Upload
duonghanh
View
225
Download
5
Embed Size (px)
Citation preview
1
PEAK-TO-PEAK AMPLITUDE OF EEG FOR THE MONITORING OF BRAIN FUNCTION
Michael NavakatikyanBrainZ Instruments, Auckland,
New Zealand
Aims
To introduce:
• A new measure of peak-to-peak amplitude (ppA)
And to describe:
• Its differences and similarities with another ppA measure, namely, aEEG.
peak
trough
ppA
2
aEEG: A Representation of ppA
Example of cardiac arrest
Designed for continuous long-term monitoring of ppA in unconscious adults
Maynard, Prior & Scott (1969) BMJ. 4: 545-6
aEEG: Application for Neonates
0
10
100
µV
Quiet Sleep
Awake/Active Sleep
• Found application in neonatal EEG monitoring, producing distinct patterns for different:
a) background statesb) seizures andc) sleep-wake cycling
• Proved to be sensitive to high frequency artefact
3
But does aEEG measure ppA in neonates accurately?
aEEG: ppA Distortion that Depends on Frequency
aEEG
0
10 Hz
100 µV
1 s
57 µV
5 Hz 1 Hz
8 µV
Raw EEG
100µV
4
How aEEG depends on frequency of cEEG?
• Reduces the values of ppA greatly around α-rhythm• α-rhythm is not present in neonates, its average frequency is much lower• Thus values of aEEG are about 3 to 6 times lower than ppA of raw EEG signal, but we never know how much
FREQUENCY, HZ
0 10 20 30
aEEG, % of Input ppA of sine wave
0
50
100 Effectively – a detector of α-rhythm, which is not present in neonates
The aim: to create a method for measuringppA without frequency distortion
New method transforms raw EEG into another signal, called rEEG
rEEG
0
10 Hz
100 µV 100 µV
5 Hz 1 Hz
100 µV
Raw EEG
100µV
5
What is the main idea behind our method of
ppA calculation?
rEEG: The Idea
There are 3 waves in a selected interval. Values of ppA range from 34 to 45 µV
45µV42µV 34µV
max
min
47µV
If we subtract min from max on this interval we will get a ppA-estimate of 47 µV
It is higher than average ppA on interval but close enough
6
rEEG: Calculation
New method transforms cEEG into another signal, called rEEG, from word “range”.(Note: Range in statistics is difference between max and min)
EEG
rEEG
0
2-sec
max
min
max
min
min
max
rEEG = Max – Min on time interval
0
rEEG: Calculation Example – Upper Estimate of ppA
ppAµV
0
30
0 2 4 6 8 10 12 14
time, sManualrEEG
EEGµV
-15
15
ppAµV
0
30
0 2 4 6 8 10 12 14
time, sManualrEEG
EEGµV
-15
15
ppAµV
0
30
0 2 4 6 8 10 12 14
time, sManualrEEG
ppAµV
0
30
0 2 4 6 8 10 12 14
time, sManualrEEG
EEGµV
-15
15
ppAµV
0
30
0 2 4 6 8 10 12 14
time, sManualrEEG
EEGµV
-15
15
ppAµV
0
30
0 2 4 6 8 10 12 14
time, sManualrEEG
EEGµVEEGµV
-15
15
7
rEEG – Examples of recording
rEEG is also (as aEEG) a compressed representation, it has a semi-logarithmic scale with cut off value of 50 µV
34 weeks
25 weeks
0
50
250
25
500
0
50
250
25
500
1:00:00 2:00:00 3:00:00 4:00:000
Cut off50 µV
rEEG : we can discern conventional EEG amplitude from it rather accurately
LM = 5-15 µV?
UM = 100-200 µV?rEEG, 3 hr
µV
0
50
250
25
500
100
10
8
μV
0
100
-100
Raw EEG
rEEG0
50
25
500
100
10
On 5-min stretch: Yes – ppA = 10 to 250 µV But does rEEG produce distinct patterns we used to see in neonates?
9
0
50
250
25
0
10
50μV aEEG
rEEG
rEEG - aEEG: Sleep-wake Cycles Similarity
The reason for similarities in aEEG and rEEG pattern is that they both represent ppA, albeit
with a different degree of accuracy
DNV LV FTCNV BS
0
10
50μV 0
50
250
25rEEG
aEEG
rEEG – aEEG: Background Patterns Similarity
Thresholds of 25 and 10 µV are suggested as roughly equivalent to 10 and 5 µV thresholds for aEEG
Classification by Hellstrőm-Westas, de Vries and Greisen (2006) NeoReviews, 7: c76-c87
10
rEEG - aEEG: Seizures Similarity
Seizures look like the narrowing amplitude band with elevated lower margin, because ppA becomes more
uniform during seizure activity
Seizures
rEEG - aEEG: High frequency artefact mimicking seizures is less pronounced
EMG artefact
11
Conclusions
• An accurate measure of ppA (peak-to-peak amplitude)
• Has similar to aEEG patterns
• Can bridge the gap between cEEG-based and aEEG-based assessment of brain function
• For example, it can be used to assess discontinuity and IBI (inter-burst interval) on raw EEG
rEEG:
Any questions, please?