ADOpy

ADOpy implements Adaptive Design Optimization to compute optimal experimental designs in real time and maximize the informativeness of collected data for statistical inference.

Key Features:

• Adaptive Experimentation: Implements the Adaptive Design Optimization (ADO) framework to select experimental conditions dynamically based on ongoing results.
• Real-time Optimal Design Computation: Computes optimal designs during experiments to guide stimulus selection and experimental parameter choices.
• Dynamic Parameter Adjustment: Adjusts experimental parameters such as stimulus intensity and timing based on real-time analysis of participant responses.
• Bayesian and Machine Learning Methods: Employs Bayesian statistics and machine learning approaches to maximize the informativeness of collected data.
• Modular Python Implementation: Provided as a modular Python package for integration into experimental workflows.

Scientific Applications:

• Psychology: Optimizes stimulus presentation and trial selection to improve efficiency in behavioral experiments.
• Neuroscience: Guides experimental parameter choice to maximize information from neural response measurements.
• Behavioral Economics: Designs trials to efficiently estimate decision-making parameters and preferences.
• Psychometric Function Estimation: Refines estimation of psychometric functions by optimizing stimulus presentation based on participant responses.
• Delay Discounting Experiments: Designs experiments to efficiently capture valuation of immediate versus delayed rewards.
• Risky Choice Analysis: Optimizes experimental conditions to improve inference about risk preferences under uncertainty.

Methodology:

Implements the ADO framework using Bayesian statistics and machine learning to maximize data informativeness, with algorithms that dynamically adjust experimental parameters (e.g., stimulus intensity or timing) based on real-time analysis of participant responses.