Entropic Explanation Autism Sameness

---
name: entropic-explanation-autism-sameness
description: "Information theory-based framework explaining insistence on sameness in autism through entropy minimization. Uses metric D_H(R, M) = H(R|M) + H(M|R) to quantify surprise and uncertainty reduction strategies. Provides formal foundations for autism therapies, robotic caregivers, and computational models of cognitive limitations."
metadata:
  arxiv_id: "2608.04616"
  published: "2026-08-05"
  authors: "Przemysław Śliwiński"
  tags: [autism, information-theory, entropy, computational-neuroscience, cognitive-modeling, therapy-design]
license: Complete terms in LICENSE.txt
---

# An Entropic Explanation of Insistence on Sameness in Autism

## Overview
This framework proposes an information theory-based explanation for insistence on sameness in autism as a general behavior pattern where individuals attempt to reduce surprise and uncertainty. The core insight defines autism as an impairment where cognitive functions are restricted to discrimination, memorization, and prediction of tangible environmental properties.

## Core Mathematical Framework

### Entropy Metric
The framework centers on the entropy metric:
**D_H(R, M) = H(R|M) + H(M|R)**

Where:
- **R**: Sequences of random stimuli from the environment
- **M**: Memory that stores and retrieves environmental patterns  
- **H(R|M)**: Conditional entropy interpreted as **surprise**
- **H(M|R)**: Conditional entropy interpreted as **uncertainty**

### Behavioral Strategy
To minimize D_H(R, M), an individual can either:
1. **Learn about R** (store knowledge in M) - adaptive learning
2. **Restrict R to known M** - insistence on sameness

The framework concludes that **insistence on sameness is a manifestation of strategy #2**.

## Key Contributions

### Quantified Concepts
The framework provides formal definitions for previously qualitative concepts:
- **Surprise**: H(R|M) - unexpected environmental changes
- **Uncertainty**: H(M|R) - inability to predict environmental patterns  
- **Sensory overload/deprivation**: Extreme values of entropy components
- **Anxiety**: High combined entropy metric
- **Comfort zone**: Low entropy regions in environmental space
- **Aberrant precision**: Overconfident predictions leading to high surprise

### Therapeutic Applications
1. **Learning therapies**: Defined as optimization algorithms minimizing D_H(R, M)
2. **Daily care routines**: Structured to maintain low entropy environments
3. **Robotic live-in caregivers**: Programs implementing entropy-minimizing behaviors
4. **Self-reliance improvement**: Formal guidelines for basic activities of daily living

### Validation Methodology
- **Turing test-like approach**: Validation without involving individuals with autism
- **Computational simulation**: Testing framework predictions in silico
- **Behavioral modeling**: Comparing model outputs with observed autistic behaviors

## Implementation Guidelines

### When to Use This Framework
- **Autism research**: Computational modeling of autistic behaviors
- **Therapy design**: Developing evidence-based intervention strategies  
- **AI/robotics**: Creating supportive systems for individuals with autism
- **Cognitive science**: Understanding entropy-based decision making
- **Neuroscience**: Linking information theory to neural processing constraints

### Computational Modeling
1. **Define environmental space**: Characterize R as stochastic processes
2. **Model memory constraints**: Implement M with limited capacity/storage
3. **Calculate entropy metrics**: Compute H(R|M) and H(M|R) over time
4. **Simulate behavioral strategies**: Compare learning vs. restriction approaches
5. **Validate against observations**: Match model outputs to real-world data

### Therapeutic Algorithm Design
- **Entropy monitoring**: Track D_H(R, M) in real-time environments
- **Intervention triggers**: Activate support when entropy exceeds thresholds
- **Adaptive learning**: Gradually expand M to handle more complex R
- **Environmental structuring**: Modify R to match current M capabilities

## Pitfalls and Considerations

### Model Limitations
- **Individual variability**: Autism spectrum heterogeneity may require personalized parameters
- **Environmental complexity**: Real-world R is high-dimensional and non-stationary
- **Memory dynamics**: M may have complex temporal dependencies beyond simple storage
- **Social factors**: Framework focuses on individual cognition, may miss social influences

### Ethical Considerations
- **Avoid pathologizing**: Frame as adaptive strategy rather than deficit
- **Respect autonomy**: Ensure interventions support rather than control
- **Individual preferences**: Balance entropy reduction with personal interests
- **Cultural context**: Consider cultural variations in "sameness" expectations

### Validation Challenges
- **Ground truth**: Difficulty obtaining objective measures of internal states
- **Longitudinal effects**: Short-term entropy reduction may not improve long-term outcomes
- **Generalization**: Success in controlled environments may not transfer to real world

## Related Work
- **Predictive coding theories**: Bayesian brain hypothesis and precision weighting
- **Information bottleneck**: Optimal compression under constraints
- **Free energy principle**: Minimizing surprise through action and perception
- **Computational psychiatry**: Mathematical models of mental disorders

## References
- Original Paper: [arXiv:2608.04616](https://arxiv.org/abs/2608.04616)
- Journal Publication: Frontiers in Computational Neuroscience. 19:1714428
- DOI: https://doi.org/10.3389/fncom.2025.1714428

## Activation Keywords
- autism entropy framework
- insistence on sameness
- D_H(R, M) metric
- surprise uncertainty autism
- information theory autism
- robotic autism caregivers
- entropy minimization behavior
- computational autism model