D.M. Groppe 1,2,3 , S. Bickel 4 , C. Keller 1,2,3,4 , L. Entz 5,6 , A.D. Mehta 1,2,3 1 Departments of Neurology and Neurosurgery, North Shore Long Island Jewish Health System; 2 Feinstein Institute for Medical Research, Hofstra University School of Medicine; 3 Harvey Cushing Institutes of Neuroscience; 4 Albert Einstein College of Medicine; 5 Institute for Psychology of the Hungarian Academy of Sciences; 6 Hungarian National Institute of Neuroscience Questions •What are the dominant frequencies of oscillatory resting activity of different human neocortical areas as measured by the electrocorticogram (ECoG)? •How reliable are such dominant frequencies within an individual? Participants & Procedure •11 subjects with pharmacologically intractable epilepsy (5 female) •Mean subject age: 32 (SD=14) years •Two clips of ECoG recordings were taken from 24 hour clinical recordings of each patient based on video footage of the subjects •During each clip, patients were awake and resting, passively watching TV, or potentially listening to the conversations of others •Mean clip duration: 6.2 (SD=1.7) minutes •Median time between clips: 33.8 (SIQR=25.3) hours • ~7 Hz oscillations, are generally the dominant mode of sub-13 Hz activity of the cortical areas sampled. This contradicts previous estimates from scalp EEG the scalp which generally found predominant ~10 Hz activity 1,2 and muddies the conventional distinction between theta and alpha bands. • Theta peaks most reliably found in fusiform, inferior-middle temporal, parahippocampal, lateral orbitofrontal cortex. • Beta peaks found most reliably in middle temporal, parsopercularis, parstriangularis, pre & post central cortex. It is not clear to what extent these peaks are harmonics of lower frequencies • Gamma peaks found in entorhinal cortex. May be related to pathology • Peaks reproduce across days and are evident in .5 min of data. But reproducibility might be improved by better control of subject state and environment. Dominant frequencies of resting human neocortical oscillations as measured by the electrocorticogram Introduction Methods Results Conclusions Background Neocortical electromagnetic activity often exhibits oscillations at different frequencies. These oscillations are apparent in the activity of the resting brain and are modulated by sensory stimulation, behavior, and cognitive state. Previous research on humans has found that the most dominant frequencies of these oscillations tend varies across neocortical areas. For example, using EEG recordings of the resting brain, Nunez et al. have found alpha activity (8-12 Hz) to be widespread and theta activity (4-7 Hz) to be somewhat frontally distributed 1,2 . Such systematic differences in neural dynamics likely reflect differences in connectivity between brain areas and resting function and may be a useful biomarker of neural disorders 3 . However characterizing the dominant frequencies of neocortical areas in humans has been difficult due the poor spatial resolution of noninvasive measures of brain electromagnetic activity. Our Study We measured magnitude of neocortical oscillatory activity in the resting brain using ECoG recordings obtained via intracranial electrodes implanted on the surface of the brain in patients with pharmacologically intractable epilepsy undergoing evaluation for surgery. ECoG Recording & Preprocessing •On average 109 (SD=10) subdural electrodes were implanted in each subject •ECoG recorded at 500-2000 Hz using an XLtek clinical system •ECoG referenced to electrodes over interhemispheric fissure and re-referenced to average reference after electrodes over ictal onset regions or with poor contacts were rejected (leaving a mean of 88, SD=10, usable electrodes per subject) •Time periods in which the patient moved or that contained interictal spikes were manually identified and rejected, leaving a mean of 5.6 (SD=1.1) minutes of data for analysis •ECoG whitened by taking its temporal derivative to heighten spectral power peaks 1. P. L. Nunez, B. M. Wingeier, R. B. Silberstein, Spatial-temporal structures of human alpha rhythms: theory, microcurrent sources, multiscale measurements, and global binding of local networks, Hum. Brain Mapp. 13, 125 (2001). 2. P. L. Nunez, L. Reid, R. G. Bickford, The relationship of head size to alpha frequency with implications to a brain wave model, Electroencephalogr. Clin. Neurophysiol. 44, 344 (1978). 3. D. Zhang, M. E. Raichle, Disease and the brain's dark energy, Nat. Rev. Neurol. 6, 15 (2010). References A B C D E Spectral Analysis •Segments of data divided into 1 sec epochs with 0.5 sec overlap •Power spectrum density (PSD) for each epoch estimated with the mean of two Slepian tapers and discrete Fourier transform •Mean PSD for each electrode derived by taking the 5% trimmed mean of all epochs to ignore outliers •PSD normalized to unit area ignoring line noise frequencies and frequencies above 164 Hz MRI Coregistration and Cortical Parcellation •(A) Postimplant CT scan, (B) Manual identification of electrodes, (C) Coregistration of CT scan to preimplant MRI, (D) Coregistration of electrode locations to FreeSurfer pial surface, (E) Electrode locations snapped to pial surface •Cortical areas automatically defined by FreeSurfer using the Desikan-Killiany atlas 1-12 Hz Peak Frequency Mean PSD Across Subjects per Cortical Area: Shaded regions represent 95% confidence intervals. n indicates the number of subjects contributing to each average. Dashed lines indicate peaks. Max Power in Bands: Mean frequency of maximal power value in select bands. Error bars=95% CI. Solid lines indicate frequency range boundaries. Bars that differ from boundaries indicate PSD peaks. 13-30 Hz Peak Frequency 31-127 Hz Peak Frequency Peak frequencies across all electrodes: Histogram of PSD peaks across all electrodes. Blank regions indicate frequencies ignored due to line noise. Uniform predominance of peaks above ~50 Hz reflects residual 1/f PSD distribution. 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 0 0.05 0.09 0.14 0.19 0.24 0.28 Hz Relative Frequency 7 Hz 14 Hz 19 Hz 23 Hz ! " #$ %& $! #&" &’$ ’#& ( ()# ()& ()% ()! ()’ ()$ ()* ()" +,-./012 14 50-6/ 7809 :6;<12=6> 7-166!7809 71--;8./012 #!#& ?@ #%!%( ?@ Contact: [email protected] Test-retest peak reliability: Mean correlation between maximal PSD frequencies in theta-alpha and beta bands across the two clips of data for each patient. Various amounts of the data in the first clip were used to estimate the amount data necessary to get reliable estimates. 0 2 4 6 8 10 12 14 superior frontal rostral middle frontal medial orbitofrontal lateral orbitofrontal parsorbitalis parstriangularis parsopercularis caudal middle frontal precentral postcentral middle temporal inferior temporal entorhinal parahippocampal fusiform precuneus inferior parietal lingual lateral occipital Hz Alpha/Theta Boundary 8 10 12 14 16 18 20 22 24 superior frontal rostral middle frontal medial orbitofrontal lateral orbitofrontal parsorbitalis parstriangularis parsopercularis caudal middle frontal precentral postcentral middle temporal inferior temporal entorhinal parahippocampal fusiform precuneus inferior parietal lingual lateral occipital Hz 20 30 40 50 60 70 80 90 superior frontal rostral middle frontal medial orbitofrontal lateral orbitofrontal parsorbitalis parstriangularis parsopercularis caudal middle frontal precentral postcentral middle temporal inferior temporal entorhinal parahippocampal fusiform precuneus inferior parietal lingual lateral occipital Hz
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
D.M. Groppe1,2,3, S. Bickel4, C. Keller1,2,3,4, L. Entz5,6, A.D. Mehta1,2,3
1Departments of Neurology and Neurosurgery, North Shore Long Island Jewish Health System; 2Feinstein Institute for Medical Research, Hofstra University School of Medicine; 3Harvey Cushing
Institutes of Neuroscience; 4Albert Einstein College of Medicine; 5Institute for Psychology of the Hungarian Academy of Sciences; 6Hungarian National Institute of Neuroscience
Questions•What are the dominant frequencies of oscillatory resting activity of different human neocortical areas as measured by the electrocorticogram (ECoG)?
•How reliable are such dominant frequencies within an individual?
Participants & Procedure•11 subjects with pharmacologically intractable epilepsy (5 female)
•Mean subject age: 32 (SD=14) years
•Two clips of ECoG recordings were taken from 24 hour clinical
recordings of each patient based on video footage of the subjects
•During each clip, patients were awake and resting, passively
watching TV, or potentially listening to the conversations of others
•Mean clip duration: 6.2 (SD=1.7) minutes
•Median time between clips: 33.8 (SIQR=25.3) hours
•~7 Hz oscillations, are generally the dominant mode of sub-13
Hz activity of the cortical areas sampled. This contradicts previous
estimates from scalp EEG the scalp which generally found
predominant ~10 Hz activity1,2 and muddies the conventional
distinction between theta and alpha bands.
•Theta peaks most reliably found in fusiform, inferior-middle temporal,
parahippocampal, lateral orbitofrontal cortex.
•Beta peaks found most reliably in middle temporal, parsopercularis,
parstriangularis, pre & post central cortex. It is not clear to what extent
these peaks are harmonics of lower frequencies
•Gamma peaks found in entorhinal cortex. May be related to
pathology
•Peaks reproduce across days and are evident in .5 min of data.
But reproducibility might be improved by better control of subject state
and environment.
Dominant frequencies of resting human neocortical oscillations as measured
by the electrocorticogram
Introduction Methods Results
Conclusions
BackgroundNeocortical electromagnetic activity often exhibits oscillations at
different frequencies. These oscillations are apparent in the activity
of the resting brain and are modulated by sensory stimulation,
behavior, and cognitive state. Previous research on humans has
found that the most dominant frequencies of these oscillations tend
varies across neocortical areas. For example, using EEG
recordings of the resting brain, Nunez et al. have found alpha
activity (8-12 Hz) to be widespread and theta activity (4-7 Hz) to be
somewhat frontally distributed1,2. Such systematic differences in
neural dynamics likely reflect differences in connectivity between
brain areas and resting function and may be a useful biomarker of
neural disorders3. However characterizing the dominant
frequencies of neocortical areas in humans has been difficult due
the poor spatial resolution of noninvasive measures of brain
electromagnetic activity.
Our StudyWe measured magnitude of neocortical oscillatory activity in the
resting brain using ECoG recordings obtained via intracranial
electrodes implanted on the surface of the brain in patients with
pharmacologically intractable epilepsy undergoing evaluation for
surgery.
ECoG Recording & Preprocessing•On average 109 (SD=10) subdural electrodes were implanted in
each subject
•ECoG recorded at 500-2000 Hz using an XLtek clinical system
•ECoG referenced to electrodes over interhemispheric fissure and
re-referenced to average reference after electrodes over ictal onset
regions or with poor contacts were rejected (leaving a mean of 88,
SD=10, usable electrodes per subject)
•Time periods in which the patient moved or that contained interictal
spikes were manually identified and rejected, leaving a mean of 5.6
(SD=1.1) minutes of data for analysis
•ECoG whitened by taking its temporal derivative to heighten
spectral power peaks
1. P. L. Nunez, B. M. Wingeier, R. B. Silberstein, Spatial-temporal structures of human
alpha rhythms: theory, microcurrent sources, multiscale measurements, and global
binding of local networks, Hum. Brain Mapp. 13, 125 (2001).
2. P. L. Nunez, L. Reid, R. G. Bickford, The relationship of head size to alpha
frequency with implications to a brain wave model, Electroencephalogr. Clin.
Neurophysiol. 44, 344 (1978).
3. D. Zhang, M. E. Raichle, Disease and the brain's dark energy, Nat. Rev. Neurol. 6, 15
(2010).
References
A B C
D E
Spectral Analysis•Segments of data divided into 1 sec epochs with 0.5 sec overlap
•Power spectrum density (PSD) for each epoch estimated with the
mean of two Slepian tapers and discrete Fourier transform
•Mean PSD for each electrode derived by taking the 5% trimmed
mean of all epochs to ignore outliers
•PSD normalized to unit area ignoring line noise frequencies and
frequencies above 164 Hz
MRI Coregistration and Cortical
Parcellation
•(A) Postimplant CT scan, (B) Manual identification of electrodes,
(C) Coregistration of CT scan to preimplant MRI, (D) Coregistration
of electrode locations to FreeSurfer pial surface, (E) Electrode
locations snapped to pial surface
•Cortical areas automatically defined by FreeSurfer using the
Desikan-Killiany atlas
1-12 Hz Peak Frequency
Mean PSD Across Subjects per Cortical Area: Shaded regions represent 95%
confidence intervals. n indicates the number of subjects contributing to each average.
Dashed lines indicate peaks.
Max Power in Bands: Mean frequency of maximal
power value in select
bands. Error bars=95% CI.
Solid lines indicate
frequency range
boundaries. Bars that differ
from boundaries indicate
PSD peaks.
13-30 Hz Peak Frequency
31-127 Hz Peak Frequency
Peak frequencies across all electrodes: Histogram of PSD
peaks across all electrodes. Blank regions indicate frequencies ignored
due to line noise. Uniform predominance of peaks above ~50 Hz reflects
residual 1/f PSD distribution.
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 1600
0.05
0.09
0.14
0.19
0.24
0.28
Hz
Rela
tive F
req
uen
cy
7 Hz
14 Hz
19 Hz
23 Hz
! " #$ %& $! #&" &'$ '#&(
()#
()&
()%
()!
()'
()$
()*
()"
+,-./012314350-6/378093:6;<12=6>
7-166!7809371--;8./012
#!#&3?@
#%!%(3?@
Contact:[email protected]
Test-retest peak reliability: Mean
correlation between maximal PSD
frequencies in theta-alpha and beta bands
across the two clips of data for each patient.
Various amounts of the data in the first clip
were used to estimate the amount data
necessary to get reliable estimates.
0
2
4
6
8
10
12
14
su
peri
or
fro
nta
l
rostr
al m
idd
le f
ron
tal
med
ial o
rbit
ofr
on
tal
late
ral o
rbit
ofr
on
tal
pars
orb
italis
pars
tria
ng
ula
ris
pars
op
erc
ula
ris
cau
dal m
idd
le f
ron
tal
pre
cen
tral
po
stc
en
tral
mid
dle
tem
po
ral
infe
rio
r te
mp
ora
l
en
torh
inal
para
hip
po
cam
pal
fusif
orm
pre
cu
neu
s
infe
rio
r p
ari
eta
l
lin
gu
al
late
ral o
ccip
ital
Hz
Alpha/ThetaBoundary
8
10
12
14
16
18
20
22
24
su
peri
or
fro
nta
l
rostr
al m
idd
le f
ron
tal
med
ial o
rbit
ofr
on
tal
late
ral o
rbit
ofr
on
tal
pars
orb
italis
pars
tria
ng
ula
ris
pars
op
erc
ula
ris
cau
dal m
idd
le f
ron
tal
pre
cen
tral
po
stc
en
tral
mid
dle
tem
po
ral
infe
rio
r te
mp
ora
l
en
torh
inal
para
hip
po
cam
pal
fusif
orm
pre
cu
neu
s
infe
rio
r p
ari
eta
l
lin
gu
al
late
ral o
ccip
ital
Hz
20
30
40
50
60
70
80
90
su
peri
or
fro
nta
l
rostr
al m
idd
le f
ron
tal
med
ial o
rbit
ofr
on
tal
late
ral o
rbit
ofr
on
tal
pars
orb
italis
pars
tria
ng
ula
ris
pars
op
erc
ula
ris
cau
dal m
idd
le f
ron
tal
pre
cen
tral
po
stc
en
tral
mid
dle
tem
po
ral
infe
rio
r te
mp
ora
l
en
torh
inal
para
hip
po
cam
pal
fusif
orm
pre
cu
neu
s
infe
rio
r p
ari
eta
l
lin
gu
al
late
ral o
ccip
ital
Hz
Related Documents
