Abstracts 2005 - The Psychonomic Society

Saturday Morning Papers 195–199
10:20–10:35 (195)
Rapid Lexical–Semantic Integration in Speech Processing: Evidence
From Pause Detection. SVEN L. MATTYS & CHRISTOPHER W.
PLEYDELL-PEARCE, University of Bristol—In this study, we use
pause detection (PD) as a new tool for studying the online integration
of lexical and semantic information during speech comprehension.
When listeners were asked to detect 200-msec pauses inserted into the
last word of a spoken sentence, their detection latencies were influenced
by the lexical–semantic information provided by the sentence.
Pauses took longer to detect when they were inserted within a word
that had multiple potential endings in the context of the sentence than
when inserted within words with a unique ending. An event-related
potential variant of the PD procedure revealed brain correlates of
pauses as early as 101–125 msec following pause onset and patterns
of lexical–semantic integration that mirrored those obtained with PD
within 160 msec. Thus, both the behavioral and the electrophysiological
responses to pauses suggest that lexical and semantic processes
are highly interactive and that their integration occurs rapidly during
speech comprehension.
10:40–10:55 (196)
Spatiotemporal Properties of Brain Activation Underlying Lexical
Influences on Speech Perception. DAVID W. GOW & CHRISTINA
CONGLETON, Massachusetts General Hospital, & SEPPO P. AHL-
FORS & ERIC HALGREN, MGH/MIT/HMS Athinoula A. Martinos
Center—Behavioral evidence from a variety of paradigms suggests
that there is a two-way relationship between fine-grained phonetic
factors and lexical activation. Within-category phonetic variation affects
the course of lexical activation, and lexical knowledge affects the
interpretation of phonetic ambiguity. We examined this relationship
in a study combining MEG and fMRI to provide high spatiotemporal
resolution imaging data while participants performed a phonetic categorization
task. In each trial, listeners heard a token of “_ampoo” or
“_andal” in which the initial segment was /s/, /ʃ/, or a fricative judged
to be intermediate between /s/ and /ʃ/. Listeners showed a robust behavioral
Ganong effect, interpreting the ambiguous fricative as /s/ in
“_ampoo” and /ʃ/ in “_andal.” Physiological data showed a distinctive
pattern of activation reflecting the interaction between phonetic and
lexical activation. The results are discussed in the context of the general
problem of interaction between top-down and bottom-up perceptual
processes.
11:00–11:15 (197)
Entering the Lexicon: Form and Function. LAURA LEACH & AR-
THUR G. SAMUEL, SUNY, Stony Brook (read by Arthur G. Samuel)—
Lexical entries contain semantic, syntactic, phonological, and orthographic
information. However, they are not static repositories; lexical
entries dynamically interact. We are studying the acquisition and development
of new lexical entries (e.g., for “figondalis”). Adult participants
either incidentally learned new “words” while doing a phoneme
monitoring task or were trained to associate each word with a picture
31
of an unusual object. We then measured both form learning (recognizing
words in noise) and functionality (the ability of the word to
support perceptual learning of its constituent phonemes). Across
5 days of training and testing, form learning increased steadily and
substantially in both training conditions. However, new words were
not well integrated into the lexicon when trained via phoneme monitoring;
their ability to support perceptual learning was small and did
not increase over training. In contrast, learning words as the names of
objects rapidly produced lexical entries with both form and function.
11:20–11:35 (198)
Auditory Language Perception Is Unimpaired by Concurrent Saccades
and Visuospatial Attention Shifts. WERNER SOMMER, Humboldt
University, Berlin, OLAF DIMIGEN, University of Potsdam and
Humboldt University, Berlin, ULRIKE SCHILD, Humboldt University,
Berlin, & ANNETTE HOHLFELD, Universidad Complutense de
Madrid—Language perception at the semantic level—as indicated by
the N400 component in the event-related brain potential—can be severely
delayed in time when other tasks are performed concurrently.
This interference could be explained by a central processing bottleneck
or by attention shifts elicited by the additional tasks. Here, we
assessed whether the additional requirement of performing saccades
of 10º to the left and right would aggravate the N400 delay. In a first
experiment, delays of N400 were more pronounced when the additional
task involved saccades than when it did not. However, when the
saccade-induced delay of information input in the additional task was
compensated by correcting SOA, the effects of additional tasks with
and without saccades on N400 latency were indistinguishable. It is
concluded that language perception is unimpaired by exogenously
triggered visuospatial attention shifts, lending further support to the
bottleneck account of concurrent task effects.
11:40–11:55 (199)
The Time Course of Shifts in Overt Attention Toward Visual Objects
During Language-Mediated Visual Search. JAMES M. MCQUEEN,
Max Planck Institute for Psycholinguistics, & FALK HUETTIG, Ghent
University—A visual world study is reported investigating the time
course of how phonological, visual shape, and semantic information
accessed from spoken words is used to direct gaze toward objects in
a visual environment. During the acoustic unfolding of Dutch words
(e.g., “kerk,” church), eye movements were monitored to pictures of
phonological, shape, and semantic competitors (e.g., a cherry [“kers”],
a house [“huis”], and a grave [“graf ”]) and to pictures of objects unrelated
on all three dimensions. Time course differences were observed
with attentional shifts to phonological competitors preceding shifts to
shape competitors, which in turn preceded shifts to semantic competitors.
These data suggest that, during language-mediated visual search,
attention is directed to the item with the currently highest priority ranking.
This ranking is codetermined by the type of lexical information
that becomes available as the spoken word unfolds and by the types of
featural information that are copresent in the visual display.