Information Integration

library(consciousnessModelR)

Purpose

This article explains the theoretical background behind simulate_information_integration(). The function is inspired by information-integration approaches to consciousness, especially the idea that consciousness may depend on both integration and differentiation.

The function does not calculate formal phi and should not be presented as an implementation of Integrated Information Theory.

Integration and differentiation

A system can be highly differentiated if its parts carry distinct patterns of activity. A system can be highly integrated if its parts influence one another and behave as a coordinated whole.

A common intuition in information-integration theories is that consciousness requires both. A system that is completely fragmented lacks unity. A system that is completely uniform lacks differentiated content.

Relation to the package

simulate_information_integration() creates a simple network of components. Components influence one another through connections. The function then computes an educational integration score based on average connectivity, shared activity among components, and differentiation among component states.

This score is not a clinical, neuroscientific, or philosophical measure of consciousness.

Basic simulation

info <- simulate_information_integration(
  n_components = 8,
  steps = 100,
  connection_probability = 0.30,
  seed = 42
)

info$summary
#>   mean_connectivity shared_information differentiation integration_score
#> 1            0.5625          0.3072903       0.3431688        0.05931701

plot_consciousness_sim(
  info$time_series,
  x = "step",
  y = "activation",
  group = "component"
)

Comparing sparse and dense systems

sparse <- simulate_information_integration(
  connection_probability = 0.10,
  seed = 42
)

moderate <- simulate_information_integration(
  connection_probability = 0.30,
  seed = 42
)

dense <- simulate_information_integration(
  connection_probability = 0.70,
  seed = 42
)

rbind(
  sparse = sparse$summary,
  moderate = moderate$summary,
  dense = dense$summary
)
#>          mean_connectivity shared_information differentiation integration_score
#> sparse             0.34375          0.2009751       0.3645807        0.02518713
#> moderate           0.56250          0.3072903       0.3431688        0.05931701
#> dense              0.75000          0.1635655       0.3287365        0.04032746

Interpretation

Sparse systems may show weak integration because components are not strongly connected. Very dense systems may show high coupling but reduced differentiation if all components behave similarly. Moderate systems may sometimes balance integration and differentiation.

This illustrates a central theoretical tension: a system must be unified enough to function as a whole, but differentiated enough to support rich content.

Relation to other functions

This model complements broadcast_network(). Broadcast models emphasize access and availability. Integration models emphasize internal structure and causal organization.

Limitations

This simulation is only an educational proxy. Formal Integrated Information Theory is mathematically complex and has specific definitions that are not implemented here.