Information transfer rate in fMRI experiments measured using mutual information theory

Area of Science:

• Neuroscience
• Information Theory
• Biomedical Engineering

Background:

• fMRI experiments can be analyzed using information theory.
• Information transfer rate is key for brain-computer interface (BCI) design.
• Maximizing information transfer rate is essential for effective BCI implementation.

Purpose of the Study:

• To apply information theory to analyze fMRI time-series.
• To evaluate the efficiency of fMRI experimental designs using mutual information rate (MIR).
• To determine optimal experimental parameters for maximizing information transfer in fMRI-BCI.

Main Methods:

• Modeling fMRI experiments as communication systems.
• Utilizing mutual information rate (MIR) to assess experimental design efficiency.
• Estimating channel capacity and comparing it to the Hartley-Shannon Theorem limit.

Main Results:

• Block lengths of 3-5 seconds yield maximum information transfer rates in fMRI.
• Shorter block lengths are limited by channel capacity.
• Longer block lengths are constrained by lower source information transmission rates.

Conclusions:

• Information theory provides a robust framework for fMRI analysis and BCI design.
• Optimal fMRI experimental design, specifically block length, is critical for maximizing BCI performance.
• The study identifies specific block lengths (3-5s) for enhanced information extraction in fMRI-BCI applications.