2025 Neuroscience PhD Sample Statements

EKATERINA A. KOULAKOVA

e.koulakova@nyu.edu | Huntington, NY

I seek to understand how the brain’s biological constraints shape cognition. Neural systems don’t merely observe the environment; they actively interpret it, constructing meaningful representations based on past experiences and structural biases to form coherent models of the world with which we interface. My goal is to understand how these systems and their architectures implement statistical processes underlying learning, memory, and decision making.

As a physics major, my drive to describe natural phenomena through quantitative organizing principles led me to join Dr. Kolmogorov’s computational materials science lab. There, we applied neural networks to predict stable chemical compositions not yet synthesized experimentally. To support this project, I compiled datasets to train our neural network (NN) model. This required me to conduct a comprehensive literature review on the physical properties of alkali-tin compositions. Using experimentally derived properties, I trained our NN on these compositions to identify the lowest free-energy states or the most stable configurations. To enhance model accuracy, I helped derive a theoretical correction term for the free energy equation to simulate substitutional defects in crystalline solids, a crucial factor in assessing thermodynamic stability [1]. This analysis enabled us to evaluate whether potential compositions lay on the free-energy convex hull, indicating stability. Based on these findings, I proposed new structures for stability screening that had not yet been experimentally synthesized. Our evolutionary searches ultimately uncovered two previously unknown stable phases demonstrating the model’s predictive strength [2]. This work introduced me to research fundamentals - data analysis and literature reviews - and allowed me to contribute to analytical derivations grounded in thermodynamic principles, essential for interpreting and validating our NN results.

Intrigued by the effectiveness of neural networks in extracting meaningful patterns from complex data, as an independent summer project, I designed a convolutional neural network (CNN) to examine how genomic constraints shape sensory perception. Although the CNN I developed and optimized ultimately fell short of predicting gene expression, this experience deepened my understanding of key neuroscience concepts. It also honed my skills in neural network optimization, coding in MATLAB, and data management.

To understand how the brain’s architecture implements perception-related algorithms, I joined the labs of Dmitry Rinberg and Shy Shoham at the NYU Neuroscience Institute. My current project addresses the "figure-ground separation" problem in olfaction, investigating how signals are extracted amid dynamic background odors. I am focused on the single glomerulus, a fundamental unit of olfactory encoding, where olfactory receptor neurons excite mitral cells (MCs) that relay sensory information to higher brain regions. To explore the encoding mechanisms of this principal circuit motif, I employ optogenetics to selectively stimulate a single glomerulus while concurrently using two-photon (2P) imaging to monitor the activation of its associated MCs while mice are engaged in a stimulus detection task. Thus far, I have developed a paradigm that demonstrates mice can detect single glomerulus stimulation at various latencies within the sniff cycle. My next objective is to assess whether this detection remains stable when a background odorant is introduced. I hypothesize that once the mouse learns to detect this minimal stimulus, its detection rate will remain high despite the presence of a masking odorant. Overall, this project explores the limitations of current models of temporal olfactory encoding by determining whether stimulus detection is correlated with its MC activation or if additional pathways are recruited through feed-back and learning. In independently executing, managing, and refining our experiment has allowed me to gain skills in training mice, conducting cranial-window surgeries, data analysis with Python and MATLAB, and an in-depth

understanding of 2P microscopy. Over the past two years in the lab, I have deepened my interest in systems neuroscience and learned to approach challenges with persistence and incremental progress, embracing the nonmonotonic yet rewarding nature of research.

During my graduate studies, I aim to explore how the brain’s biological constraints shape learning, memory, and decision-making through an integrated computational and systems neuroscience approach. Cornell’s Neurobiology and Behavior program stands out to me for its strong faculty expertise aligned with my research interests.

Dr. Weinan Sun’s research on how the hippocampus learns a set of discrete states and transition rules aligns with my goal of uncovering deterministic principles that guide behavior. Implying a predictive relationship between the current state, a sensory input, and a subsequent state greatly appeals to me, especially in a naturalistic behavioral paradigm reducing constraints on behavioral output. Furthermore, his assessment that learned manifolds could be applied to novel conditions echo insights that I gained during my legal internship, as discussed in my personal statement.

Building on my experience with two-photon imaging, I am drawn to Dr. Ian Elwood’s research on decision-making and working memory, especially the computational role of dopamine in the basal ganglia. I am particularly intrigued by the possibility of uncovering nuances in dopamine’s function beyond reward prediction error, perhaps exploring its role in planning behaviors within the dorsomedial striatum. To this end, I appreciate the lab’s use of virtual reality to bypass stringent behavioral tasks, enabling high-dimensional statistical analyses of behavioral outputs. Employing machine learning to analyze experimental results in unveiling relations between neural activity and behavior aligns with my core research goals.

Similarly, the work of Drs. Antonio Fernandez-Ruiz and Azahara Oliva on hippocampal memory coding in naturalistic behavioral paradigms deeply interests me. I am eager to investigate how biological neuronal networks generalize learning across domains, particularly the role of social memory traces in hippocampal-subcortical interactions during sleep and relevant wake states. Using closed-loop optogenetic manipulations to remotely modulate neural activity and test causal hypotheses would provide powerful insights into these processes.

Beyond its academic offerings, I am drawn to Ithaca for its small-town charm. During my undergraduate years at Binghamton University, I often visited to hike and study at local cafes - an environment I believe is perfectly suited for my graduate studies. Ultimately, I am eager to deepen my expertise in quantitative modeling and systems neuroscience through Cornell’s Neurobiology and Behavior program, with the goal of establishing my own line of research one day.

[1] Daeumer et al., Acta Materialia (2022), https://doi.org/10.1016/j.actamat.2022.117671

[2] Kharabadze et al., npj Computational Materials (2022), https://doi.org/10.1038/s41524-022-00825-4

EKATERINA A. KOULAKOVA

e.koulakova@nyu.edu | Huntington, NY

My interest with the subjectivity individuals experience in response to common events initially drew me to a career in law. Two internships within the political and legal fields demonstrated to me that these disciplines organize collective human behaviors rather than analyzing the fundamental mechanisms that drive individual decisions. This realization prompted a shift in my goals toward understanding and modeling the inner workings of the brain and its cognitive processes.

To broaden my perspective, as an undergraduate, I interned with a congresswoman whose ideologies differed significantly from my own. Engaging in discussions that challenged my perspectives deepened my understanding of complex issues and fostered a more nuanced worldview. This experience refined my ability to approach disagreements with openness to opposing perspectives, highlighting how individuals derive differing interpretations of the same circumstances based on their unique prior experiences - our personal ‘training sets’.

Driven to explore how subjective frameworks shape decision-making, I interned with the Broome County Attorney’s Office. Reading case files and attending court hearings, I gained an intimate glimpse into the lives of the county’s delinquent youth. Many of these individuals, shaped by challenging circumstances, acted in ways that seemed inevitable - rational, even, when viewed through the lens of evolution which has shaped the human brain over millions of years to promote our survival and well-being. This experience highlighted a central drive in my intellectual journey: to investigate the biological foundations of learning, memory, and decision making processes, particularly to understand how biological factors constrain individual autonomy. This realization prompted a pivot in my focus from the macroscopic lens of the humanities to the microscopic organization of the brain.

Since my university offered limited neuroscience resources, I pursued a second major in physics to deepen my understanding of complex phenomena through quantitative reasoning. Shifting my academic concentration required great resilience, taking numerous winter and summer courses and self-studying through online lectures. Finding this new path deeply rewarding, I joined a computational material science lab which employed neural networks extensively. I was captivated by the parallels between neural networks and biological systems - particularly how a set of initial parameters can structurally modify the ways in which an organism processes and adapts to input stimuli - reaffirming the insights that I had gained from my political and legal internships. This research experience only solidified my commitment to transition to neuroscience, providing a bridge between the deterministic principles of physics and the complexities of biological systems.

Graduating with degrees in political science and physics during a liminal COVID era, I initially pursued research assistant roles in systems neuroscience. However, many opportunities were highly specialized, often revolving around equipment management. To gain broader exposure and avoid early career pigeonholing, I joined Thorlabs, a company whose photonics tools are integral to neuroscience research. As an Application Engineer, I provided technical support to professors, students, and researchers in microscopy and imaging. This role honed my technical communication skills and offered hands-on experience in building and troubleshooting photonics systems, including lasers, fiber optics, and imaging devices. While I valued the comprehensive understanding of photonics equipment that I had gained, I longed for the creative, exploratory nature of long-term research that I experienced as an undergraduate. This realization led me to transition to the Rinberg and Shoham labs at NYU, where I am now fully engaged in systems neuroscience projects that align with my goal of understanding how neural architecture shapes cognition.

At Cornell, I am committed to fostering inclusivity by supporting my colleagues and actively seeking diverse perspectives - values that I have cultivated during my time at Thorlabs and during my political and legal internships. Drawing on my unique background in computational physics and neuroimaging, I strive to contribute meaningfully to collaborative efforts. I believe that embracing a variety of viewpoints enhances the quality and impact of our work, as the richness of our individual experiences strengthens both the depth and scope of our scientific endeavors.