Cognitive Science Colloquium Series
2023-2024 Series
October 26, 2023 (5:00 pm, SH 181)
Chris Kello (University of California, Merced)
The stunning rise of large language models
Deep learning has its roots in “connectionist” neural network modeling that is foundational to the field of cognitive science. Connectionist models were coined as “deep learning” when they became powerful enough for industrial-strength applications in, most notably, computer vision and speech recognition. Natural language processing was slower at first to benefit from deep learning, but then the transformer neural network, introduced in 2017, advanced sequence processing to a level that few if any researchers could have imagined just fifteen years ago. This talk will briefly review the history of deep learning leading up to the large language models that underlie ChatGPT, including some basics of how they work, and how they are changing cognitive science and the world.
Video of the talk is available here. PPT slides are available here.
2022-2023 Series
April 20, 2023 (5:00 pm, Science Hall 181)
Heather Sheridan (University at Albany, SUNY)
The perceptual skill of experts in music and chess: Evidence from eye movements
In this talk, I will discuss how eye tracking methodology can be used to study the remarkable perceptual skill of experts. Specifically, I will discuss how eye tracking can be used to test chunking and template theories of expertise. According to these theories, over the course of many hours of practice, experts learn how to process domain-specific visual stimuli in terms of larger patterns instead of individual features. I’ll discuss a series of experiments that tested these theories using visual search and visual change detection paradigms that were developed in the domains of chess and music. The results from these experiments indicated that both chess and music experts can rapidly shift their attention towards relevant regions of an image and can rapidly encode complex domain-specific visual patterns. I will discuss the implications of these results for chunking and template theories, and I’ll also discuss how these theories could be extended in the future to accommodate the multisensory memory representations of music experts.
2021-2022 Series
April 21, 2022 (5:00 pm, Science Hall 181)
Theo Rhodes (SUNY Oswego)
Memory as Foraging and Foraging as Memory
Performance in a free recall memory task (e.g., name as many animals as you can) is characterized by bursts and clusters of recalls interspersed with longer pauses. Animal foraging behavior (e.g., squirrels looking for nuts) is also characterized by bursts of short movements interspersed with longer paths. The similarity between these behaviors goes beyond analogy, and the shared statistical properties of the behaviors suggest a common underlying cause. I will present results from human memory tasks as well as human foraging tasks, showing how each improves our understanding of the other and how they may inform our understanding of cognition more broadly.
2019-2020 Series
March 3, 2020 (5:00 pm, CSB Auditorium)
Language and cognition in d/Deaf children: Solving the right problems, and solving the problems right
Matthew Hall (Temple University)
Maximizing developmental outcomes in d/Deaf and hard of hearing (DHH) children requires a deeper understanding of the reasons why disparities between DHH and hearing children persist. Deficits in implicit learning and executive function are one possible factor. In contrast to views that link difficulties in these domains to auditory deprivation, my research shows that reported deficits in implicit learning may have been an artifact of poor methodology. I find that deficits in executive function are real, but more likely to stem from diminished access to language rather than diminished access to sound. According to this view, deaf children from hearing families are expected to benefit from early access to American Sign Language (ASL), provided they have sufficient access to ASL input. The crucial question then is: how much access is enough? To answer it, we first need a method of quantifying the distribution of input to which DHH children have access, especially during infancy and toddlerhood. I have begun to develop such a method, building on similar work in hearing multilinguals. I will share results from our first two waves of data collection families of DHH children from across the nation, discuss analytical challenges, and present my vision for how this method -once fully developed- will be useful for both research and clinical/educational practice. The ultimate aim of this research is to ensure that all DHH children arrive at school with the linguistic, cognitive, and social-emotional competence they need to achieve their full potential.
2018-2019 Series
February 25, 2019 (5:00 pm, CSB Auditorium)
What doesn’t kill you makes you stronger: Psychological trauma and its relationship to enhanced memory control
Justin C. Hulbert (Bard College)
Students enter college with a wide variety of life experiences that help shape and prepare them for the challenges they’ll encounter in school and beyond. Unfortunately, statistics suggest that many of those experiences could be considered traumatic, such as witnessing or experiencing major accidents or acts of violence. While these traumas are, without question, awful, could previous exposure to certain types of trauma enhance future resilience by essentially training individuals to more effectively cope with a range of unwanted memories? I will present the results of two studies that explored this possibility, highlighting along the way some of the causes and consequences of motivated forgetting in everyday life.
The video of the lecture can be found here.
November 19, 2018 (5:00 pm, LC 104)
From Psychology to Language and back: The role of intentions in utterance comprehension
Ira Noveck (Institut des Sciences Cognitives — Marc Jeannerod)
Paul Grice famously made a distinction between sentence meaning(the properties of a sentence assigned to it by the grammar) and speaker’s meaning (what the speaker intended to communicate by uttering a sentence). This distinction is arguably at the core of Experimental Pragmatics, which is a discipline devoted to (a) testing pragmatic theories (and often between two or more) and, more generally, (b) investigating pragmatic phenomena, with the rigorous methods of experimental psychology. One goal of this talk is to describe how an experimental approach has added value to debates and discussions concerning pragmatics. A second goal is to describe how Grice indirectly augured a different sort of divide, i.e. between researchers who have employed his approach in order to propose what are essentially linguistic rules that lead to a speaker’s meaning and those who have focused on the role played by intention-reading in accessing a speaker’s meaning. In retrospect, intention-reading has not received nearly as much attention in experimental pragmatics as, say, the hypothesized step-by-step accounts of scalar inferences (e.g. how a listener interprets Some cabs are yellowto mean Some but not all cabs are yellow). I will thus describe how Grice envisioned communication as a means of accessing a speaker’s intentions. From there, I will turn to areas of experimental pragmatic research that highlight the role of intention-reading in communication.
2017-2018 Series
May 7, 2018 (5:00 pm, Science Hall 181)
Men Hunt and Women Nest: Challenging Claims about Intrinsic Gender Differences
Alison Nash (SUNY New Paltz)
Scientific claims about intrinsic gender differences in behavior and cognition have been widespread for almost four decades, beginning with the backlash after the Women’s Movement of the 1970s. For more than 30 years, Dr. Nash’s scholarship has challenged the theoretical underpinnings that eclipse the importance of social context in understanding human behavior. Her theoretical papers have critiqued aspects of sociobiology, attachment theory, and evolutionary psychology, illustrating how flawed logic and methodology can serve to reify gender stereotypes. At the same time, her empirical research has taken her from Great Gull Island to the play spaces of young children as she critically examined assumptions about gendered behavior that are derived from these theories. From birds to babies to bubbies (i.e. grandmothers), Dr. Nash’s diverse methodologies have revealed surprising complexity and variability in behaviors that may be overlooked when they are simplistically described as innate.
March 27, 2018 (5:00 pm, CSB Auditorium)
Productive Struggle: Using Cognitive Science to Enhance Learning
Nate Kornell (Williams College)
If you want to perform well tomorrow, you should make things hard on yourself today; as Frederick Douglass said, “without a struggle, there can be no progress.” I will talk about the value of struggle in education. My focus will be strategies that succeed in the long term by making students to do worse in the short term. These strategies are often unpopular, with both students and teachers, because they reveal students’ weaknesses and seem to slow learning down. Interventions that make things easier, like the use of “learning styles” in education, are more popular but less effective. I will talk about “desirable difficulties” like spacing learning sessions apart and taking tests as a way to learn. I will also talk about how to make these strategies more appealing to students by decreasing their failure rate without decreasing their learning.
The video of the talk can be found here.
2016-2017 Series
December 5, 2016 (5:00 pm, CSB Auditorium)
Learning to move and moving to learn
Karen Adolph (New York University)
Infants learn to move in the context of continual development. Moreover, developmental changes in motor skills generate new opportunities for learning. A fruitful way to study these processes is to consider learning as embodied in the reality of infants’ growing and changing bodies, embedded in the practical exigencies of an ever-expanding physical environment, and enculturated by social interactions and culturally determined childrearing practices. In adopting this perspective, surprising findings have emerged that provide new insights into the relations between perception and action, and between learning and development.
The video of the talk can be found here.
October 25, 2016 (CSB Auditorium, 2:00 pm)
The sweet taste of childhood: From basic biology to health implications
Julie Mennella (Monell Chemical Senses Center, Philadelphia)
From the age of two years, an American is more likely to consume a sugar- sweetened product than a fruit or vegetable on any given day—a troubling statistic, given that food preferences are established early in childhood, as well as the strong association between this dietary pattern and increased risk of developing a number of chronic diseases. In this talk, I will review the ontogeny and biopsychology of sweet taste, highlighting how a biological drive to prefer sweetness at high concentrations during childhood, which would have conferred an advantage in environments of scarcity, predisposes them to overconsumption and obesity in the modern food environment replete with added sweetness.
October 14, 2016 (LC 104, 3:30 pm)
Speech: A signal for communication?
Athena Vouloumanos (New York University)
Like many animals, human infants have biases for the vocalizations of their own species, preferring speech to many non-speech sounds just hours after birth. How do these early proclivities develop and how do they contribute to human communicative development? I will draw from behavioral and neural data to discuss how infants come to recognize properties of speech that are relevant to its use for intentional communication by humans. Before preverbal infants produce or understand many words, they recognize how speech is used by others to communicate about different types of entities in the world. This early communicative competence may provide infants with a channel for learning from others and lay a foundation for our social and cultural life as humans.
2015-2016 Series
April 28, 2016 (LC 104, 5:00 pm)
Electrophysiology of face and object processing in deaf signers
Teresa Mitchell (University of Massachusetts Medical School, Brandeis University)
Cross-modal plasticity refers to changes within sensory modalities that are the result of the absence of functioning in one sensory system. In the case of congenital deafness, studies have shown that both behavioral and neural responses to visual motion are changed in specific ways. Behavioral sensitivity to motion in the periphery is greater in congenitally deaf than typically hearing individuals, and activity in brain regions typically devoted to analyzing visual motion are enhanced. This talk will describe research designed to characterize how deafness and native sign language use affect face processing and the neural network that subserves it. Face processing is a phylogenetically old system, with a long developmental timecourse within individuals. This sets up the possibility for congenital hearing loss to change the developmental outcome of behavioral and brain responses to this class of stimuli, as compared to other classes of visual forms, over the lifespan. Effects of native sign language use on face and object processing will also be examined.
March 16, 2016 (CSB Auditorium, 5:00 pm)
The myth of the Lehman Sisters? Sex, testosterone, and financial risk-taking
Cordelia Fine (University of Melbourne)
There is growing scientific interest in the role of testosterone in financial risk-taking – a topic of considerable public interest too, with complaints of there being ‘too much testosterone on Wall Street’. Both research and debate are often grounded in an implicit model in which testosterone is presumed to be the proximal mechanism underlying the evolved masculine trait of risk-taking. This talk will identify a number of conceptual and empirical problems with the underlying assumptions of this model, and report preliminary findings from a meta-analytic study on testosterone and financial risk-taking.
November 12, 2015 (LC 104, 3:30 pm)
Lay Theories of Mind and Brain
Diego Fernandez-Duque (Villanova University)
What are some of the common-sense beliefs that people hold about the mind, the brain, and their relationship? To address this question, I will describe two lines of research from my lab.
1. The Allure of Neuroscience. We found that explanations of psychological phenomena become more appealing when accompanied by irrelevant neuroscience information. I will discuss possible reasons why.
2. The Self is Brain-Made. We found that people think of their ‘true’ self as brain-based and unchangeable through willful control. I will discuss this in relation to determinism and moral responsibility.
The video of the talk is available here; the PPT slides of the talk are available here.
2014-2015 Series
April 29, 2015 (CSB Auditorium, 3:30 pm)
Zebrafish in Neuroscience: Modern Approaches to Studying Neuronal Dynamics in a Live Animal
Joseph DiPietro (Cornell University)
The various genetic tools available for larval zebrafish, as well as the transparency of their body make these animals a great model for studying vertebrate nervous system. The ability to target specific classes of cells and directly image, monitor, and manipulate them in an intact live animal open up the possibility to answer fundamental questions about the structure and function of the nervous system as well as the neural basis of behavior. In this talk I will discuss the use of zebrafish in modern neuroscience, the various experimental techniques available, and the types of questions that can be answered. Specifically, I will focus on the work I have done trying to understand how the dendrites of motor neurons in the spinal cord search for new inputs. Understanding the strategy that neurons use to find new inputs (and remove old ones) could provide clues into how the nervous system homeostatically regulates the formation and breaking of synaptic contacts. Because the tools available for zebrafish allow for noninvasive analysis of individual neurons in a live intact animal, it makes it possible answer questions about how dendritic and synaptic dynamics might change throughout the circadian cycle. This could provide clues about the role that sleep and circadian rhythms might have on the regulation of neuronal circuits.
The video of the talk is available here.
November 7, 2014 (LC 104, 2:30 pm)
Cognitive skill building in infancy and its influence on later development
Lisa S. Scott (University of Massachusetts, Amherst)
Using a combination of cross-sectional and longitudinal-training designs, behavioral measures of looking time, eye-tracking, and electrophysiological recordings of neural activity (event-related potentials; ERPs) we have begun to elucidate the perceptual and cognitive experiences that enhance or bias learning during the first year of life. Our work suggests that infants carefully learn from their surrounding environment and that this learning influences cognitive, perceptual and social processing. Here, I will present work that examines the role experience plays in shaping infant learning about people and objects and how parental labeling during the first year of life can serve as a springboard for cognitive skills in childhood. For example, when infants hear parents label two different monkey faces in a storybook with individual-level names like ‘Oliver’ or ‘Suzie’ they learn that it is likely important for them to attend to the visual details necessary to tell the two monkeys apart. However, if parents label all monkeys, “monkey” infants learn to group them into a category and focus on the features that the two monkeys share. These differences can be identified both in behavior and in the brain. Our recent findings suggest that individual-level learning in infancy results in skills lasting into early childhood (i.e., 4 years). Specifically, these skills benefit faces and/or objects that are perceived and recognized at the individual level (e.g., human faces). These results are noteworthy because they link early learning, prior to the onset of productive language and several years prior to formal education, with later cognitive skills and neural responses.
The video of the talk is available here.
October 23, 2014 (CSB auditorium, 3:30 pm)
Which aspects of language and cognition depend on linguistic input? Insights from homesign gesture systems
Marie Coppola (University of Connecticut)
Researchers in the cognitive sciences have long debated the relationships between linguistic input and language structure, as well as the relationships between language and cognition. Homesign systems offer a unique window into these relationships. Homesigns are gesture systems developed by deaf individuals who are not exposed to conventional sign or spoken language input. Homesign systems exhibit a number of linguistic properties, but appear to lack others, which depend on access to a linguistic model and/or interaction within a language community. I will show that homesign systems have structure at a variety of levels of linguistic analysis, including phonology and discourse structure. I will describe some of the developmental consequences of linguistic (but not social) deprivation, particularly with respect to number cognition. Finally, I will discuss my work with Manos Unidas (www.manos-unidas.org), a non-profit organization whose mission is to promote equal access to language and education for deaf individuals in Nicaragua.
The video of the talk is available here.
2013-2014 Series
May 1, 2014 (CSB Auditorium, 4:30 pm)
Individual differences in the shape bias in word learning in typically developing children and children with Specific Language Impairment
James S. Magnuson (University of Connecticut and Haskins Laboratories)
By the time children produce about 150 nouns, they exhibit a “shape bias” when generalizing object names. When children at this stage are shown a novel object and asked to find another like it (“look at this! can you find another one?”), they will randomly choose objects similar to the novel object in shape, texture, color, or size. When the object is given a name, however (“look at this wug! can you find another wug?”), they predominantly choose objects with the same shape. Linda Smith and her colleagues have argued that children have learned that there is coherent covariation between visible object properties and linguistic frames involving names; that is, that children have picked up on a regularity in the world that nameable solid objects are more likely to share shape than other dimensions. Given evidence that children with Specific Language Impairment (SLI) have subtle deficits in visual perception, we hypothesized that such subtle deficits would impair their ability to detect the visual-linguistic regularities that afford the shape bias and accelerate object word learning. We compared children ages 4-6 (well beyond the 150 noun stage) with typical language (TL) development and children with SLI in classic shape bias paradigms, and also assessed them on multiple linguistic and nonlinguistic abilities using standardized tests. We also assessed them with a simple paired visual associate (PVA) learning task repeated over 4 days. At the group level, children with TL showed a robust shape bias, and significant learning in the PVA task, while children with SLI failed to exhibit a shape bias and failed to learn in the PVA task. Individual differences analyses of all children revealed that PVA performance was the best predictor of shape bias; in fact, a regression model including only PVA could not be improved by including any other linguistic or nonlinguistic assessment. This suggests that subtle deficits in nonlinguistic domains may contribute significantly to word learning deficits and other deficits observed in SLI.
The video of the talk is available here.
March 27, 2014 (LC 104, 4:30 pm)
Audiovisual speech perception in children with autism spectrum disorders
Julia Irwin (Southern Connecticut State University)
In this talk I will present data comparing children with autism spectrum
disorders (ASD) to those with typical development (TD) on auditory, visual
and audiovisual speech perception. Using eye tracking methodology, we
assessed group differences in visual influence on heard speech and pattern
of gaze to speaking faces. There were no differences in perception of
auditory syllables /ma/ and /na/ in clear listening conditions or in the
presence of noise. In addition, there were no differences in perception of
a non-speech, non-face control. However, children with ASD were
significantly less visually influenced than TD controls in mismatched AV
and speech reading conditions, and showed less visual gain (AV speech in
the presence of auditory noise). Further, to examine whether differential
patterns of gaze may underlie these findings, we examined participant gaze
to the speaking faces. The children with ASD looked significantly less to
the face of the speaker overall. When children with ASD looked at a
speaker’s face, they looked less at the mouth of the speaker and more to
non-focal areas of the face during the speech reading and AV speech in
noise conditions. No group differences were observed for pattern of gaze to
non-face, non-speech controls.
December 2, 2013 (CSB Auditorium, 4:30 pm)
Development Evolving: From Innateness to Epigenesis
Mark Blumberg (University of Iowa)
With their shared interests in time and change, developmental scientists and evolutionary biologists should be natural allies. But for most of the 20th century, development was banished from mainstream evolutionary thinking. This rift can be traced to the belief that genes determine the bodies and behaviors of adults, thereby rendering development uninteresting and unimportant. But times have changed: Embryos and infants are now widely viewed as vital intermediaries between the adults of successive generations, thereby focusing attention back on the mechanisms of development. Critically, closer inspection of behaviors that were once considered hardwired and innate is yielding a deeper appreciation for the complex cascade by which genetic and non-genetic factors jointly, interchangeably, and reliably reproduce behavior across generations. But still, even as processes of developmental change move to the forefront of biology, many developmental scientists continue to embrace ideas held over from an earlier time. Using examples drawn from a variety of domains, including sensory neuroscience and motor behavior, I will show in this talk how a full consideration of the processes of typical and atypical development, including so-called freaks of nature, will engender a unified and interdisciplinary understanding of change across developmental and evolutionary time.
The video of the talk is available here.
2012-2013 Series
April 3, 2013 (LC 108, 4:00 pm)
Incidence, Essence, and Developmental Systems: Biologos in Action
Susan Oyama (John Jay College, CUNY)
Biologos, the “language of language” (information, transcription, translation, etc.) that dominates contemporary gene-talk, is problematic in several ways. Among other things, it supports covert forms of preformationism and essentialism, even though both are supposed to have been banished from biology. By implying that the DNA contains representations of the organism, Biologos works against attempts to deal effectively with nature-nurture contrasts, as well as ongoing attempts to synthesize development with evolution. I argue that the concept of the developmental system offers a fruitful alternative to this complex of deeply traditional habits of thought.
The video of the talk is available here.
March 14, 2013 (LC 108, 4:30 pm)
Perceiving objects by feeling them: Understanding dynamic touch
Alen Hajnal (University of Southern Mississippi)
Haptic perception is the ability to perceive properties of objects through touch and movement. Stephen and Hajnal (2011) have shown that haptic skills gained in one effector, such as the arm, transfer to other effectors, such as the leg. I ask whether haptic perception of length transfers from rigid rods to compliant ropes. Fractal analyses of three-dimensional hand motion recordings revealed an intricate picture of exploration styles that contribute to the prediction of trial-by-trial changes in perceived length.
February 28, 2013 (LC 104, 4:00 pm)
How did I pick up that accent? Phonetic convergence in spoken communication
Jennifer Pardo (Montclair State University)
Acoustic and phonetic variability within and between talkers does not arise by chance. Instead, such variation is attributable to a variety of causes including talker physiology, dialect, affect, and social/situational factors. Among the important social functions that speech serves are marking a talker’s identity, signaling their orientation to an interactive situation, and indexing the situational roles of interacting talkers. Studies of conversational interaction demonstrate the influence of such functions on many aspects of spoken communication. In particular, talkers sometimes become more similar over the course of a conversational interaction. The degree of inter-talker phonetic convergence has been found to be influenced by social factors such as talker role and sex. This talk provides an overview of the current state of research on phonetic convergence.
Spring 2012 Series
April 27, 2012 (CSB Auditorium, 4:00 pm)
Tales of Instinctive Behavior and Experiential Co-dependency
David B. Miller (University of Connecticut)
In recent decades, the emerging field known as “evo-devo” (as in “evolution-development) has generated exciting research on the importance developmental influences on evolution, and vice versa. The interrelationship between evolution and development necessitates a better understanding of underlying developmental mechanisms. I shall discuss views of development from a historical perspective and how earlier views have “evolved” into what some refer to as the transactional view of development, a way of thinking about development that captures the inherent complexity of emergent developmental systems. Part and parcel of this view is the role of nonobvious (or nonlinear) experiential influences on the development of instinctive behavior. I’ll present examples of how such subtle influences greatly influence behavior and then illustrate this further with my own research on alarm call responsivity of mallard ducklings, which not only portrays nonlinear experiential influences but also multiple pathways toward a common developmental outcome.
The video of the talk is available here (please note that the media stream uses the original Windows Media format and will not work in Chrome browsers, only Internet Explorer and FireFox).
March 5, 2012 (LC 108, 3:00 pm)
The Fallacies of the Five Senses
Michael Gordon (William Paterson University)
More than 2000 years ago, Aristotle proposed the, now, traditional divisions between the five senses. While these divisions have been perpetuated in modern culture, numerous behavioral and physiological studies would seem to challenge the validity of such a model. Several alternative theories have since been suggested, including divisions based on differences in the types of media detected, the sensory organs, neurophysiological pathways and structures, and phenomenology. Evidence from developmental, neurophysiological, and multimodal behavioral studies is evaluated against these theories to clarify their sufficiency. In conclusion it is proposed that there are ecological constraints on perception and action that have guided what we perceive as the natural sensory divisions, suggesting that sensory divisions are derived from the environments in which we evolved.