[Paper Review] Measuring Consciousness

Siddhartha Sen|arXiv (Cornell University)|Sep 20, 2016

Neural Networks and Applications3 references3 citations

TL;DR

This paper proposes a theoretical framework to measure consciousness using EEG-derived brain wave correlation functions and Shannon's information theory. By computing a time-dependent information function $ C(t) $ and its derivative $ D(t) $, the model quantifies consciousness as dynamic information processing in the brain, with higher values indicating greater awareness and a strong link to gamma wave activity.

ABSTRACT

A measurement based formula for consciousness, C, as a function of time t, is constructed. The formula depends on identifying a natural relevant self-generated, time-dependent dynamical process inherent in any entity. For human beings the relevant dynamical process identified is the ensemble of brain waves, observed in EEG measurements, that are represented in the model by their measured time dependent correlation functions. These correlation functions define the accessible dynamical state of the brain at any moment of time. From them a time dependent probability function, P(t), is extracted by using a mathematical identity. According to information theory, -P(t) log P(t), is a measure of the information contained in the brain waves. Consciousness, C, is defined by this information theory formula: it is not localized, does not depend on specific hardwire details of the brain, but reflects the information content present in brain waves. Justifications, based on observational evidence, are given for the formula and it is shown that C reflects the degree of "awareness" that a person has at a given moment of time. The model explains the observed time delay between when a brain wave is seen to initiate an action and when there is awareness that the action has been initiated, in terms of the way brain waves processes information. Some testable consequences of the model, including the role dreaming sleep plays in long term memory storage, are discussed. It is also shown that non living entities have C=0.

Motivation & Objective

• To develop a testable, measurement-based model of consciousness grounded in experimental EEG data.
• To define consciousness not as a fixed brain structure but as dynamic information processing in neural ensembles.
• To establish a quantitative link between brain wave patterns (e.g., gamma, delta) and levels of awareness.
• To provide a theoretical justification for using correlation functions and information theory in measuring subjective experience.
• To differentiate conscious from unconscious states using time-varying information metrics, such as $ D(t) $, based on neural dynamics.

Proposed method

• Extracts time-dependent probability distributions $ P_N(t) $ from EEG correlation functions at $ N $ brain points.
• Applies Shannon’s information formula $ C(t) = -\sum P_N(t) \ln P_N(t) $ to quantify information content related to awareness.
• Defines the rate of information processing as $ D(t) = \frac{dC}{dt} $, representing the dynamic aspect of consciousness.
• Uses a mean field approximation to model contributions from delta, theta, alpha, beta, and gamma waves with weights $ c_i $ summing to one.
• Relies on empirical EEG data to assign $ c_i $ values based on brain state (e.g., $ c_4 $ dominant in wakefulness, $ c_1 $ in deep sleep).
• Links $ D(t) $ to brain wave frequencies via $ D(t) = \sum_{i=1}^5 c_i \omega_i \frac{dC_i(x)}{dx} $, with $ x = \omega_i t $, to connect dynamics to consciousness levels.

Experimental results

Research questions

• RQ1 Can consciousness be quantified as a time-dependent information content derived from EEG correlation functions?
• RQ2 How does the rate of information processing $ D(t) $ correlate with known states of awareness (e.g., REM sleep, deep sleep, epileptic absence seizures)?
• RQ3 To what extent do gamma wave components $ c_5 $ distinguish conscious from unconscious states despite high neural activity?
• RQ4 Can the model differentiate between vegetative states and minimally conscious states using information-theoretic metrics?
• RQ5 How do memory access and time-dependent neural signals contribute to the emergence of a continuous sense of self in the model?

Key findings

• The model defines consciousness via $ C(t) $, the information content of brain waves, and $ D(t) $, the rate of change of this information, both derived from EEG correlation functions.
• States with low $ D(t) $, such as REM sleep or epileptic absence seizures, are classified as unconscious, even when neural activity is high.
• Gamma wave activity ($ c_5 $) is consistently present in wakeful and REM states but absent in epileptic absence seizures, explaining their lower consciousness levels.
• The mean field approximation allows the model to assign state-dependent weights $ c_i $ to different brain wave types, with $ c_4 $ dominating in wakefulness and $ c_1 $ in deep sleep.
• The derivative $ D(t) $ is shown to depend on frequency components $ \omega_i $, linking neural oscillation dynamics directly to consciousness metrics.
• The model supports the idea that consciousness arises from dynamic, time-dependent information processing rather than static brain wiring, with continuity of self tied to memory and persistent information flow.