A new computational technique could make it easier to engineer useful proteins

To engineer proteins with useful functions, researchers usually begin with a natural protein that has a desirable function, such as emitting fluorescent light, and put it through many rounds of random mutation that eventually generate an optimized version of the protein.

This process has yielded optimized versions of many important proteins, including green fluorescent protein (GFP). However, for other proteins, it has proven difficult to generate an optimized version. MIT researchers have now developed a computational approach that makes it easier to predict mutations that will lead to better proteins, based on a relatively small amount of data.

Using this model, the researchers generated proteins with mutations that were predicted to lead to improved versions of GFP and a protein from adeno-associated virus (AAV), which is used to deliver DNA for gene therapy. They hope it could also be used to develop additional tools for neuroscience research and medical applications.

Woman gestures with her hand in front of a glass wall with equations written on it.
MIT Professor of Brain and Cognitive Sciences Ila Fiete in her lab at the McGovern Institute. Photo: Steph Stevens

“Protein design is a hard problem because the mapping from DNA sequence to protein structure and function is really complex. There might be a great protein 10 changes away in the sequence, but each intermediate change might correspond to a totally nonfunctional protein. It’s like trying to find your way to the river basin in a mountain range, when there are craggy peaks along the way that block your view. The current work tries to make the riverbed easier to find,” says Ila Fiete, a professor of brain and cognitive sciences at MIT, a member of MIT’s McGovern Institute for Brain Research, director of the K. Lisa Yang Integrative Computational Neuroscience Center, and one of the senior authors of the study.

Regina Barzilay, the School of Engineering Distinguished Professor for AI and Health at MIT, and Tommi Jaakkola, the Thomas Siebel Professor of Electrical Engineering and Computer Science at MIT, are also senior authors of an open-access paper on the work, which will be presented at the International Conference on Learning Representations in May. MIT graduate students Andrew Kirjner and Jason Yim are the lead authors of the study. Other authors include Shahar Bracha, an MIT postdoc, and Raman Samusevich, a graduate student at Czech Technical University.

Optimizing proteins

Many naturally occurring proteins have functions that could make them useful for research or medical applications, but they need a little extra engineering to optimize them. In this study, the researchers were originally interested in developing proteins that could be used in living cells as voltage indicators. These proteins, produced by some bacteria and algae, emit fluorescent light when an electric potential is detected. If engineered for use in mammalian cells, such proteins could allow researchers to measure neuron activity without using electrodes.

While decades of research have gone into engineering these proteins to produce a stronger fluorescent signal, on a faster timescale, they haven’t become effective enough for widespread use. Bracha, who works in Edward Boyden’s lab at the McGovern Institute, reached out to Fiete’s lab to see if they could work together on a computational approach that might help speed up the process of optimizing the proteins.

“This work exemplifies the human serendipity that characterizes so much science discovery,” Fiete says.

“This work grew out of the Yang Tan Collective retreat, a scientific meeting of researchers from multiple centers at MIT with distinct missions unified by the shared support of K. Lisa Yang. We learned that some of our interests and tools in modeling how brains learn and optimize could be applied in the totally different domain of protein design, as being practiced in the Boyden lab.”

For any given protein that researchers might want to optimize, there is a nearly infinite number of possible sequences that could generated by swapping in different amino acids at each point within the sequence. With so many possible variants, it is impossible to test all of them experimentally, so researchers have turned to computational modeling to try to predict which ones will work best.

In this study, the researchers set out to overcome those challenges, using data from GFP to develop and test a computational model that could predict better versions of the protein.

They began by training a type of model known as a convolutional neural network (CNN) on experimental data consisting of GFP sequences and their brightness — the feature that they wanted to optimize.

The model was able to create a “fitness landscape” — a three-dimensional map that depicts the fitness of a given protein and how much it differs from the original sequence — based on a relatively small amount of experimental data (from about 1,000 variants of GFP).

These landscapes contain peaks that represent fitter proteins and valleys that represent less fit proteins. Predicting the path that a protein needs to follow to reach the peaks of fitness can be difficult, because often a protein will need to undergo a mutation that makes it less fit before it reaches a nearby peak of higher fitness. To overcome this problem, the researchers used an existing computational technique to “smooth” the fitness landscape.

Once these small bumps in the landscape were smoothed, the researchers retrained the CNN model and found that it was able to reach greater fitness peaks more easily. The model was able to predict optimized GFP sequences that had as many as seven different amino acids from the protein sequence they started with, and the best of these proteins were estimated to be about 2.5 times fitter than the original.

“Once we have this landscape that represents what the model thinks is nearby, we smooth it out and then we retrain the model on the smoother version of the landscape,” Kirjner says. “Now there is a smooth path from your starting point to the top, which the model is now able to reach by iteratively making small improvements. The same is often impossible for unsmoothed landscapes.”

Proof-of-concept

The researchers also showed that this approach worked well in identifying new sequences for the viral capsid of adeno-associated virus (AAV), a viral vector that is commonly used to deliver DNA. In that case, they optimized the capsid for its ability to package a DNA payload.

“We used GFP and AAV as a proof-of-concept to show that this is a method that works on data sets that are very well-characterized, and because of that, it should be applicable to other protein engineering problems,” Bracha says.

The researchers now plan to use this computational technique on data that Bracha has been generating on voltage indicator proteins.

“Dozens of labs having been working on that for two decades, and still there isn’t anything better,” she says. “The hope is that now with generation of a smaller data set, we could train a model in silico and make predictions that could be better than the past two decades of manual testing.”

The research was funded, in part, by the U.S. National Science Foundation, the Machine Learning for Pharmaceutical Discovery and Synthesis consortium, the Abdul Latif Jameel Clinic for Machine Learning in Health, the DTRA Discovery of Medical Countermeasures Against New and Emerging threats program, the DARPA Accelerated Molecular Discovery program, the Sanofi Computational Antibody Design grant, the U.S. Office of Naval Research, the Howard Hughes Medical Institute, the National Institutes of Health, the K. Lisa Yang ICoN Center, and the K. Lisa Yang and Hock E. Tan Center for Molecular Therapeutics at MIT.

Exposure to different kinds of music influences how the brain interprets rhythm

When listening to music, the human brain appears to be biased toward hearing and producing rhythms composed of simple integer ratios — for example, a series of four beats separated by equal time intervals (forming a 1:1:1 ratio).

However, the favored ratios can vary greatly between different societies, according to a large-scale study led by researchers at MIT and the Max Planck Institute for Empirical Aesthetics and carried out in 15 countries. The study included 39 groups of participants, many of whom came from societies whose traditional music contains distinctive patterns of rhythm not found in Western music.

“Our study provides the clearest evidence yet for some degree of universality in music perception and cognition, in the sense that every single group of participants that was tested exhibits biases for integer ratios. It also provides a glimpse of the variation that can occur across cultures, which can be quite substantial,” says Nori Jacoby, the study’s lead author and a former MIT postdoc, who is now a research group leader at the Max Planck Institute for Empirical Aesthetics in Frankfurt, Germany.

The brain’s bias toward simple integer ratios may have evolved as a natural error-correction system that makes it easier to maintain a consistent body of music, which human societies often use to transmit information.

“When people produce music, they often make small mistakes. Our results are consistent with the idea that our mental representation is somewhat robust to those mistakes, but it is robust in a way that pushes us toward our preexisting ideas of the structures that should be found in music,” says Josh McDermott, an associate professor of brain and cognitive sciences at MIT and a member of MIT’s McGovern Institute for Brain Research and Center for Brains, Minds, and Machines.

McDermott is the senior author of the study, which appears today in Nature Human Behaviour. The research team also included scientists from more than two dozen institutions around the world.

A global approach

The new study grew out of a smaller analysis that Jacoby and McDermott published in 2017. In that paper, the researchers compared rhythm perception in groups of listeners from the United States and the Tsimane’, an Indigenous society located in the Bolivian Amazon rainforest.

pitch perception study
Nori Jacoby, a former MIT postdoc now at the Max Planck Institute for Empirical Aesthetics, runs an experiment with a member of the Tsimane’ tribe, who have had little exposure to Western music. Photo: Josh McDermott

To measure how people perceive rhythm, the researchers devised a task in which they play a randomly generated series of four beats and then ask the listener to tap back what they heard. The rhythm produced by the listener is then played back to the listener, and they tap it back again. Over several iterations, the tapped sequences became dominated by the listener’s internal biases, also known as priors.

“The initial stimulus pattern is random, but at each iteration the pattern is pushed by the listener’s biases, such that it tends to converge to a particular point in the space of possible rhythms,” McDermott says. “That can give you a picture of what we call the prior, which is the set of internal implicit expectations for rhythms that people have in their heads.”

When the researchers first did this experiment, with American college students as the test subjects, they found that people tended to produce time intervals that are related by simple integer ratios. Furthermore, most of the rhythms they produced, such as those with ratios of 1:1:2 and 2:3:3, are commonly found in Western music.

The researchers then went to Bolivia and asked members of the Tsimane’ society to perform the same task. They found that Tsimane’ also produced rhythms with simple integer ratios, but their preferred ratios were different and appeared to be consistent with those that have been documented in the few existing records of Tsimane’ music.

“At that point, it provided some evidence that there might be very widespread tendencies to favor these small integer ratios, and that there might be some degree of cross-cultural variation. But because we had just looked at this one other culture, it really wasn’t clear how this was going to look at a broader scale,” Jacoby says.

To try to get that broader picture, the MIT team began seeking collaborators around the world who could help them gather data on a more diverse set of populations. They ended up studying listeners from 39 groups, representing 15 countries on five continents — North America, South America, Europe, Africa, and Asia.

“This is really the first study of its kind in the sense that we did the same experiment in all these different places, with people who are on the ground in those locations,” McDermott says. “That hasn’t really been done before at anything close to this scale, and it gave us an opportunity to see the degree of variation that might exist around the world.”

A grid of nine different photos showing a researcher working with an individual at a table. The individuals are wearing headphones.
Example testing sites. a, Yaranda, Bolivia. b, Montevideo, Uruguay. c, Sagele, Mali. d, Spitzkoppe, Namibia. e, Pleven, Bulgaria. f, Bamako, Mali. g, D’Kar, Botswana. h, Stockholm, Sweden. i, Guizhou, China. j, Mumbai, India. Verbal informed consent was obtained from the individuals in each photo.

Cultural comparisons

Just as they had in their original 2017 study, the researchers found that in every group they tested, people tended to be biased toward simple integer ratios of rhythm. However, not every group showed the same biases. People from North America and Western Europe, who have likely been exposed to the same kinds of music, were more likely to generate rhythms with the same ratios. However, many groups, for example those in Turkey, Mali, Bulgaria, and Botswana showed a bias for other rhythms.

“There are certain cultures where there are particular rhythms that are prominent in their music, and those end up showing up in the mental representation of rhythm,” Jacoby says.

The researchers believe their findings reveal a mechanism that the brain uses to aid in the perception and production of music.

“When you hear somebody playing something and they have errors in their performance, you’re going to mentally correct for those by mapping them onto where you implicitly think they ought to be,” McDermott says. “If you didn’t have something like this, and you just faithfully represented what you heard, these errors might propagate and make it much harder to maintain a musical system.”

Among the groups that they studied, the researchers took care to include not only college students, who are easy to study in large numbers, but also people living in traditional societies, who are more difficult to reach. Participants from those more traditional groups showed significant differences from college students living in the same countries, and from people who live in those countries but performed the test online.

“What’s very clear from the paper is that if you just look at the results from undergraduate students around the world, you vastly underestimate the diversity that you see otherwise,” Jacoby says. “And the same was true of experiments where we tested groups of people online in Brazil and India, because you’re dealing with people who have internet access and presumably have more exposure to Western music.”

The researchers now hope to run additional studies of different aspects of music perception, taking this global approach.

“If you’re just testing college students around the world or people online, things look a lot more homogenous. I think it’s very important for the field to realize that you actually need to go out into communities and run experiments there, as opposed to taking the low-hanging fruit of running studies with people in a university or on the internet,” McDermott says.

The research was funded by the James S. McDonnell Foundation, the Canadian National Science and Engineering Research Council, the South African National Research Foundation, the United States National Science Foundation, the Chilean National Research and Development Agency, the Austrian Academy of Sciences, the Japan Society for the Promotion of Science, the Keio Global Research Institute, the United Kingdom Arts and Humanities Research Council, the Swedish Research Council, and the John Fell Fund.

Researchers uncover new CRISPR-like system in animals that can edit the human genome

A team of researchers led by Feng Zhang at the McGovern Institute and the Broad Institute of MIT and Harvard has uncovered the first programmable RNA-guided system in eukaryotes — organisms that include fungi, plants, and animals.

In a study in Nature, the team describes how the system is based on a protein called Fanzor. They showed that Fanzor proteins use RNA as a guide to target DNA precisely, and that Fanzors can be reprogrammed to edit the genome of human cells. The compact Fanzor systems have the potential to be more easily delivered to cells and tissues as therapeutics than CRISPR/Cas systems, and further refinements to improve their targeting efficiency could make them a valuable new technology for human genome editing.

CRISPR/Cas was first discovered in prokaryotes (bacteria and other single-cell organisms that lack nuclei) and scientists including Zhang’s lab have long wondered whether similar systems exist in eukaryotes. The new study demonstrates that RNA-guided DNA-cutting mechanisms are present across all kingdoms of life.

“This new system is another way to make precise changes in human cells, complementing the genome editing tools we already have.” — Feng Zhang

“CRISPR-based systems are widely used and powerful because they can be easily reprogrammed to target different sites in the genome,” said Zhang, senior author on the study and a core institute member at the Broad, an investigator at MIT’s McGovern Institute, the James and Patricia Poitras Professor of Neuroscience at MIT, and a Howard Hughes Medical Institute investigator. “This new system is another way to make precise changes in human cells, complementing the genome editing tools we already have.”

Searching the domains of life

A major aim of the Zhang lab is to develop genetic medicines using systems that can modulate human cells by targeting specific genes and processes. “A number of years ago, we started to ask, ‘What is there beyond CRISPR, and are there other RNA-programmable systems out there in nature?’” said Zhang.

Feng Zhang with folded arms in lab
McGovern Investigator Feng Zhang in his lab.

Two years ago, Zhang lab members discovered a class of RNA-programmable systems in prokaryotes called OMEGAs, which are often linked with transposable elements, or “jumping genes”, in bacterial genomes and likely gave rise to CRISPR/Cas systems. That work also highlighted similarities between prokaryotic OMEGA systems and Fanzor proteins in eukaryotes, suggesting that the Fanzor enzymes might also use an RNA-guided mechanism to target and cut DNA.

In the new study, the researchers continued their study of RNA-guided systems by isolating Fanzors from fungi, algae, and amoeba species, in addition to a clam known as the Northern Quahog. Co-first author Makoto Saito of the Zhang lab led the biochemical characterization of the Fanzor proteins, showing that they are DNA-cutting endonuclease enzymes that use nearby non-coding RNAs known as ωRNAs to target particular sites in the genome. It is the first time this mechanism has been found in eukaryotes, such as animals.

Unlike CRISPR proteins, Fanzor enzymes are encoded in the eukaryotic genome within transposable elements and the team’s phylogenetic analysis suggests that the Fanzor genes have migrated from bacteria to eukaryotes through so-called horizontal gene transfer.

“These OMEGA systems are more ancestral to CRISPR and they are among the most abundant proteins on the planet, so it makes sense that they have been able to hop back and forth between prokaryotes and eukaryotes,” said Saito.

To explore Fanzor’s potential as a genome editing tool, the researchers demonstrated that it can generate insertions and deletions at targeted genome sites within human cells. The researchers found the Fanzor system to initially be less efficient at snipping DNA than CRISPR/Cas systems, but by systematic engineering, they introduced a combination of mutations into the protein that increased its activity 10-fold. Additionally, unlike some CRISPR systems and the OMEGA protein TnpB, the team found that a fungal-derived Fanzor protein did not exhibit “collateral activity,” where an RNA-guided enzyme cleaves its DNA target as well as degrading nearby DNA or RNA. The results suggest that Fanzors could potentially be developed as efficient genome editors.

Co-first author Peiyu Xu led an effort to analyze the molecular structure of the Fanzor/ωRNA complex and illustrate how it latches onto DNA to cut it. Fanzor shares structural similarities with its prokaryotic counterpart CRISPR-Cas12 protein, but the interaction between the ωRNA and the catalytic domains of Fanzor is more extensive, suggesting that the ωRNA might play a role in the catalytic reactions. “We are excited about these structural insights for helping us further engineer and optimize Fanzor for improved efficiency and precision as a genome editor,” said Xu.

Like CRISPR-based systems, the Fanzor system can be easily reprogrammed to target specific genome sites, and Zhang said it could one day be developed into a powerful new genome editing technology for research and therapeutic applications. The abundance of RNA-guided endonucleases like Fanzors further expands the number of OMEGA systems known across kingdoms of life and suggests that there are more yet to be found.

“Nature is amazing. There’s so much diversity,” said Zhang. “There are probably more RNA-programmable systems out there, and we’re continuing to explore and will hopefully discover more.”

The paper’s other authors include Guilhem Faure, Samantha Maguire, Soumya Kannan, Han Altae-Tran, Sam Vo, AnAn Desimone, and Rhiannon Macrae.

Support for this work was provided by the Howard Hughes Medical Institute; Poitras Center for Psychiatric Disorders Research at MIT; K. Lisa Yang and Hock E. Tan Molecular Therapeutics Center at MIT; Broad Institute Programmable Therapeutics Gift Donors; The Pershing Square Foundation, William Ackman, and Neri Oxman; James and Patricia Poitras; BT Charitable Foundation; Asness Family Foundation; Kenneth C. Griffin; the Phillips family; David Cheng; Robert Metcalfe; and Hugo Shong.

 

Unraveling connections between the brain and gut

The brain and the digestive tract are in constant communication, relaying signals that help to control feeding and other behaviors. This extensive communication network also influences our mental state and has been implicated in many neurological disorders.

MIT engineers have designed a new technology for probing those connections. Using fibers embedded with a variety of sensors, as well as light sources for optogenetic stimulation, the researchers have shown that they can control neural circuits connecting the gut and the brain, in mice.

In a new study, the researchers demonstrated that they could induce feelings of fullness or reward-seeking behavior in mice by manipulating cells of the intestine. In future work, they hope to explore some of the correlations that have been observed between digestive health and neurological conditions such as autism and Parkinson’s disease.

“The exciting thing here is that we now have technology that can drive gut function and behaviors such as feeding. More importantly, we have the ability to start accessing the crosstalk between the gut and the brain with the millisecond precision of optogenetics, and we can do it in behaving animals,” says Polina Anikeeva, the Matoula S. Salapatas Professor in Materials Science and Engineering, a professor of brain and cognitive sciences, director of the K. Lisa Yang Brain-Body Center, associate director of MIT’s Research Laboratory of Electronics, and a member of MIT’s McGovern Institute for Brain Research.

Portait of MIT scientist Polina Anikeeva
McGovern Institute Associate Investigator Polina Anikeeva in her lab. Photo: Steph Stevens

Anikeeva is the senior author of the new study, which appears today in Nature Biotechnology. The paper’s lead authors are MIT graduate student Atharva Sahasrabudhe, Duke University postdoc Laura Rupprecht, MIT postdoc Sirma Orguc, and former MIT postdoc Tural Khudiyev.

The brain-body connection

Last year, the McGovern Institute launched the K. Lisa Yang Brain-Body Center to study the interplay between the brain and other organs of the body. Research at the center focuses on illuminating how these interactions help to shape behavior and overall health, with a goal of developing future therapies for a variety of diseases.

“There’s continuous, bidirectional crosstalk between the body and the brain,” Anikeeva says. “For a long time, we thought the brain is a tyrant that sends output into the organs and controls everything. But now we know there’s a lot of feedback back into the brain, and this feedback potentially controls some of the functions that we have previously attributed exclusively to the central neural control.”

As part of the center’s work, Anikeeva set out to probe the signals that pass between the brain and the nervous system of the gut, also called the enteric nervous system. Sensory cells in the gut influence hunger and satiety via both the neuronal communication and hormone release.

Untangling those hormonal and neural effects has been difficult because there hasn’t been a good way to rapidly measure the neuronal signals, which occur within milliseconds.

“We needed a device that didn’t exist. So, we decided to make it,” says Atharva Sahasrabudhe.

“To be able to perform gut optogenetics and then measure the effects on brain function and behavior, which requires millisecond precision, we needed a device that didn’t exist. So, we decided to make it,” says Sahasrabudhe, who led the development of the gut and brain probes.

The electronic interface that the researchers designed consists of flexible fibers that can carry out a variety of functions and can be inserted into the organs of interest. To create the fibers, Sahasrabudhe used a technique called thermal drawing, which allowed him to create polymer filaments, about as thin as a human hair, that can be embedded with electrodes and temperature sensors.

The filaments also carry microscale light-emitting devices that can be used to optogenetically stimulate cells, and microfluidic channels that can be used to deliver drugs.

The mechanical properties of the fibers can be tailored for use in different parts of the body. For the brain, the researchers created stiffer fibers that could be threaded deep into the brain. For digestive organs such as the intestine, they designed more delicate rubbery fibers that do not damage the lining of the organs but are still sturdy enough to withstand the harsh environment of the digestive tract.

“To study the interaction between the brain and the body, it is necessary to develop technologies that can interface with organs of interest as well as the brain at the same time, while recording physiological signals with high signal-to-noise ratio,” Sahasrabudhe says. “We also need to be able to selectively stimulate different cell types in both organs in mice so that we can test their behaviors and perform causal analyses of these circuits.”

The fibers are also designed so that they can be controlled wirelessly, using an external control circuit that can be temporarily affixed to the animal during an experiment. This wireless control circuit was developed by Orguc, a Schmidt Science Fellow, and Harrison Allen ’20, MEng ’22, who were co-advised between the Anikeeva lab and the lab of Anantha Chandrakasan, dean of MIT’s School of Engineering and the Vannevar Bush Professor of Electrical Engineering and Computer Science.

Driving behavior

Using this interface, the researchers performed a series of experiments to show that they could influence behavior through manipulation of the gut as well as the brain.

First, they used the fibers to deliver optogenetic stimulation to a part of the brain called the ventral tegmental area (VTA), which releases dopamine. They placed mice in a cage with three chambers, and when the mice entered one particular chamber, the researchers activated the dopamine neurons. The resulting dopamine burst made the mice more likely to return to that chamber in search of the dopamine reward.

Then, the researchers tried to see if they could also induce that reward-seeking behavior by influencing the gut. To do that, they used fibers in the gut to release sucrose, which also activated dopamine release in the brain and prompted the animals to seek out the chamber they were in when sucrose was delivered.

Next, working with colleagues from Duke University, the researchers found they could induce the same reward-seeking behavior by skipping the sucrose and optogenetically stimulating nerve endings in the gut that provide input to the vagus nerve, which controls digestion and other bodily functions.

Three scientists holding a fiber in a lab.
Duke University postdoc Laura Rupprecht, MIT graduate student Atharva Sahasrabudhe, and MIT postdoc Sirma Orguc holding their engineered flexible fiber in Polina Anikeeva’s lab at MIT. Photo: Courtesy of the researchers

“Again, we got this place preference behavior that people have previously seen with stimulation in the brain, but now we are not touching the brain. We are just stimulating the gut, and we are observing control of central function from the periphery,” Anikeeva says.

Sahasrabudhe worked closely with Rupprecht, a postdoc in Professor Diego Bohorquez’ group at Duke, to test the fibers’ ability to control feeding behaviors. They found that the devices could optogenetically stimulate cells that produce cholecystokinin, a hormone that promotes satiety. When this hormone release was activated, the animals’ appetites were suppressed, even though they had been fasting for several hours. The researchers also demonstrated a similar effect when they stimulated cells that produce a peptide called PYY, which normally curbs appetite after very rich foods are consumed.

The researchers now plan to use this interface to study neurological conditions that are believed to have a gut-brain connection. For instance, studies have shown that autistic children are far more likely than their peers to be diagnosed with GI dysfunction, while anxiety and irritable bowel syndrome share genetic risks.

“We can now begin asking, are those coincidences, or is there a connection between the gut and the brain? And maybe there is an opportunity for us to tap into those gut-brain circuits to begin managing some of those conditions by manipulating the peripheral circuits in a way that does not directly ‘touch’ the brain and is less invasive,” Anikeeva says.

The research was funded, in part, by the Hock E. Tan and K. Lisa Yang Center for Autism Research and the K. Lisa Yang Brain-Body Center, the National Institute of Neurological Disorders and Stroke, the National Science Foundation (NSF) Center for Materials Science and Engineering, the NSF Center for Neurotechnology, the National Center for Complementary and Integrative Health, a National Institutes of Health Director’s Pioneer Award, the National Institute of Mental Health, and the National Institute of Diabetes and Digestive and Kidney Diseases.

Computational model mimics humans’ ability to predict emotions

When interacting with another person, you likely spend part of your time trying to anticipate how they will feel about what you’re saying or doing. This task requires a cognitive skill called theory of mind, which helps us to infer other people’s beliefs, desires, intentions, and emotions.

MIT neuroscientists have now designed a computational model that can predict other people’s emotions — including joy, gratitude, confusion, regret, and embarrassment — approximating human observers’ social intelligence. The model was designed to predict the emotions of people involved in a situation based on the prisoner’s dilemma, a classic game theory scenario in which two people must decide whether to cooperate with their partner or betray them.

To build the model, the researchers incorporated several factors that have been hypothesized to influence people’s emotional reactions, including that person’s desires, their expectations in a particular situation, and whether anyone was watching their actions.

“These are very common, basic intuitions, and what we said is, we can take that very basic grammar and make a model that will learn to predict emotions from those features,” says Rebecca Saxe, the John W. Jarve Professor of Brain and Cognitive Sciences, a member of MIT’s McGovern Institute for Brain Research, and the senior author of the study.

Sean Dae Houlihan PhD ’22, a postdoc at the Neukom Institute for Computational Science at Dartmouth College, is the lead author of the paper, which appears today in Philosophical Transactions A. Other authors include Max Kleiman-Weiner PhD ’18, a postdoc at MIT and Harvard University; Luke Hewitt PhD ’22, a visiting scholar at Stanford University; and Joshua Tenenbaum, a professor of computational cognitive science at MIT and a member of the Center for Brains, Minds, and Machines and MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL).

Predicting emotions

While a great deal of research has gone into training computer models to infer someone’s emotional state based on their facial expression, that is not the most important aspect of human emotional intelligence, Saxe says. Much more important is the ability to predict someone’s emotional response to events before they occur.

“The most important thing about what it is to understand other people’s emotions is to anticipate what other people will feel before the thing has happened,” she says. “If all of our emotional intelligence was reactive, that would be a catastrophe.”

To try to model how human observers make these predictions, the researchers used scenarios taken from a British game show called “Golden Balls.” On the show, contestants are paired up with a pot of $100,000 at stake. After negotiating with their partner, each contestant decides, secretly, whether to split the pool or try to steal it. If both decide to split, they each receive $50,000. If one splits and one steals, the stealer gets the entire pot. If both try to steal, no one gets anything.

Depending on the outcome, contestants may experience a range of emotions — joy and relief if both contestants split, surprise and fury if one’s opponent steals the pot, and perhaps guilt mingled with excitement if one successfully steals.

To create a computational model that can predict these emotions, the researchers designed three separate modules. The first module is trained to infer a person’s preferences and beliefs based on their action, through a process called inverse planning.

“This is an idea that says if you see just a little bit of somebody’s behavior, you can probabilistically infer things about what they wanted and expected in that situation,” Saxe says.

Using this approach, the first module can predict contestants’ motivations based on their actions in the game. For example, if someone decides to split in an attempt to share the pot, it can be inferred that they also expected the other person to split. If someone decides to steal, they may have expected the other person to steal, and didn’t want to be cheated. Or, they may have expected the other person to split and decided to try to take advantage of them.

The model can also integrate knowledge about specific players, such as the contestant’s occupation, to help it infer the players’ most likely motivation.

The second module compares the outcome of the game with what each player wanted and expected to happen. Then, a third module predicts what emotions the contestants may be feeling, based on the outcome and what was known about their expectations. This third module was trained to predict emotions based on predictions from human observers about how contestants would feel after a particular outcome. The authors emphasize that this is a model of human social intelligence, designed to mimic how observers causally reason about each other’s emotions, not a model of how people actually feel.

“From the data, the model learns that what it means, for example, to feel a lot of joy in this situation, is to get what you wanted, to do it by being fair, and to do it without taking advantage,” Saxe says.

Core intuitions

Once the three modules were up and running, the researchers used them on a new dataset from the game show to determine how the models’ emotion predictions compared with the predictions made by human observers. This model performed much better at that task than any previous model of emotion prediction.

The model’s success stems from its incorporation of key factors that the human brain also uses when predicting how someone else will react to a given situation, Saxe says. Those include computations of how a person will evaluate and emotionally react to a situation, based on their desires and expectations, which relate to not only material gain but also how they are viewed by others.

“Our model has those core intuitions, that the mental states underlying emotion are about what you wanted, what you expected, what happened, and who saw. And what people want is not just stuff. They don’t just want money; they want to be fair, but also not to be the sucker, not to be cheated,” she says.

“The researchers have helped build a deeper understanding of how emotions contribute to determining our actions; and then, by flipping their model around, they explain how we can use people’s actions to infer their underlying emotions. This line of work helps us see emotions not just as ‘feelings’ but as playing a crucial, and subtle, role in human social behavior,” says Nick Chater, a professor of behavioral science at the University of Warwick, who was not involved in the study.

In future work, the researchers hope to adapt the model so that it can perform more general predictions based on situations other than the game-show scenario used in this study. They are also working on creating models that can predict what happened in the game based solely on the expression on the faces of the contestants after the results were announced.

The research was funded by the McGovern Institute; the Paul E. and Lilah Newton Brain Science Award; the Center for Brains, Minds, and Machines; the MIT-IBM Watson AI Lab; and the Multidisciplinary University Research Initiative.