Beyond the Paper: A Conversation with Dr. Samantha White

Interviewed by Dr. Paige N. McKeon, June 26, 2024

White et al. investigated mechanisms underlying reward-related decision-making in rats. Previous research focused on decision-making after learning a task, but here White and colleagues investigated decision-making during the initial stages of learning. Rats were trained to associate visual cues with 4% or 16% sucrose and then had to choose between these reward-related cues. Rats made choices slowly at first, but deliberated less as they spent more time in the task and became more experienced decision-makers. This behavior was then modeled computationally with two different approaches: exponential gaussian or drift diffusion. Preference for a more rewarding option was correlated with drift rate, which suggests that there is a direct relationship between subjective valuation of reward and decision-making. These findings advance our understanding of decision-making by demonstrating that experience significantly enhances decision-making. They also bring us closer to the underlying mechanisms of decision-making.

“I think being a scientist is so cool and I love the variety of my day to day, and spending time talking to people who know way more than I do about things I find deeply fascinating. I [...] feel grateful for the opportunity to keep exploring the breadth and depth of knowledge neuroscience research has to offer.”

Dr. Samantha White

Congratulations on your recent PhD from American University in Washington D.C. and your postdoctoral fellow position at the National Institute of Mental Health! Can you tell us a little about what led you to pursue a PhD in neuroscience and about your interest in studying decision-making?

It’s a bit of a long running story, but I think it’s a funny one. My junior year of high school I decided I was tired of having the same unrequited crush on a guy on the wrestling team. I’d recently discovered the science section of my nearest Barnes and Noble and found a great book entitled "The Chemistry Between Us: The Neuroscience of Sex, Love, and Attraction" by Larry Young, PhD, and Brian Alexander. In my naivety I thought if I understood the neuroscience of having a crush, I would be able to make it vanish. I learned two important things from reading that book: 1) understanding a few aspects of the neurobiology of having a crush (or of any brain-based phenomenon) doesn’t lead to a eureka-like resolution of that crush/phenomenon; and 2) there were so many fascinating things left to learn about the brain and so many revolutionary techniques available to answer them.

When I got to college, this book (and a handful of others) led me to pursue a major in neuroscience. I got hooked on neurophysiology in my first neuroscience course, The Neuron, with Mark Laubach. After the course, I was invited to join his lab for the summer between sophomore and junior year as an undergrad research assistant. In my experience in the lab, I fell in love with the pursuit of scientific knowledge, and I truly found that doing science is so much fun! These two things, of course, led naturally to pursuing a PhD. This publication actually features some of the first data I collected in my second research summer, and it’s so fun to see it finally out all these years later.

Everyone’s research journeys are different, how was your PhD experience? What challenges did you face along the way and what helped guide and motivate you through them? At what point did you know you wanted to continue doing research as a postdoc? Can you talk about how the transition from PhD candidate to postdoc has been?

Earning my PhD is one of my proudest accomplishments so far, and I’m very grateful to have had the opportunity and resources to make it happen. Overall, I had so much fun. While it’s a really serious accomplishment and took so much work and growth, it was a silly and joyful experience – at least that’s how I remember it after the post-defense fog cleared. There were definitely rough times (especially in the thick of the pandemic), but I had a strong support system to push and pull me through it.

In the beginning of the PhD, I faced a lot of technical challenges – how do I properly handle a rat? How do I get my data from one format to another? How will we rebuild the lab when we move from one building to another? How do I set up a brand new electrophysiology system? Each of these things had their own near vertical learning curves, and looking in retrospect, I kind of can’t believe I figured half of it out. But I think that’s ultimately what a PhD is – learning and figuring it out, and knowing when to leverage the knowledge of people who know way more than I do. Towards the latter half of the PhD, the challenges were less technical, and more intellectual and personal. I was surprised to realize how much personal growth comes hand in hand with academic growth, and for a while it felt really difficult to grow in both ways at the same time. It took a lot of good conversations with my PI, my lab mates, my friends and my family to keep everything straight and balanced to do this really hard thing. Without them, I couldn’t have done it (two full pages of my dissertation are dedicated to thanking each one of them).

I want to be honest about the fact that I thought about quitting graduate school two or three times, because I think that feeling is more common than it can seem, and we should talk about it more. In each of the moments where I deeply considered walking away, I asked myself three questions: 1) Why did I want to do a PhD in the first place? 2) What was making me want to quit in this exact moment? 3) What could I do about that particular reason to quit so I can get back to enjoying the process? I realized most of the reasons that made finishing feel impossible were addressable with hard conversations, reorienting to my goals, and taking time and space to think through the roadblocks. After those moments, I returned to doing the things I loved to do. It made writing the dissertation a bittersweet process to remember the times I almost quit, and I feel all the more proud of the fact that I decided to finish what I started.

I ultimately decided to do a postdoc for the same reasons I went to graduate school: I love science and the brain and learning about it. I think being a scientist is so cool and I love the variety of my day to day, and spending time talking to people who know way more than I do about things I find deeply fascinating. I wanted to continue that as a postdoc, and so far it’s been exactly what I was hoping for in that sense. I admit it was a rough start because I forgot what it’s like to have a steep learning curve again! (I chose a lab with techniques and topics all new to me). But after a few months, I’ve settled in a bit more and am gaining confidence in new skills and feel grateful for the opportunity to keep exploring the breadth and depth of knowledge neuroscience research has to offer.

“I was surprised to realize how much personal growth comes hand in hand with academic growth, and for a while it felt really difficult to grow in both ways at the same time. It took a lot of good conversations with my PI, my lab mates, my friends and my family to keep everything straight and balanced to do this really hard thing.”

You were invited to join the inaugural Early Career Researcher Advisory Board for JNeurosci, can you speak about that experience? What advice do you have for early career researchers? Has speaking to others as a member on this board influenced your career decisions at all?

I’m a big fan of the peer review process, and think it’s so fundamental to encouraging quality, clarity, and especially accessibility of academic research publications. I feel lucky that my first experiences with peer review came from assisting my graduate mentor with reviews. I found the process of reviewing a manuscript really challenging and rewarding, and I think it encouraged me to be more critical of my own work in a positive way. As a senior graduate student, I earned more independence as a reviewer by completing SfN's Reviewer Mentor Program, which I highly recommend to every late-stage grad student or early-stage postdoc. I can’t recommend it enough. It was such an outstanding experience. I gained a lot of confidence in my reviewing skills, and it taught me how to be more helpful, careful, and clear.

I wanted to join the Early Career Researcher Advisory Board because I wanted to learn more about the full publication process, and I wanted to be able to represent the views of ECRs to the Editorial Board. I wanted to share how my experience reviewing with my graduate adviser was such a fundamental aspect of my training, and brainstorm ways to make this a more accessible opportunity for other ECRs. I also wanted to learn about other ways graduate students and postdocs might be able to be involved in the publishing process because I think sometimes these kinds of experiences are what lead to interest in different career paths, should someone want to pursue a career outside of academia – I felt like access to the process in this way might improve the way I mentor students towards their careers in the future. I’ve had a great time so far serving with my fellow members, and we’re looking forward to adding new members to the board soon. Together we’ve already been able to contribute more than I anticipated, and as the inaugural board, it’s been fascinating to be able to set up a bit of precedent for the future, and I’m grateful to be a part of that.

Were the outcomes of your experiments in this paper surprising? In your view, what are the most valuable contributions of this paper to the field? Where do you see the project going?

As we discuss in the paper, Alex Kacelnik and colleagues made critical observations about differences in how animals in the wild learn their environment versus how we train animals to learn in the laboratory. We took inspiration from their observations and from their laboratory studies with starlings. I think the concept behind the present study is one of the takeaways I wanted to offer for the field: the way we train subjects could have major implications for behavior and for neural measures, and the training in and of itself is important to study. Many times in decision-making studies, we focus on subjects with extensive training to reduce behavioral and neural variability. However, our results here suggest something dynamic occurs from early to later instances of decision-making behavior. We know the brain is equally important for learning as it is for decision-making, and I feel that studying extensively trained animals might lead to an underestimation of how the brain is involved in making decisions early in the process. The behavioral design in this study allows for investigation of early decision making independent of other types of learning. Additionally, the model fitting in the study leads to interesting ideas about how the brain might be representing the decision-making process as it learns the process. We followed up on the role of the frontal cortex to early instances of decision making with pharmacological perturbation studies and in vivo electrophysiology recordings to investigate predictions by these models, which will come out in forthcoming publications. 

"[...] the way we train subjects could have major implications for behavior and for neural measures, and the training in and of itself is important to study. [...] We know the brain is equally important for learning as it is for decision making, and I feel that studying extensively trained animals might lead to an underestimation of how the brain is involved in making decisions early in the process."

Were there any challenging hurdles you had to face as you performed experiments for this paper?

One of the challenges specific to this work was I wasn’t initially sure how to analyze the response time data, because the histogram distributions did not meet assumptions of normality. Every math and statistics teacher I’ve had since fifth grade has drilled the importance of these assumptions, however, I didn’t ever get a good grasp of what to do if data isn’t normal, besides transform it. As seen in the study, the response time data doesn’t follow a normal distribution, and it had such a skew that a log or 1/RT transform wasn’t sufficient to normalize it. So, I challenged myself to find the best way to analyze the data as it was, which is what ultimately led to the ex-gauss model fitting and then the DDM analysis. This was all in my early days of working with R and Python, so it was another steep curve to learn how to fit these models. There’s also a huge body of literature debating how the models should be interpreted, which created a challenging intellectual exercise of how to interpret my own data. I’m glad I went through the trouble because I think it added a compelling spin to the paper, revealed more information than standard summary statistics could, and led to some interesting predictions for what may be going on in the brain.

One of the other challenges that I faced was working with my past-self since this data was collected over the span of a few years. There’s a tweet that went viral when I was at the beginning of my PhD and it said the following: “Finishing a PhD is like finishing a group project where your partner made a ton of mistakes at the beginning of the assignment. Except your partner is just you 4 years ago. -@_JohnMola” It didn’t resonate then, but I thought back to it often as I was revisiting the initial cohorts of data that I collected from the beginning to four years later because, wow, was it true. I was trying to decipher scribbles and find where I stored earlier versions of analyses and so, if you’re an early career researcher reading this: please, for the sake of yourself in the future, think about future-you trying to analyze your data and what they might really want to know! Label your files well, and document the heck out of your code!

How was your experience with the eNeuro review process? Do you have any thoughts on the science review process in general that you’d like to share?

This is my fourth eNeuro publication, and every time I’m always pleased with how the review process goes. It’s always been extremely straightforward, and the reviews have always contributed to improving my work, so I always appreciate the time that the reviewers take putting into it. As I said before, I’m generally fond of the peer review process because I think it’s important for making sure scientific manuscripts are written in an accessible way and ensuring conclusions are supported with well-designed experiments. I’m especially a fan of the recent decision by many journals to publish reviews because, as a trainee, I think there’s so much to learn by examining what experts in the field pay attention to when reviewing papers. It helps me orient to which aspects of a paper may be important to be thoughtful of when I’m reading subject matter outside my own immediate expertise. It’s sort of like attending a good journal club and absorbing the knowledge of the senior people in the room. And, as a student of a small and topic-diverse program, access to reviews contributed to how I learned to think critically about manuscripts I was reading that were outside my department, but within my research topic.

Are there any projects you are currently working on that you would like to give readers a heads up about?

About half of my dissertation has been published up to this point, so I’m excited to get the remaining studies out because it’s work I feel very proud of (as most PhDs feel about their dissertations, of course). It’s in the works, but in the meantime my colleague, Jensen Palmer, has a great study that came out recently as a preprint (and will be published shortly). In it, she examines the role of prelimbic cortex in this task. Specifically, she used reversible inactivations in female and male rats and found they sped up their choice latencies, all without affecting their choice accuracy. Drift diffusion modeling on the response latency data found selective effects of prelimbic cortex inactivation on the decision threshold, suggesting that this region of frontal cortex may mediate the amount of information necessary to enact a choice. I think her study pairs nicely with the work published here and makes further predictions about how the frontal cortex might be representing information over the course of learning and during instances of decision making. I’m working on writing up the findings from these neural recordings, so if you’re interested in this line of work, stay tuned!

“I wasn’t initially sure how to analyze the response time data, because the histogram distributions did not meet assumptions of normality. [...] I challenged myself to find the best way to analyze the data as it was [...] I’m glad I went through the trouble because I think it added a compelling spin to the paper, revealed more information than standard summary statistics could, and led to some interesting predictions for what may be going on in the brain.”

Read the full article:

Learning to Choose: Behavioral Dynamics Underlying the Initial Acquisition of Decision-Making
Samantha R. White, Michael W. Preston, Kyra Swanson, and Mark Laubach