Under each condition, auditory evoked potentials were also collected for 15 minutes (the results of the latter were not reported in this study). The total acquisition time is usually half an hour per state. Including the positioning of 256 electrode hd-EEG cap and the recovery time of subjects, the whole experiment process is about 5 hours. Since it is impossible to use this experimental procedure as part of routine surgery, the experiment was conducted in healthy volunteers in the medical environment of the cyclotron research center. All data collection and sedation processes were continuously monitored by a certified anesthetist who was responsible for the safety of the subjects. Complete resuscitation equipment is always available.
During the study and recovery period, ECG, blood pressure, pulse oxygen saturation (SpO2), end tidal partial pressure of carbon dioxide (ETCO2) and respiratory rate (In vivo Research, Inc., magnitude, model 3150M) were continuously monitored. Propofol is injected into the right hand or forearm vein through an intravenous catheter. Throughout the study, the subjects spontaneously breathed and gave extra oxygen (5 liters/minute) through loose plastic masks. Computer controlled intravenous infusion of propofol (Alaris TIVA) to obtain a constant concentration of effector sites. A three compartment pharmacokinetic model was used to estimate the plasma and effector site concentrations of propofol. After reaching the appropriate concentration of the site of action, a 5-minute balance period is allowed to ensure the balance of propofol redistribution between chambers. Simple constant rate propofol infusion was used with a computerized Marsh model to predict when the infusion rate would be manually adjusted to maintain the desired steady state propofol level, although the primary goal was to reach the range of clinical conditions (Ramsay score). In each clinical state, arterial blood samples were also collected immediately before and after scanning for subsequent propofol concentration determination and blood gas analysis. The mean arterial blood concentration of propofol was 1.91 ± 0.52 mcg/mL in sedation group and 3.87 ± 1.39 mcg/mL in LOC group.
In order to compare with the anesthesia data, we selected amplitude matched spontaneous sleep slow waves from the slow wave population of six male subjects (aged 24-35 years) at the University of Wisconsin, and analyzed them as part of the previous study.
We recorded the hd-EEG of each subject for several minutes before receiving propofol, after propofol injection but before and after sedation. Sedation is accompanied by spindle (12 - 15hz) and β (15 - 25hz) Power increase. LOC is accompanied by an almost 20 fold increase in incremental power. Alpha (8-12) and θ There is also a small increase in power (4-8 Hz). These changes were reversed when consciousness returned. These power increases are most obvious in the front - inside electrode cluster.
We note that during sedation, the gamma (25-40 Hz) power is more than twice the baseline level, and this increase persists even during LOC. Although it has been proved that EEG power above 20 Hz can be adjusted by muscle or eye movements, several reasons indicate that these factors do not affect our results. First, we found that after sedation, the EMG power had a downward trend (P<0.08, paired t-test). Although this does not rise to the level of statistical significance, it is in the opposite direction. We expect to see whether the EMG activity drives the gamma EEG signal. In addition, previous studies reported that propofol anesthesia would reduce facial electromyography and inhibit eye movements. Secondly, we found that there was no correlation between the gamma power of Fz channel and the gamma power of EMG or EOG signal, and the EOG did not change after LOC (P>0.7, student t-test). In addition, the terrain where the gamma power increases under sedation is different from that of eye or muscle derived artifacts, which may be more lateral. Finally, unlike the scanning related gamma activity, the gamma activity observed during LOC is not composed of transient bursts, but is increasing.
