Propofol Sedation Alters Perceptual and Cognitive Functions in Healthy Volunteers as Revealed by Functional Magnetic Resonance Imaging

Abstract

Background: Elucidating networks underlying conscious perception is important to understanding the mechanisms of anesthesia and consciousness. Previous studies have observed changes associated with loss of consciousness primarily using resting paradigms. The authors focused on the effects of sedation on specific cognitive systems using task-based functional magnetic resonance imaging. The authors hypothesized deepening sedation would degrade semantic more than perceptual discrimination.

Methods: Discrimination of pure tones and familiar names were studied in 13 volunteers during wakefulness and propofol sedation targeted to light and deep sedation. Contrasts highlighted specific cognitive systems: auditory/motor (tones vs. fixation), phonology (unfamiliar names vs. tones), and semantics (familiar vs. unfamiliar names), and were performed across sedation conditions, followed by region of interest analysis on representative regions.

Results: During light sedation, the spatial extent of auditory/motor activation was similar, becoming restricted to the superior temporal gyrus during deep sedation. Region of interest analysis revealed significant activation in the superior temporal gyrus during light (t [17] = 9.71, P < 0.001) and deep sedation (t [19] = 3.73, P = 0.001). Spatial extent of the phonologic contrast decreased progressively with sedation, with significant activation in the inferior frontal gyrus maintained during light sedation (t [35] = 5.17, P < 0.001), which didn't meet criteria for significance in deep sedation (t [38] = 2.57, P = 0.014). The semantic contrast showed a similar pattern, with activation in the angular gyrus during light sedation (t [16] = 4.76, P = 0.002), which disappeared in deep sedation (t [18] = 0.35, P = 0.731).

Conclusions: Results illustrate broad impairment in cognitive cortex during sedation, with activation in primary sensory cortex beyond loss of consciousness. These results agree with clinical experience: a dose-dependent reduction of higher cognitive functions during light sedation, despite partial preservation of sensory processes through deep sedation.

Figure 1.

Moving window (20 trials) of proportion of responses in each participant, along with sedation classifications assigned to the data. Black lines denote functional magnetic resonance imaging run breaks. Blue shading = Pre; green shading = Light-Sedation; teal shading = transitional (not analyzed); red shading = Deep-Sedation.

Figure 2.

Accuracy and reaction time to the perceptual and semantic tasks across levels of sedation. Accuracy decreased (due to increasing missed responses) while reaction times increased under deep levels of sedation.

Figure 3.

Contrast of activation to tones vs. baseline across levels of sedation. Activation to tones primarily (in orange) was centered around motor and sensory systems. Activation during baseline (in blue) highlights semantic processing in between task events.

Figure 4.

Combination of the three task contrasts, designated by color, to illustrate relative overlap. Blue: AudMotor, green: Phoneme, red: Semantic; orange: AudMotor+ Phoneme, yellow: Phoneme+ Semantic.

Figure 5.

Main effect of sedation level across all task contrasts. Regions in the motor (supplementary motor area, premotor cortex), auditory and phonological (superior temporal gyrus, inferior frontal gyrus), and semantic (posterior cingulate, precuneus), all showed consistent reduction in activation across all task contrasts with increasing sedation.

Figure 6.

Interaction effect of task contrast by sedation level. Sedation had different effects on task contrasts in most of the regions studied (detailed in the region of interest analysis in Figure 7).

Figure 7.

Average activity within regions of interest derived from the voxel-wise interaction of task contrast by sedation level. Activity in auditory cortex diminished with sedation, but remained active under deep sedation. Phonologic and semantic effects were reduced in light sedation, and were not detected in deep sedation. Regions with activity significantly different than zero (p<0.002) are denoted with ‘*’; regions with significant differences between foil and target stimuli (p<0.002) are denoted with ‘^’. Error bars denote standard error of the mean.