Power changes and cognitive performance

1
Part DThe functional meaning of theta ERS and
evoked theta and alpha oscillations
2
Studies focusing on event-related changes in
theta and alpha band power generally show that
theta synchronizes whereas alpha desynchronizes
Whole power (Gabor wavelet power estimates),
Sternberg task, load 4 P4
3
Frequency specificity and functional meaning of
theta for working memory (WM). Data from Klimesch
et al (2001). Episodic retrieval is reflected by
a process specific increase in human theta
activity. Neuroscience Letters, 302, 49-52.
Picture encoding (hits, dotted) and recognition
(hits, bold correct rejections dashed ). O1 IAF
10.3
A) Theta 4.3-6.3Hz
B) Lower-1 alpha 6.3-8.3 Hz
Theta old/new - effect
ERD ERS 50 0
50 100
ERD ERS 50 0
50 100
-1 µV
-1 µV
Evoked theta recogn. hits
Evoked lower-1 alpha recogn. hits
-1000 -500 0 500
1000 ms
-1000 -500 0 500
1000 ms
D) Upper alpha 10.3-12.3 Hz
C) Lower-2 alpha 8.3-10.3 Hz
-1 µV
-1 µV
Evoked lower-2 alpha recogn. hits
Evoked upper alpha recogn. hits
ERD ERS 50 0
50 100
ERD ERS 50 0
50 100
-1000 -500 0 500
1000 ms
-1000 -500 0 500
1000 ms
4
MEG Study Jensen Tesche (2002) Sternberg task
with visually presented digits
Fig. 2 from Jensen, O., Tesche, C. (2002).
Frontal theta activity in humans increases with
memory load in a working memory task. European
Journal of Neuroscience. 15 (8), 1395-1399.
5
Are early ERP components generated by evoked
theta and alpha oscillations?Do they reflect a
co-activation of theta and alpha?This question
is motivated by the previously described
findings if theta is related to working memory
(WM) and upper alpha to long-term memory (LTM)
one would expect some sort of co-activiation
between theta and upper alpha because during most
cognitive processes, WM and LTM will interact.
6
During an early time window post-stimulus alpha
is still synchronized whereas theta starts to
synchronize. Both frequencies show pronounced
phase locking
Sternberg task, load 4 O2 Whole power (Gabor
wavelet power estimates) Phase
locking index (PLI)
PLI
7
The basic properties of the P1-N1 complex can be
described by a superposition of an evoked theta
and alpha wave
8
Two models of ERP generation
Only the phase reset model is able to predict
that intertrial phase variability and amplitude
decreases
9
Picture recognition task (Klimesch et al. 2004),
Extent of phase locking (phase resetting) as
measured by the phase locking index (PLI) cf.
Schack Klimesch (2002)
PLI
Fz Hits
Fz Corr. Rejection

Gemittelter PLI
, Ch
4 (Fz)
, Mk
140,
Int.
100 von 102

0
10
20
Hz
30
40
-200 0 200 400 600 ms
-200 0 200 400 600 ms
0.2
0.4
0.6
0.8
1
sec
O2 Hits
O2 Corr. Rejection
Gemittelter PLI
, Ch
26 (O2)
, Mk
139,
Int.
36 von 37

Gemittelter PLI
, Ch
26 (O2)
, Mk
140,
Int.
100 von 102

0
0
10
10
20
20
Hz
Hz
30
30
40
40
0.2
0.4
0.6
0.8
1
-200 0 200 400 600 ms
-200 0 200 400 600 ms
0.2
0.4
0.6
0.8
1
sec
10
Dissociation between increase in phase locking
and decrease in power
Fz
O1
PLI
Whole
Evoked
Whole
Evoked
PLI
Theta
Upper Alpha
P1 Time window
N1 Time window
Reference
11
Good memory performers (M) show a significantly
larger phase locking in the N1 time window as
compared to bad performers (M-)
12
Correlations between individual alpha frequency
(IAF) and Peak Latency (L) and between Peak
Latency and Memory Performance (hits) during
Encoding (E) and Recognition (R) Data from
Klimesch et al (2004)
13
Instantaneous Phase Alignment (IPA)between
frequencies predicts P1 and N1 latency and
amplitude
14
Picture encoding and retrieval task. Data and
methods from Gruber,W., Klimesch, W., Sauseng,
P. Doppelmayr M. (2004). Alpha phase
synchronization predicts P1 peak latency and
amplitude size. Cerebral Cortex, in press.
Recording site O1, example for one subject
a
b
Event-related potential
Evoked power
max
N1
-20
-10
0
Gabor estimates
Voltage µV
Frequency Hz
10
20
P1
min
Time ms
Time ms
c
Significant PLI (a 10 )
d
Whole power
max
max
Sign. PLI
Gabor estimates
Frequency Hz
Frequency Hz
min
n.s.
min
Time ms
Time ms
15
Phase alignment between frequencies, example for
one subject (data from Gruber et al. 2004)
e
Phase angle pos. peak, neg. peak
Frequency Hz
16
Phase angle of theta (6 Hz) and alpha (13 Hz) are
aligned at the N1 component recording site O1,
example for one subject
b
17
h
Phase alignment and ERP over all subjects
18
Prediction of P1, N1 latency and amplitude on
basis of significant phase alignment
For the selected frequency bins at each time
point the difference between the measured angle
and a positive (0/360 ) and a negative peak
(180) was calculated. These distances were
averaged over frequencies for each time point.
(If the measured angle was between 90 and 270
the distance to 180 was calculated. If the
measured angle was between 270 and 90 the
distance to 0/360 was calculated).
Two final values were selected, one closest to
0/ 360 (reflecting a positive peak) and another
closest to 180 (reflecting a negative peak).
The latencies of these two values were selected
within the broad time interval of 60 180 ms
post-stimulus but blind with respect to the
specific latency windows of the P1 and N1 used
for peak detection in ERP analyses.
O1 O2 P3
P4 T5
T6
P1 N1 P1 N1 P1
N1 P1 N1 P1 N1
P1 N1
19
Sites O1 and O2 Percentage of frequency
binswithin traditional frequency bands at the P1
and N1Contribution of Alpha is particularly
large at P1, Contribution of Theta is larger at
N1

20
Dissociation between two event-related measures,
PLI and Power is consistent with the phase reset
but inconsistent with the evoked model
a
5-6 Hz
3.5 1.5
0.6 0.1
Theta
Power
PLI
Prestimulus Poststimulus
b
11-12 Hz
3.5 1.5
0.6 0.1
Upper Alpha
Power
PLI
Prestimulus Poststimulus
21
Conclusions
The P1 and N1 component is generated by a
significant alignment in absolute phase of
frequencies in the alpha, theta and slow beta
range (8-16 Hz)
Significant phase alignment predicts P1 and N1
latency
Number of alpha frequency bins predict are
related to P1 and N1 amplitude size
Findings provide strong support for oscillatory
phase reset model
22
A)
0,4
B)
ERS
triggered by latency of reversal
triggered by stimulus
0
Band power changes in z-values
ERD
-0,4
23
Synchronization of a network by an oscillation
Synchronized, rhythmic activation
Not synchronized activation of a network
24
Prestimulus Stimulus Poststimuus
Theta
Upper Alpha