Frontiers in Psychology,
Journal Year:
2022,
Volume and Issue:
13
Published: Nov. 15, 2022
In
natural
reading,
the
reader's
processing
of
a
word
starts
when
is
located
in
parafoveal
vision.
Relative
to
situation
with
an
invalid
preview,
fixations
are
significantly
shorter
after
fully
valid,
identical
preview.
Although
research
on
identity
preview
benefit
has
been
ongoing
for
more
than
40
years,
mechanism
this
benefit,
level
lexical
it
occurs,
and
its
relationship
sentential
context
remain
unclear.
This
study
employed
EEG
brain
component
analysis
technology
address
these
questions
Chinese
sentence
reading.
We
manipulated
make
target
plausible
or
not
present
results
showed
that
can
affect
only
early
positivity,
reflecting
orthographic
words,
but
also
N400
LPC
components,
late
in-depth
semantic
words.
Conflicting
context,
which
implausible
cannot
be
integrated
into
sentence,
interfere
temporarily
processes.
These
findings
suggest
process
promote
subsequent
reading
at
multiple
levels,
role
modulated
by
contextual
information.
Humans
can
read
and
comprehend
text
rapidly,
implying
that
readers
might
process
multiple
words
per
fixation.
However,
the
extent
to
which
parafoveal
are
previewed
integrated
into
evolving
sentence
context
remains
disputed.
We
investigated
processing
during
natural
reading
by
recording
brain
activity
eye
movements
using
MEG
an
tracker
while
participants
silently
one-line
sentences.
The
sentences
contained
unpredictable
target
word
was
either
congruent
or
incongruent
with
context.
To
measure
processing,
we
flickered
at
60
Hz
measured
resulting
responses
(i.e.
Rapid
Invisible
Frequency
Tagging,
RIFT
)
fixations
on
pre-target
words.
Our
results
revealed
a
significantly
weaker
tagging
response
for
were
previous
compared
ones,
even
within
100ms
of
fixating
immediately
preceding
target.
This
reduction
in
also
found
be
predictive
individual
speed.
conclude
semantic
information
is
not
only
extracted
from
parafovea
but
before
fixated.
early
extensive
supports
rapid
required
reading.
study
suggests
theoretical
frameworks
should
incorporate
concept
deep
processing.
Scientific Data,
Journal Year:
2024,
Volume and Issue:
11(1)
Published: Oct. 9, 2024
This
paper
introduces
the
DERCo
(Dublin
EEG-based
Reading
Experiment
Corpus),
a
language
resource
combining
electroencephalography
(EEG)
and
next-word
prediction
data
obtained
from
participants
reading
narrative
texts.
The
dataset
comprises
behavioral
collected
500
recruited
through
Amazon
Mechanical
Turk
online
crowd-sourcing
platform,
along
with
EEG
recordings
22
healthy
adult
native
English
speakers.
experiment
was
designed
to
examine
context-based
word
by
large
sample
of
participants,
while
developed
extend
validation
predictability.
Online
were
instructed
predict
upcoming
words
complete
entire
stories.
Cloze
probabilities
then
calculated
for
each
so
that
this
predictability
measure
could
be
used
support
various
analyses
pertaining
semantic
context
effects
in
recordings.
revealed
significant
differences
between
high
low
predictable
words,
demonstrating
one
important
type
potential
analysis
necessitates
close
integration
these
two
datasets.
material
is
valuable
researchers
neurolinguistics
due
word-level
context.
Language and Cognition,
Journal Year:
2023,
Volume and Issue:
16(2), P. 380 - 400
Published: Sept. 13, 2023
Abstract
Studies
on
sentence
processing
in
inflectional
languages
support
that
syntactic
structure
building
functionally
precedes
semantic
processing.
Conversely,
most
EEG
studies
of
Chinese
do
not
the
priority
syntax.
One
possible
explanation
is
language
lacks
morphological
inflections.
Another
may
be
presentation
separate
components
individual
screens
disrupts
framework
construction
during
reading.
The
present
study
investigated
this
using
a
self-paced
reading
experiment
mimicking
rapid
serial
visual
and
an
eye-tracking
reflecting
natural
In
both
experiments,
‘ba’
sentences
were
presented
to
young
adults
four
conditions
differed
across
dimensions
congruency.
Evidence
supporting
functional
syntax
over
semantics
was
limited
only
context,
which
violations
blocked
semantics.
Additionally,
we
observed
later
stage
integrating
plausible
with
failed
Together,
our
findings
extend
highlight
importance
adopting
more
ecologically
valid
methods
when
investigating
Humans
can
read
and
comprehend
text
rapidly,
implying
that
readers
might
process
multiple
words
per
fixation.
However,
the
extent
to
which
parafoveal
are
previewed
integrated
into
evolving
sentence
context
remains
disputed.
We
investigated
processing
during
natural
reading
by
recording
brain
activity
eye
movements
using
MEG
an
tracker
while
participants
silently
one-line
sentences.
The
sentences
contained
unpredictable
target
word
was
either
congruent
or
incongruent
with
context.
To
measure
processing,
we
flickered
at
60
Hz
measured
resulting
responses
(i.e.,
Rapid
Invisible
Frequency
Tagging,
RIFT)
fixations
on
pre-target
words.
Our
results
revealed
a
significantly
weaker
tagging
response
for
were
previous
compared
ones,
even
within
100
ms
of
fixating
immediately
preceding
target.
This
reduction
in
RIFT
also
found
be
predictive
individual
speed.
conclude
semantic
information
is
not
only
extracted
from
parafovea
but
before
fixated.
early
extensive
supports
rapid
required
reading.
study
suggests
theoretical
frameworks
should
incorporate
concept
deep
processing.
Humans
can
read
and
comprehend
text
rapidly,
implying
that
readers
might
process
multiple
words
per
fixation.
However,
the
extent
to
which
parafoveal
are
previewed
integrated
into
evolving
sentence
context
remains
disputed.
We
investigated
processing
during
natural
reading
by
recording
brain
activity
eye
movements
using
MEG
an
tracker
while
participants
silently
one-line
sentences.
The
sentences
contained
unpredictable
target
word
was
either
congruent
or
incongruent
with
context.
To
measure
processing,
we
flickered
at
60
Hz
measured
resulting
responses
(i.e.,
Rapid
Invisible
Frequency
Tagging,
RIFT)
fixations
on
pre-target
words.
Our
results
revealed
a
significantly
weaker
tagging
response
for
were
previous
compared
ones,
even
within
100
ms
of
fixating
immediately
preceding
target.
This
reduction
in
RIFT
also
found
be
predictive
individual
speed.
conclude
semantic
information
is
not
only
extracted
from
parafovea
but
before
fixated.
early
extensive
supports
rapid
required
reading.
study
suggests
theoretical
frameworks
should
incorporate
concept
deep
processing.
Humans
can
read
and
comprehend
text
rapidly,
implying
that
readers
might
process
multiple
words
per
fixation.
However,
the
extent
to
which
parafoveal
are
previewed
integrated
into
evolving
sentence
context
remains
disputed.
We
investigated
processing
during
natural
reading
by
recording
brain
activity
eye
movements
using
MEG
an
tracker
while
participants
silently
one-line
sentences.
The
sentences
contained
unpredictable
target
word
was
either
congruent
or
incongruent
with
context.
To
measure
processing,
we
flickered
at
60
Hz
measured
resulting
responses
(i.e.,
Rapid
Invisible
Frequency
Tagging,
RIFT)
fixations
on
pre-target
words.
Our
results
revealed
a
significantly
weaker
tagging
response
for
were
previous
compared
ones,
even
within
100
ms
of
fixating
immediately
preceding
target.
This
reduction
in
RIFT
also
found
be
predictive
individual
speed.
conclude
semantic
information
is
not
only
extracted
from
parafovea
but
before
fixated.
early
extensive
supports
rapid
required
reading.
study
suggests
theoretical
frameworks
should
incorporate
concept
deep
processing.
Cognition,
Journal Year:
2024,
Volume and Issue:
255, P. 105991 - 105991
Published: Nov. 26, 2024
In
recent
years,
several
influential
computational
models
and
metrics
have
been
proposed
to
predict
how
humans
comprehend
process
sentence.
One
particularly
promising
approach
is
contextual
semantic
similarity.
Inspired
by
the
attention
algorithm
in
Transformer
human
memory
mechanisms,
this
study
proposes
an
"attention-aware"
for
computing
relevance.
This
new
takes
into
account
different
contributions
of
parts
expectation
effect,
allowing
it
incorporate
information
fully.
The
attention-aware
also
facilitates
simulation
existing
reading
their
evaluation.
resulting
relevance
can
more
accurately
fixation
durations
Chinese
tasks
recorded
eye-tracking
corpus
than
those
calculated
approaches.
study's
findings
further
provide
strong
support
presence
preview
benefits
naturalistic
reading.
Furthermore,
relevance,
being
memory-based,
possess
high
interpretability
from
both
linguistic
cognitive
standpoints,
making
them
a
valuable
tool
modeling
eye-movements
gaining
insight
language
comprehension.
Our
emphasizes
potential
these
advance
our
understanding
language.
PeerJ,
Journal Year:
2023,
Volume and Issue:
11, P. e15291 - e15291
Published: May 3, 2023
Background
Numerous
studies
have
confirmed
that
skilled
readers
can
benefit
from
a
semantically
related
preview
word
(
i.e
.,
semantic
benefit,
SPB),
suggesting
extract
information
the
parafovea
to
achieve
efficient
reading.
It
is
still
under
debate
whether
occurrence
of
this
because
association
between
and
target
words
or
contextual
fit
in
sentence
context.
Methods
Two
independent
factors,
plausibility
(preview
plausible/implausible)
relatedness
(semantically
related/unrelated),
were
manipulated,
we
further
strictly
controlled
for
syntactic
present
study.
Results
The
results
showed
first-pass
reading
times
significantly
shorter
plausible
condition
than
implausible
condition.
However,
main
effect
was
found
only
gaze
duration
measure.
Discussion
pattern
revealed
affects
preferentially
Chinese
reading,
supporting
account.
Our
findings
implications
better
understanding
parafoveal
processing
provide
empirical
support
eye-movement
control
model.
Humans
can
read
and
comprehend
text
rapidly,
implying
that
readers
might
process
multiple
words
per
fixation.
However,
the
extent
to
which
parafoveal
are
previewed
integrated
into
evolving
sentence
context
remains
disputed.
We
investigated
processing
during
natural
reading
by
recording
brain
activity
eye
movements
using
MEG
an
tracker
while
participants
silently
one-line
sentences.
The
sentences
contained
unpredictable
target
word
was
either
congruent
or
incongruent
with
context.
To
measure
processing,
we
flickered
at
60
Hz
measured
resulting
responses
(i.e.
Rapid
Invisible
Frequency
Tagging,
RIFT)
fixations
on
pre-target
words.
Our
results
revealed
a
significantly
weaker
tagging
response
for
were
previous
compared
ones,
even
within
100ms
of
fixating
immediately
preceding
target.
This
reduction
in
RIFT
also
found
be
predictive
individual
speed.
conclude
semantic
information
is
not
only
extracted
from
parafovea
but
before
fixated.
early
extensive
supports
rapid
required
reading.
study
suggests
theoretical
frameworks
should
incorporate
concept
deep
processing.
bioRxiv (Cold Spring Harbor Laboratory),
Journal Year:
2024,
Volume and Issue:
unknown
Published: March 16, 2024
Abstract
Inferences
are
an
indicator
of
a
greater
reading
comprehension,
as
they
imply
combination
implicit
and
explicit
information
that
usually
combines
textual
representation
with
background
knowledge
the
reader.
The
aim
this
study
is
to
explore
costs
benefits
time
course
inferences
in
university
students
comprehension
difficulties
at
3
stages
during
narration.
method
used
was
event-related
potential
(ERP)
technique
order
register
brain
activity
63
teaching
program
while
read
familiar,
less-familiar
neutral
stories.
Results
show
slow
negativity
component
words
coming
from
familiar
contexts
when
compared
less
ones
first
locus;
N400
Post-N400
second
locus,
reflecting
ones;
and,
lastly,
through
use
lexical
decision
task,
FN400
components
were
found
third
especially
for
pseudowords.
These
results
interpreted
preferably
bottom-up
processing,
which
characterized
by
access
less-skilled
readers.