Scientific Background¶

Self from Coupled Predictive Loops Hypothesis

1. Note¶

This document outlines the relationship between this hypothesis and existing work in cognitive science, neuroscience, and related theories. Its goals are to:

Clarify the empirical foundations
Distinguish established findings from theoretical inferences
Demonstrate the compatibility and integrative nature of the model

This framework is not constructed from scratch, but represents a structural integration of multiple existing directions.

2. Overall Correspondence¶

The core mechanisms involved in this model include:

Multi-prediction structure
Prediction–error minimization
Trajectory compression and memory
Cross-temporal continuity constraints
Multi-signal coupling and indistinguishability
Behavioral attribution mechanisms (agency attribution)

These mechanisms are supported across different research domains. However:

They have not yet been unified into a single structural framework centered on “trajectory selection.”

3. Predictive Systems and Multi-Prediction Structure¶

Related theory: Predictive Processing

Established Findings¶

The brain continuously generates predictions about sensory input
Prediction errors are used to update internal models
Predictive mechanisms exist across multiple levels of neural organization

Supporting Evidence¶

Forward models predict the outcomes of self-generated actions
Sensory consequences of self-generated actions are partially attenuated
The brain represents probability distributions rather than single outcomes

Role in This Model¶

Multi-prediction structure → corresponds to probabilistic expansion
Predictive trajectories → temporal unfolding of states and actions
Predictive consistency → foundation of agency and attribution

4. Information Compression and Trajectory Representation¶

Related theories:

Predictive Coding
Information theory (compression / efficient coding)

Established Findings¶

Neural systems reduce redundancy and extract statistical structure
Neural encoding exhibits compression properties

Supporting Evidence¶

Memory is reconstructive rather than frame-by-frame storage
Autobiographical memory exhibits structured and pattern-based organization

Role in This Model¶

Compression defines trajectory space
Provides the basis for cross-temporal matching
Supports the structural continuity of the self

5. Temporal Continuity and Error Minimization¶

Related directions:

Prediction error minimization
Temporal integration mechanisms

Established Findings¶

Systems continuously compare current states with predictions
Perception and cognition operate within temporal integration windows

Supporting Evidence¶

Temporal consistency is linked to self-stability
Disruptions in temporal continuity affect identity and continuity experience

Role in This Model¶

Continuity error → deviation across temporal trajectories
Self-stability → result of error minimization
Dominant trajectory → most stable path

6. Multi-Signal Integration and Unified Experience¶

Related theory: Integrated Information Theory

Established Findings¶

Consciousness levels correlate with the degree of network integration
Coordinated activity across brain regions relates to conscious content

Supporting Evidence¶

Reduced integration is associated with diminished consciousness
Highly integrated systems correspond to more unified experience

Role in This Model¶

Supports signal coupling mechanisms
Supports the indistinguishability principle
Corresponds to the unity of experience

7. Competition Mechanisms and Global Access¶

Related theory: Global Workspace Theory

Established Findings¶

Information competes for limited processing resources in the brain
Only a subset of information reaches global accessibility

Supporting Evidence¶

Attention modulates information selection
Global broadcasting influences multiple subsystems

Role in This Model¶

Competition among predictive trajectories
Selection of the dominant trajectory
Stability of a single self

8. Self-Model Theory¶

Related theory: Self-Model Theory of Subjectivity Proposed by: Thomas Metzinger

Core Ideas¶

The self is a dynamically generated model
The system does not label it as a “model”
Subjective experience corresponds to the model’s content

Correspondence in This Model¶

Self = dominant trajectory
Experience = signal integration along the trajectory
First-person perspective = output form

9. Attribution Mechanisms and Agency¶

Related directions:

Sense of agency
Comparator models

Established Findings¶

Attribution of action depends on consistency between prediction and feedback
The brain compares predicted outcomes with actual feedback

Supporting Evidence¶

Prediction match → self-attribution
Prediction mismatch → external attribution

Role in This Model¶

Agency = attribution mechanism
Based on prediction–evolution matching
Bound to the dominant trajectory

10. Abnormal States¶

Related condition: Schizophrenia

Established Findings¶

Abnormalities in action prediction
Disturbances in self-attribution
Reduced network integration
Impaired temporal continuity

Explanation in This Model¶

Disruption of prediction–matching
Increased continuity error
Instability of the dominant trajectory
Breakdown of signal coupling

Important note:

These correspondences are structural interpretations and do not constitute clinical explanations.

11. Theoretical Background of the First-Person Perspective¶

Related position: Daniel Dennett

Core Ideas¶

The first-person perspective is not an independent entity
It arises from narrative and reporting mechanisms within the system

Correspondence in This Model¶

First-person perspective = unified output of the dominant trajectory
Arises from indistinguishable signal integration

12. Summary of Support Strength¶

Mechanism	Level of Support
Predictive systems / multi-prediction	High
Compression / representational structure	Medium–High
Temporal continuity	Medium
Signal integration	Medium
Attribution mechanisms (agency)	Medium–High
Indistinguishability principle	Strong theoretical support
Trajectory selection mechanism	Speculative (core innovation)
Unified framework	No direct precedent

13. Summary¶

The scientific basis of this hypothesis can be summarized as:

Multiple key mechanisms have been independently studied
But have not yet been unified into a single generative structure

The primary contribution of this model is:

To integrate multi-prediction, trajectory compression, continuity constraints, signal coupling, and attribution mechanisms into a unified framework of “trajectory selection.”

Current status:

Highly compatible with existing science
Possesses explanatory potential
Still requires formalization and experimental validation