Anne Churchland

Anne Churchland

Neuron Q&A Anne Churchland Anne Churchland’s scientific experience has focused on understanding the neural circuits behind multisensory decision maki...

315KB Sizes 2 Downloads 53 Views

Neuron

Q&A Anne Churchland Anne Churchland’s scientific experience has focused on understanding the neural circuits behind multisensory decision making. In an interview with Neuron, she discusses large-scale recordings in behaving animals, communication between experimental and theoretical labs, and the creation of her website anneslist.net to highlight women in systems and computational neuroscience in an effort to help close the gender gap in conference speakers. Anne Churchland is an Associate Professor at Cold Spring Harbor Laboratory. Her lab aims to understand the neural circuits that support multisensory decision making. Lab members train rodents to judge auditory, visual, or multisensory stimuli, and they measure and manipulate neurons during these decisions. By repurposing multisensory behaviors from the human perception literature, the lab leverages behavioral paradigms with established theoretical frameworks. These frameworks assist greatly in interpreting behavioral and neural data and have revealed, for example, that rodents can optimally weight sensory signals for multisensory integration. Further, this approach makes it possible to identify behaviors that are shared across animals and are not idiosyncratic to a particular species. Dr. Churchland’s postdoctoral training was in the laboratory of Michael Shadlen. During that time, she was funded by a K99 award that aimed to test multiple models of decision making on data she collected from the parietal cortex of nonhuman primates. She was coadvised by XiaoJing Wang and Alex Pouget, and the combined guidance of her three mentors provided training in both experimental and theoretical neuroscience. This training built on expertise she developed as a graduate student working on motion processing and oculomotor control with Stephen Lisberger. In addition to being a professor, Dr. Churchland is on the executive committee for the Computational and Systems Neuroscience (Cosyne) conference and is the director of the undergraduate research program in bioinformatics and computational neuroscience held at Cold Spring Harbor Laboratory. 940 Neuron 92, December 7, 2016

ysis tools to interpret the data. The coming years will bring exciting discoveries.

Anne Churchland Cold Spring Harbor Laboratory

What do you think are the big questions to be answered next in your field? The field has made a lot of headway in understanding how sensory signals are processed, and we are also beginning to understand movement generation at a deep level. But what happens in between remains very mysterious. Major outstanding questions include: How are incoming sensory signals interpreted, given the constraints of a particular context? How are these incoming signals incorporated with ongoing internal dynamics? How do emotions, memories, and innate drives shape neural signals when different decisions are being considered? The field is primed to begin tackling these questions: we have the technology for large-scale neural recordings and the anal-

Which aspect of science, in your field or in general, do you wish the general public knew more about? I wish the public more fully understood that basic research in animals supports clinical discoveries, in part because of the similarities among the brains of many animals. When I tell nonscientists about our discovery that rat and human subjects similarly combine sensory information in a statistically optimal fashion, they are sometimes shocked. This is because many of us grew up learning that our human brains are fundamentally different from other animals—certainly from a lowly rat! However, both rats and humans have proliferated because their brains evolved ways to process information that are adaptive and drive decisions that further survival and reproductive success. Much of the information that rats and humans encounter is similar: many of the same sights, sounds, and smells are experienced by rats and humans, and both species must use these signals to avoid danger, find food, and secure mates. These shared goals likely drive many common computations across animals. By understanding these computations in laboratory animals, we can often gain deep insights into human brain function. To tackle your favorite research question, is there a tool that either needs to be developed or is currently available that could be implemented in a novel way? We need more targeted and precise ways to manipulate neurons. Neuroscience underwent a revolution when tools from molecular biology and genetics made it possible to manipulate neurons in new ways. However, to truly advance our

Neuron

Q&A understanding of brain function, we need tools with even more precision and flexibility than those currently available. For instance, it has become clear that neurons in many areas encode variables and drive behavior using complex and heterogeneous population responses. This is certainly true in the posterior parietal cortex, where we measure responses during decision making. Only when our causal tools allow us to mimic these complex patterns of activity can we truly harness the power of causal manipulations to understand the relationship between neural activity and behavior. What has been the highlight of your career? Starting my own lab as a principal investigator (PI) was an exhilarating experience. I loved defining a large question and then thinking of the smaller projects that would collectively begin to answer it. I also found it incredible to see the students and postdocs in my lab come to share my enthusiasm for these questions and bring new insights to each problem. One of our early discoveries was that the same principles that govern multisensory processing in humans were present in rodents. I had a hunch that this might be the case because I believe that many cortical computations related to sensory processing are shared across diverse animals. But to see this in action—to observe our rat subjects and our human subjects using the same strategy to assemble multisensory information—was thrilling. Who were your key early influences? There are two: at first, I was very influenced by my family’s enthusiasm for neuroscience. My parents were philosophers who thought deeply about how discoveries in neuroscience have the potential to change how we think about the mind. Their enthusiasm helped me to understand at an early stage the potential of neuroscience. The field has changed a lot since I was a teenager talking at the dinner table with my parents and brother during the ‘‘Decade of the Brain’’ Back then, neuroscience was exciting but also restricted in its scope because the tools available at the time were fairly limited. To have gained in those early years an appreciation for the brain’s mysteries,

and to now make strides in unraveling them, feels extraordinarily lucky. Later on, as a graduate student, I was deeply influenced by my mentor, Steve Lisberger, and also by the lab’s many collaborators, especially Tony Movshon. These two set a high standard for science: rigorous experimental design, carefully conducted experiments, and thorough analysis. Their appreciation for a combined experimental and theoretical approach to neuroscience was very influential and had a big impact on my own trajectory. I also learned from them the power of a collaborative approach. The longstanding collaboration between the Lisberger and Movshon labs meant that personnel often travelled from UCSF to NYU to collect data. Meeting students, postdocs, and faculty from another institution was an eye-opening experience for me because of their different perspectives on problems in the field. I still remember conversations among Lisberger and Movshon lab students, very late at night during physiology experiments, as we listened to neurons spiking and talked about the future of neuroscience. What motivated you to become a scientist? My love of collecting and analyzing data! My appreciation for the big questions, and my ability to design the right experiments to tackle them, developed later. My first passion was coming into the lab every day, setting up an experiment, watching the data come in, and then playing around with it on MATLAB. My first experience with this was as a graduate student in Steve Lisberger’s lab at UCSF. The experience was very much enriched by the fabulous environment there. My lab members—Jennifer Raymond, Nicholas Priebe, I-han Chou, Megan Carey, Justin Gardner, Javier Medina, Sascha du Lac, and Mark Churchland—inspired me constantly. The conversations and collaborations with that group were very motivating and played a big role in my decision to pursue a career in neuroscience. What is your view on big datagathering collaborations as opposed to hypothesis-driven research by small groups? There are two questions here: (1) hypothesis-driven versus data gathering

and (2) collaborative versus individual approaches. (1) I tend to favor hypothesis-driven research in my own lab, but I see the need for data-gathering research as well. My appreciation for the need for data-gathering research comes in part from the fact that sometimes my own data neither supports nor refutes my hypothesis; instead, it tells me something completely different about the brain that I didn’t expect. (2) I favor collaborative approaches. The field of neuroscience must evolve beyond the individual-PI approach. This approach was feasible when neuroscience was a small and developing field, but as the field matures, the individual-PI approach limits the ability to rapidly make discoveries on a large enough scale to truly understand the brain. The success of other large-scale collaborations, such as LIGO, makes it clear that a coordinated approach is effective at tackling large problems. The key question is how neuroscience can evolve to support larger collaborations. Transitioning from single-PI research teams to larger teams such as those supported by the BRAIN initiative and the DOD is a critical first step and has been very successful. However, these teams are mainly limited to five or six PIs, usually from one or two countries. The next step is to have collaborations that include 20–30 teams and span many countries. These coordinated efforts will lead the way in developing standardized experimental setups alongside efficient data- and code-sharing infrastructure. Developing a coordinated vision for how to solve a major scientific problem will be a challenge and will require a new way of thinking. Organizations that can begin to consider funding these larger teams will pave the way for a new era of collaborative neuroscience that will be transformative. What do you think are the biggest problems or challenges that science as a whole is facing today? First, there is a lack of coordination across labs doing similar research. This leads to duplicated efforts and nonstandardized approaches. Second, there is too little communication between experimental and theoretical labs. The magnitude and complexity of modern neuroscience data necessitate close collaborations between experimentalists and theorists. Neuron 92, December 7, 2016 941

Neuron

Q&A To facilitate this, there must be a cultural shift in which both groups find a way to bridge this gap: standardized data formats, more training in mathematical techniques for experimentalists so they can identify and work closely with the right theoretical collaborator, and more experimental training for theorists so they can understand the data at a deep level. In your opinion, what are the most pressing questions for the field? How do neural populations encode information and guide behavior? The recent years have ushered in a change in how we think about and analyze neural activity. There has been a dramatic change in our ability to record many neurons simultaneously and a growing appreciation for population-level analyses. These have been transformative, but they are still at an early stage. As we acquire more large-scale recordings, especially in behaving animals and in conjunction with appropriate neural perturbations, our understanding of the neural basis of cognition will expand tremendously. Where do you see the strongest potential for progress and new breakthroughs in neuroscience? The greatest potential comes from investigators who are willing to join forces with one another and collaborate to solve major outstanding questions in the field. These collaborations need to occur both between experimentalists who are tackling related questions and between experimentalists and theorists. These close collaborations will support the ability to make fast progress on big problems. How do you find inspiration? I find inspiration by talking with colleagues about science, both the big questions and the small details that are so critical to answering them. Sometimes, when I am frustrated with a problem, I’ll have a collaborator come and spend time in my lab. Recently, I was frustrated by our limited understanding of some anatomical divisions in rat visual and parietal cortex. A collaborator of mine, Heather Read, came to collaborate. She lived here on campus, so we became immersed in the

science—sometimes talking about the problem from morning till night. It renewed my enthusiasm for getting to the bottom of the problem and furnished me with a more powerful set of tools for doing so than I had before her visit. Do you have a role model in science? If so, who and why? My current work was inspired by a partnership of scientists: Greg DeAngelis and Dora Angelaki. They fearlessly brought a paradigm from the human psychology literature, optimal multisensory integration, into the realm of experimental physiology. By bringing this paradigm to an animal model and combining theory with experimental results, their work offered major insights into the neural mechanisms that support this ability. Seeing their success motivated me to bring these same paradigms to rodents, a model system with tools that make it possible to add a neural-circuits approach that builds on the understanding they have established. Importantly, my decision to tackle those questions in rodents was a difficult one. In 2009, when I decided to start a rodent lab, many doubted the potential of rodents as a model for higher cognitive functions like decision making. I was inspired at that time by the success of a small number of groups who had begun to take behavioral paradigms developed in primates and adapt them to rodents. The creative work done by Tony Zador, Zach Mainen, Matteo Carandini, Carlos Brody, David Tank, and Karel Svoboda was very inspiring to me. Their willingness to share their ‘‘tricks of the trade’’ was instrumental in my being able to take the multisensory approach developed in human and nonhuman primates and bring it to a powerful animal model. What do you do when you’re not in the lab? When I am not actively engaged in my own lab’s scientific agenda, I devote effort to leveling the playing field in science. As an expert in decision making, I have seen firsthand that subjects often use irrelevant information to guide decisions. This happens in the lab, with perceptual judgements about sensory stimuli, and it happens in

the greater scientific arena, with judgments about what should be published, who should be hired, and who should be invited to speak. Even the most well-meaning among us can be swayed by information that is irrelevant to a person’s scientific ability, such as their gender, their race, or a host of other characteristics. As a field, we will do better science when we improve at evaluating scientific content in a way that is free from these biases. As a step towards this goal, I created a website, anneslist.net, that highlights women in systems and computational neuroscience. I originally created it for my own purposes as an organizer of Cosyne. By creating a database of women, I made it possible for conference and seminar organizers to easily identify suitable speakers. Using this database allows organizers, such as those for Cosyne, to achieve a gender ratio that meets or exceeds the base rate of women in the field. Have you encountered particular difficulties? How did you overcome them? When I was a postdoc, I struggled to define and articulate my vision as a scientist. This was surprising because in graduate school I was opinionated and outspoken and would tell anyone (often unprompted) about why my project was the most important one for moving the field forward. But during my years as a fellow a number of things changed, both personally and professionally, and the strong voice that defined my early years faded. I remember once, about 4 years into my postdoc, a senior colleague heard my tepid answer to an engaging question and said, ‘‘What the heck happened to you? Why don’t you say what you really think anymore?’’ I realized that the words were true and that the time was right for me to redevelop my own vision and ideas and to say them out loud, with gusto. This really took hold when I became an independent investigator at Cold Spring Harbor Laboratory. Like any new investigator, I had ups and downs, but I have never wavered in my confidence about the importance of the work we are doing and its relevance to the larger goals of neuroscience. It feels great to be my true, outspoken, opinionated self again! http://dx.doi.org/10.1016/j.neuron.2016.11.028

942 Neuron 92, December 7, 2016