The document presents a study on Unbiased Markov Chain Monte Carlo (MCMC) methods, focusing on a proposed methodology involving coupled chains to achieve unbiased estimators for Bayesian inference. It discusses the implementation of MCMC algorithms, specifically the Metropolis-Hastings method, and highlights the importance of maximizing coupling to improve efficiency and reduce bias. Additionally, the paper includes applications in neuroscience, showcasing its relevance in estimating parameters using particle filters and controlled sequential Monte Carlo methods.