This repository contains the official implementation of the paper: "Biologically-Inspired Emotional Processing for Adaptive Decision-Making in Non-Stationary Environments"
The Emotional-Cognition Integration Architecture (ECIA) is a biologically-inspired reinforcement learning framework designed to manage environmental uncertainty. Unlike traditional methods that utilize statistical thresholds, this study demonstrates that emotion-like mechanisms (External Limbic System) function as adaptive heuristics for rapid uncertainty management.
This codebase reproduces the large-scale experimental replication (N=3,600 runs) reported in the manuscript, comparing ECIA against state-of-the-art non-stationary baselines.
- Biologically-Inspired Architecture: Implements 8 computational emotions (Plutchik's wheel), episodic memory retrieval, and dopamine-modulated learning rates.
- Non-Stationary Environments:
Environment A: Sudden Strategy Reversal (Shock)Environment B: Predictable Alternation (Cyclic)Environment C: Stochastic Disruptions (High Uncertainty)
- Improved Baselines:
- Sliding Window UCB (SW-UCB)
- Adaptive Thompson Sampling (Adaptive TS)
- Context-Aware Epsilon-Greedy
- Computational Tractability: Optimized for standard research workstations.
All experiments in the paper were conducted using the following specifications:
- OS: Windows 10 Pro
- CPU: Intel Core i7-6700 CPU @ 3.40GHz
- RAM: 16GB
- Python Version: 3.8
The simulation relies on 12 specific Fibonacci seeds (34, 55, ..., 6765) to ensure statistical robustness and reproducibility of the reported results.
-
Clone the repository:
git clone https://github.com/DrKDH/PeerjCS.git cd PeerjCS -
Install dependencies:
pip install -r requirements.txt
To run the full simulation suite and generate the performance metrics:
python ECIA_improved.py- Execution Time: Approximately 20-30 minutes for the full suite (3,600 runs).
- Outputs: The script will generate CSV files containing mean rewards and recovery metrics, corresponding to the Tables and Figures in the manuscript.
Running this code will reproduce the following key findings:
- Functional Specialization: ECIA statistically outperforms baselines in uncertain/stochastic environments (Env A & C).
- Cost of Complexity: Simpler baselines (like SW-UCB) outperform ECIA in strictly predictable environments (Env B).
- Synergistic Integration: Ablation studies demonstrate that removing components (Emotion, Memory, Dopamine) causes non-additive performance degradation.
This project is open-source and available for academic and research purposes.