In Reply:—
We thank Dr. Jäntti et al. for their interest in our study on the differential dynamic of action on cortical and subcortical structures of anesthetic agents during induction of anesthesia.1The location of the signal recorded through deep-brain electrodes is clearly a concern for all stereoelectroencephalography measurements. In epilepsy surgery, neurophysiologists and neurosurgeons rely on deep-brain recordings to precisely delimit the brain region to remove surgically. If such a gross contamination due to volume conduction were to occur, the clinical results would certainly be poor. We agree that one problematic issue in combined scalp cortical (electroencephalographic) and subcortical (electrosubcorticographic) electrogenesis recording is that of volume conduction from cortical to subcortical regions or the opposite. But first, we observed different data from each site showing that we observed synaptic activity from different regions of the brain (as shown in fig. 2 in the article). Second, we used a bipolar recording for the scalp and the deep-brain electrode. Wennberg et al. 2demonstrated that scalp potential recording in the subthalamic nucleus with monopolar montage totally disappears with a bipolar montage.
We do not believe that the illustration provided by Dr. Jäntti et al. supports their claim that we recorded only cortical activity both on the scalp electrodes and from the deep-brain electrode. First, the authors suggested that plot 4 of the deep-brain electrode was “close to the vertex,” which is not correct. As shown in figure 1 in the article, this plot is far from the cortex and close to the thalamus. Second, it shows a burst constituted of one slow wave superimposed with rapid activities of various frequencies. These rapid activities are more complex than classic spindles. In addition, the illustration shows, at the beginning of the burst, high-frequency activity in the cortex but not on the depth electrode. The interpretation made to explain this discrepancy is purely speculative. In fact, it demonstrates that cortical and subcortical electrogenesis were different.
We were also surprised to read, in their letter and in the published material from which the patient data came,3that spindle generation occurred in the cortex, contrary to a large body of data showing the crucial role of the thalamus.4This is not new, because Morrison and Bassett5showed in 1945 that spindles survived in the thalamus after bilateral decortication. We also published in another article that spindles appeared in the thalamus during physiologic sleep before they appeared in the temporal cortex in epileptic patients undergoing stereoelectroencephalography.6In this study, the same surgically implanted electrode was used to record, using bipolar montage, both cortical and thalamic activity. This clearly shows that we recorded deep-brain electrophysiology and not only the electroencephalogram through the implanted electrode.
We found the topographic interpretation of Dr. Jäntti et al. confusing. In their article, they state that the phase reversal of the initial component of the burst in figure 2 of this study (Cz–D1/D1–C7, where D1 is the depth electrode) is an argument for a thalamic origin. That is correct according to the classic rule of electroencephalographic bipolar interpretation,7but they do not use the same rule for the latter component of the electroencephalographic data (δ activity and “spindle” in fig. 1), and they conclude that there is a cortical origin of this component. We believe this interpretation is incorrect and question their comments regarding topography in our data.
Dr. Jäntti et al. also suggest that cortical and subcortical activities are similar at a deep anesthetic state in our study (fig. 2 of the article), demonstrating that we recorded the same signal in both settings. As we write in the article, the activities are similar during deep anesthesia (T4). What were important in our study were the differences in electrophysiology between the electroencephalogram and the electrosubcorticogram we observed during induction of anesthesia, which were obvious from our figures and data, demonstrating that we recorded different activities. We discarded periods with a burst suppression pattern, which in our opinion cannot be interpreted, because electrophysiology tends to be uniform throughout the brain at a very deep anesthetic state (ultimately identical when the electroencephalogram is flat).
Finally, the scalp derivation we used did not minimize slow wave activity. It did not maximize the amplitude of slow waves, but it did not change the dynamics of slow wave appearance. This is clear in our article (figs. 2 and 4) showing the early appearance of δ waves.
We agree with Dr. Jäntti et al. that a thorough understanding of basic electrophysiology is mandatory to interpret the electroencephalogram during anesthesia.8We disagree, however, with their interpretation and believe that routine electrophysiologic recordings with depth electrodes, work in our laboratory,9and evidence from the literature support the validity of our data.
*Centre Hospitalier Universitaire Timone, Université de la Méditerranée, Marseille, France. lionel.velly@ap-hm.fr