To the Editor:—
Plourde et al. 1recently reported the effects of propofol on the brain’s response to different forms of auditory stimuli during sedation and anesthesia. This holistic approach is the only rational option to understand such phenomena as learning during anesthesia and perioperative dreaming.
One of the most interesting findings from this study is the extreme diminution of cortical response to auditory stimuli during the first episode of “sedation” before anesthesia. The cortical response to auditory stimuli during the period of “sedation” before anesthesia was much less than at baseline or at an equivalent blood level of sedation during recovery. The authors offer the explanation of acute tolerance to propofol in the interpretation of these results, which I would suggest is unnecessary.
We too have noted an increase in brain activation during propofol sedation, most evident in the statistical maps used to define brain activity in neuroimaging studies. During propofol or thiopental sedation, a very visible increase in statistical activation to auditory stimulation is present, although in actuality a small decrease in regional cerebral blood flow occurs.2Plourde et al. also document a similar discrepancy between blood oxygenation level–dependent signal magnitudes, which remain similar in many regions to baseline, and the increase in statistical significance, from roughly t = 14 at baseline to 19 in recovery, clearly visible in their brain map figures. There is no doubt that the subjects in the study of Plourde et al. were awake and sedated during recovery, because they remembered some 60% of the stimuli presented to them. Therefore, the phenomenon of increased statistical significance seems to be a marker of the effect of a sedative drug on auditory stimulation in an awake cortex. The new, paradoxical finding in the study of Plourde et al. , however, is that this situation is absent in the initial “sedation” condition, where both blood oxygenation level–dependent signal and significance are similar to the anesthesia condition. The only behavioral measure prospectively collected in this study is memory performance, and it too is noticeably absent in the initial “sedation” condition, being the same as during anesthesia. However, at recovery, the memory performance is at the expected 60% of control. As noted by Plourde et al. , these sedative concentrations of propofol should produce performance of approximately 50% of baseline/control, as we have documented in our studies.3
The most parsimonious explanation of the paradoxical findings during the initial “sedation” condition is that the subjects were indeed asleep, whereas they were awake and truly sedated in the recovery condition. The contention that sedation was equivalent at the beginning and at recovery is based solely on measured concentrations of propofol, not on behavioral state. Clear interpretation of the imaging findings is difficult because volunteers were performing no active behavioral task (e.g. , button press to hearing a word) to measure their degree of sedation and responsiveness while inside the scanner. Although the authors state that “it is difficult to fall asleep in the cramped and noisy scanner environment,” our experience has been that, even without drug, it is actually easy to fall asleep in a 3-T magnetic resonance scanner in the absence of an active task. Other investigators at our institution have had similar experiences; being confined in a constant position with the arms tucked in under warm blankets, particularly if the eyes are closed, is a situation similar to swaddling a baby to induce sleep. The “close monitoring” of the subjects is not defined, and an “impression” of whether the subject is awake or asleep is difficult to assess while in the scanner, particularly in retrospect. Any subjects sleeping after initial sedation would wake up when moved out of the scanner in preparation for induction of anesthesia.
On another point, some caution must be exercised regarding which auditory processes are affected most by propofol. The authors state that higher level processing for speech and voice is abolished during anesthesia, whereas the response to nonspecific auditory stimuli is not. However, the cortical response to nonspecific auditory stimuli was determined using all measured data versus a rest condition, whereas response to speech and voice were measured using only a subset of the data collected, with the comparison being between two active conditions. The effect of this change in data handling is evident in baseline. Whereas maximal t values for the nonspecific auditory stimulation are on the order of 14, those of the speech- and voice-related activation are only approximately 3.5 (the significance threshold is 3.2). Congruently, blood oxygenation level–dependent signal magnitudes at baseline for nonspecific auditory stimuli are approximately 30, whereas for speech and voice they are approximately 5. Therefore, the presence of decreased but still present brain activity in response to voice or speech during anesthesia would not likely be detected in this study.
Plourde et al. have presented a more refined evaluation of auditory processing during sedation and anesthesia than has heretofore been published. The interesting questions raised by these findings clearly elucidate paths for further study. However, the conclusions drawn from the current study are limited by lack of behavioral data and by issues of statistical significance. I would suggest that the evidence for acute tolerance to propofol is confounded by sleeping subjects. The issue of inhibition of higher order but not general auditory processes by propofol is still an open one. It seems likely that decoding of auditory stimuli during anesthesia is still functioning at some level, because numerous studies have demonstrated at least an implicit form of memory for these stimuli, albeit in much larger sample sizes (e.g. , Deeprose et al. 4).
Memorial Sloan-Kettering Cancer Center and Weill Medical College of Cornell University, New York, New York. firstname.lastname@example.org