1 1 2 Oscillatory phase synchronisation: A brain mechanism of memory matching and attention Paul Sauseng 1,2 , Wolfgang Klimesch 1 , Walter R. Gruber 1 , Niels Birbaumer Conjunct COST B27 and SAN Scientific Meeting, Swansea, UK, 16-18 September 2006
Feb 03, 2016
1
1
2
Oscillatory phase synchronisation: A brain mechanism of memory matching and attention
Paul Sauseng1,2, Wolfgang Klimesch1, Walter R. Gruber1, Niels Birbaumer2
Conjunct COST B27 and SAN Scientific Meeting, Swansea, UK, 16-18 September 2006
2
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
Top Down
Bottom Up
Matching
3
How is matching between top-down and bottom-up information implemented in the human brain?
Herrmann et al. (2004a)
Busch et al. (2004)
Herrmann et al. (2004b) Herrmann et al. (2004a)
MUM
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
4
Theta oscillations (around 5 Hz), fronto-parietal network and top-down activation
Sarnthein et al. (1998) Sauseng et al. (2005)
Sauseng et al. (2006)
von Stein & Sarnthein (2000): Integrative brain processes like top-down activation are reflected by long-range theta and/or alpha activity
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
5
Herrmann et al. (2004a)
MUM
localdistributed
Interaction
between
GAMMA
and
THETA
?
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
6
Experimental Paradigm
+
+
+
Time [ms]
2500 ms
34 ms
600-800 ms
50 ms
valid trialinvalid trialIntroduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
7
EEG data acquisition and analysis
22 healthy subjects32-channel EEG at 250 Hz
Time-frequency analysis
Beamformer source analysis of 35-45 Hz activity
Local sources (dipole clusters) into maxima of gamma sources
Raw EEG at scalp level transformed into ongoing source activity using the dipolar model derived from gamma activity
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
8
Claims and measures
Memory matching is reflected by gamma:theta interaction and amplified by attention=> cross-frequency phase synchronisation index (cfPSI)
consistency of phase differences between different frequencies over trials
Is gamma:theta phase synchronisation phase locked to stimulus onset? And does evoked gamma activity reflect memory matching (MUM)?
=> phase locking index (PLI)consistency of phase angle of one frequency in respect to stimulus onset over trials
Is top-down activation reflected by a fronto-parietal theta network?=> phase coherence
consistency of phase differences between electrode sites over trials at one frequency
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
9
Local Gamma sources (BESA)
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
10
Cross-frequency Phase Synchronisation Index (40 Hz : n Hz)
+
+
+
+
+
+
cfPSI (target right – target left)
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
11
Cross-frequency Phase Synchronisation Index (40 Hz : 4-8 Hz)
4-way interaction F4,84 = 3.35, p<.05 no sign. effects in other frequency bands
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
12
Phase Locking Index
Valid conditionInvalid condition
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
13
Phase Coherence (on the scalp level)
a b
c d
Theta
Gamma
+
+
+
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
14
Conclusion: a cascade of brain oscillatory mechanisms of memory matching
Herrmann et al. (2004a)
MUM
AT
TE
NT
ION
evoked gamma during first 100 ms poststimulus
gamma : theta phase synchronization100 – 200 ms
evoked theta between 100 and 300 ms
Fronto-parietal theta network pre- and poststimulus
Introduction
Methods- Exp. Pardigm-Data Processing
Results- Gamma Sources- cfPSI- PLI- Phase Coh.
Conclusions
15
This research was supported by the Deutsche Forschungsgemeinschaft DFG (Project BI 195/51-1)
16
t
t
t
t
µV
µV
0 pi
2 pi
0 pi
2 pi
Cross-frequency Phase Synchronisation Index (cfPSI)
17
Cross-frequency Phase Synchronisation Index (cfPSI)
t0 pi
2 pi
pi
-pi
18
Cross-frequency Phase Synchronisation Index (cfPSI)
pi
-pi
pi
-pi
pi
-pi
pi
-pi
pi
-pi
19
difference
Cross-frequency Phase Synchronisation Index (cfPSI)
20
Phase Locking Index (PLI)
stimulus onset
t
21
Phase Coherence
Phase locking value (PLV) Lachaux et al. (1999)
Consitency of phase difference between 2 electrodes / sensors over trials
cfPSI, PLI, Phase Coherence:
0 = no phase synchronisation / locking
1 = absolute phase synchronisation / locking