Amsterdam Brain and Cognition (ABC)

Category formation

Research priority area Brain and Cognition

Models and tests of early category formation: interactions between cognitive, emotional, and neural mechanisms

Models and tests of early category formation: interactions between cognitive, emotional, and neural mechanisms

Interview: Searching for the basis of categorization

by Vittorio Busato

‘How do babies who don't yet speak a language learn to differentiate between types of categories such as dogs and cats; between sounds and speech sounds such as ba and da or p, d and g; between primary emotions such as angry and happy, safe and unsafe? How are such categories represented in the brain? How do babies generalize knowledge they acquire to other categories? Is there overlap in that learning process? What do children learn from feedback about mistakes made? How do these young children make a developmental leap and suddenly acquire the insight to accurately differentiate between categories? What are the characteristics of those categories that have emerged? Does categorical perception take place, or would we rather speak of a continuous perception based on which a child comes to a categorical classification? What about the individual differences between babies in the speed at which they familiarize themselves with a categorization process? These are all important questions that are at the basis of how we learn to make various types of differentiations in the world around us.'

These are the words of Maartje Raijmakers. She is an associate professor with the Department of Psychology at the University of Amsterdam (UvA). Together with Paul Boersma and Susan Bögels, professor of phonetics and professor of orthopedagogics at the same university, respectively, she is responsible as the lead researcher for the project Models and tests of early category formation: interactions between cognitive, emotional and neural mechanisms, one of four projects awarded in the interdisciplinary and interfaculty university research priority program Brain & Cognition of the UvA.

New model

All three are among the few researchers at the UvA who already do scientific research with babies and test them here, too. What's more, Raijmakers and Boersma have substantial experience in modeling these types of categorization processes and have published on this. Raijmakers has developed the adaptive resonance model for category learning, a model that-briefly put-tries to explain how feedback helps children to differentiate between categories. Boersma is the spiritual father of the so-called bidirectional multi-level language acquisition model, a model that describes how we acquire a language through listening and speaking. What exactly it will look like they don't yet know in this phase of the project, but based on experiments Raijmakers and Boersma hope to eventually combine their two models into a new model.

In this recently started project they not only unite practical knowledge they gained while conducting baby research, but also their expertise in the area of phonetics, psychology and pedagogy with respect to learning and forming categories. Especially for this project a new baby lab will be set up, complete with state-of-the-art equipment to register eye movement in babies and record EEGs, among others.

Paola Escudero Neyra, who is involved with this university research priority program as a postdoc, has gained some experience recording EEGs in babies and has completed several exploratory studies at the baby lab of American Scott P. Johnson in Los Angeles. Johnson, an internationally renowned expert in the area of cognitive neuroscientific baby research, will be involved in this project as a visiting professor at a later stage. Together with their postdocs, graduate students and undergraduate students, Raijmakers, Boersma and Bögels hope to find, as part of this extensive research project, answers to fundamental questions about the inception and forming of categories. As an added practical advantage, they will be able to join efforts to recruit babies for their experiments.

Bögels thinks that with this project they have started a rather unique interdisciplinary collaboration, from an international perspective. ‘My research is predominantly focused on the origin of fears and the specific role of parents in this-social referencing is the scientific jargon. There are evolutionary aspects to primary emotions such as fear and anger, but we also partly learn those through parents and other experts. Parents will send signals as to whether something is safe or not. Babies learn to classify those stimuli in categories such as safe or dangerous and will base their behavior on those: proceed or avoid. The added value of our collaboration that I'm most looking for is that we will discover more about the possible overlap in learning different types of categories. By bundling our knowledge and expertise we will be able to fine tune these processes much more in terms of methodology.

Boersma has a similar opinion. ‘Although our project involves basic science research, part of it could most certainly be characterized as exploratory. That's why, frankly speaking, I don't have very concrete expectations, other than that as researchers from different disciplines we will certainly learn a lot from each other.'

New paradigm

The usual paradigm to study category learning in babies is the so-called habituation paradigm. Babies will repeatedly get to see pictures of one type of category, such as dogs, or they will keep hearing a similar sound, such as ba. At some point babies will start to find that boring. They'll get used to those similar stimuli and will therefore likely recognize them; they've learned something. They will barely pay attention to a new dog, but they will when they see a picture of a cat or hear a different sound. Research shows that this shift in attention can be seen not only at the level of behavior, but also in the brain activity of babies.

‘We're not only interested in whether a baby learns something but particularly in how he or she learns something,' Raijmakers says, after explaining the habituation paradigm. But this requires a different paradigm that is closer to how adults learn to categorize. ‘We will study, among other things, how children learn two categories at the same time,'Raijmakers continues. ‘We have developed a task in which babies get to see moving dogs or cats on a screen. The dogs briefly disappear from the screen and reappear on the right side, while the cats reappear on the left side. During this task we will track the eye movements of the baby and will figure out whether a baby learns to anticipate where cats or dogs will reappear. When a baby correctly predicts this, as indicated by the eye movements, he or she has learned something about the categorization. Subsequently, you can use new pictures to test whether the baby can generalize this categorization. So we're looking at predictive ability; we don't settle for an observation of whether or not a child is bored.'

Another technique researchers will use is the so-called morphing of photos and sounds. Using special software, photos of facial expressions can slowly transition from angry to happy, for example. A program developed by Boersma can very gradually morph artificially generated sounds into one another. Bögels: ‘This technology enables us to zoom in much more precisely on the attention- and categorization process of babies and any shifts in that process.'

Psychoeducation

All in all it's an ambitious and interdisciplinary research project that Raijmakers, Boersma and Bögels are planning. Nevertheless the three scientists are cautious not to have exaggerated expectations of what the project will eventually yield, especially because of the exploratory aspects of their research. Boersma hopes that they will at least come to a new model of how the brain learns to categorize and classify. ‘Those processes, in my view, are in large part symbolic for what makes us human. And perhaps it will yield applied knowledge that will eventually benefit dyslectics, who have trouble forming sound categories, or autistic children, who have trouble reading emotions on faces."

Bögels expects that their fundamental research will yield practical knowledge on the learning and unlearning of fears in young children as part of interactions with their parents. ‘If our research project yields new implications on this, that's important for psychoeducation on the origin and treatment of fears. It's important for us to teach parents how they can think about certain fears in a more nuanced way, but especially how they can impart that on their children.'

Raijmakers is mainly focusing her hopes on better theory development on how we learn to categorize. ‘We won't stop at descriptions, but will try to come to explanations. How, for example, does a baby determine whether an unexpected event means danger or is interesting to explore? When we understand more about the developmental process of early category formation, we also learn to better understand the functioning of adults.'

Contact

Prof. dr. Paul Boersma (ACLC)
Dr. Maartje Raijmakers (Psychology)
Prof. dr. Susan Bögels (CDE)

Participating institutes:

The following institutes participate:

Interview: Searching for the basis of categorization

by Vittorio Busato

‘How do babies who don't yet speak a language learn to differentiate between types of categories such as dogs and cats; between sounds and speech sounds such as ba and da or p, d and g; between primary emotions such as angry and happy, safe and unsafe? How are such categories represented in the brain? How do babies generalize knowledge they acquire to other categories? Is there overlap in that learning process? What do children learn from feedback about mistakes made? How do these young children make a developmental leap and suddenly acquire the insight to accurately differentiate between categories? What are the characteristics of those categories that have emerged? Does categorical perception take place, or would we rather speak of a continuous perception based on which a child comes to a categorical classification? What about the individual differences between babies in the speed at which they familiarize themselves with a categorization process? These are all important questions that are at the basis of how we learn to make various types of differentiations in the world around us.'

These are the words of Maartje Raijmakers. She is an associate professor with the Department of Psychology at the University of Amsterdam (UvA). Together with Paul Boersma and Susan Bögels, professor of phonetics and professor of orthopedagogics at the same university, respectively, she is responsible as the lead researcher for the project Models and tests of early category formation: interactions between cognitive, emotional and neural mechanisms, one of four projects awarded in the interdisciplinary and interfaculty university research priority program Brain & Cognition of the UvA.

New model

All three are among the few researchers at the UvA who already do scientific research with babies and test them here, too. What's more, Raijmakers and Boersma have substantial experience in modeling these types of categorization processes and have published on this. Raijmakers has developed the adaptive resonance model for category learning, a model that-briefly put-tries to explain how feedback helps children to differentiate between categories. Boersma is the spiritual father of the so-called bidirectional multi-level language acquisition model, a model that describes how we acquire a language through listening and speaking. What exactly it will look like they don't yet know in this phase of the project, but based on experiments Raijmakers and Boersma hope to eventually combine their two models into a new model.

In this recently started project they not only unite practical knowledge they gained while conducting baby research, but also their expertise in the area of phonetics, psychology and pedagogy with respect to learning and forming categories. Especially for this project a new baby lab will be set up, complete with state-of-the-art equipment to register eye movement in babies and record EEGs, among others.

Paola Escudero Neyra, who is involved with this university research priority program as a postdoc, has gained some experience recording EEGs in babies and has completed several exploratory studies at the baby lab of American Scott P. Johnson in Los Angeles. Johnson, an internationally renowned expert in the area of cognitive neuroscientific baby research, will be involved in this project as a visiting professor at a later stage. Together with their postdocs, graduate students and undergraduate students, Raijmakers, Boersma and Bögels hope to find, as part of this extensive research project, answers to fundamental questions about the inception and forming of categories. As an added practical advantage, they will be able to join efforts to recruit babies for their experiments.

Bögels thinks that with this project they have started a rather unique interdisciplinary collaboration, from an international perspective. ‘My research is predominantly focused on the origin of fears and the specific role of parents in this-social referencing is the scientific jargon. There are evolutionary aspects to primary emotions such as fear and anger, but we also partly learn those through parents and other experts. Parents will send signals as to whether something is safe or not. Babies learn to classify those stimuli in categories such as safe or dangerous and will base their behavior on those: proceed or avoid. The added value of our collaboration that I'm most looking for is that we will discover more about the possible overlap in learning different types of categories. By bundling our knowledge and expertise we will be able to fine tune these processes much more in terms of methodology.

Boersma has a similar opinion. ‘Although our project involves basic science research, part of it could most certainly be characterized as exploratory. That's why, frankly speaking, I don't have very concrete expectations, other than that as researchers from different disciplines we will certainly learn a lot from each other.'

Psychoeducation

All in all it's an ambitious and interdisciplinary research project that Raijmakers, Boersma and Bögels are planning. Nevertheless the three scientists are cautious not to have exaggerated expectations of what the project will eventually yield, especially because of the exploratory aspects of their research. Boersma hopes that they will at least come to a new model of how the brain learns to categorize and classify. ‘Those processes, in my view, are in large part symbolic for what makes us human. And perhaps it will yield applied knowledge that will eventually benefit dyslectics, who have trouble forming sound categories, or autistic children, who have trouble reading emotions on faces."

Bögels expects that their fundamental research will yield practical knowledge on the learning and unlearning of fears in young children as part of interactions with their parents. ‘If our research project yields new implications on this, that's important for psychoeducation on the origin and treatment of fears. It's important for us to teach parents how they can think about certain fears in a more nuanced way, but especially how they can impart that on their children.'

Raijmakers is mainly focusing her hopes on better theory development on how we learn to categorize. ‘We won't stop at descriptions, but will try to come to explanations. How, for example, does a baby determine whether an unexpected event means danger or is interesting to explore? When we understand more about the developmental process of early category formation, we also learn to better understand the functioning of adults.'

Participating institutes:

The following institutes participate:

New paradigm

The usual paradigm to study category learning in babies is the so-called habituation paradigm. Babies will repeatedly get to see pictures of one type of category, such as dogs, or they will keep hearing a similar sound, such as ba. At some point babies will start to find that boring. They'll get used to those similar stimuli and will therefore likely recognize them; they've learned something. They will barely pay attention to a new dog, but they will when they see a picture of a cat or hear a different sound. Research shows that this shift in attention can be seen not only at the level of behavior, but also in the brain activity of babies.

‘We're not only interested in whether a baby learns something but particularly in how he or she learns something,' Raijmakers says, after explaining the habituation paradigm. But this requires a different paradigm that is closer to how adults learn to categorize. ‘We will study, among other things, how children learn two categories at the same time,'Raijmakers continues. ‘We have developed a task in which babies get to see moving dogs or cats on a screen. The dogs briefly disappear from the screen and reappear on the right side, while the cats reappear on the left side. During this task we will track the eye movements of the baby and will figure out whether a baby learns to anticipate where cats or dogs will reappear. When a baby correctly predicts this, as indicated by the eye movements, he or she has learned something about the categorization. Subsequently, you can use new pictures to test whether the baby can generalize this categorization. So we're looking at predictive ability; we don't settle for an observation of whether or not a child is bored.'

Another technique researchers will use is the so-called morphing of photos and sounds. Using special software, photos of facial expressions can slowly transition from angry to happy, for example. A program developed by Boersma can very gradually morph artificially generated sounds into one another. Bögels: ‘This technology enables us to zoom in much more precisely on the attention- and categorization process of babies and any shifts in that process.'

Contact

Prof. dr. Paul Boersma (ACLC)
Dr. Maartje Raijmakers (Psychology)
Prof. dr. Susan Bögels (CDE)

Published by  Amsterdam Brain and Cognition (ABC)

1 September 2015