University Relations

FALL 2016 Colloquia

Thursdays, 4:00 - 5:30 pm, Elliott N119


Virginia HeinenSeptember 29

Virginia Heinen, Ecology, Evolution and Behavior

"The Emergence of Conventional Signals in European Starlings"

Human communication is conventional: the relationship between a word and the meaning it conveys is largely arbitrary, and signaller and receiver must both agree on this relationship for communication to tae place. Humans can also quickly develop novel communication systems. Conventional signals are present in some cases of animal communication as well, but they are not well understood. i will present results from simulation models and experiments that investigate how conventional signalling systems develop, and possible barriers to their success. In an experimental setting, pairs of European starlings can quickly develop signalling conventions, and that these signals are non less reliable, or no more difficult to learn, than non-conventional signals. These conventional signals appear to be influenced by the signaller's initial signalling preferences, in contrast to theories of signal evolution that ascribe more control to receivers.

Suggested readings:


Jason M. FordOctober 6

Jason Ford, Philosophy, UofM Duluth

"Time and the Gorilla: How Changes in Time on Task Affect Inattentional Blindness"

I examine how changing various features of Simons and Chabris's Gorilla experiment impact the rates at which subjects report seeing the person in the gorilla suit. The results I would like to present will provide evidence that inattentional blindness is not related to memory, that subjects are paying more attention to the task at the outset than they are as it progresses, and that if subjects are peripherally aware of the person in the gorilla suit, they are not seeing her as a gorilla.

Suggested readings:


Al YonasOctober 20

Al Yonas, Child Development

"What does the perception of 3-D shape from shading tell us about the development of perception?"


Matthew ChafeeOctober 27

Matthew Chafee, Neuroscience

"Neural and synaptic dynamics in prefrontal networks for cognition and their disruption in animal models of schizophrenia"

Prefrontal networks compute by transmitting information between neurons in circuits. Disruption of this communication downstream of synaptic malfunction is thought to lead to cognitive deficits in human neuropsychiatric diseases such as schizophrenia. I will discuss data from our lab that describes how temporal dynamics in prefrontal circuits can be probed to recover the normal pattern of communication between neurons in prefronta circuits during cognition, and how this pattern of communication may fail as a result of synaptic malfunction in animal models of schizophrenia.

Suggested readings:


Apostolos GeorgopoulosNovember 3

Apostolos Georgopoulos, UMN Neuroscience and Brain Sciences Center

"A (human) followup on Matt Chafee's talk on schizophrenia"

In this colloquium I will discuss neural interactions in schizophrenia based on our recent results from resting-state magnetoencephalographic (MEG) studies.

Suggested readings:


Geoff GhoseNovember 10

Geoff Ghose, UMN Neuroscience

"Probability representations in visual cortex"

Numerous psychophysical studies have suggested that Bayesian like operations, in which evidence and inputs are weighted according to likelihood, underlie perceptually guided judgments. However, how such likelihoods are represented physiologically within the brain, the accuracy and physiological constraints of these representations, and the means by which they alter or influence sensory representations has been unclear. We have approached this problem by training non-human primates to perform challenging visual detection tasks in which the spatiotemporal likelihood of a behaviorally relevant event happening within a trial is fixed across many training sessions, but no explicit cues of likelihood are ever presented. We have found, in several different tasks, that animals automatically form implicit representations of likelihood schedules, such that performance is higher for more likely events than less likely events. These representations are accurate on a scale of 100s of milliseconds and with a spatial resolution on the order of a degree. Physiological recordings during task performance suggest that these precise behavioral effects arise from a targeting of gain-like modulations to those neurons which most reliably reflect likely events. These findings suggest optimal behavior cannot be simply explained by classic decoding models, such as "winner-take-all" and "population vector" readouts, in which the representations to be decoded are immutable across tasks. Rather, our results suggest that probability representations exist at the earliest levels of sensory representation within the cortex, and that these probability based modulations are sufficient to explain Bayesian-like perception.


November 17

Robert KruegerRobert Krueger, Psychology

"Toward Empirical Classification of Psychopathology"

Psychopathology has historically been classified based on authority. However, it is also possible to develop classification systems based on data. I will review recent efforts by ourselves and others to develop an empirical, data-based approach to psychopathology classification.

Suggested readings:


Sarah KaufmanDecember 1

Special Lecture: "The Art of Grace: A Cognitive, Cultural & Neuroscientific Perspective"

Sarah Kaufman, Senior Arts Writer & Dance Critic, The Washington Post; 2010 Pulitzer Prize for Criticism

Mayo Memorial Auditorium, University of Minnesota Medical School


Fall Institute


Fall Institute 2016

"Computing and Cognition"

Thursday, December 8, 1:00 - 5:00pm
401/402 Walter Library - University of Minnesota - Twin Cities Campus


Updated April 3, 2017->