right side of the brain were combined because there was no statistically significant difference evident between the two brain hemispheres.

Table 2
Power Spectrum Variables (Adapt Tech Overview 1992)

table2.gif (11290 bytes)


Data collected from the two fronto-temporal channels were used in the statistical analysis.  Means and standard deviations for all processed EEG parameters are shown in Table 3. BIS, SEF, MPF, Absolute alpha, Absolute beta, and EMG data were significantly different between the responders and non-responders.  Linear regression analysis demonstrates that BIS correlates with RSS (Figure 1). Ramsay Sedation Scores between 0 and 3 indicate deeply sedated patients ("non-responders") and correspond to lower BIS scores (55 to 70). Similarly, less sedated patients ("responders") have higher RSS (4 to 6) and higher BIS (70 to 100) scores.

Table 3
Means and standard deviations for all EEG parameters.

table3.gif (7134 bytes)

The correlation of other EEG parameters with RSS was also computed (Table 4).  Of all parameters, BIS correlated the best.

Table 4
Correlation of EEG parameters with RSS.

table4.gif (1130 bytes)



The primary conclusion of this experiment is that BIS correlates well with RSS.  In the past, the only method of assessing sedation was the RSS system.  RSS is subjective because different observers may give the same patient different scores.  Furthermore, RSS is not precise in determining sedation because it is difficult to differentiate if a responder has a score of 4, 5, or 6.  Likewise, it is difficult to distinguish if a non-responder is a 0, 1, 2, or 3.  Therefore, RSS can differentiate a responder from a non-responder, but it fails to provide a precise measurement of the patient's level of sedation.

figure1.gif (5132 bytes)
Figure 1
Linear Regression analysis of BIS with RSS.  This graph shows the range of BIS at each RSS number. Central point represents the mean BIS score at each RSS number.

On the other hand, BIS is more precise than RSS because BIS can distinguish a deeply sedated non-responder from a lightly sedated non-responder.  Such patients would have different BIS scores, but their RSS may be the same.  Using only RSS, one may over-sedate the patient, giving more sedative than is necessary to make the patient comfortable.  However, for a patient who is being continuously monitored by an EEG machine, the drug can be titrated to a dose that maintains the patient's BIS score between 60 and 80.  At this point, the patient will be slightly sedated and comfortable.   A BIS score of less than 60 is probably an indication of over-sedation, since this score correlates with RSS numbers of 3 or less (Figure 1).

back.gif (221 bytes)


Page 3


Susan Khera - Bispectral Analysis of EEG Waves Provides... [1] [2] [3] [4]