|
|
|
|
Abstract 1/10:
Organizational Meeting
1/17: Gregor Schoener
Dynamic Field Theory as a framework for understanding embodied cognition.
Understanding embodied and situated cognition means understanding how cognitive processes are closely linked to sensory and motor processes and depend on the behavioral and environmental history and context in which they unfold. Such understanding must be based on principles of neural function. Although neurons are discrete units, their discreteness is unrelated to discreteness in behavior, such as when people respond categorically to stimulus or task continua. Similarly, the discrete time structure of neural spiking events is unrelated to discrete behavioral events, such as the initiation of a motor act. The neuronal level of description appropriate for understanding behavior is thus spatio-temporally continuous. Dynamical field theory is a neurally inspired theoretical framework which accounts for how decision events emerge from continuos time processes, how cognitive functions emerge from neuronal interaction, and how experience structures behavior. The talk will illustrate these ideas by references to models of movement planning, working memory, and discrimination as well as by showing how such models enable robots to acquire simple perceptual representations.1/24: Randy Beer
TBA
TBA1/31: Christoph Weidemann, David Huber & Richard Shiffrin
How predictive information affects object identification
We use short time visual priming (prime word followed by a brief and
masked target word, followed by two choices) to investigate visual
object identification. In previous research we showed that priming
could be changed from positive to negative by changing prime
durations (attention) from short to long. We presented a ROUSE model
to explain such results: Features from primes join the target
percept, and then cannot be distinguished from features produced by
the target. The system that evaluates evidence to form object
identification deals with this source confusion by discounting
evidence from features known to have been in primes. Short primes
produce too little discounting (causing positive priming) and long
primes produce too much discounting (causing negative priming). An
alternative approach suggests the reversal of priming is simply a
strategic shift induced by the change in presentation of the primes.
As one attack on this possibility we decided to use situations in
which the primes are not neutral, and indeed predict the correct
answer (a situation that in fact occurs in many extant priming studies).
We found that the direction and strength of the predictive
information strongly affected performance, albeit in unexpected ways.
In general the effects of diagnosticity were superimposed on the
previous patterns we had observed. We showed using the ROUSE model
that the findings were consistent with the assumption that diagnostic
information affected the rate of discounting: lowering the
discounting rate when the primes predict the correct answer (so the
prime features tend to govern the choice), and the reverse when the
primes predict the wrong answer.2/7: Tim Pleskac & Jerry Busemeyer
A Dynamic and Stochastic Theory of Choice, Response Time, and Confidence
The three most basic performance measures used in cognitive research are choice, response time, and confidence. We present a diffusion model that accounts for all three using a common underlying process. The model uses a standard drift diffusion process to account for choice and decision time. To make a confidence judgment, we assume that evidence continues to accumulate after the choice. Judges then interrupt the process to categorize the accumulated evidence into a confidence rating. The fully specified model is shown to account qualitatively for the most important interrelationships between all three response variables found in past research.2/28:
Neural mechanisms of haptic object recognition
In humans and many other primates, vision plays the major role in
object recognition. But objects can also be recognized by touch. To
do this, visual and tactile (or haptic) systems must represent
volumetric space. It has been suggested that because of this overlap
in processing requirements, vision and haptics may also share some
common neural substrates. Studying haptic object recognition not only
informs us about the neural mechanisms of haptic object recognition,
but also helps to constrain theories describing the nature of object
processing in general. We have been exploring the neural substrates
of visual and haptic object recognition for both generic objects and
more specialized object categories, such as faces. Data will be
presented from behavioral, neuropsychological and neuroimaging
studies. First, these studies converge to suggest that mechanisms for
volumetric shape recognition that are shared across vision and touch
are instantiated in the lateral occipital complex (LOC). Second,
these studies also call for a reassessment of the neural mechanisms
involved in visual face recognition and instantiated in the fusiform
gyrus. Taken together, these findings suggest that models of object
recognition that are based solely on visual empirical data may be
omitting key aspects of human object recognition.3/7: Chris Honey, Indiana University
The anatomical connections between regions of the cerebral cortex
form a structural network upon which neural activity unfolds.
Cortical regions dynamically couple, forming "functional networks"
that are associated with perception, cognition and action as well as
with the so-called "default "or “resting" state. Functional networks
extracted from higher frequency dynamics undergo rapid
reconfiguration in, e.g., perceptual binding or sensorimotor
coordination. Functional networks extracted from lower frequency
spontaneous cortical dynamics are organized into anti-correlated
clusters, and it appears that the transient activation of these
clusters is related to processes of attention and spatial
reorientation. In this talk, I will discuss our ongoing attempts to
relate patterns of functional association observed in neuroimaging
experiments to the underlying anatomical connectivity of the cerebral
cortex.3/21: Dean Bennett Bertenthal
TBA
TBA3/28: Adam Sandborn
TBA
TBA4/4: Joe Anderson
TBA
TBA4/11: Noah Silbert
TBA
TBA4/18: Todd Gureckis
TBA
TBA4/25: Ji Son
TBA
TBA |
|
|
