Joshua is a PhD candidate in the Natural Computation Lab (PI Virginia de Sa) at the University of California, San Diego Cognitive Science department. His research investigates the role of human perception and insight in the data analysis process, and attempts to improve how researchers interact with and understand machine learning algorithms. His research approach combines psychophysical studies of both novice and expert data analysts with new algorithms and software tools that give human researchers an intuitive view into the behavior of cutting-edge machine learning techniques. Such intuitions will lead to better research and better utilization of machine learning techniques in unrelated fields.
Joshua's other academic interests include infant visual development, human computation, the cognitive benefits and demands of video gaming, spatial cognition, and multi-modal stimulus processing. Over the course of his graduate career he has mentored five undergraduate research assistants across three projects and received two awards for teaching. He hopes to inspire a love of science and reason in his students while he is in graduate school and in the future as a professor.
How does the brain represent visual input? We know a great deal about the coding of simple features in isolation, such as the orientation of lines, but we know very little about how these simple features are combined into more complex, useful representations. For instance, how would the brain represent the color of a large uniform surface? Many researchers have assumed that the population activity of simple feature detectors is itself the higher-level representation, but this ignores the fact that perception is an active process: some kind of computation must eventually be performed to detect that a uniform surface is present. In my work I have argued for a multiscale representation, where different neurons code for the appearance of different sized regions of space, thus allowing many fewer neurons to be active, and furthermore explicitly coding for uniformity. This kind of representation would be useful for any surface property (orientation, depth, etc), not just color.
This conclusion has empirical support: my own psychophysics experiments are consistent with it, and I have conducted computational modeling research showing that multiscale representations do a good job of explaining many visual illusions that people experience. The implications of this work have little real-world utility at present.
It does have large implications for understanding how the visual system works, however, as it significantly constrains the kinds of inputs later, more high-level visual areas receive. Given the utility of understanding the visual system (particularly as a guide to researchers developing artificial visual systems for robots, cars, etc), the potential payoff of this kind of basic science to our everyday life is huge.
I am continuing this work and other inquiries into basic visual representations as a postdoc, and hope to start applying for faculty positions in the next 1-2 years. Failing a long-term academic appointment, I am interested in pursuing brain-inspired approaches to real-world artificial vision problems in industry.
The Glushko Fellowship came to me in my final year of graduate education, and cobbled together with other funding sources, it allowed me to focus on finishing my dissertation without the distraction of TAing. Given the heavy workload of finishing the dissertation on time, I doubt I could have done a good job of that and TAing at the same time, and I am deeply grateful that I did not end up having to prioritize one over the other.
Rachel studies lexical storage and access in language production and comprehension, with an eye to understanding how the different components of a lexical representation are activated and used when forming words and sentences. Notwithstanding the extreme ease with which most humans can produce fluent speech, language is a cognitively complex and opaque process. Having a better understanding of how, precisely, speech is processed and produced in real life is a critical component of understanding cognition as a whole. Mental processes do not exist individually in the mind, and thus fully mapping how the mind and the brain work requires a thorough understanding of all of the cognitive systems. Rachel hopes to figure out these mental processes underlying language and eventually go into research full-time.
I was thrilled to be accepted to UC San Diego and to win a fellowship because this department is exactly where I want to be. San Diego is such a nexus of cognitive and neuroscience research, and it is amazing for me to have the opportunity to work and interact with so many brilliant and disparate researchers here. I was drawn to UCSD thanks to the incredible diversity of the departmental research programs, and the large number people who do research on exactly the sort of questions that interest me. In fact, I have three separate advisers and I sincerely intend to work with all of them. My ideal course of research combines all three areas of focus and expertise, and allows me to investigate areas of language processing that lie at the intersection of language production, imaging research, and lexical access.
Kensy’s research explores the conceptual processes involved in everyday face-to-face communication. He investigates these processes by looking at the spontaneous motor movements— or gestures— people produce in natural interactions. When considered along with their co-produced speech, gestures offer rich insights into conceptual structure as it unfolds in real-time. Kensy’s dissertation zooms in on co-speech pointing gestures and examines the particulars of how and when they are produced. Other current projects include collaborations on cross-cultural studies of time-related gestures and on how headshakes relate to discourse structure. In future research, Kensy hopes to use gesture analysis— complemented with the more traditional tools of linguistics and experimental psychology— to better understand how concepts differ across cultures and how they develop in children. This research has the potential to impact a number of areas, especially education, and might also have implications for development of computer-mediated interaction technologies.
My research focus is the on the visualization of multiple streams of time-based data. As more and more research involves collection of rich sources of multimodal data, such as video, computer logs, and sensor readings, there is an increased need for ways of understanding these complex data sets. My research explores the use of interactive information visualization for the purpose of analyzing this type of data. There are two main components of this research that I work on in parallel. First, I collect and analyze data about how researchers interact with visualized multimodal time-based data, focusing on temporal patterns of movement in the data and interaction with different visualizations. Second, as a way of enabling and applying the data collection and analysis, I develop tools and techniques for interactive visualization.
To date, I have developed a tool that enables researchers to visualize and annotate multiple streams of data. This tool is currently in use by several researchers both within UCSD and at other institutions, and collects data about how the researchers explore the data and generate annotations. In addition, video is taken of the researchers using the tool, so that I can collect richer information than can be collected by digital logs alone. By using the tool as a starting point, it has also enabled research of a collection of novel techniques for interacting with visualized temporal data, including the use of computational techniques to identify important segments of data and the use of digital pen technology to integrate information recorded on paper during initial observation with later analysis.
The Glushko fellowship has been instrumental in enabling and forwarding my research. It has allowed me to acquire the tools I need to develop visualization techniques and to support studying interaction with those techniques. It has also enabled me to travel to conferences to interact with other researchers, including the ACM CHI conference, where I have had the chance to discuss my research with others in the field of information visualization, as well as the Measuring Behavior conference, where I gave a talk on the tools that I have developed to support my research.
I was born in La Plata, Argentina and raised primarily throughout the Indian subcontinent, South America, and Europe. I obtained B.A.s in Computer Science and Cognitive Science (Highest Honors) from UC Berkeley and conducted research at Helen Wills Neuroscience Institute and Palo Alto Research Center (PARC). There I developed several projects examining information flow in distributed cortical networks during language processing as well as novel applications of wearable brain-computer interface technology, particularly in the domain of human-computer interaction. My graduate research at UCSD has continued along these two paths. One branch of research has focused on the development of software tools for estimating and visualizing directed information flow and synchronization between multiple spatially-localized neuronal sources estimated from scalp and intracranial EEG and using these tools to study distributed cortical network dynamics underlying cognitive processes such as response inhibition, visual attention, and error processing. I am an active contributor to the widely-used open-source EEGLAB software suite for single-trial EEG analysis and am the developer of EEGLAB’s Networked Source Information Flow Toolbox (NetSIFT) for EEG/ECoG information flow analysis and visualization, which will be officially released at Society for Neuroscience 2010. I have also been invited to lecture on information flow and causality at several EEGLAB workshops and symposiums held worldwide.
I firmly believe that a more complete understanding of functional brain connectivity will not only transform human neuroscience – allowing us to elucidate how our brains work in health, illness, youth, and old age – but will also lead to vast improvements in brain-computer interface (BCI) technology. Cognition is a distributed phenomenon, requiring the dynamic cooperation of multiple neuronal structures to achieve a desired outcome. By developing cutting-edge methods for noninvasive realtime analysis of information flow dynamics between neuronal structures and incorporating these into practical, wearable, BCI systems, I hope to push the boundaries of what we currently can achieve with BCI technology.
Outside academia, I combine my passion for science/technology and the Arts in the form of BioArt. Since 2007, I have collaborated with other musicians, composers, and scientists to develop new-media installations and performances that, through interactive audiovisual representations of electrical signals recorded from the brain/body, invite the participant to explore and reflect on the intricate coupling between mind and matter in social environments. Some recent works – several of which utilize cutting-edge BCI technology – include Mindchill, Mindchill 2.0, EEG Ocean, In Tones: Music for Online Performer, and Just: A Suite for Violin, Flute, Cello, and Brain with several more exciting installations and concerts planned in San Diego and San Francisco next year!
The Glushko Fellowship has been instrumental in providing me the financial freedom to collaborate with multiple research groups at UCSD. This has been key to expanding my research horizons and allowing me to develop software tools and algorithms which can be used effectively by the broader scientific community to answer some of the most pressing questions in cognitive neuroscience today. I am immensely appreciative of this opportunity and I truly hope that future students will continue to benefit, as I have, from these generous donations.
Although we tend to think of emotion and cognition as two separate and fundamentally different kinds of processes, in fact they are both essentially (a series of) brain states. While a great deal of research has already contributed to our understanding of the ways in which certain kinds of thinking can affect our moods, I am interested in the opposite question: how might our moods affect our thinking? In particular, my work investigates how positive and negative moods differentially affect the processing of language. We used electrophysiological measures (ERPs) and found that compared to a negative mood, a positive mood seems to have a ‘broadening’ effect on the range of items that are (pre)activated in semantic memory. This was observed both when people were processing words that were plausible (but somewhat unlikely) in a given context (e.g. “Something you wear on your feet… skis”), as well as when people were asked to produce a target word in response to three moderately related clue words (e.g. “swiss, cottage, cake… cheese”). Being in a positive mood thus facilitates both the processing and the retrieval of words that are distantly related to the context or to one another. Future work may shed light on how being in a positive or negative mood differentially affects our construction of memories and our ability to learn.
The receipt of the Glushko fellowship was extremely important to me, as I received it at a time in which I was undergoing some financial and personal hardship. Not only did the fellowship give me the financial means to continue my studies, the formal recognition of my contributions to the department reminded me of what I’d already accomplished and strengthened my resolve to continue to be an active and engaged member of the Cognitive Science community here at UCSD.
Ross’s research focuses on language comprehension, and in particular on how knowledge derived from real-world experience is evoked by, and serves to guide, real-time comprehension processes. He also has a complementary interest in the role that sensory and motor systems, traditionally believed to be irrelevant to language processing, play in comprehension. Recent research has shown that language comprehension engages brain areas central to perceiving and acting in the world, highlighting that engagement with the outside world may systematically impact concurrent language processing, and vice versa. Ross’s research on the relationship between language comprehension and real-world experience has implications for our understanding of language learning and dysfunction and of the interplay between linguistic and sensorimotor processes, as well as for development of language processing systems.
A personal message from Ross: “When I was applying to PhD programs, UCSD Cognitive Science was my top choice academically. When I was notified of receiving the Glushko Fellowship, my choice of graduate school became a no-brainer. After just over two years in the program, I can’t imagine being anywhere else, and the Glushko Fellowship is one of the reasons I chose to come in the first place. Many thanks to Dr. Glushko for his generosity.”
I study mathematics from the perspective of embodied cognition. Why is mathematics so stable, successful, certain, widespread – but difficult to teach? To answer these questions, I use methods from cognitive linguistics, gesture studies, cognitive ethnography, and the history and philosophy science to document the ways in which mathematical practice involves the body, material artifacts, and culture. By documenting expert practice, I hope to inform the teaching and learning of mathematics at all levels, from kindergarten to graduate school.
Before arriving in San Diego, I was a national team athlete in Canada. For years I competed on the Canadian National Wrestling Team, and I was an alternate for the 2008 Beijing Olympics. After retiring from sport – at the ripe old age of 24! – I knew I was ready to pursue my research interests in earnest. My questions required methods and results from diverse fields, from philosophy to neuroscience. Moreover, I knew I wanted the kind of high-performance environment that had served me so well in sport. This drew me to the Department of Cognitive Science at UCSD, where I hoped to find a challenging but supportive intellectual home. And what a home I found, full of top flight researchers who care about mentoring the next generation of scientists. I hope one day to represent UCSD Cognitive Science as a professor, as I continue to study mathematical reasoning while, in turn, nurturing a new crop of cognitive scientists.
If not for the Glushko Fellowship, I probably would not have come to San Diego – nor, in fact, would I have become a cognitive scientist. Until the last minute, I was entertaining enticing offers from departments of philosophy; UCSD Cognitive Science was an exciting, if expensive, option. It was only through the generous support of the Glushko Fellowship that I was able to afford to attend UCSD. I'm incredibly happy that I was able to make that choice. Cognitive science – as both a department at UCSD and a discipline more generally – is exactly the kind of exciting, challenging, supportive, and inspiring community I hoped to find. And for that I thank you, Prof. Glushko.
Sandra studies brain mechanisms supporting episodic memory in the parietal cortex and the hippocampus. The work is carried out through multiple single neuron recordings in rats executing navigational sequences. Specifically, she examines how spatially-specific neuronal firing patterns change when two highly familiar and oft-repeated episodes, having the form of traversals across a path, are linked to form a single full episode. The research has implications for understanding how episodic memories are formed in the hippocampus and parietal cortex. The work also addresses the extent to which overlapping portions of two otherwise different episodic memories are distinguished by their associated hippocampal and parietal cortical firing patterns. Finally, for the first time, the work tackles the question of how neural mappings of episodic memories adapt when two previously independent episodes are linked to form a single sequence of experiences. Ongoing experiments are testing the hypothesis that parietal and/or hippocampal firing patterns occurring at the juncture between two previously independent paths must adapt in one of three ways (i.e., appearance/disappearance of place fields, stretching of existing place fields, and no change) to enable fluid movement across the full length of the joined paths.
Sandra would like to thank Robert Glushko and the Cognitive Science Department for this very generous award.
Micah is a Ph.D. student in the Department of Cognitive Science where he studies the perception and memory of complex acoustic signals in birds and humans. Prior to attending UCSD, he attended Pomona College, where he earned a Bachelors degree in Mathematics and Music. After graduation, Micah received a Fulbright Fellowship to study music perception and cognition at the University of Jyväskylä in Finland. His current research projects investigate how birds and humans use pitch in the perception of conspecific communication signals. These studies comparatively investigate how birds use pitch to recognize individuals using song and how humans use pitch to recognize individual human talkers. Micah's research has also investigated how birds perceive and respond to music, particularly whether they are able to "hear the beat" by synchronizing movements to the beat of a musical stimulus.
These studies may be informative for those studying disorders that affect auditory processing, including autism. Autism has long been associated with a propensity toward unusually accurate absolute pitch perception. Studying the role pitch plays in the recognition of communication signals is an important step for those trying to understand how this tendency may lead to communicative disorders.
I am interested in the questions about how language emerged and how it is transmitted from individual to individual. Language is a complex system involving inputs and interactions at many levels—from cognitive biases arising from intrinsic properties of the nervous system, to pragmatic properties that only emerge from the interaction of agents in social networks. Our use of language recruits processes as diverse such as perception, motor systems, and memory and learning in comprehension, production, and acquisition. More often than not, this has inspired provincialism and fragmented ways of approaching the study of language. Addressing the question of how language—and communication systems in general—evolved through computational modeling and laboratory experiments offers a framework that could help bring together and test hypotheses from different sub-fields, and hopefully increase our understanding of language.
While I chose UCSD because of UCSD’s unique, multi-level approach to cognitive science, the Glushko fellowship made it an easier choice for me to turn down larger stipends at other schools. Prof. Glushko’s award has allowed me to focus on my research interests at my first-choice school without worrying about how to pay the bills. But beyond the benefits I have received individually, however, Prof. Glushko’s generous funding of cognitive science research has had an enormous impact on the field of cognitive science in general, by increasing incentives for students and senior researchers alike to tackle big questions in cognitive science without being hampered by funding considerations. I cannot express how sincerely and deeply I appreciate Prof. Glushko’s generosity. Thank you.
My research focuses on how the brain acquires language and processes words. I have been studying adults whose native language is Spanish, and who learned English as a second language when they started elementary school. Using a combination of magnetoencephalography (MEG) and structural magnetic resonance imaging (MRI), we have analyzed the millisecond-by-millisecond changes in brain activity in the two languages. In a series of studies, we reported that the less proficient language (regardless of whether it was learned first or second) recruits bilateral brain regions in addition to the classical left hemisphere fronto-temporal areas when processing words. This activity begins during the earliest stages of word encoding and continues through later meaning processing stages.
Recently, we have begun to look at the recruitment of non-classical language areas in other populations using our MEG/MRI methods. What happens when the brain is deprived of auditory input and language is acquired in the visual modality (as is the case in American Sign Language)? In hearing individuals who learn ASL as a second language, are these bilateral regions recruited when ASL is the less proficient language? Furthermore, how are signs encoded as meaningful stimuli in the brain?
These are questions that I am pursuing as part of my dissertation, however many of them will continue beyond graduate school. The Glushko Fellowship allowed me to start my graduate career with an open mind toward the questions that I could ask in my research, which gave me the freedom to explore various different literatures that have since been incorporated into my work. Funding agencies view broad perspectives as essential to successful research, and I have since been awarded fellowships and research grants from the Kavli Institute for Brain and Mind at UCSD, the UCSD Institute for Neural Computation, the Center for Research in Language, and the National Science Foundation.
My research interests in the broadest sense stem from a desire to understand the origins of our minds. My academic and research experience in Evolutionary Psychology as undergraduate introduced to me idea that biological variation, mediated by genetics, could contribute to understanding variation in and the origins of behavior and cognitive functioning. Bolstered by pre-graduate work experience in a lab studying the genetics responsible for biological variations responsible for hypertension, I entered my graduate study with the hope to apply similar techniques to the study of cognition.
My current research interests are trying to understand how and to what extent biological variation in the brain can translate the effects of genetic variation into cognitive and behavioral difference among individuals. UCSD offers the unique opportunity, through work in and the collaborations of Dr. Terry Jernigan’s lab, to examine this question directly and from a developmental perspective. Modeling the effects of genetic variation on cognition and behavior through the course of development has far reaching implications from understanding the genetic signatures of psychiatric and neurological disorders to success in school. There is no earlier symptom or better chance for early intervention than an understood genetic signature.
Given the budget cuts and economic struggles faced by the University of California, I feel both humbled and thankful to have received the generous Glushko fellowship award. From a practical perspective, life of a graduate student can at times be a struggle but the extra security provided by the Glushko award helps enormously to alleviate some of the burden. Also, and even more importantly, in times when education and science funding is cut, the commitment of our donor to support science and graduate study is inspiring. I hope to continue in academics as a career beyond graduate school and it will be the direct result of fellowships such as the Glushko award that make this possible. From one person to another, I sincerely thank you for your support of my education, our program and university as well as science in general.