UC San Diego

We investigated phytoplankton growth and micro- and mesozooplankton grazing patterns in the California Current Ecosystem (CCE) during summer 2021. Two water parcels, followed over a duration of 4-5 days using satellite-tracked drogued drifter for quasi- Lagrangian experimental cycles were investigated for inshore and offshore differences. Phytoplankton growth rates and microzooplankton grazing rates were determined using the two-point dilution method, and daily Bongo tows were deployed for mesozooplankton collection, for biomass and grazing estimates based on gut fluorescence. Instantaneous rates of growth and grazing between the two cycles were µ = 0.45 (± 0.13) d-1 for Cycle 2 (inshore) and 0.60 (± 0.1) d-1 for Cycle 3 (offshore), and microzooplankton grazing rates were 0.36 (± 0.21) d-1 for Cycle 2 and 0.37 (± 0.11) d-1 for Cycle 3. Mesozooplankton contributed much less to grazing for both cycles, grazing 0.05 (± 0.02) d-1 inshore and 0.025 (± 0.006) d-1 offshore, removing 4% and 2% of phytoplankton standing stock, respectively. In both cycles, the net calculated phytoplankton growth was positive, but this was only statistically significant for the offshore region. The dominant grazers within the mesozooplankton community were not consistent between the two regions of the CCE: the three smallest size classes (0.2-2 mm) contributed the most to grazing in Cycle 2, while in Cycle 3, the dominant grazers were the two smallest size classes (0.2-1 mm). Mesozooplankton grazing showed changes related to diel vertical migration. These analyses contribute to our understanding of growth and grazing dynamics in plankton food webs, and to understanding carbon cycling in the CCE.

This dissertation explores the multifaceted roles of rare-earth doped hydroxyapatite in biomedical applications, bridging the gap between luminescence studies and osteogenic potential in the context of bone tissue engineering and regenerative medicine. In the second chapter, we look into the luminescence properties of terbium, cerium, and europium doped hydroxyapatite scaffolds when immersed in simulated body fluid (SBF) over a four-week period. Our comprehensive study reveals a consistent decrease in luminescence emission intensity across all samples, accompanied by a reduction in the concentration of rare-earth ions within the scaffolds, as confirmed by energy dispersive spectroscopy. Furthermore, fluorescence spectroscopy demonstrates the translocation of these ions into the SBF, indicating the scaffolds' partial dissolution over time. The employment of rare-earth ions as luminescence markers offers profound insights into apatite formation mechanisms, presenting significant implications for the development of safer and more durable materials in biomedical applications.Expanding upon the foundational knowledge established in the first and second chapters, the third chapter investigates the osteogenic potential of rare-earth doped hydroxyapatite scaffolds using a murine pre-osteoblastic cell line. This study assesses the effects of ytterbium, terbium, cerium, and europium doping on osteoblast differentiation, gauged by alkaline phosphatase activity and the expression of osteogenic marker genes such as Runx2, OCN, ALP, OPN, and BMP2. Our findings indicate a notable enhancement in differentiation activity with the incorporation of rare-earth elements, with europium and ytterbium doped HAp showing superior performance. Cathodoluminescence spectroscopy further corroborates these results by revealing distinct emission peaks specific to the Eu2+/Eu3+ and Yb2+ ions, underscoring the role of valence state incorporation in augmenting osteoblast differentiation. Collectively, this dissertation contributes expanding the field of biomaterials by elucidating the dual utility of rare-earth doped hydroxyapatite scaffolds in promoting osteogenic differentiation and providing luminescence-based insights into scaffold behavior in physiological environments. By combining the luminescence stability studies with the exploration of osteogenic potential, our research underscores the importance of integrating multifunctional elements into scaffold design to enhance their performance in bone tissue engineering. The insights garnered from both studies not only pave the way for the development of novel biomaterials but also highlight the potential of rare-earth elements as pivotal components in the advancement of regenerative medicine and biomedical engineering.

A calorie-restricted diet robustly improves skeletal muscle insulin sensitivity and is a cornerstone lifestyle intervention for preventing and treating clinical hyperglycemia. While studies demonstrate that insulin- stimulated activation of phosphatidylinositol 3-kinase (PI3K) and RAC-beta serine/threonine-protein kinase (Akt) are fundamental to enhanced muscle insulin sensitivity after calorie restriction (CR), the downstream signaling steps remain to be fully elucidated. We have previously demonstrated that the lysine acetyltransferases, E1A binding protein (p300) or cAMP response element binding protein binding protein (CBP) are required for insulin-stimulated glucose uptake by skeletal muscle, at both sub-maximal and maximal concentrations of insulin. Moreover, p300/CBP are phosphorylated by Akt, which increases their acetyltransferase activity. Considering these findings together, the objective of this project was to determine whether p300 or CBP are required for the ability of CR to enhance skeletal muscle insulin sensitivity. We hypothesized that combined inhibition of p300 and CBP acetyltransferase activity would prevent CR-induced enhancement of skeletal muscle insulin sensitivity. To address this objective, 8-week-old male C57BL/6NJ mice were either fed ad libitum (AL) or were fed a CR diet (60% of AL intake) for 21 days. Then, an ex vivo [3H]-2-deoxyglucose (2DOG) approach was used to assess basal and insulin-stimulated (60 μU/mL [0.36 nM]) 2DOG uptake in paired extensor digitorum longus (EDL) and soleus from fasted (4 h) mice. To define the role of p300/CBP acetyltransferase activity, paired muscles from AL- or CR-fed mice were pre-incubated for 60 minutes with either Vehicle (DMSO; [AL-Vehicle, CR-Vehicle]) or the p300/CBP acetyltransferase activity inhibitor, iP300w (25mM; [AL-iP300w, CR-iP300w]). As expected, insulin-stimulated 2DOG uptake (i.e. Insulin 2DOG minus basal 2DOG) in Vehicle incubated muscles was ~2.9-fold and ~2.0-fold higher in the soleus and EDL, respectively, from CR- vs. AL-fed mice. Remarkably, iP300w not only blocked insulin-stimulated 2DOG uptake in AL-fed mice, but it also abrogated the insulin-sensitizing effects of CR on muscle insulin sensitivity. To further define the importance of p300 and CBP to CR-induced enhancement of skeletal muscle insulin sensitivity, we studied mice with tamoxifen-inducible and skeletal muscle-specific knockout of CBP and heterozygous (HZ) loss of p300 (referred to as, i-mPZ/mCKO) or knockout of p300 and HZ loss of CBP (referred to as, i-mCZ/mPKO). In these mice, the AL diet or 21 days of CR was initiated 3 weeks after ~10 week-old mice were dosed with tamoxifen (5 consecutive days, 2 mg/day); Cre recombinase-negative littermates, who were also dosed with tamoxifen, were the experimental controls (referred to as wildtype [WT]). Interestingly, oral glucose tolerance and EDL insulin sensitivity were comparably enhanced by CR in WT and both i-mPZ/mCKO and i-mCZ/mPKO mice, thus suggesting that just one allele of p300 or CBP is sufficient for CR to enhance skeletal muscle insulin sensitivity. Taken together, these results demonstrate that p300 or CBP are essential for the ability of calorie restriction to enhance skeletal muscle insulin sensitivity.

“How do I bring my full self into what I do?” This was the question I began asking myself at UCSD. I received the note from my movement professor, Stephen Buescher, “Your blackness is something to embrace and is, even, playable. Why do you not tap into that more often? Do you not feel comfortable doing that here?” Growing up in predominantly white schools, I learned code switching at an early age. It had not dawned on me that code switching had manifested in my work. I only allowed my vernacular to show up in work that was specifically for people that looked like me.

In my final UCSD show, I was fortunate enough to play Juliet in Romeo and Juliet. Now, for a bit of context, my only Shakespeare experience prior to graduate school was a poorly done high school rendition of A Midsummer Night's Dream. I played a fairy; I could not tell you which. Needless to say, I was terrified. I am thankful for the training I received from Marco Barricelli and Ursula Meyer. I found, however, that in the beginning I made acting choices based on how I thought Juliet should sound and act. What I found throughout was someone that looked, sounded, and acted much more like me. Juliet is perceived as innocent and naïve, but is much smarter and more powerful than expected. She enjoys control and stability in her life and when she’s not in control it gets a little messy. I am talking about Juliet remember? Not the actor that did not stop looking at her script between every scene until show #6. Definitely not talking about me. She is passionate and playful and, when I play Juliet, she is Black. She may roll her neck, pucker her lips, and flick her wrist to accentuate her words. Maybe she does code-switch when she is in mixed company or maybe she doesn’t. I learned through this process that that is okay. In fact, it is great because who I am can also be who Juliet is. I can live in Shakespeare’s canon, and I can bring all of me into every role I embody.

“You’re going to do this.” Marco Barricelli, my Shakespeare professor, once said to me. He meant Shakespeare, but what I took from that is that I can do this, I’m going to do this, and I can do it in my own way.

This dissertation comprises two papers that examine the psychological mechanisms underlying consumer decision making. In the first paper, I investigate consumers’ behavior surrounding a novel payment method, buy now, pay later (BNPL), and show that paying with BNPL increases consumers’ purchase of products above their usual (“ordinary-for-self”) price, and that consumers also prefer to use BNPL when paying for purchases above their usual price. I find that this is due, in part, to the way BNPL is able to transform the mental categorization of purchases above one’s usual price in a product category into prices that are more acceptable to one’s financial self-control. In the second paper, I examine buyers’ requests for revisions of final deliverables on a large online freelance platform, and find that buyers are more likely to request platform-vetted high-quality female sellers revise their final deliverables than equivalent male sellers. I discuss and test for alternative explanations for this finding, and propose that this pattern emerges because people have overly high expectations of high-achieving women but not men, driven by the assumption that successful women have advanced despite repeatedly being held to higher standards.

As the demand for high-speed, reliable wireless communication among interconnected devices rises, the need for robust next-generation wireless MIMO systems becomes crucial.This dissertation is motivated by the need to address the challenges inherent in the development of such systems, specifically focusing on two key aspects: (i) Robust block-sparse mmWave channel modeling and (ii) Robust detection in the presence of few-bit MIMO systems. Central to this thesis is the integration of model-based methods with deep neural network (DNN)-aided approaches, leveraging the synergy between these two paradigms to enhance system performance and mitigate the impact of model inaccuracies and mismatches.

The first part of this dissertation focuses on the spatial modeling of mmWave channels, to capture the heterogeneous scattering behavior, through block-sparse signal recovery.Despite the promise of block-sparse signal processing for channel modeling in the angular domain, a key challenge is block-patterned estimation without knowledge of block sizes and boundaries. This work propose a novel total variation sparse Bayesian learning (TV-SBL) method for block-sparse signal recovery under unknown block patterns. Unlike conventional approaches that employ block-promoting regularization on signal components, this method introduces two classes of hyperparameter regularizers for the SBL cost function inspired by total variation (TV) denoising. The first class relies on a conventional TV difference unit, allowing iterative SBL inference through convex optimization, thus facilitating the use of various numerical solvers. The second class integrates a region-aware TV penalty to penalize signal and zero blocks differently, thereby enhancing performance. An alternating optimization algorithm based on expectation-maximization is derived for computationally efficient parallel updates for both regularizer classes. Going beyond model-based methods, this work also presents a basis for extension to DNN-aided block-sparse signal recovery for 1-D and 2-D signals.

The second part of this dissertation focuses on designing detection algorithms for signal recovery in few-bit MIMO systems, beginning with a detailed analysis of one-bit MIMO systems. This begins by analyzing the smoothness and convexity of the one-bit likelihood function, based on the Gaussian CDF, for signal recovery. This culminates in an improved gradient descent (GD) algorithm for one-bit MIMO, and ensuing convergence analysis. The accelerated GD method is applied to one-bit MIMO recovery, further improving convergence. The analysis is extended to an effective surrogate function for the Gaussian CDF, i.e., the logistic regression (LR), explaining the enhanced performance when utilized as a surrogate likelihood. Constrained optimization, incorporating detection from a finite M-QAM constellation, is addressed by the introduction of a \textit{learnable} Gaussian denoiser to project detected symbols onto the M-QAM subspace.

Another class of DNN-aided regularizers is proposed for one-bit MIMO, utilizing a regularized gradient descent update. A novel constellation-aware loss function is incorporated to tailor the DNN loss function to M-QAM symbol recovery. The key utility of a generalized DNN-aided GD update is for detection in mmWave channels, where there is a higher contrast in per-user channel powers, presenting a challenge for joint multi-user detection. Leveraging a general parametric DNN structure enables the development of a novel hierarchical detection training algorithm, ensuring network design for equitable detection in mmWave channels, where users with higher channel powers experience improved recovery performance.

The final part of this dissertation extends the research to two-bit MIMO detection. In few-bit MIMO systems like the two-bit MIMO receiver, the quantization noise exhibits distinct characteristics positioned between the fully saturated one-bit scenario and the independently additive noise observed in higher resolutions. However, limited research has focused on accurately characterizing this unique quantization noise profile. These properties of quantization noise, along with the constraints of optimization for MIMO signal recovery, make DNNs ideally suited for the signal recovery. The DNN-augmented receiver algorithm developed attempts to learn this noise behavior to dequantize the signal, without the need for explicit analytical characterization, thereby enhancing signal recovery in few-bit MIMO systems.

First Generation College Students (FGCS) account for one third of all doctoral degree recipients in the USA, however, very little is known about their educational experiences. Using a narrative inquiry approach, this study uses a Cultural Proficient framework to explore the barriers that prevent FGCS from pursuing post-undergraduate opportunities, how they’ve overcome them and provide next steps for educators and leaders. In an attempt to expand asset-based research, this dissertation applies Resiliency Lens, Critical Race Theory and Socialization Lens to the experiences of six participants. Research on FGCS suggests some of the challenges FGCS encounter in college is more debt, lack of guidance in the admissions process, feelings of disconnect in their communities back home and in higher education. Stories, however, can be an important tool in helping students reflect on their experiences, to stay engaged in the classroom, build agency, develop growth mindsets, and create important networks with formal and informal leaders to succeed in academia.

This dissertation presents three essays in the design of randomized experiments in economics. Chapter 1 "Testing for Underpowered Literatures" proposes a novel estimator consistent for the expected number of statistically significant results that a set of experiments would have reported had their sample sizes all been counterfactually increased by a chosen factor. An application to randomized controlled trials (RCTs) published in top economics journals finds that doubling every experiment’s sample size would only increase the power of two-sided t-tests by 7.2 percentage points on average. Chapter 2 "Linear Estimation of Global Average Treatment Effects" studies the problem of estimating the average causal effect of treating every member of a population, as opposed to none, using an experiment that treats only some. We provide recommendations for experimental designs and estimation procedures. Chapter 3 "Evaluation of a Teacher Training in Uganda: Specifications for Endline" is a midline report for a multi-year field experiment in Uganda. The experiment evaluates the impact of a general skills teacher training program in rural Uganda. In the trial, we offer the program to a sample of 640 teachers in 39 secondary schools. The chapter describes data collected so far, proposes hypotheses and estimands for endline, and demonstrates the precision of these methods by running them on midline data.

This thesis explores my experience directing X by Alistair McDowell as a lens through which I identify, meditate on, and codify directorial challenges and experiences that feel unprecedented, but are in fact essential and deeply rooted parts of my process.

Analyzing the dynamics underlying various neural phenotypes is key to understanding the systems, cellular, and subcellular processes that guide them. However, while there exist many robust methods for object detection in biological settings, such as SLEAP and CellPose, there lacks a universal approach for linking these detections through time. Specifically, given the complex, variable imaging conditions that most biological videography occurs under, current deep learning approaches which exploit only local information may not be suitable for this setting. This is because biological videos contain edge-cases that are not typically seen in the applications that most multiple object tracking (MOT) approaches are designed for such as pedestrian and automotive tracking. Here, we introduce DREEM (DREEM Reconstructs Every Entity’s Motion), a deep learning framework which leverages a transformer-based architecture to directly learn the associations between objects in a large temporal context. We demonstrate that DREEM enables the training of state-of-the-art models for biological MOT. Then, we show that DREEM is sample efficient and can transfer seamlessly across a variety of diverse settings, enabling use in a wide variety of fields. Our code and pretrained models will be released at https://github.com/talmolab/biogtr