Announcer:
Welcome to the Salk Institute’s Where Cures Begin podcast, where scientists talk about breakthrough discoveries with your hosts, Allie Akmal and Brittany Fair.

Allie Akmal:
Tanya Sharpee is a professor in Salk’s Computational Neurobiology Laboratory. She uses methods from mathematics, statistics and physics to chart the principles by which the brain’s billions of neurons exchange information and energy. Her work can get pretty theoretical, but what makes it so intriguing is that it’s all about the very concrete way we sense our environment through sight, sound and smell. Dr. Tanya Sharpee, welcome to Where Cures Begin.

Tanya Sharpee:
Thank you for having me here.

Allie Akmal:
I think one of the earliest conversations we had, I was asking you about that big internet debate about whether the dress was blue or gold, or maybe it was blue and yellow—I can’t remember the two colors that people saw and they saw it so differently and they felt so strongly about it. Can you talk about that, and what was behind those differences in perception?

Tanya Sharpee:
Yeah, so the issue is that it’s amazing that the differences between perception were almost black and white. So what was black or blue for one person was yellow or white for another person. So how can we be looking at the same image, perceive a very different aspect of it. And it turns out that our perception is influenced by our belief about the structure of the light, for example, around us. And we can be completely certain that that’s the answer. Two people can read the same article and form completely different opinions or interpret that evidence in completely different ways based on what they believe is true or not.

Allie Akmal:
And in the case of the dress, when you talk about beliefs, these are not conscious beliefs people have, but they’re computations their brain is making in how the object is lit or where the light is coming from. That kind of thing.

Tanya Sharpee:
That’s right. So that’s basically based on our experience with the world. So in the case of the dress, the explanation was if the person spends more time inside or more time of outdoors, they have different lighting conditions and that influences their perception.

Allie Akmal:
So if the photograph is taken outside, their brain is going to make certain assumptions about what the lighting will do and therefore what the color of the dress will be.

Tanya Sharpee:
That’s right.

Allie Akmal:
Okay. Well, that’s utterly fascinating.
How would you describe the field you’re in or what your profession is? Are you a biologist? Are you a physicist? How do you describe yourself?

Tanya Sharpee:
You know, they say that if you ask a physicist, then they will tell you that everything is physics. If you ask a computer scientist, they will say everything is computer science. If you ask a chemist and they will say, well, everything is really chemistry and all of these views are correct. You know, my training was in physics. So I could say I’m a physicist who studies biological systems. And that, that probably would be the best answer.

Allie Akmal:
Your work, if I’m understanding this correctly, seeks to quantify the activity of brain cells called neurons to better understand how our senses function, in other words, how we see and hear things and how we smell odors.

Tanya Sharpee:
Yeah. How the brain works and what are the algorithms that it uses. There are many people across the globe who are working towards understanding that code. And I think that’s exciting.

Allie Akmal:
Well, some of the applications of your work are in vision, for example, for self-driving cars…

Tanya Sharpee:
Yes, artificial vision, but also for brain machine interface. That’s another hot area of research. So if we understand more about the neural code, then we can make better devices. So we had a paper, actually, it was almost two and a half years ago. You know, vision is one of the mysteries of sensory processing. Even though we know more about vision than perhaps other senses, but it’s the case that the more we know, the more we realize that we don’t know, and we were trying to figure out how different parts of the visual scene come together, to yield our coherent perception. Of course it wasn’t a full solution to the problem, but, you know, we made progress elucidating how different types of edges are perceived, edges that are defined by differences in texture, such as between, for example, you can have a furry animal sitting on a tree branch and they can be similarly colored, but still the texture will be different and our eyes would be able to detect that edge. They’re called second-order edges.

And they’re more difficult to detect. And we saw that the secondary visual areas have neurons that specialize in that.

Allie Akmal:
in understanding the differences in texture…

Tanya Sharpee:
Or detecting edges that are defined by changes and changes in texture.

Allie Akmal:
So do you take data from actual neurons and then use statistical methods and algorithms to sort of organize them and from that to take rules of how they may be operating?

Tanya Sharpee:
Both. So we work on idealized codes and it’s a pure theory and then see how they can be implemented in the brain. And we look at the data and see what are the properties of the models that might best describe neural responses, and then try to match the two lines of research.

Allie Akmal:
Sort of see where they fit together?

Tanya Sharpee:
Yes, I mean, it’s like in a textbook, you know, they give you a problem. And then, at the end there is a solution. So you’re trying to solve it without looking at the solution, but then on occasion for difficult problems, you might want to look in the solution at least, you know, what is the number? Are you getting the correct number? So that’s how I view it. We do ideal theoretical solutions. And then we look in the brain to see what is happening and compare our solutions to the sort of answers that are provided by nature. So the reason we use tools from statistical physics or information theory or statistics in general is that we’re searching for ideal code that can be used to transform information without loss. And that can provide us with a guide to understanding the brain, if we understand the ideal solution that neural circuits might be implementing.

Allie Akmal:
In addition to studying vision Sharpee also studies the sense of smell known as olfaction.

Tanya Sharpee:
So one can actually view olfaction as a communication system between plants and animals or between animals and each other. But it’s interesting how we can use the statistics of signals in the natural world to learn clues about organization of the nervous system.

Allie Akmal:
And one of the reasons you took on this study was because if I recall correctly, you said that you can’t really tell very much from an odor molecule’s structure, what it’s going to smell like. So you have to look at it in other ways, is that correct?

Tanya Sharpee:
Yes. And so we think that methods such as those that we used—or similarly they’re called embeddings or mappings of molecules, assigning them coordinates—then the coordinates can be a better a predictor of how things might smell than the chemical structure itself.

Allie Akmal:
So what the lab did was to use statistics, to map combinations of odors that are found together in nature. And the locations of coordinates on the map could tell them things about those odors. Whether, for example, people would find particular odors, pleasant, or unpleasant.

Tanya Sharpee:
So for example, if we smell food and we need to know whether it’s spoiled or not, and we need to know whether there are dangerous bacteria that are present in there. So by experience, we’ll learn that certain bacteria produce certain kinds of molecules and we detect those molecules and maybe we shouldn’t eat this food. So that’s certainly intuition behind the co-occurrence measure. So bacteria can produce any number of signatures, but if I detect one or two of them that’s sufficient warning for me not to eat what they have found and pleasant to me in the past. But some products have only a few molecules that identify them and others can have up to 30 molecules that contribute to our perception. So it turns out one of the most complicated mixtures is Cognacs. So they have about 30 different molecules that influence our perception.

Allie Akmal:
Let’s see, you have done, since I’ve worked with you, vision and odor. And are you working on hearing as well?

Tanya Sharpee:
Yes. we do work on hearing. We have an ongoing collaboration with groups in France and Israel to study hearing. So that’s an ongoing project.

Allie Akmal:
Switching gears. Did you grow up in this country?

Tanya Sharpee:
No, I grew up in Ukraine. I came here for graduate school.

Allie Akmal:
So you grew up in a family of scientists. Can you tell us a little bit about that?

Tanya Sharpee:
I guess I was fortunate to, to see examples of successful scientists in front of my eyes; how they worked. My grandfather was a mathematician. You know, even though he was in his seventies, he would say, he would always say, I have to go learn how to become a better mathematician. So he would wake up at five in the morning and work before everybody was awake. He would work. And I would try to wake up as early as possible, maybe at six, but not as early he would wake up and we would work together at the table on our problems. He would work on his problems. I would work on my problems.

Allie Akmal:
And was this when you were in grade school or older?

Tanya Sharpee:
Yeah, or even younger.

Allie Akmal:
Oh my goodness.

Tanya Sharpee:
So he would give me these large multiplication problems of 10-digit numbers and everybody in the family would be, “Oh no, how can you ask a child to multiply such big numbers?” And he’d say “It’s okay. She can look, there is a multiplication table right next to her. And if you know the algorithm, you can multiply any number.” I liked that. I admired him very much. And then, you know, you would take a break, you know, chop some vegetables for soup, then continue doing your mathematics and so on.

Allie Akmal:
So, and so you had this mathematician grandfather, but you also had some physicists in the family.

Tanya Sharpee:
Yeah. Both of my parents are physicists and my father is an experimentalist. And he’s the reason why I’m a theorist. I looked at, you know, his experiment, very complicated. He dealt with metals. Then they had to be heated to high temperatures. You know, there was a furnace over here and the cooling water over there and it’s all running and his constantly fixing his rig. And I said, well, no way I can do that. Who’s going to fix the rig for me? So I better become a theorist.

Allie Akmal:
And was he disappointed?

Tanya Sharpee:
I don’t know, but we do have this experimentalists-theorists debate…

Allie Akmal:
A friendly rivalry perhaps?

Tanya Sharpee:
I’m not so sure. You know, he’s kind of dismissive—”These theorists, they’re just making up a number. Experimentalists have to really think to ask nature the question and the theorist can just come up with some kind of a theory. who knows if that’s right or wrong.” And you know, I try to hold my own by saying that it’s the theory that determines what is experimental-observable. This is one of the quotes by Einstein, because without the theory, you will know what experiment to do. You know, when he predicted that the light will bend by this amount, if you don’t have a theory that the light bends, you will never do the experiment. It’s a never ending saga.

Allie Akmal:
Do you have any advice you would give to girls specifically who are interested in a career in science?

Tanya Sharpee:
Well, I mean I think there’s this notion that women are not good at mathematics or at least they’re afraid of theoretical physics. So I think it’s certain doable where a woman to be specialized in physics and mathematics. I wouldn’t exclude the possibility that the ways we think about the problem are different between men and women. You know, one person can be faster, but the other person can see different connections. So this is what is emphasized in, for example, in this book called A Beautiful Mind. So this was a person with schizophrenia or some mental disorder, and he could see connections between unrelated observations that other people would miss. So I think the message from that book and from other experiences that we have different viewpoints and can be useful. I think the art is to find a combination or an area where ones particular strengths and approaches are useful, but one shouldn’t exclude mathematics and theoretics. There are unique challenges for women in mathematical and physical and in general in sciences, but that doesn’t mean that women should pursue them.

Allie Akmal:
Well, Dr. Sharpee, thank you so much for joining us.

Tanya Sharpee:
Thank you so much.

Announcer:
Join us next time for more cutting-edge Salk science. At Salk, world-renowned scientists work together to explore big bold ideas from cancer to Alzheimer’s aging to climate. Where Cures Begin is a production of the Salk Institute’s Office of Communications. To learn more about the research discussed today, visit Salk dot EDU slash podcast.