Electroencephalogram signatures of loss and recovery of consciousness from propofol

Abstract

Unconsciousness is a fundamental component of general anesthesia (GA), but anesthesiologists have no reliable ways to be certain that a patient is unconscious. To develop EEG signatures that track loss and recovery of consciousness under GA, we recorded high-density EEGs in humans during gradual induction of and emergence from unconsciousness with propofol. The subjects executed an auditory task at 4-s intervals consisting of interleaved verbal and click stimuli to identify loss and recovery of consciousness. During induction, subjects lost responsiveness to the less salient clicks before losing responsiveness to the more salient verbal stimuli; during emergence they recovered responsiveness to the verbal stimuli before recovering responsiveness to the clicks. The median frequency and bandwidth of the frontal EEG power tracked the probability of response to the verbal stimuli during the transitions in consciousness. Loss of consciousness was marked simultaneously by an increase in low-frequency EEG power (<1 Hz), the loss of spatially coherent occipital alpha oscillations (8-12 Hz), and the appearance of spatially coherent frontal alpha oscillations. These dynamics reversed with recovery of consciousness. The low-frequency phase modulated alpha amplitude in two distinct patterns. During profound unconsciousness, alpha amplitudes were maximal at low-frequency peaks, whereas during the transition into and out of unconsciousness, alpha amplitudes were maximal at low-frequency nadirs. This latter phase-amplitude relationship predicted recovery of consciousness. Our results provide insights into the mechanisms of propofol-induced unconsciousness, establish EEG signatures of this brain state that track transitions in consciousness precisely, and suggest strategies for monitoring the brain activity of patients receiving GA.

Fig. 1.

Dynamics of the behavioral responses to the verbal and click stimuli during induction and emergence from propofol-induced unconsciousness. (A) The time course of the target effect-site concentrations of propofol during induction and emergence for subject 7. (B) Time courses of the response-probability curves for the click (blue, P_clicks) and verbal (red, P_verbal) stimuli and their respective 95% credibility intervals (shaded areas) for subject 7. The vertical black lines, from left to right, show E_ON, LOC, ROC, and E_OFF. (C) The curve of the difference between the verbal and click response-probability curves (P_verbal − P_clicks) and its associated 95% credibility interval for subject 7. (D) Group-level click (blue, P_clicks) and verbal (red, P_verbal) response-probability curves computed by aligning the individual subjects' response data during induction with respect to LOC and during emergence with respect to ROC. (E) Group-level curve for the difference between the verbal and click response-probability curves (P_verbal−P_clicks) and its 95% credibility intervals computed from the group-level curves in D. These results show that during gradual induction and emergence with propofol, loss and recovery of consciousness occur not instantaneously but gradually, and the probability of response depends critically on stimulus saliency.

Fig. 2.

Dynamics of the EEG spectrogram during induction and emergence from propofol-induced unconsciousness. (A) Group-level click (blue, P_clicks) and verbal (red, P_verbal) response-probability curves, as shown previously in Fig. 1D. (B) Group-level baseline-normalized spectrograms from a frontal channel (approximately Fz, nearest-neighbor Laplacian reference) aligned with respect to LOC and ROC. The white contour circumscribes the regions where power differs significantly from baseline (P < 0.05, sign test) and indicates significant increases in power spanning low-frequency (0.1–1Hz) through gamma (25–35 Hz) bands. (C) Group-level time course of power in low-frequency, alpha (8–12 Hz), and gamma bands aligned with respect to LOC and ROC. (D) Group-level spatial distribution of low-frequency, alpha, and gamma power during unconsciousness (LOC + 15 min). These analyses show that changes in broad-band gamma/beta power coincide with the behavioral changes before LOC and after ROC, whereas changes in slow and alpha power coincide with LOC and ROC.

Fig. 3.

Time course of the traveling peak, the continuous transformation in median frequency and bandwidth spanning the gamma, beta, and alpha bands during the transitions into and out of unconsciousness. (A) Group-level spectrograms computed between 2 and 40 Hz for a single frontal channel (approximately Fz, nearest-neighbor Laplacian reference), aligned with respect to LOC (Left) and ROC (Right) and normalized by the baseline spectrum. The 25th, median, and 75th percentiles within this frequency range are overlaid in white. The median represents the center frequency of the traveling peak, while the interquartile range (i.e., the difference between the 75th and 25th percentiles) represents the bandwidth of the traveling peak. (B) Spatial distribution of power at the median frequency at different behavioral end points. Pre-LOC is the midpoint between E_ON and LOC. Unconscious refers to the midpoint between LOC and ROC. Post-LOC is the midpoint between ROC and E_OFF.

Fig. 4.

Spatially coherent alpha oscillations at LOC and ROC. (A) At baseline, spatially coherent alpha oscillations are concentrated in occipital channels, represented by the principal mode of the cross spectral matrix. (B) Activity within the principal mode, represented by the modal projection (the proportion of total power captured by the principal mode) dissipates at LOC (Left) and returns at ROC (Right). (C) In the unconscious state, spatially coherent alpha oscillations are concentrated in frontal channels. (D) Activity within this frontal alpha mode, characterized by the modal projection, begins after LOC and ceases at ROC. Statistical significance for the modal projection was assessed using a permutation procedure (Materials and Methods); areas shaded in gray were not significant (P > 0.05). These analyses reveal that there is a change in coherent alpha oscillations at LOC, when a spatially coherent occipital alpha mode shuts off and a spatially coherent frontal alpha mode engages. These changes reverse at ROC.

Fig. 5.

Two distinct patterns of phase–amplitude modulation that develop asynchronously with LOC and ROC. (A) Group-level click (blue, P_clicks) and verbal (red, P_verbal) response-probability curves as shown in Fig. 1D. (B) The phase–amplitude histogram showing the relationship between the low-frequency (0.1–1 Hz) phase (y-axis, shown with reference sinusoid) and mean-normalized alpha/beta (8–14Hz) oscillation amplitude (color map) as a function of time (x-axis) relative to LOC (Left) and ROC (Right). The trough-max pattern, in which the alpha oscillation amplitude is maximal at the low-frequency troughs, occurs during transitions into and out unconsciousness. The peak-max pattern, in which the alpha oscillation amplitude is maximal at the low-frequency peaks, occurs during periods of profound unconsciousness. (C) The trough-max pattern observed in the time-domain EEG trace of an individual subject. (D) The peak-max pattern observed in the time-domain EEG trace of an individual subject. Modulograms and time-domain traces are from a frontal channel (approximately Fz, nearest-neighbor Laplacian reference). The peak-max modulation identifies a profound state of unconsciousness, whereas the trough-max pattern appears during transitions into and out of unconsciousness and therefore can be used to predict ROC.

Fig. 6.

Summary of behavioral and EEG signatures during induction of and emergence from propofol-induced unconsciousness. (A) Responses to auditory stimuli show continuous changes in probability of response and a salience dependence during the transition to unconsciousness and during ROC. (B) The frontal EEG spectrogram shows a traveling peak that begins as broad-band beta-gamma power at the onset of behavioral effects and decreases in frequency and bandwidth into the alpha range toward LOC. (C) Spatially coherent posterior alpha oscillations disappear and spatially coherent frontal alpha oscillations appear at LOC. At ROC, spatially coherent posterior alpha oscillations reappear, and spatially coherent frontal alpha oscillations disappear. (D) Low frequency (<1 Hz) power increases at LOC and decreases at ROC. (E) Two patterns of low-frequency phase modulation of alpha/beta amplitude. The trough-max pattern appears at the transition into and out of LOC. The peak-max pattern appears at profound unconsciousness. These results establish EEG signatures that characterize unconsciousness, track the transitions into and out of unconsciousness, and provide a means to monitor and predict the brain states of patients receiving propofol for GA or sedation.