reinforcing HPA axis hyperactivity, which is characteristic of MDD and postpartum depression (PPD) in adulthood. This thesis presents a series of experiments designed to test each step of the proposed pathway using a rodent model of ELA. This thesis also emphasizes the importance of understanding the mechanisms underlying ELA-induced psychiatric disorders in adulthood by drawing a comparison to PPD, a condition for which recent advances in mechanistic understanding have allowed for the development of effective treatment. This comparison suggests ELA-induced depression may represent a mechanistically distinct subtype of MDD, and thus requires a comprehensive understanding to best treat those affected. Ultimately, this work proposes a novel mechanistic model linking ELA to adult psychiatric disorders and emphasizes the importance of understanding unique neural pathways to effectively treat individuals.

In motor skill learning, two different types of prediction errors have been identified: Sensory prediction errors (SPEs) and reward prediction errors (RPEs). SPEs occur when the predicted sensory consequence of a movement does not align with the observed consequence, while RPEs occur when the result of said movement deviates from the desired goal. These errors were previously believed to be processed independently, with SPEs driving implicit sensorimotor adaptation in the cerebellum and RPEs driving explicit decision-making in the striatum. Recent research, however, has suggested otherwise, finding that SPEs influence decision-making by elevating risk tolerance, while RPEs influence implicit adaptation by increasing recalibration. What remains unclear is whether RPEs influencing SPE-driven adaptation extends to the neuroanatomical level. To explore this, we utilized the visuomotor reach adaptation task, which enables isolation of SPEs and RPEs and their subsequent effects on implicit adaptation behavior. Using fMRI, we asked if cerebellar activity associated with SPEs is modulated by RPEs when both error signals are experienced in tandem compared to just SPEs alone. From this, we replicated previous behavioral findings that RPEs are insufficient to drive implicit adaptation while SPEs can regardless of RPE presence. We also observed a numerical trend consistent with literature that the combination of SPEs and RPEs drive greater adaptation than SPEs alone. While strictly exploratory and below threshold due to a limited sample size, we observed preliminary evidence of cerebellar activity associated with RPEs when both error signals were experienced. This pattern may suggest that RPEs modulate SPE processing, which could explain the enhanced adaptation when both error signals are experienced together. These results overall support the viability of our MRI-adapted task for future prediction error imaging studies, and that the current trends may manifest under more conventional thresholds with a complete sample size.

Preterm birth is a global neonatal concern, being one of the leading causes of death in children under 5 years old along with birth defects and Sudden Infant Death Syndrome (SUID) (World Health Organization (WHO), n.d.; March of Dimes, n.d.). Based on one statistical report about preterm birth in the United States, the rate of preterm birth increased from 9.6% to 10.5% from from 2014-2021, decreasing to 10.4% in 2022 (March of Dimes, n.d.; Martin & Osterman, 2019). Over the past couple of decades, along with the maintained rates of preterm birth, medical advancements have contributed to increased survival rates of preterm infants. According to a US study, the survival rates until discharge or one year after birth of preterm infants born at weeks 22-28 of gestation increased from 76% in 2008-2012 to 78.3% in 2013-2018 (Bell et al., 2022). This difference was even more stark for earlier gestational ages; in preterm infants born at 22 weeks, the survival rate increased from 6.6% to 10.9%. With consistent rates of preterm birth and increasing survival rates, it is pertinent to continually work towards a better understanding of the complications that are associated with preterm birth, including neurodevelopmental impairments. Studies have found increased risk for neurodevelopmental disorders (NDDs) in all stages or prematurity, even late prematurity, when compared with term-born counterparts (Nivins et al., 2026; Kong et al., 2025). The exact increase in risk varies between the different disorders, but one study found that, in considering a number of behavioral and emotional disorders, the hazard ratios ranged from 1.20 to 10.7 (Kong et al., 2025). In addition, there are also observed sex-differences in the presentation of multiple NDDs, and these differences often lead to disparities in diagnosing for women. As of yet, the exact mechanisms behind these differences aren’t completely known (Kim et al., 2026). This thesis seeks to explore existing knowledge surrounding the signaling pathways by which preterm birth mediates structural alterations in neural networks that are associated with NDDs. The present work will go through typical neurodevelopment, provide an overview of preterm birth and neurodevelopmental impairments and disorders well known for their sex differences, detail signaling pathways associated with inflammation and neuroprotection, and finish with an exploration of recent research regarding the role of placental endocrine function in sex-differences seen in NDDs.

The ability to perform body movements, especially during unexpected environmental changes, requires the transformation of sensory information into precise motor movement. Drosophila melanogaster can be leveraged as a model organism to investigate rapid motor control, namely wing-steering muscle activity on a sub-millisecond time interval. This level of motor precision can be credited to their haltere organs which serve as gyroscopic sensors, detecting the rotational movement, and stimulating equilibrium reflexes to maintain flight stability. Recent connectomics work in the organization of the haltere sensory afferents shows that these neurons project both directly onto wing-steering muscles but also indirectly via interneurons. One subclass of interneurons, the contralateral haltere interneurons (cHINs), have been long established anatomically. Connectomics work now provide predictions about the role of these neurons in flight, yet these hypotheses remain untested. Here, I harness the power of genetic drivers, optogenetics, and quantitative behavior to study the behavioral effects of these interneurons along with the newly identified contralateral wing interneurons (cWINs). Results first indicate that optogenetic activation of the cHINs considerably decreases body saccade magnitude and duration while no significance was observed under cWIN activation. More notably, both cHIN and cWIN activation substantially increase compensatory head yaw magnitude following peak angular body velocity. These saccade performances in flies, i.e., rapid turns, can further prompt us in understanding the visual saccades we perform as primates and the implications of disynaptic connections to motor circuitry.

Schizophrenia is a chronic psychiatric disorder characterized by positive and negative symptoms. While recent research has extensively explored working memory deficits in schizophrenia, few studies have investigated how continuous memories evolve within working memory. This study aims to address this gap by investigating how task performance errors and attractor dynamics (stable memory states within mnemonic space) in visual working memory relate to symptom severity in schizophrenia. We used a continuous color report task with varying memory loads (1-3 items) and delay durations (500ms, 4000ms) in 47 individuals with schizophrenia and 33 healthy controls. Results demonstrated that both memory load and delay duration significantly increased angular error, with each additional memory item increasing error by approximately 7 degrees. Mean angular error significantly predicted symptom severity as measured by the Positive and Negative Syndrome Scale (PANSS) (p = 0.0207). Analysis of bias patterns across color space revealed greater variability in individuals with schizophrenia, suggesting possible disrupted attractor dynamics that warrant further research.

Adult onset hypothyroidism, characterized by a decreased level of circulating thyroid hormones, is robustly associated with depressive- and anxiety-like phenotypes/features in both clinical and rodent populations, implicating thyroid hormone signaling as a broad modulator of affective states and behaviors. However, how it contributes to shaping behavioral and neural responses to ethologically relevant stressors, such as social defeat, remains poorly characterized. Here, male C57BL/6J mice with propylthiouracil (PTU)-induced hypothyroidism exposed to acute social defeat stress paradigm enabled comparisons of behavioral coping strategy and whole-brain c-Fos as a neural activity proxy between controls. Manual behavioral scoring for active and passive coping strategies revealed no significant differences in retaliation and fleeing responses, or their compositions and switching dynamics, suggesting appropriate retention of defense under acute social threat. Individual variation in coping strategies correlated with c-Fos activity in most notably hypothalamic and amygdalar nuclei, regardless of condition. Whole-brain iDISCO+ c-Fos mapping identified nominally significant brain region activity differences that partake in social behaviors and stress responses (e.g. periaqueductal gray, basolateral and basomedial amygdalar nuclei, dentate gyrus), though no regions survived FDR correction. Cross-validation of the periaqueductal gray using traditional immunohistochemistry and FIJI-based cell counting elicited a similar weaker neural recruitment trend in PTU mice and partial correlation with individual-based active coping. Together, these findings provide evidence for behaviorally intact but preliminary diverse brain nodes where thyroid hormone status and social stress processing may interact, offering insight for future exploration of hypothyroidism may be shaping stress-circuitry and vulnerability to social adversity.

In this thesis, I developed and evaluated a deep learning approach for classifying estrous stage from vaginal cytology images in a lab-specific rodent dataset. Using a VGG16-based convolutional neural network trained with class balancing, augmentation, and cross-validation, the final model achieved stable three-class performance (~65% accuracy), substantially improving over earlier baselines. However, the most important findings were not about overall accuracy alone. Model errors were concentrated in biologically adjacent stages, particularly between diestrus and proestrus, and these same regions showed the highest levels of annotator disagreement. This suggests that classification difficulty is not just technical and that it reflects underlying biological ambiguity and limitations in how continuous endocrine states are represented with discrete labels. I also tested whether adding short-term temporal context through adjacent-day image pairs could improve classification, but this approach did not resolve ambiguity and instead introduced a strong bias toward diestrus. The results show that estrous-stage classification is shaped by an interaction between model design, annotation structure, and the dynamic nature of the biology itself. Rather than building a perfect classifier, this thesis demonstrates that classifier performance can be used as a tool to identify where biological boundaries are most difficult to define and where current labeling frameworks may fall short.

Addiction remains a complex and persistent public health crisis, often framed through the Brain Disease Model of Addiction (BDMA), which conceptualizes addiction as a chronic, relapsing brain disorder rooted in pathological changes. While the BDMA has contributed valuable insights into neurobiological mechanisms like dopamine dysregulation, prefrontal cortex impairments, and stress system activation, it has been critiqued for reducing addiction to individual pathology and sidelining social and environmental factors. This thesis critically evaluates the BDMA’s core assumptions– not to reject the model, but to extend it– arguing that addiction is best understood as a multi-systemic condition shaped by dynamic interactions between neurobiological, social, environmental, and stress-related systems. Drawing on neuroscience, epigenetics, and stress physiology, this thesis explores how these systems co-regulate each other through recursive feedback loops that drive both the biological and behavioral dimensions of addiction Rather than viewing social and environmental factors as peripheral risk contributors, this work positions them as core mechanisms that actively shape addiction’s neurobiology and behavioral patterns. By integrating these perspectives, the thesis moves beyond linear disease frameworks, proposing a dynamic systems model that aims to reflect the complex realities of addiction and offer a new pathway for research and intervention.

Bisphenol A (BPA), an endocrine-disrupting chemical found in numerous consumer products, has raised concerns about its potential neurotoxicity. We employed an integrative multi-omics approach, using the Neuro-2a mouse neuroblastoma cell line, to investigate the neurotoxic effects of BPA at human-relevant concentrations. This pilot study specifically aimed to determine biologically relevant doses of BPA for the Neuro2a model and characterize an in-vitro neurotoxicity profile. Initial findings indicated that high-dose BPA (100-500 µM) induced significant cytotoxicity, while lower doses (0.1-1 nM), representative of human exposure levels, did not significantly impact cell growth. Metabolomic analysis of chronic BPA exposure at such low doses revealed significant disruptions in energy homeostasis and lipid metabolism. Transcriptomic analysis of chronic low-dose BPA exposures similarly implicated genes involved in metabolic pathways, as well as in methylation and Wnt signaling. In conjunction, these findings highlight potential risk factors for neurological dysfunction and potentially neurodegenerative disorders such as Alzheimer’s disease and Parkinson’s disease given similarly implicated dysregulation in these pathways. Overall, our findings support the idea that BPA may be a significant risk factor for neurological dysfunction and potentially neurodegenerative disorders by affecting key metabolic pathways, even at low, human relevant doses. The final aim of this work was to analyze EDC exposure as a mechanism of environmental racism and slow violence, drawing upon a specific case study involving Indigenous communities living near the Athabasca tar sands region of Alberta, Canada.

Working memory is vital to human cognition, providing us with the ability to temporarily hold and work with information. Deficits in working memory can manifest as attentional issues such as those present in cognitive disorders like ADHD. Higher working memory capacity (WMC) has previously been shown to be protective against the negative effects that auditory distractions can have on our cognition, meaning that the development of working memory has a direct effect on one’s susceptibility to auditory distractions in adulthood. This study specifically investigated how chronic noise exposure during the development of working memory affected one’s susceptibility to different kinds of auditory distractions in adulthood. College-aged participants who grew up in noisy city environments with consistent chronic noise exposure and quieter rural/suburban environments with less chronic noise exposure were recruited and completed an online auditory working memory complex span task with and without auditory distraction. Data analysis focused on participant distractibility in the form of worsened task performance when an auditory distractor was present. The findings showed no consistent significant differences in distractibility between those who grew up in the presence of more chronic noise exposure in cities and those who grew up in quieter environments. All participants, regardless of developmental environment, showed greater distractibility when less familiar, more artificial auditory distractors were present, hinting at a possible desensitization mechanism to familiar auditory distractors. In addition, higher baseline working memory capacity (WMC) was shown to be modestly correlated with increased susceptibility to auditory distractions, contrary to previous literature.

Sensorimotor learning involves the dynamic interplay between implicit recalibration and explicit strategic control. While prior work has distinguished between cognitive strategies such as mental rotation (MR) and response caching (RC), it remains unclear how task structure shapes their engagement during adaptation. This study examined how target set size and perturbation schedule influence strategy use in a visuomotor rotation task. Eighty-five participants performed center-out reaching movements under conditions varying in the number of target locations (2 vs. 12) and the rotation schedule (abrupt vs. gradual). Response times and movement angles were analyzed across distinct learning phases.

As predicted, low set size conditions exhibited substantial reductions in response time across training, consistent with a shift from MR to RC strategies. High set size conditions maintained elevated response times, suggesting sustained reliance on parametric computation. All groups improved in movement accuracy over time; however, larger aftereffects were observed in the low set size groups, aligning with use-dependent learning. Rotation schedule had comparatively smaller effects on both response times and accuracy.

These findings underscore the role of environmental structure in guiding cognitive strategy selection during motor adaptation and suggest implications for optimizing performance in applied settings such as neurorehabilitation and brain-machine interface design.

The medial entorhinal cortex (mEC) and hippocampus (HPC) are both critically involved in the formation and maintenance of cognitive maps. Although historically studied during spatial navigation, such mental maps may be an instance of a common mechanism for organizing knowledge across many domains. For example, the hippocampal formation encodes not only locations in navigational space but also 'locations' in other continuous task-relevant variables, such as visual space, sound frequency, and sensory evidence, among others. Yet, whether or how cognitive maps conjointly represent dimensions across different domains is less well understood. A recent study demonstrated integration of spatial location and sensory evidence in the HPC, but whether there is similar joint encoding across these dimensions in mEC, a major input to HPC, is unknown. Understanding whether such conjunctive coding exists in the mEC is essential for uncovering how cognitive maps are constructed and utilized to guide complex behaviors. To close this knowledge gap, we use an established virtual-reality navigation decision-making task – the accumulating towers task (ATT) – in which mice integrate pulsatile visual evidence during virtual navigation to make a choice about which way to turn to obtain a reward. Through acute electrophysiological recordings (via Neuropixel probes), we demonstrate that mEC neurons encode both navigational space and accumulated evidence conjointly during the ATT, with firing pattern and speed analyses underlining the possibility that this population of entorhinal conjunctive neurons includes grid cells. Our evidence of conjoint encoding in the entorhinal circuit supplements the growing body of work examining cognitive maps within the hippocampal formation that not only guide us through physical space, but also structure learning, memory, and decision making.

Fluctuations in estradiol are known to modulate social behavior, cognition, and affect, yet the circuit-level mechanisms through which estrogen shapes behavioral output remain poorly understood. This thesis investigates the role of estrogen receptor alpha (ERα) signaling in the medial prefrontal cortex (mPFC) in regulating behavior and behavioral persistence. Using a combination of ovariectomy with estradiol replacement in mice, region-specific ERα knockdown, and ERα overexpression, this work tests the necessity and sufficiency of mPFC estrogen receptor signaling across social and non-social behavioral domains. Baseline experiments demonstrate that estradiol enhances behavioral persistence and structured engagement across both social interaction and exploratory contexts. ERα knockdown in the mPFC reveals that while local receptor signaling contributes to the organization of social behavior, estradiol-dependent increases in persistence in the non-social domain are largely preserved, suggesting that estrogen-dependent behavioral persistence may be mediated by distinct circuit mechanisms, with mPFC ERα signaling preferentially regulating social behavior, and other circuits supporting persistence in non-social contexts. In contrast, ERα overexpression in mPFC selectively enhances the temporal stability of behavior without significantly altering social preference, indicating that increased receptor availability is sufficient to modulate behavioral persistence but not social preference. Together, these findings support a model in which estrogen acts as a domain-general modulator of behavioral stability, with mPFC ERα signaling contributing specifically to flexibility of behavior and behavioral persistence. These results further suggest that estrogenic modulation of prefrontal circuits may operate within a nonlinear, potentially inverted U-shaped framework, providing new insight into how hormonal state dynamically shapes neural computation and behavior.

Adeno-associated virus (AAV) vectors are widely used to deliver genes to the central nervous system, but they still come with important limitations. One major issue is that they can activate the immune system and, in some cases, contribute to neuronal toxicity. Recent studies have shown that this response can be triggered when Toll-like receptor 9 (TLR9) detects unmethylated CpG motifs within the AAV genome. In this study, I analyzed CpG frequency and distribution across wild AAV serotypes, recombinant AAV components, and other DNA viruses. CpG enrichment was first assessed based on observed versus expected CpG frequency, and I further used the RF3 metric to evaluate CpG content in the context of TLR9 immunostimulatory potential. I compared wild AAV genomes to randomized control sequences that preserve nucleotide composition but remove biological structure. In addition, wild AAV genomes showed reduced RF3 values, suggesting a depletion of CpG motifs associated with TLR9 activation. This pattern was consistent across all serotypes and was statistically significant. When I analyzed different regions of the AAV genome, I found that CpG distribution is not uniform. Coding regions (rep and cap) showed consistent and constrained RF3 values, while inverted terminal repeats (ITRs) showed much higher variability. Then, I compared these patterns to recombinant AAV components and a panel of DNA viruses. I also included DNA viruses as a comparison group to see whether the patterns seen in AAV were part of a broader viral trend. This was important because different DNA viruses have different life cycles, genome structures, and relationships with the host. Recombinant AAV elements showed a wide range of RF3 values depending on the type of component. In particular, transgenes often had higher CpG content, suggesting that they may contribute more strongly to predicted TLR9 activation in engineered vectors. In contrast, oncogenic DNA viruses had significantly lower RF3 values than nononcogenic viruses. This suggests that CpG depletion may be related to viral persistence and immune evasion, especially in viruses that remain in the host long term. Importantly, this pattern did not seem to be explained by genome length or GC content alone. Although CpG frequency can be affected by overall nucleotide composition, viruses with similar GC content still showed different RF3 values. Oncogenic viruses also continued to show lower CpG enrichment even when compared within similar GC ranges. Overall, these results suggest that wild AAV genomes may have been shaped by evolutionary pressure to reduce CpG motifs and avoid TLR9 mediated immune detection. In contrast, recombinant AAV vectors are built from different engineered parts, some of which may reintroduce CpG high sequences. These findings show why CpG content should be considered in AAV vector design and may help support the development of safer and more effective gene therapy strategies.