Thalamocortical model for a propofol-induced alpha-rhythm associated with loss of consciousness

Abstract

Recent data reveal that the general anesthetic propofol gives rise to a frontal α-rhythm at dose levels sufficient to induce loss of consciousness. In this work, a computational model is developed that suggests the network mechanisms responsible for such a rhythm. It is shown that propofol can alter the dynamics in thalamocortical loops, leading to persistent and synchronous α-activity. The synchrony that forms in the cortex by virtue of the involvement of the thalamus may impede responsiveness to external stimuli, thus providing a correlate for the unconscious state.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Results from propofol infusion in a single representative subject. (A) Time course of propofol infusion and behavioral response. The green traces indicate the latency of correct responses to auditory stimuli. Cessation of response indicates anesthetic-induced loss of consciousness. The red trace describes the estimated propofol blood concentration. (B) Whole-study spectrogram from a frontal electrode. Loss of consciousness coincides with the emergence of broadband β-activity, which strengthens and slows into the α-range as the infusion increases. (C) Coherent EEG activity. At the deepest levels of general anesthesia the α-band exhibits high global coherence (2); being frontal, it is distinct from the classic occipital α in anatomical location.

Fig. 2.

Schema of model network and mechanism. (A) Network consists of separate populations of pyramidal cells (denoted PY) and INs (here, both FS and LTS are lumped into a single icon) that interact with a mutual population of RE and TC cells. (B) Propofol perturbs the network by potentiating the GABAA synaptic current and decay time from INs onto PY cells and from RE neurons onto TC cells. Note that all model cells are single compartments.

Fig. 3.

Mean simulated EEG spectrogram and coherence (n = 10) as a function of GABA A synaptic strength. (A) At baseline, the EEG reflects γ-activity in the 40-Hz range. As the inhibition increases to threefold the baseline level, the frequency decreases to a 10- to 13-Hz α-rhythm. (B) Coherence (0–50 Hz) between thalamic and cortical spiking is shown as a function of dose level (color bar indicates strength of coherence). High coherence is observed from 10 to 13 Hz when propofol increases to threefold the baseline level. (C) Raster of spiking activity in the model near baseline level of propofol. The cortical cells are divided into two populations (E: 1–45, 46–90; LTS: 91–98, 99–106; FS: 107–114, 115–122). RE cell activity is sparse and irregular, whereas TC cells do not spike at all. (D) At 250% of baseline, both thalamic cell types participate in the spiking activity at α-frequency.

Fig. 4.

Voltage traces for a four-cell model. Propofol changes from a low to high dose at t = 5 s; at that time, cortical cells transition from a β- to α-spiking frequency. Simultaneously, thalamic cells are recruited into the cortical activity, thus engaging a rhythmic thalamocortical loop.

Fig. 5.

Response of TC-RE pair to cortical drive. (A) TC-RE model cell spiking at the baseline (gGABA = 0.04). Spiking is sparse and uncorrelated. (B) TC-RE model cell spiking at a high-dose level (gGABA = 0.14). Spiking is entrained to the 12-Hz drive (the drive arrives at the dashed vertical lines). (C) Detail of activity from 4,160–4,260 ms. (Top to Bottom) RE voltage, TC voltage, gh (in TC cell), and IGABA (in TC cell) for baseline (thin line) and propofol (thick line) conditions. The increased IGABA and, consequently, the Ih are important in promoting the rhythmic thalamic response. (D) Mean Synchronization (Sync) Index obtained over different noise realizations for the single TC-RE pair with a 12-Hz cortical drive (n = 20). As GABA increases beyond twofold the baseline value, the SI increases, reflecting entrainment of the thalamic cells to the cortical excitation.

Fig. 6.

Frequency response of thalamic cells at high-dose levels shows that entrainment is constrained to the α-range. (A) Power spectrum of thalamic cell spiking in response to random excitation. The interspike intervals of the excitation are drawn from a uniform distribution. (B) Thalamic cell spiking frequency in response to periodic excitation at different frequencies (n = 20). As a result of the time scales of the inhibition and intrinsic currents, the thalamus always responds near 11 Hz. (C) In the absence of cortical excitation, TC-RE cells produce waxing and waning oscillations similar to the thalamic spindles modeled by Destexhe et al. (15, 20).

Fig. 7.

Simulation of small network with connectivity between RE cells. (A) Network connectivity. (B) Power spectra of cortical spiking activity with (Upper) and without (Lower) RE cell coupling. In the latter case, as a result of different parameterizations and noise, the activity displays disparate frequencies. (C) Cross-covariance between E-cell spiking showing clear synchrony in the case of RE coupling (Upper) and a lack of synchrony otherwise (Lower).