Unveiling age-independent spectral markers of propofol-induced loss of consciousness by decomposing the electroencephalographic spectrum into its periodic and aperiodic components

Abstract

BackgroundInduction of general anesthesia with propofol induces radical changes in cortical network organization, leading to unconsciousness. While perioperative frontal electroencephalography (EEG) has been widely implemented in the past decades, validated and age-independent EEG markers for the timepoint of loss of consciousness (LOC) are lacking. Especially the appearance of spatially coherent frontal alpha oscillations (8–12 Hz) marks the transition to unconsciousness.Here we explored whether decomposing the EEG spectrum into its periodic and aperiodic components unveiled markers of LOC and investigated their age-dependency. We further characterized the LOC-associated alpha oscillations by parametrizing the adjusted power over the aperiodic component, the center frequency, and the bandwidth of the peak in the alpha range.MethodsIn this prospective observational trial, EEG were recorded in a young (18–30 years) and an elderly age-cohort (≥ 70 years) over the transition to propofol-induced unconsciousness. An event marker was set in the EEG recordings at the timepoint of LOC, defined with the suppression of the lid closure reflex. Spectral analysis was conducted with the multitaper method. Aperiodic and periodic components were parametrized with the FOOOF toolbox. Aperiodic parametrization comprised the exponent and the offset. The periodic parametrization consisted in the characterization of the peak in the alpha range with its adjusted power, center frequency and bandwidth. Three time-segments were defined: preLOC (105 – 75 s before LOC), LOC (15 s before to 15 s after LOC), postLOC (190 – 220 s after LOC). Statistical significance was determined with a repeated-measures ANOVA.ResultsLoss of consciousness was associated with an increase in the aperiodic exponent (young: p = 0.004, elderly: p = 0.007) and offset (young: p = 0.020, elderly: p = 0.004) as well as an increase in the adjusted power (young: p < 0.001, elderly p = 0.011) and center frequency (young: p = 0.008, elderly: p < 0.001) of the periodic alpha peak. We saw age-related differences in the aperiodic exponent and offset after LOC as well as in the power and bandwidth of the periodic alpha peak during LOC.ConclusionDecomposing the EEG spectrum over induction of anesthesia into its periodic and aperiodic components unveiled novel age-independent EEG markers of propofol-induced LOC: the aperiodic exponent and offset as well as the center frequency and adjusted power of the power peak in the alpha range.

Electroencephalogram Biomarkers from Anesthesia Induction to Identify Vulnerable Patients at Risk for Postoperative Delirium

“…We applied a similar method in a previous investigation with a detailed methodic explanation. 19 The following EEG parameters were computed: the spectral edge frequency (the frequency under which 95% of the power is located), the mean power of the power peak in the alpha range (8 to 12 Hz), the alpha peak frequency (the frequency with the highest power within the alpha band), the alpha power difference between baseline and LOC_1/2/15, the mean power of the beta range before decomposition, the mean power of the power peak in the beta range (12 to 30 Hz), beta peak frequency (the frequency with the highest power within the beta band), and the beta ratio. The beta ratio calculation was derived from the method of Rampil: beta ratio = log[(P30 to 47 Hz)/ (P11 to 20 Hz)].…”

“…11,25,26 During the anesthesia-induced transition to unconsciousness, elderly patients also show a reduced power in the alpha range (8 to 12 Hz) compared to young adults. 19 We hypothesized that a postinduction reduced alpha power might be a sign of a vulnerable brain, leading to a higher risk to develop POD. 24 Furthermore, we observed that POD patients exhibit a reduced postinduction beta arousal, associated with a lower spectral edge frequency within the first minute after LOC.…”

Aperiodic and Hurst EEG exponents across early human brain development: a systematic review

“…In adulthood, β significantly declines with age (Voytek, et al ., 2015; Waschke, Wöstmann and Obleser, 2017) although the physiological origin of this age-related change is unclear. EEG 1/ f measures also display behavioural and clinical relevance, particularly in conditions thought to relate to shifts in E:I balance, such as those affecting attention and behaviour (Waschke et al ., 2021; Robertson et al ., 2019), states of consciousness (Leroy et al ., 2023), and functional recovery from stroke (Leemburg et al ., 2018). Prior to future research utilising 1/ f measures to explore possible atypical brain E:I or as a biomarker in clinical populations, we must first characterise 1/ f measures across the typically developing (TD) lifespan, from infancy to early adulthood (other studies have begun to chart this for later adulthood, see Finley et al ., 2022).…”

The role of etiology in the identification of clinical markers of consciousness: comparing EEG alpha power, complexity, and spectral exponent

“…Indeed, the electroencephalogram (EEG) spectral exponent (i.e., exponential decay of power over frequency) has gained increasing attention as a novel marker of consciousness (Colombo et al 2019; Lendner et al 2020; Maschke, Duclos, Owen, et al 2022; Colombo et al 2023; Leroy et al 2023). Especially in the absence of oscillatory peaks – which is a common phenomenon in disorders of consciousness (DOC) – our group has previously demonstrated diagnostic power of the spectral exponent above and beyond oscillatory power (Maschke, Duclos, Owen, et al 2022).…”