makeig ucsd i 10
TRANSCRIPT
!"#"#$%&'$#"()*%+,-"#%./#-0"12%3!
41'5%!-6*"$%
"#$%&'&(!)*+!,('+-.!/*01'&-%*#!
2#34(+$3&5!*)!/-.3)*+#3-!6-#!73(8*!
789:%;;<=>+%?',62:'@%A%B&4.%
C')*0D*,E%8F7F!
4G%!-6*"$%8F7F!
99:;<=!>3$&*+5!•!!?@@A!B!9C6D!EF-G(38H!
•!!?@@I!B!"#)*0-J!"/<!)*+!99:!EF-G(38K!=(..K!L'#8K!6(M#*N$G3H!
•!!?@@O!P!99:Q"/<!R**.S*J!ET#.U$-.GU(V'HK!"R/!W!9C/!
•!!?@@@!P!9CDP30-8(!1.*&$!EL'#8!W!F-G(38H!
•!!XYYY!P!R**.S*J!:2"!V($38#!E7(.*+0(H!
•!!XYYX!P!99:;<=!E$TT#U'T$VU(V'H!
•!!XYYZ!P!,">!$'11*+&K!-#V!E7(.*+0(!W!F-G(38H!+()(+(#T(!1-1(+!
•!!XYYZ!P!?$&!!99:;<=!N*+G$[*1!E2/67K!;-!L*..-K!/-.3)*+#3-H!!
•!!XYYZ!P!?$&!99:;<=!1.'8P3#$!
•!!XYYI!B!\*+G$[*1$!3#!D*+&*!ED*+&'8-.H!-#V!;3S*#!E6DCH!
•!!XYY]!B!\*+G$[*1!3#!63#8-1*+(!
•!!XYY]!P!?$&!6R27^!$&+'T&'+(!-#V!T*01*#(#&!T.'$&(+3#8!&**.$!
•!!XYYO!P!\*+G$[*1$!3#!<$1(&!E_+-#T(HK!;-!L*..-!E/-.3)*+#3-HK!!6-#%-8*!E/[3.(H!
•!!XYY`!P!,">!$'11*+&!+(#(N(Va!
•!!XYY@!B!,_R!B!,('+*(.(T&+*0-8#(%T!_*+N-+V!>(-V!F*V(.3#8!R**.S*J!
•!!XYY@!P!\*+G$[*1$!3#!=.**03#8&*#!"#V3-#-K!<$1(&!_+-#T(K!;-!L*..-!/-.3)*+#3-K!!65V#(5!<'$&+-.3-!
•!!XYY@!B!9C"/<!7-&-C34(+QD+*V'T(+!B!9J1(+30(#&-.!C(-.P%0(!"#&(+-T%4(!/*#&+*.!W!<#-.5$3$!
•!!XY?Y!B!\*+G$[*1$!3#!L54-$G5.-!_3#.-#VK!,/R2K!R-3N-#K!D*+&.-#V!b+(8*#K!-#V!2/67!
•!!XY?Y!P!>(-V"R!V-&-S-$(!%(V!&*!99:;<=!!
•!!XY??!B!99:;<=!/[-&!3#&+*V'T(V!
S. Makeig 2010
EEGLAB downloads for 20/06/2007
Total count is 34
Username Email Comments
Russia @mail.ru eeg, erp, bci
Company @nexstim.com EEG develop e r
Indonesia @tf.itb.ac.id Brain Computer Interface
Finland @psyka.jyu.fi
Australia @newcastle.edu.au Auditory Psychophysics Psychopathology
La Jolla @gmail.com Cogneuro
EEGlab is great!
Chinc a @126.com hi!
? @yahoo.com LFP in DBS patients
US Gov @pnl.gov
US EDU @bethelks.edu EEG and ERP responses to music stimuli
US EDU @wjh.harvard.edu Neuroscienc e
US EDU @wlu.edu olfaction ERP
Switzerland @student.ethz.ch
Sweden @neuro.gu.se EEG
Germany @med.uni-
muenchen.de
China? @163.com Signal Processing
China @sina.com ic a
Finland @helsinki.fi cognitive brain research
Spain @ugr.es
Netherland s @sdf.nl dfg
Company? @tom.com BCI
Franc e @hotmail.fr
Biomedical engineering
movement-related cortical potentials
brain-computer interfaces
•! ~200 EEGLAB downloads a week
… all together to at least 90 country domains
•! > 3,500 on the ‘eeglablist’ discussion list
•! 20++ EEGLAB plug-ins available
S. Makeig 2007
99:;<=!\*+G$[*1$!90 Participants (2007):
•! Canada
•! USA
•! L-1-#!
•! R-3N-#!
•! 6U!c*+(-!
•! <'$&+-.3-!
•! Germany Austria
•! Italy
•! Norway
•! Ireland
•! England
S. Makeig 2010
Portland
La Jolla
Bloomington
Santiago
Newcastle
Singapore
Taiwan
Porto Aspet
Jyväskylä
I jumped ...
I swerved …
I smiled …
I threw ….
I pointed …
I held …
I reached …
I ran …
I shot …
I ducked
I tossed …
I gaped …
Who
am I? S. Makeig 2001
All of a sudden ...
I looked to see if …
I searched the scene for …
I looked again at ….
It occurred to me that …
I wondered if …
I noticed that …
I decided that …
I imagined …
The feeling hit me like …
It struck me that …
I realized that …
?S. Makeig 2001
“When I went to the back of the house for a second
time, I looked around more carefully. Some light
came from the neighbors’ on the other side of the
grape-stake fence. I noticed that the back door of
Stanley Broadhurst’s house was slightly ajar. I
opened it all the way and turned on the kitchen
lights. There were marks around the lock which
showed that it had been jimmied. It occurred to me
that the guy who did the job might still be inside. …
I turned off the kitchen light and waited. The house
was silent. From outside I could hear the pulsing
hum of the arterial boulevard I had just left.”
- Ross MacDonald The Underground Man
Dynamic Brain Events
S. Makeig 2001
Brain processes
have evolved and function
to optimize the outcome
of the behavior
the brain organizes
in response to
perceived challenges
and opportunities.
Embodied Cognition & Agency
+,-"#2%0**9%9:*%1:-HH*#$*%'I%
9:*%0'0*#9J%
perception action
evaluation
Functional Brain Imaging
6U!F-G(38K!XYY@!4G%!-6*"$%8F7F!
4'0*%:K0-#%D,-"#%"0-$"#$%0"H*29'#*2!
!!!!?@X]!!!d?$&!!['0-#!99:!+(T*+V3#8!
EEG (+-%
!!!!?@A`!!!!!?$&!99:!$1(T&+-.!-#-.5$3$!
!!!!?@]X!!!d?$&!T*01'&(+!9CD!-4(+-83#8!E/<RH!
ERP (+-%
!!!!?@O@!!!!!?$&!(4(#&P+(.-&(V!V($5#T[+*#3e-%*#!
!!!!?@@A!!! !?$&!)FC"!=b;7!+(T*+V3#8$!
fMRI (+-%
!!!!?@@A!!!!!!?$&!S+*-VS-#V!9C6D!
!!!!?@@I!!!!!?$&!0'.%$*'+T(!99:!f.&(+3#8!S5!"/<!
!!!!XYY@!!!d?$&!T*00(+T3-.!V+5!(.(T&+*V(!99:!&*5$!
fEEG & BMI (+-!a%
Makeig (2006)
What is EEG?!
EEG (scalp surface fields)
ECOG (larger cortical
surface fields) Local
Extracellular
Fields
Intracellular and
peri-cellular fields
Synaptic and
other trans-
membrane potentials
Brain dynamics are
inherently multi-scale
At each spatial recording scale, the
signal is produced by active partial
coherence of distributed activities at the next smaller scale.
Scott Makeig 2007
Cross-scale coupling
is bi-directional!
Larger
Smaller
EEG (scalp surface fields)
ECOG (larger cortical
surface fields) Local
Extracellular
Fields
Intracellular and
peri-cellular fields
Synaptic and
other trans-
membrane potentials
Brain dynamics are
inherently multi-scale
At each spatial recording scale, the
signal is produced by active partial
coherence of distributed activities at the next smaller scale.
Scott Makeig 2007
Cross-scale coupling
is bi-directional!
Larger
Smaller
SCALE CHAUVINISM!
&',(1'1',(1-H%
&',(1'9:-H-0"1%
L-HK*%4/29*0%
!'MKH-('#2%
;@:-@(1%
!'MKH-('#2%
N:*%$*#*,-('#%-#M%0'MKH-('#%'I%;;<%O%=PQ%
"2%&R!Q=;S%-#M%#'9%T*HH%29KM"*M%
4G%!-6*"$%8FFU!
C*K,'$H"-HV%
;W1"9-9',/%3#:"D"9',/%
X*0'M/#-0"1%"0-$"#$%
g!30-83#8!S+-3#!
;#*,$/!
73+(T&!AP7!3#4(+$(!0*V(.K!
S'&!h'3&(!$.*N!W!3#V3+(T&!
Functional Brain Imaging
;H*19,'0-$#*(1%"0-$"#$%
g!30-83#8!.*T-.!T*+%T-.!
4/#1:,'#/!
AP7!30-83#8!+(h'3+($!0*V(.K!
S'&!h'3&(!)-$&!W!V3+(T&!0(-$'+(!!
*)!*#(!-$1(T&!*)!#('+-.!-T%43&5U!
6U!F-G(38K!XYY@!4G%!-6*"$%8F7F!
!D[-$(!T*#($!E_+((0-#H!
!<4-.-#T[($!ED.(#eH!
4G%!-6*"$%8F7F!
Scott Makeig 2008
Macro field dynamics are
spontaneous emergent
dynamic patterns – in both
outer space and cortex.
Alan Friedman
Big Bear Solar Observatory/NJIT
Alan Friedman
The spatiotemporal field
dynamics of cortex and brain
have not yet been imaged on
multiple spatial scales!
YG%>6-H"#%>1-,E%ZG%Q-H0*,E%<G%?',,*HHE%[%!-6*"$%8F7F!
21-H@%2"$#-H2%\%2'K,1*%2"$#-H2%J%
/*+&(J!
6G'..!
;*T-.!
65#T[+*#5!
;*T-.!
65#T[+*#5!
6G3#!C(.-%4(!
"#V(1(#V(#T(!
4G%!-6*"$%8FFU!
Brain EEG " Scalp EEG!
21-H@%2"$#-H2%\%2'K,1*%2"$#-H2%J%
/*+&(J!
6G'..!
;*T-.!
65#T[+*#5!
;*T-.!
65#T[+*#5!
6G3#!C(.-%4(!
"#V(1(#V(#T(!
4G%!-6*"$%8FFU!
Naïve 2-D interpretation of EEG signals?
?
? ?
?
?
? ?
Actual cortical source volume
conduction patterns (cartoon)
Cortical EEG signal projection
patterns as point processes
“Surely, Dr. Lowe, if there were gravity waves, we would have detected them by now.”
The very broad EEG point-spread function
iU!<G-.3#!<T-+!W!6U!F-G(38!XY?Y!
The very broad EEG point-spread function
iU!<G-.3#!<T-+!W!6U!F-G(38!XY?Y!Single spatially labile source
The Dome of the Sky
Scott Makeig 2008
Stephen L. Alvarez
3-D structure of the Universe
NASA 2009
The very broad EEG point-spread function
iU!<G-.3#!<T-+!W!6U!F-G(38!XY?Y!Spatially static cortical patch source
!D[-$(!T*#($!E_+((0-#H!
!<4-.-#T[($!ED.(#eH!
Scott Makeig 2007
@10 Hz, 20 cm
0° 360°
18°
The very broad EEG point-spread function
iU!<G-.3#!<T-+!W!6U!F-G(38!XY?Y!
Phase lag, center to edge: 18°
The very broad EEG point-spread function
iU!<G-.3#!<T-+!W!6U!F-G(38!XY?Y!
Phase lag, center to edge: 0°
MRI Segmentation
Solve the forward problem
using realistic
head models (BEM)
Mesh generation
EEG/MEG
Source
Image
Inverse
Problem
Signal
Processing Sensor
Localization
Electromagnetic
source localization
Zeynep Akalin Acar, & Scott Makeig ‘06
Simple
Map
ERP !" EEG !" LFP !" #Spikes
1960 " Response
averaging
2000" 1993"
Brain Electrophysiology
S. Makeig, TINS 2002
?
MICRO
MACRO
RT
SPIKES
LFP
ECOG
EEG Recorded !?
~1,000,000 GHz
~1 Hz
BEHAVIOR BRAIN ?
~1 MHz
S. Makeig 2007
?
ERP
Scott Makeig, 2008
Studying ‘cognitive perception’ using ERPs
Data ! Average + “Background”
But, this linear decomposition is veridical
if & only if:
1. The Average appears in each trial.
2. The “Background” is not perturbed in other ways by the time locking events.
The response averaging model:
ERP EEG “noise” EEG
BOLD ERB BOLD “noise”
Not True / Not Defined
Not True
S. Makeig 2004
Conceptual legacies of
single sensor response/rate averaging
-! Reduction of the time series data at each channel
to a single average response time series.
-! Reduction of the data collected at each channel
to an isolated spatial point process.
How to capture more of the event-related brain
dynamics contained in high-density EEG data?
CSF
EEG Cocktail Party
Blind EEG Source Separation by
Independent Component Analysis
S. Makeig (2000)
Spatial Source Filtering!
CSF
EEG Cocktail Party
Blind EEG Source Separation by
Independent Component Analysis
S. Makeig (2000)
Makeig et al., NIPS95
Infomax ICA
Cortex
Domains
of Local Synchrony
Independent
Thalamus
Are EEG source outputs (nearly) independent?
Freeman - phase cones
Plenz - avalanches
S. Makeig (2007)
L'.3(!b#&*#!W!6U!F-G(38!EXYY]H!
3#M*@*#M*#9%D,-"#%;;<%2'K,1*2%
7'-.P$500(&+3T!V31*.(!
$*'+T(!
63#8.(!V31*.(!
$*'+T(!
9h'34-.(#&!V31*.($!
Independent muscle signals
S. Makeig, J. Onton 2005
Onton & Makeig, submitted
Distributed
muscle /
movement
events
N. BigdelyS. Makeig, 2009
ICA is a linear data decomposition method
W * Channel_Data = Activations
W-1 * Activations= Channel_Data
Independent Components of Human EEG are Dipolar
"/<!3#!
1+-T%T(!
(100 channels,
~500k time points)
b#&*#!W!F-G(38K!XYY]!
24 Subjects –Frontal Midline Theta Sources
Onton & Makeig 2006
Equivalent dipole density
Visual Working Memory
Onton et al., 2005
Sternberg
letter
memory task
Onton et al., ‘05
Onton et al., 2005
Auditory
oddball
plus novel
sounds
Onton et al., ‘05
Equivalent dipole density
Auditory Novelty
Onton et al., 2005
Emotion
imagery
task
Onton et al., ‘05
Equivalent dipole density
Emotion Imagination
Onton et al., 2005
Word
memory
(old/new) task
Onton et al., ‘05
Equivalent dipole density
Task A – Old/New Word Memory
dipoledensity()
Onton et al., 2005
Visually
cued
button press
task
Onton et al., ‘05
Equivalent dipole density
Task B – Cued finger movements
dipoledensity()
CPT Flanker FAST EC EO EO EC
Modeling Spatiotemporal Variability
3-Model AMICA Decomposition
Time-on-Task (1.5 hours)
Mod
el Log
Lik
elih
ood
Grainne MacLoughlin & Jason Palmer, 2010
1.! Consider, in so far as possible, the multi-dimensional dynamics of
the brain as expressed in the whole recorded signals.
2.! Un-mix source (and artifact) contributions of individual source
areas using independent component analysis (ICA).
3.! Visualize trial-by-trial relationships of source component activities
to experimental variables (using 2-D ‘ERP-image’ plots).
4.! Model the event-related dynamics of the source components
(using time/frequency analysis).
5.! Localize the separated source areas using biophysical inverse
modeling.
6.! Compare similarities in source dynamics and locations across
subjects using cluster analysis.
7.! Model transient source network dynamics and the contexts in
which they appear.
Mining Event-Related Brain Dynamics
S. Makeig 2010
Mining Event-Related Brain
Dynamics!
1.! Consider, in so far as possible, the multi-dimensional dynamics of
the brain as expressed in the whole recorded signals.
2.! Un-mix source (and artifact) contributions of individual source
areas using independent component analysis (ICA).
3.! Visualize trial-by-trial relationships of source component activities
to experimental variables (using 2-D ‘ERP-image’ plots).
4.! Model the event-related dynamics of the source components
(using time/frequency analysis).
5.! Localize the separated source areas using biophysical inverse
modeling.
6.! Compare similarities in source dynamics and locations across
subjects using cluster analysis.
7.! Model transient source network dynamics and the contexts in
which they appear.
Mining Event-Related Brain Dynamics
S. Makeig 2010
99:;<=!/*#T(1&$!Q!F(-$'+($!
•! 9C6D!B!(4(#&P+(.-&(V!$1(T&+-.!1*N(+!
•! "R/!B!3#&(+P&+3-.!T*[(+(#T(!E1[-$(!.*TG3#8H!
•! 9C/!B!(4(#&P+(.-&(V!T*[(+(#T(!
99:;<=!F(-$'+($!Q!j3$'-.3e-%*#$!
•! (+130-8(EH!B!$*+&(V!&+3-.PS5P&+3-.!V5#-03T$!
•! (#4&*1*EH!B!9CD$!-#V!T*01*#(#&$!
•! k&*1*EH!B!(4(#&P+(.-&(V!$1(T&+-.!1*N(+!T[-#8($!
S. Makeig 2004
ERP-Image Plotting
1.! Display single trials as
color-coded horizontal
lines (e.g., red is +!V,
blue is -!V, green is 0).
2.! Sort all trials according
to some variable of
interest (here, subject
RT).
3.! Smooth vertically.
Jung et al., Human Brain Mapping, 2001.
dB
Time (ms)
Fre
qu
ency (
Hz)
10
ERSP
Makeig et al., PLOS ‘04
timef() tftopo()
EVENT LOCAL PHASE
Inter-trial Coherence (ITC) (“phase-locking factor”)
•! Significant consistency of local phase of a
physiological waveform across successive trials.
delay
EVENT LOCAL PHASE
EVENT LOCAL PHASE
frequency
PHASE LOCKING
AVERAGE ERP
SINGLE TRIALS µV
P = 0.02
P = 0.02
INTER-TRIAL COHERENCE
NO AMPLITUDE INCREASE
400 SIM. TRIALS ...
ERP-IMAGE PLOT
INTER-TRIAL COHERENCE
(phase resetting)
erpimage()
NO AMPLITUDE INCREASE
ITC / PHASE LOCKING
TIME
FR
EQ
UE
NC
Y
ERSP
ITC
dB
µV2
r timef()
Space of time / frequency changes …
ERS ERD Baseline
ITC
N
o IT
C
ERPs are produced by ITC>0, not by power increases
“True” ERP (visual
P1)
ITC
ERP
S. Enghoff
“True” PPR (visual ‘alpha
ringing’)
Event-Related Coherence (ERC)
•! Significant consistency of local phase difference
between two concurrent physiological waveforms.
delay
frequency
PHASE1 EVENT PHASE2
"
PHASE1 EVENT PHASE2 "
PHASE1 EVENT PHASE2 "
Event-related Coherence
TWO SIMULATED THETA PROCESSES
FR
EQ
UE
NC
Y
TIME
EVENT-RELATED COHERENCE
COHERENCE LAG
r
deg
ERC
crossf()
J Klopp, K Marinkovic, P Chauvel, V Nenov, E Halgren Hum Br Map
11:286-293 (2000)
Ant. Cing.
Fusiform
Post. Cing.
Rossen, Makeig, et al..!
I;;<]%PK#1('#-H%;;<%4'K,1*%30-$"#$!
N,-#2"*#9%;;<%C*9T',6%&'##*1()"9/!
Tim Mullen, S. Makeig et al. unpublished
43PN]%>%9''HD'W%I',%0'M*H"#$%
;;<%2'K,1*%"#I',0-('#%^'T%
1 ! Alertness
2 ! Attention
3 ! Arousal
4 ! Anticipation
5 ! Affect
6 ! Awareness
7 ! Agency
8 ! Aha!
What can EEG measure?
Uses for EEG?!
1 ! Alertness
2 ! Attention
3 ! Arousal
4 ! Anticipation
5 ! Affect
6 ! Awareness
7 ! Agency
8 ! Aha!
What can EEG measure?
EEG-based Cognitive-State Monitoring
DSP and Display module
2.5 x 1.5 in
Lin at. al., Proc. IEEE, July 2008.
Estimating Cognitive
states of the drivers
Sample Results
Attended Location
left center right
Fre
qu
necy (
Hz)
10
20
40
left center right
left center right left center right
10
20
40
10
20
40
10
20
40
Baseline Spectrum
(1 s before stimulus)
Right Alpha Left Alpha
Central Alpha
Frontocentral
Component
EEG and Attention
Westerfield & Makeig, 2001
Load 3
Load 5
Load 7
dB
Onton & Makeig, 2006
EEG and Attention
K. Gramann, J. Onton, & S. Makeig, 2008
Clusters distinguishing Turners & Nonturners
Julie Onton & Scott Makeig, Frontiers in Human Neuroscience, 2009
Changes in distribution of broadband high-frequency
EEG power with imagined emotion
+&3=>+]%>%9''HD'W%I',%DK"HM"#$%D,-"#_
1'0@K9*,%"#9*,I-1*%0'M*H2%
_*'+&[!"#&(+#-%*#-.!=/"!F((%#8!<$3.*0-+!F((%#8!:+*'#V$K!D-T3fT!:+*4(K!/<!
L'#(K!XY?Y!
cU!:+-0-##K!RPD!L'#8!W!6!U!F-G(38!!EXYY`H!
!'D"H*%+,-"#O+'M/%30-$"#$%N*1:#'H'$/%
•! /'++(#&!99:K!(5(!&+-TG3#8K!-#V!
0*%*#!T-1&'+(!$5$&(0$!-,*%
@',9-DH*U!!
•! "#!-!)(N!5(-+$K!!"#$#%&"'
03T+*(.(T&+*#3T%!'+3%2/29*02!N3..!
S(!)(-$3S.(U!
Q-,-M"$0%2:"`%
%>=;aN%JJ%!!!!! ! %%b;;<%"2%2'0*9:"#$%9'%K2*E%
%%% % % %#'9%cK29%2'0*9:"#$%9'%29KM/de%
AlanBauer.com
<.&[*'8[!&+-V3%*#-.!99:!-#-.5$3$!0(&[*V$![-4(!'$(V!
+(.-%4(.5!.3l.(!*)!&[(!-S'#V-#&!3#)*+0-%*#!-S*'&!
V3$&+3S'&(V!S+-3#!V5#-03T$!T*#&-3#(V!3#!99:!V-&-!a!
a!#(N!+(T*+V3#8!&(T[#*.*85K!-#V!#(N!-#-.5$3$!
0(&[*V$!S-$(V!3#!S+-3#!1[5$3*.*85!-..*N!03#3#8!
0'T[!0*+(!*)!&[3$!3#)*+0-%*#K!-#V!0-G(!1*$$3S.(!
&[(!-#-.5$3$!*)!#(N!(J1(+30(#&-.!1-+-V380$U!
C*T%I*-2"DH*%-@@H"1-('#2%'I%;;<%3#T.'V(m!
•!!)*+!99:!)((VS-TG!)*+!-l(#%*#!&+-3#3#8!
•!!)*+!>38[PV(f#3%*#!)99:!30-83#8!)*+!T.3#3T-.!+($(-+T[!
•!!)*+!nP[38[PV(f#3%*#!(.(T&+*T*+%T-.!$*'+T(!30-83#8!
•!!)*+!F*S3.(!S+-3#QS*V5!30-83#8!EF*="H!
New dimensions of EEG research and application