Dr. Kanaka Rajan
Associate Professor Department of Neurobiology, Blavatnik Institute, Harvard Medical School
Faculty Kempner Institute for the Study of Natural and Artificial Intelligence, Harvard University
Biophysics Theory Postdoctoral Fellow Princeton University
PhD in Neuroscience Columbia University
MS Brandeis University
The teen-aged Kanaka followed a typical trajectory of an engineering student in India. She studied engineering in high school, and specialized in biochemical engineering as an undergraduate. But when she took a summer research internship position at the National Institute of Mental Health in Bangalore, she found herself adventuring into a new world--one of clinicians discussing mental illnesses, their pathophysiologies and their interactions with society. She was thrilled by the novelty of her environment, but also by a sense of familiarity. Kanaka was raised by her grandmother who had schizophrenia and neuroscience research sparked a feeling of connection to her grandmother’s lived experiences. Upon returning to engineering school and completing her undergraduate studies, she felt that she was approaching a fork in the road of her journey. Taking the leap to study neuroscience in graduate school was uncharted territory, but she was determined to recapture the thrill she felt in the clinic.
Kanaka began her neuroscience PhD at Brandeis University, where she enthusiastically learned how to perform complex and cutting-edge experimental neuroscience techniques during a rotation in Dr. Eve Marder’s laboratory, such as electrically recording from the lobster stomatogastric ganglion. Although her technical skills were excellent, designing experiments felt less intuitive to her. Kanaka used this potential roadblock as an opportunity to identify a skill she wanted to learn in graduate school. That skill, which Eve implements so well, is to distill a problem into its most minimal parts. During her training, Kanaka also became an expert in this style of intuition, and it guides her research to this day.
Like a fish returning to water, Kanaka ultimately decided to join Dr. Larry Abbott’s laboratory for her PhD and explore her interests in theoretical neuroscience using her engineering background. Here, she could build her intuition for understanding the brain in an environment that felt like home. In the Abbott lab, she stumbled upon a question that she could not get out of her head. In retrospect, Kanaka sees this nagging nature of a scientific question as a requisite for working on it: “it has to be stuck in your teeth.” It’s her own visceral and delightful description of a question that is at the same time thoroughly fascinating and entirely bothersome. Graduate-school-Kanaka fixated on the question: How does the brain reconcile incoming inputs with its ongoing internal dynamics? To get at a possible solution, she designed artificial neural network models with ongoing, intrinsically generated dynamics. These models had the ability to turn off internal noise in order to pay attention to incoming drives. Although Kanaka’s question was inspired by biology, her model did not incorporate many real biological limitations. At this stage of her intellectual journey, she was satisfied with this style of model. It wasn’t until her postdoctoral work that she started to weave thick biological threads through her research.
As a Biophysics Theory Postdoctoral Fellow at Princeton University, Kanaka had the intellectual freedom to pursue her theory interests by collaborating with other theorists and experimentalists. She began working with Dr. William Bialek on a theory of sensory signal processing that suggests that neurons may be sensitive to correlations in incoming stimuli. Afterwards, she approached these ideas with Dr. David Tank and Dr. Chris Harvey in the context of their work on neural sequential firing properties in the parietal cortex of mice making decisions. They had observed that during certain decision-making behaviors, neurons fire at different times within a larger, wave-like sequence of activity--like fans doing “the wave” at a sports game. During decisions, neural activity varied between firing in these wave-like sequences or ramping up to a level “fixed-point.” Her work in the Tank lab focused on building a recurrent neural network model to generate sequences and thus behavioral choices, and model mouse decision-making behaviors observed in the lab. Although it is common to correlate neural firing to behavioral variables like an animal’s running speed or the accuracy of its decisions, Kanaka feels that neural dynamics could underscore more fundamental features of the task itself, like measuring time or marking the beginning or end of events. She has maintained a fascination with these abstract neural processes as the building blocks of cognition; it is a major part of her laboratory’s current research.
Kanaka now heads a theoretical neuroscience laboratory in New York City at the Friedman Brain Institute within the Icahn School of Medicine at Mount Sinai. Her laboratory collaborates with experimentalists to create theories that explain how neural processes comprise cognitive behaviors. One of her research arcs builds upon her PhD and postdoctoral work on fixed-point and sequential neural firing patterns, or neural “dynamical motifs.” Kanaka and her colleagues observed neural firing ramping up to a “fixed-point” when animals performed more complex tasks, and firing in a wave-like sequence when animals performed more naturalistic behaviors. In other words, the task complexity affected the observed activity dynamics. Furthermore, they saw sequential firing across brain regions with different anatomies and across very different tasks. If these firing patterns are not specific to brain region, behavioral variables, or task features, then what are they signaling? Her hypothesis harmonizes with her PhD question: “How does the brain reconcile incoming inputs (switching states) with its ongoing internal dynamics (current state)?” Fixed-point neurons could steadily represent a state while you’re performing a task, while sequential neurons could allow you to switch between states and string together your actions. In this paradigm, silencing a brain region and seeing a disruption of behavior doesn’t necessarily mean that particular region is important for performing the task. Instead, we could interpret this as the animal is no longer able to string together their actions. Kanaka’s idea is poignant; even just being aware of it should change how we interpret behavioral studies.
The other arc of her lab’s research has grown out of her collaboration with Dr. Karl Deisseroth’s laboratory; their groups work together to analyze the whole brain activity of the larval zebrafish. Kanaka asserts that any region’s neural activity is a function of its connections to the rest of the brain. Therefore, activity from that single region is not very informative of its function. The whole larval zebrafish work sparked her thinking about the modularity of the brain and why it needs multiple regions to solve problems. Did modularity arise as behaviors became more complex? More specifically, how can we understand multi-regional interactions? Her laboratory is interested in the mechanistic underpinnings of the problem-solving abilities of populations of neurons; what kinds of inputs they require, and what sizes are optimal, for example.
Aside from her stellar scientific talents, Kanaka is dedicated to cultivating an academic community that is diverse, empathetic, and supportive. Beyond hiring women scientists of color, she challenges academic institutions to give them the tools and promote the environment that they need to succeed once they are there. She leads by example as part of grant review and faculty search committees. She applies her empathetic and intersectional approach in these positions, noting that diverse people have diverse lived experiences--and these should be celebrated when building a scientific community. After all, it's those lived experiences that lead us to uniquely approach the questions that get “stuck in our teeth.”
Find out more about Kanaka and her lab’s research here.
Check out Dakota’s full interview with Kanaka on February 24th 2021 below or wherever you get your podcasts!
Andalman, Aaron S., Vanessa M. Burns, Matthew Lovett-Barron, Michael Broxton, Ben Poole, Samuel J. Yang, Logan Grosenick, et al. 2019. “Neuronal Dynamics Regulating Brain and Behavioral State Transitions.” Cell 177 (4): 970–85.e20.
Rajan, Kanaka and William Bialek. 2013. “Maximally Informative ‘Stimulus Energies’ in the Analysis of Neural Responses to Natural Signals.” PloS One 8 (11): e71959.
Rajan, Kanaka, Christopher D. Harvey, and David W. Tank. 2016. “Recurrent Network Models of Sequence Generation and Memory.” Neuron 90 (1): 128–42.
