University Relations

Spring 2012 Colloquia


photo of michael manserJan 26
Michael Manser, Mechanical Engineering, HUMANfirst program

Assessing driver behavioral adaptation to a rural intersection driver support system

Driver support systems have the potential to improve driving safety. However, most research only evaluates initial performance with the system and does not evaluate continued adaptation to the system to determine if the benefit continues or is negated by unintended use of the system. The efficacy of a previously evaluated rural intersection driver support system was examined in a simulated driving environment relative to system introduction, continued use, and potential positive transfer/carry over effects. Participants drove through a simulated rural intersection twelve times each day for a week with an intersection decision support system turned off during days one and five and turned on days two, three, and four. This experimental design allowed for an examination of the efficacy of the driver support system upon initial introduction, after continued use, and whether there were any carry-over effects. Results indicated drivers benefited from the rural intersection driver support system and that the benefit continued as exposure to the system continued. In addition, drivers continued to benefit from system use even after the system was no longer available. Results are discussed in terms of driver performance while using the system


photo of Maria GiniFeb 2
Maria Gini, Computer Science

How does a shared mental model affect performance in robot teams?

A critical part of teamwork in human teams is the existence of a shared team mental model. Shared team mental models are specially important to make distributed teamwork effective and appear to be even more critical for mixed teams of humans, agents, and robots. We are interested in measuring the extent to which teams have a shared mental model and to measure how sharedness relates to team performance. We present preliminary results in measuring performance for a team of simulated robots that operate in a simple simulation scenario, called the Blocks World for Teams (BW4T), and analyze how team performance is affected by sharedness of mental models. We outline how to extend the work to mixed teams of humans, agents, and robots.

Suggested reading:


photo of Nate PowellFeb 9
Nate Powell

Learn only what you have to: Motion model extrapolation in noisy environments

Accurately predicting what happens next is important in everyday life. Trajectory extrapolation is ecologically ubiquitous from predators intercepting their prey in the wild, to air traffic controllers plotting courses, to drivers navigating busy highways, to athletes intercepting projectiles, to stock brokers planning trades — all prediction tasks require the individual to anticipate a future state on the basis of incomplete data. Successful prediction thus entails formulating the prediction problem as a statistical estimation problem, which combines the current sensory information with an internal model of the world in deriving the best estimate of the world's future state. How we acquire internal models remains largely unexplored. My talk addresses under what conditions we learn a more accurate internal model for motion extrapolation.


photo of Vasileios ChristopoulosFeb 16
Vasileios Christopoulos
, The Andersen Lab, California Institute of Technology

New insights into the unilateral neglect syndrome: An fMRI-monkey study

Unilateral neglect syndrome is a neuropsychological disorder occurring frequently after extensive parietal lesions. It is characterized by attentional deficits and impairments of spatial awareness in the space contralateral to the damaged hemisphere that cannot be explained by sensory or motor disorders. A less severe form of the unilateral neglect syndrome is the spatial extinction, in which patients are unable to perceive stimuli in the contralesional hemifield, but only when a simultaneous ipsilesional stimulus is also presented.

We used event-related fMRI in monkeys to examine behavioral and neural consequences of reversible pharmacological inactivation of lateral intraparietal (LIP) area, in an oculomotor decision making task. We inactivated the LIP area by locally injecting the GABAA agonist muscimol, while monkeys performed the task in a vertical bore 4.7T scanner. The task involved memory saccades either to one stimulus (instructed trials) or a selection between two stimuli presented simultaneously in both hemifields (choice trials). The behavioral results showed that LIP inactivation causes a decision bias towards ipsilesional targets in the choice task, but does not affect the ability to perform saccades to single targets in either hemifields.

To assess the effects of LIP inactivation to the functional connectivity of brain regions involved in the task, we performed a voxel-by-voxel cross-correlation analysis between stationary BOLD fMRI time series extracted from 18 (9/hemisphere) frontal parieto-temporal regions of interest (ROIs) in 4 conditions: instructed left and right, choice left and right. We found that the strength of positive and negative intra- and inter-hemispheric interactions were increased significantly after LIP inactivation, when monkeys selected the contralesional targets in the "choice" trials. In contrary, we did not find any significant changes on the interactions before and after inactivation in the "instructed" trials. The changes of the cortical interactions after LIP inactivation suggest a compensatory functional reorganization of the frontal parieto-temporal cortical network to alleviate the behavioral deficits.

Suggested reading:

  • Kinsbourne, M. (1977). Hemi-neglect and hemisphere rivalry. Adv Neurol 18, 41-49.
  • Vuilleumier, P. O., and Rafal, R. D. (2000). A systematic study of visual extinction. Between and within-field deficits of attention in hemispatial neglect. Brain 123 ( Pt 6), 1263-1279.
  • Koch, G., Oliveri, M., Cheeran, B., Ruge, D., Lo Gerfo, E., Salerno, S., Torriero, S., Marconi, B., Mori, F., Driver, J., et al. (2008). Hyperexcitability of parietal-motor functional connections in the intact left-hemisphere of patients with neglect. Brain 131, 3147-3155.


Feb 23
Emilie Snell-Rood, Ecology, Evolution and Behavior

Costs, benefits and constraints in the evolution of learning: lessons from butterflies (and a few mammals)

Learning and cognition allow organisms to cope with novel and varying environments. And yet, despite these benefits, learning varies widely across populations and species. This talk details several studies on how understanding costs, constraints and benefits of learning may allow us to understand variation in cognition. First, it has been hypothesized that variation in learning is associated with costly neural investment which results in changes in life history such as delayed reproduction. In support of this idea, genetic variation in learning in butterflies is correlated with variation in brain size and reproductive timing and manipulation of reproductive timing results in changes in learning ability. Second, it has also been suggested that nutrition may act as a constraint on the evolution of costly neural tissue. Studies within and across butterflies provide support for this hypothesis but suggest that neural development can also evolve to compensate for poor nutritional environments. Finally, cognition is thought to be beneficial in novel environments, and yet it is unclear to what extent cognitive abilities are evolving in human-altered environments. Preliminary data analyses suggest that relative cranial capacity has changed over the last 100 years in several species of mammals across Minnesota. These results have implications for understanding the evolution of learning in a range of systems, including humans."

Suggested reading:


photo of Dan KerstenMarch 8
Dan Kersten, Psychology

"From Lightness to Shininess"

In our daily experience, we perceive material properties at a glance--such as whether an object has light or dark pigmentation, or whether it is shiny or matte. The computational problems are well-known: a given surface could be a black pigment in bright light, or white pigment in dim light, and a complex visual pattern on a smooth object could be due to the reflections of the environment as with a shiny chrome fixture, or it could be painted on. How do we tell the difference? The first problem--lightness perception--has a long history in perception research, yet puzzles remain, such as the nature of the neural mechanisms for representing and using contextual information to discount illumination. The second--"shininess"--has a shorter history, and seems to pose even stiffer challenges to our understanding of how vision arrives at determinations of material properties. I will describe results from two approaches to these two problems. For the first problem, I will describe neuroimaging results showing that human cortical BOLD activity in retinotopic areas, including V1, is correlated with context-dependent lightness variations, even when local luminance remains constant. Further, responses to these lightness variations, measured with a dynamic version of the Craik-O'Brien illusion, are resistant to a distracting attentional task. For the second problem, I will describe an analysis of natural constraints that determine human perception of shininess given surface curvature, and object motion. One set of demonstrations shows that apparent shininess is a function of how patterns of natural illumination interact with surface curvature. A second set of demonstrations illustrates how the visual system is sensitive to the way that specularities slide across a surface.



photo of Sam MClureMarch 22
Sam McClure, Psychology, Stanford University

A decision neuroscience approach to auctions

Cognitive neuroscience offers the potential to constrain models of the processes that underlie decision-making. I will discuss some of our recent work using this approach to investigate a phenomenon in auctions known as the winner's curse. Auctions are competitive social environments in which people often bid more than theory prescribes — commonly resulting in net financial loss. I will propose that bidding is well characterized as reinforcement learning with biased reward representations based on social preferences. We developed a new learning algorithm to model trial-by-trial changes in behavior. Indicative of reinforcement learning, prediction errors estimated with this model correlated with activity in the striatum and ventromedial prefrontal cortex. Additionally, we found that winning and losing auctions evoked activity in areas related to social and emotional processes, including the temporal-parietal junction and anterior insula. Connectivity analyses suggest that monetary and social value signals are integrated in the ventromedial prefrontal cortex and striatum. Based on these results, we argue for a novel, brain-based, mechanistic account for the integration of reinforcement history and social preferences in economic decision-making.

Suggested reading:

Sam McClure's talk is sponsored by the University of Minnesota Department of Psychology and co-sponsored by the Center for Cognitive Sciences.


photo of Claire HillMarch 29
Claire Hill, Law School

Psychology and the Financial Crisis

The financial crisis is importantly a story of human behavior. Some critical actors can be characterized as behaving "badly" in well-understood ways, but most cannot. I will discuss how we might understand the behavior of three important actors in the crisis: investment bankers, rating agency employees, and 'consumers' of ratings. What narratives did they tell themselves as to what they were doing and why? Especially as to the latter two: How did the data they were receiving affect the 'theories' under which they were operating?

Suggested reading:


photo of Haris Tzagarakis April 5
Haris Tzagarakis, Brain Sciences Center

This one is for Greedo: A look at human impulsivity through motor planning and uncertainty.

Psychiatrists and lay people alike often describe impulsive people as those who `act before they think'. In other words impulsivity conjures up the image of premature, poorly considered action, dissociated from sound decision making. In disease, impulsivity is a trait that underlies a great deal of debilitating psychopathology whilst, when considering individual variability, it is part and parcel of the repertoire of human traits that makes for differing "styles" in responding to environmental challenges. In this presentation I will briefly review the importance of the concept in health and disease and techniques that have been used to explore and describe it before discussing my attempt to approach it through uncertainty in motor planning.

Suggested reading:


photo of Phil ZelazoApril 12
Phil Zelazo, Institute of Child Development

Executive function and the developing brain: A developmental systems approach



photo of Mark AranoffApril 19
Mark Aranoff, Dept of Linguistics, SUNY Stony Brook

Al-Sayyid: A Language Blooms in the Desert

Al-Sayyid Bedouin Sign Language (ABSL) emerged about 80 years ago in a village in the Negev Desert in present-day Israel when four siblings were born into a family, deaf from birth on account of a recessive gene (Connexin 26). There are now almost 150 deaf people in the village and many members of the community, both deaf and hearing, use the sign language. For the last decade, a team of linguists has been investigating the emergence of ABSL across the generations. This talk will present a survey of our research to date, showing how different components of the language have emerged in different ways and at different rates. Some components appear to be quite highly determined. Some components have become organized more quickly. The organization of some components is highly dependent on modality, others less so. The differences among these components may shed light on the nature of the linguistic system as a whole.


May 3
photo of Mark LewisMark Lewis, Educational Psychology, College of Education & Human Development

Enhancing Comprehension and Potentiating Learning from Expository Text

Expository texts are central to learning at all stages of life. But unfortunately, many children and adults struggle to comprehend the types of expository prose that fills textbooks, scholarly journals, and popular press articles. Two key reasons that readers struggle to comprehend expository texts are because they lack requisite prior knowledge and because the architecture of the human mind severely limits how effectively readers can integrate the knowledge they do have with textual information. This talk will review our recent efforts to employ fundamental principles of human learning in order to improve text comprehension and to build a transferrable foundation that also supports future learning.