Quantifying consciousness through intrinsic probability density function

Abstract

Consciousness remains a multifaceted phenomenon that is difficult to be measured by traditional quantification methods. Here we propose the intrinsic probability density function (iPDF) as a quantitative method to evaluate the dynamic inter-cortical interactions that underlie conscious states. First, the method utilizes empirical mode decomposition to derive intrinsic mode functions (IMFs) from EEG signals. Then, the method generates scale-dependent probability density functions for successive partial sums of IMFs that can capture subtle variations in neural modulation patterns. We tested the iPDF analysis across various consciousness states such as general anesthesia, distinct sleep stages (wakefulness, REM, and deep sleep), sensory conditions (eyes open versus eyes closed), and between dementia patients and healthy subjects. Our findings reveal that active neural interactions or modulations during wakefulness and REM sleep are characterized by super-Gaussian iPDF patterns. By contrast, the reduced interactions observed in anesthesia and deep sleep yield near-Gaussian iPDF profiles. We also present a classification model built on iPDF features that achieved an accuracy of approximately 87 % in distinguishing dementia patients from health controls, demonstrating the iPDF as a potential biomarker in clinical screening. This study supports the idea that consciousness emerges from complex, scale-dependent neural processes and presents a robust, quantitative framework that may enhance both our theoretical understanding and practical assessment of various states of consciousness.