Intrinsic units: identifying a system's causal grain

Area of Science:

• Neuroscience
• Theoretical Physics
• Philosophy of Mind

Background:

• Integrated Information Theory (IIT) proposes that consciousness is a fundamental property of physical systems.
• IIT quantifies consciousness using integrated information (Φ), representing a system's cause-effect power.
• The theory posits that the 'grain' of system units (e.g., neurons, synapses) must maximize Φ.

Purpose of the Study:

• To develop a computational framework for IIT that includes macro units, not just micro units.
• To identify a system's intrinsic units – those that directly constitute its experience.
• To extend the mathematical framework of IIT 4.0 for assessing cause-effect power across different scales (grains).

Main Methods:

• Formalizing the physical existence postulates of IIT for macro units.
• Extending the IIT 4.0 mathematical framework to evaluate cause-effect power across multiple grains.
• Utilizing simple, simulated systems to compute integrated information with macro units.

Main Results:

• Demonstrated that systems with macro units can exhibit higher integrated information (Φ) than systems composed solely of micro units.
• Identified specific types of macro units that can enhance a system's cause-effect power.
• Provided a method to determine the intrinsic units of a system from its intrinsic perspective.

Conclusions:

• The developed framework enables the computation of integrated information in systems with macro units.
• This approach offers a foundation for empirical testing and inferring consciousness within physical systems based on IIT.
• The identification of intrinsic units provides a more nuanced understanding of how systems generate experience.