Modeling Learning Across the Lifespan

Q&A with Catherine Hartley

At a special session on the relation between psychology and neuroscience at last year’s CNS conference in San Francisco, Catherine Hartley said: “Even if we can predict behavior, if we don’t know how it works, we likely have not achieved our goals.” While computational algorithms and tools may help researchers predict people’s behavior, it is not enough, she argued ⁠— which is why Hartley has devoted her career to understanding the cognitive and neural mechanisms that drive behavior, especially learning.

A cognitive neuroscientist at NYU, Hartley focuses on understanding how learning and decision-making work in the brain across the lifespan. For her work using computational models to elucidate changes in goal-directed learning and decision-making over the course of human development, Hartley is receiving a CNS Young Investigator Award.

We spoke with Hartley about her research, her upcoming award talk at the CNS 2020 meeting in Boston, and what it’s like to work with children in her studies.

CNS: How did you become interested in the cognitive neuroscience of learning?

Hartley: I’ve always been interested in understanding the complexity of human behavior — what drives our actions and choices. I began doing cognitive neuroscience research as an undergraduate working in John Gabrieli’s lab, and enjoyed thinking about the relation between brain and behavior at this mechanistic level. Afterwards, I worked for several years at an artificial intelligence research company, where I began to think about what specific cognitive ingredients are necessary to carry out different types of behaviors. In retrospect, I think these experiences really shaped my interest in understanding the function of learning, memory, and decision-making systems in the human brain, as well as how these systems wire together over the course of development.

CNS: What do we know now about learning across the lifespan that we didn’t know prior to your team’s work in computational modeling?

Hartley: Computational models are useful tools for testing hypotheses about which aspects of a cognitive process change with age. Across a few studies, our work has emphasized that development brings about changes not only in the operations being carried out during learning, but also in the information that appears to be represented and used in the learning process. For example, we have found that in decision-making contexts in which adults’ choices take into account explicit verbal instruction or beliefs about the structure of the environment, children’s decisions are better captured by less complex models that only represent how good or bad that action was in the past.

CNS: Some of your recent work has shown enhanced memory in learning environments that offer more agency. Can you briefly describe the key results there and how it might translate to a real-world setting?

Hartley: We learn differently in situations in which our actions are consequential, and we can use them to bring about desired outcomes. Our past work has shown that prefrontal and striatal neural circuits are engaged to a greater degree when we’re able to exert control, which we hypothesize reflects a shift into a more goal-directed learning state. Consistent with this idea, in recent work, we’ve shown that our memory is better for episodes in which we had control than for those that were uncontrollable. And this was equally true of children, adolescents, and adults. This sort of mechanism may underlie the benefits of classroom contexts that enable children to engage in active learning.

CNS: Is there a single piece of data or study you are most excited to share in your CNS 2020 award talk that we haven’t discussed yet?

Hartley: I’m hoping to have time to talk about some recent and exciting collaborative work using smartphone technology to study exploration and novelty-seeking in adolescents and adults ⁠— our first foray into studying reward processing outside of the laboratory in the real world.

Even for those who don’t have any particular interest in the periods of childhood or adolescence, as cognitive neuroscientists (and people!), we’re all interested in understanding how the brain gives rise to the complex and flexible repertoire of human behavior. And the ultimate aim of my work, and of developmental cognitive neuroscience research more broadly, is to try to understand the assembly instructions of this cognitive machinery — how all of the switches and connections get configured.

CNS: What is it like to work with children in your studies?

Hartley: Children tend to be our most engaged and excited participants. It can be challenging to get children through MRI scans, but they get particularly excited when we show them pictures of their brains. We try to make our experimental tasks game-like to hold their attention, and most kids — and adults! — seem to find them fun. Although, we once had a 4-year-old clearly express his disinterest in our task by climbing out of his seat and lying face down on the floor.

CNS: What do you most want people to understand about your work? And why should they come to your award talk?

Hartley: Even for those who don’t have any particular interest in the periods of childhood or adolescence, as cognitive neuroscientists (and people!), we’re all interested in understanding how the brain gives rise to the complex and flexible repertoire of human behavior. And the ultimate aim of my work, and of developmental cognitive neuroscience research more broadly, is to try to understand the assembly instructions of this cognitive machinery — how all of the switches and connections get configured. And as for why people should come to my talk, I will be hiding goodie bags under three randomly selected seats. Seriously!

CNS: What are you most looking forward to about the CNS meeting in Boston?

Hartley: People’s faces when they find the goodie bags! No really — there are so many exciting symposia in the scientific program this year, and it’s always great to see friends.

CNS: What are the next steps for your work?

Hartley: We have a number of ongoing neuroimaging projects and planned studies going forward that aim to relate developmental changes in learning and decision-making to the function of prefrontal and subcortical circuitry. In several of these studies, we plan to use neural decoding approaches to relate developmental shifts in planning and mnemonic processes to underlying differences in stimulus representations in the brain.

-Lisa M.P. Munoz