S1 (FriAM 1-65) - The Psychonomic Society

Papers 277–283 Sunday Morning
assessed via slowed RTs following a representation switch. Additionally
the use of SDR requires control of executive attention to keep inactive
representations from interfering with the current response. Participants
were given a category learning task composed of one- and
two-dimensional substructures. Control of executive attention was
measured using a working memory capacity (WMC) task. Participants
who used SDR showed greater slowing of responses following a substructure
switch, and a greater correlation between learning performance
and WMC. These results provide support for the principle of
SDR in category learning.
9:00–9:15 (277)
Attention Dynamics in Category Learning. AARON B. HOFFMAN
& BOB REHDER, New York University (read by Bob Rehder)—Theories
of human category learning assume that attention plays a central
role in modulating the influence of cues on learners’ behavior.
However, each theory incorporates attention in a different way. Three
experiments used eye tracking to record the attention dynamics of
human learners as they acquired novel categories. Changes in the pattern
of eye fixations were compared to predictions of two prominent
connectionist learning models: Kruschke’s (1992) ALCOVE (attention
learning covering map) and Kruschke and Johansen’s (1999)
RASHNL (rapid attention shifts n’ learning). Across the three experiments
we found that people exhibited early and rapid attention shifts.
The early shifts were coupled with subsequent suboptimal attention
profiles, in which people continued to attend to irrelevant information,
in Experiments 1 and 2, or to more information than the minimum
required, in Experiment 3. These findings are interpreted as
being more consistent with RASHNL’s predictions regarding attention
than with ALCOVE’s.
9:20–9:35 (278)
Extremely Selective Attention: Eyetracking Studies on Dynamic Attentional
Allocation to Stimulus Features. MARK BLAIR, MARCUS
WATSON, FIL MAJ, & CALEN WALSHE, Simon Fraser University—
Selective attention is an important part of most theories of category
learning. This allows them to model empirical results that show that
people learn to attend to features that are helpful for distinguishing
the categories being acquired, and learn to ignore features that are not.
Attentional allocation is thus assumed to be dynamic, because it shifts
with experience to optimize categorization performance. Typically,
however, attention is also assumed to be consistently allocated to all
categories and all stimuli. We report results that violate this assumption.
Subjects learned several categories with differing numbers of informative
features, so the optimal attentional distribution varied
across stimuli. Using an eyetracker to measure attention, our results
show that subjects readily learn to vary their attentional allocation in
response to specific stimulus features. We examine some details of
this dynamic process.
9:40–9:55 (279)
Cognitive Dynamics From Eye Movements: Representational
Change As a Phase Transition. DAMIAN G. STEPHEN, REBECCA
SULLIVAN, JAMES A. DIXON, & JAMES S. MAGNUSON, University
of Connecticut (read by James A. Dixon)—Capturing the dynamics
of cognition has been a long-standing, fundamental challenge for cognitive
science. The present research demonstrates that the phase space
of cognition can be reconstructed through a time series of action—
specifically, the changing position of eye-gaze. Participants were
asked to solve simple gear-system problems by predicting the motion
of a target gear, given the turning direction of a driving gear. Participants
initially solved the problems by simulating the motion of the
gears, but then spontaneously discovered a mathematical solution to
the problem. Using the eyetracking time series to reconstruct phase
space, we show that the emergence of the mathematical solution constitutes
a phase transition. Specifically, real-time discovery of the
mathematical relation was predicted by a decrease in system entropy,
a hallmark of phase transitions in dynamical systems. Thus, the pres-
43
ent study demonstrates the emergence of higher order cognitive phenomena
through nonlinear dynamics.
Pronouns and Reference
Beacon B, Sunday Morning, 8:00–9:20
Chaired by Amit Almor, University of South Carolina
8:00–8:15 (280)
An fMRI Investigation of Repeated Reference in Discourse.
AMIT ALMOR, VEENA A. NAIR, JEREMY L. MAY, LEONARDO
BONILHA, JULIUS FRIDRIKSSON, & CHRIS RORDEN, University
of South Carolina—Repeated reference is an important part of coherent
discourse but very little is known about the brain basis of processing
repeated reference in discourse. A recent functional magnetic
resonance imaging (fMRI) study (Almor, Smith, Bonilha, Fridriksson,
& Rorden, in press) found that reading a repeated proper name
reference (e.g., Joe) to a salient discourse referent leads to increased
activation in temporal regions and intraparietal sulcus (IPS) in comparison
to reading a pronoun reference (e.g., he). This was interpreted
as indicating the involvement of spatial brain circuits (IPS) in the
management and manipulation of representations of discourse referents
(temporal regions). The conclusions of this study were, however,
limited by the fact that it did not manipulate the salience of referents.
We report the results from an fMRI experiment which aimed to address
this limitation by comparing the brain activation associated with
reading pronoun and repeated name references to salient and nonsalient
referents.
8:20–8:35 (281)
The Role of Attention During Pronoun Comprehension. JENNIFER
E. ARNOLD & SHIN-YI C. LAO, University of North Carolina, Chapel
Hill—Discourse features (e.g., first mention) are claimed to put referents
in discourse focus, rendering them accessible to listeners during
processes like pronoun comprehension. Two experiments investigated
the relationship between such “discourse focus” and visual
focus of attention. Listeners viewed a picture of two gender-matched
characters, and decided if it matched the story they heard—for example,
“Doggy is flying a kite with Birdy . . . He . . .” The pronoun
referred to either the first or second character. A visual capture cue
(highlighting around one character) appeared briefly during the first
clause, either right before the mention of the second character (Experiment
1) or at trial onset (Experiment 2). In a combined analysis
of the two experiments, participants were more likely to say that the
story did not match the picture when the pronoun referred to the
second-mentioned character. Additionally, the visual capture cue
modulated the first-mention bias, suggesting a relationship between
discourse and visual attention.
8:40–8:55 (282)
Processing Differences for Pronouns and Reflexive Pronoun
Anaphors. SHELIA M. KENNISON, SONDRA L. NOLF, & JAMES E.
VAUGHN III, Oklahoma State University—A series of experiments
investigated how pronouns (he and she) and how reflexive pronoun
anaphors (himself and herself ) were resolved with gender ambiguous
antecedents (e.g., doctor, nurse, patient). Processing differences were
observed for the two type of referents, with the largest difference
being observed when referents were resolved with gender neutral antecedents.
In two of the four experiments, individual differences in
working memory were assessed. Participants’ processing was related
more strongly to working memory during the processing of pronouns
than during the processing of anaphors. Implications for models of
referential processing will be discussed.
9:00–9:15 (283)
Fillers As Metacognitive Collateral Signals: Evidence From Language
Comprehension. DALE J. BARR, University of California,
Riverside, & MANDANA SEYFEDDINIPUR, Stanford University—