Abstract
Humans sometimes do perceive differences where physically there are none. It is thus tenable that perception is susceptible to seemingly random fluctuations in brain activity or “neural noise”. Here, we demonstrate the potency of both trial-aggregated as well as trial-by-trial measures in the human electroencephalogram (EEG) to characterize neural noise as (i) a trait of individuals of varying age (n = 19; 19–74 years), and (ii) a brain state that predicts an individual’s impending perceptual decision. Human participants were instructed to discriminate two identical, consecutively presented pure tones. Behaviorally, all participants reported perceiving pitch differences of first versus second tone. Neurally, decisions for the first versus the second tone were preceded by more consistently phase-locked responses to the first tone in the theta (4–9 Hz) band at central scalp electrodes. Second, a trial-wise information-theoretic measure quantifying the irregularity of broadband EEG, Weighted Permutation Entropy (WPE), prior to stimulus onset allowed to classify a listener’s impending decision on this trial. Average entropy not only increased with participants’ age, but correlated with previously suggested measures of an altered excitation-inhibition balance in the aging brain. Therefore, neural noise is best conceived not only as a state variable that can shape perceptual decisions but moreover can capture trait-like changes with age.
Significance Statement Humans sometimes do perceive differences where physically there are none, a phenomenon likely due to seemingly random fluctuations in brain activity, or “neural noise”. The potency of neural noise to explain intra- and inter-individual differences is largely unresolved. Here, we show that two complementary neural noise measures from the human EEG, across-trials phase-coherence of slow neural oscillations and a within-trial entropy measure, both predict decisions of participants when they are to compare physically identical tones. We also find the overall level of neural noise to increase with participants’ age, likely reflecting changes to an excitation-inhibition balance. Thus, neural noise not only helps to characterize behaviorally relevant brain states, but can also second as an inter-individual marker of neurobiological age.
