Detecting
temporal
and
spectral
features
of
neural
oscillations
is
essential
to
understanding
dynamic
brain
function.
Traditionally,
the
presence
frequency
are
determined
by
identifying
peaks
over
1/f
noise
within
power
spectrum.
However,
this
approach
solely
operates
domain
thus
cannot
adequately
distinguish
between
fundamental
a
non-sinusoidal
oscillation
its
harmonics.
Non-sinusoidal
signals
generate
harmonics,
significantly
increasing
false-positive
detection
rate
—
confounding
factor
in
analysis
oscillations.
To
overcome
these
limitations,
we
define
criteria
that
characterize
introduce
Cyclic
Homogeneous
Oscillation
(CHO)
method
implements
based
on
an
auto-correlation
determines
oscillation's
periodicity
frequency.
We
evaluated
CHO
verifying
performance
simulated
sinusoidal
oscillatory
bursts
convolved
with
noise.
Our
results
demonstrate
outperforms
conventional
techniques
accurately
detecting
Specifically,
sensitivity
specificity
as
function
signal-to-noise
ratio
(SNR).
further
assessed
testing
it
electrocorticographic
(ECoG,
8
subjects)
electroencephalographic
(EEG,
7
recorded
during
pre-stimulus
period
auditory
reaction
time
task
(6
SEEG
subjects
6
ECoG
collected
resting
state.
In
task,
detected
alpha
pre-motor
beta
occipital
EEG
signals.
Moreover,
hippocampal
human
hippocampus
state
subjects).
summary,
demonstrates
high
precision
domains.
The
method's
enables
detailed
study
characteristics
oscillations,
such
degree
asymmetry
waveform
oscillation.
Furthermore,
can
be
applied
identify
how
govern
interactions
throughout
determine
biomarkers
index
abnormal
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Май 4, 2024
Abstract
High
Frequency
Oscillations
(HFOs)
is
an
important
biomarker
that
can
potentially
pinpoint
the
epileptogenic
zones
(EZs).
However,
duration
of
HFO
short
with
around
4
cycles,
which
might
be
hard
to
recognize
when
embedded
within
signals
lower
frequency
oscillatory
background.
In
addition,
annotating
HFOs
manually
time-consuming
given
long-time
recordings
and
up
hundreds
intracranial
electrodes.
We
propose
leverage
a
Switching
State
Space
Model
(SSSM)
identify
events
automatically
instantaneously
without
relying
on
extracting
features
from
sliding
windows.
The
effectiveness
SSSM
for
detection
fully
validated
in
EEG
recording
human
subjects
undergoing
presurgical
evaluations
showed
improved
accuracy
capturing
occurrence
their
duration.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Май 9, 2024
In
neuroscience,
understanding
how
single-neuron
firing
contributes
to
distributed
neural
ensembles
is
crucial.
Traditional
methods
of
analysis
have
been
limited
descriptions
whole
population
activity,
or,
when
analyzing
individual
neurons,
criteria
for
response
categorization
varied
significantly
across
experiments.
Current
lack
scalability
large
datasets,
fail
capture
temporal
changes
and
rely
on
parametric
assumptions.
There's
a
need
robust,
scalable,
non-parametric
functional
clustering
approach
interpretable
dynamics.
To
address
this
challenge,
we
developed
model-based,
statistical
framework
unsupervised
multiple
time
series
datasets
that
exhibit
nonlinear
dynamics
into
an
Determining
the
presence
and
frequency
of
neural
oscillations
is
essential
to
understanding
dynamic
brain
function.
Traditional
methods
that
detect
peaks
over
1/
f
noise
within
power
spectrum
fail
distinguish
between
fundamental
harmonics
often
highly
non-sinusoidal
oscillations.
To
overcome
this
limitation,
we
define
criteria
characterize
introduce
cyclic
homogeneous
oscillation
(CHO)
detection
method.
We
implemented
these
based
on
an
autocorrelation
approach
determine
oscillation’s
frequency.
evaluated
CHO
by
verifying
its
performance
simulated
oscillatory
bursts
validated
ability
in
electrocorticographic
(ECoG),
electroencephalographic
(EEG),
stereoelectroencephalographic
(SEEG)
signals
recorded
from
27
human
subjects.
Our
results
demonstrate
outperforms
conventional
techniques
accurately
detecting
In
summary,
demonstrates
high
precision
specificity
time
domains.
The
method’s
enables
detailed
study
characteristics
oscillations,
such
as
degree
asymmetry
waveform
oscillation.
Furthermore,
can
be
applied
identify
how
govern
interactions
throughout
biomarkers
index
abnormal
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2023,
Номер
unknown
Опубликована: Окт. 6, 2023
Detecting
temporal
and
spectral
features
of
neural
oscillations
is
essential
to
understanding
dynamic
brain
function.
Traditionally,
the
presence
frequency
are
determined
by
identifying
peaks
over
1/f
noise
within
power
spectrum.
However,
this
approach
solely
operates
domain
thus
cannot
adequately
distinguish
between
fundamental
a
non-sinusoidal
oscillation
its
harmonics.
Non-sinusoidal
signals
generate
harmonics,
significantly
increasing
false-positive
detection
rate
-
confounding
factor
in
analysis
oscillations.
To
overcome
these
limitations,
we
define
criteria
that
characterize
introduce
Cyclic
Homogeneous
Oscillation
(CHO)
method
implements
based
on
an
auto-correlation
determines
oscillation's
periodicity
frequency.
We
evaluated
CHO
verifying
performance
simulated
sinusoidal
oscillatory
bursts
convolved
with
noise.
Our
results
demonstrate
outperforms
conventional
techniques
accurately
detecting
Specifically,
sensitivity
specificity
as
function
signal-to-noise
ratio
(SNR).
further
assessed
testing
it
electrocorticographic
(ECoG,
8
subjects)
electroencephalographic
(EEG,
7
recorded
during
pre-stimulus
period
auditory
reaction
time
task
(6
SEEG
subjects
6
ECoG
collected
resting
state.
In
task,
detected
alpha
pre-motor
beta
occipital
EEG
signals.
Moreover,
hippocampal
human
hippocampus
state
subjects).
summary,
demonstrates
high
precision
domains.
The
method's
enables
detailed
study
characteristics
oscillations,
such
degree
asymmetry
waveform
oscillation.
Furthermore,
can
be
applied
identify
how
govern
interactions
throughout
determine
biomarkers
index
abnormal
Detecting
temporal
and
spectral
features
of
neural
oscillations
is
essential
to
understanding
dynamic
brain
function.
Traditionally,
the
presence
frequency
are
determined
by
identifying
peaks
over
1/f
noise
within
power
spectrum.
However,
this
approach
solely
operates
domain
thus
cannot
adequately
distinguish
between
fundamental
a
non-sinusoidal
oscillation
its
harmonics.
Non-sinusoidal
signals
generate
harmonics,
significantly
increasing
false-positive
detection
rate
—
confounding
factor
in
analysis
oscillations.
To
overcome
these
limitations,
we
define
criteria
that
characterize
introduce
Cyclic
Homogeneous
Oscillation
(CHO)
method
implements
based
on
an
auto-correlation
determines
oscillation's
periodicity
frequency.
We
evaluated
CHO
verifying
performance
simulated
sinusoidal
oscillatory
bursts
convolved
with
noise.
Our
results
demonstrate
outperforms
conventional
techniques
accurately
detecting
Specifically,
sensitivity
specificity
as
function
signal-to-noise
ratio
(SNR).
further
assessed
testing
it
electrocorticographic
(ECoG,
8
subjects)
electroencephalographic
(EEG,
7
recorded
during
pre-stimulus
period
auditory
reaction
time
task
(6
SEEG
subjects
6
ECoG
collected
resting
state.
In
task,
detected
alpha
pre-motor
beta
occipital
EEG
signals.
Moreover,
hippocampal
human
hippocampus
state
subjects).
summary,
demonstrates
high
precision
domains.
The
method's
enables
detailed
study
characteristics
oscillations,
such
degree
asymmetry
waveform
oscillation.
Furthermore,
can
be
applied
identify
how
govern
interactions
throughout
determine
biomarkers
index
abnormal
