Brain-machine interactions for assessing the dynamics of neural systems

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biophysics

Background:

• Establishing bidirectional communication between the nervous system and external devices is vital for brain-machine interfaces (BMIs).
• Neural population signals are complex and depend on poorly understood neural dynamics, posing a challenge for BMI development.
• Predicting neural system output requires understanding the underlying neural dynamics and their dimensionality.

Purpose of the Study:

• To introduce a novel technique for identifying the dynamics of neural populations interacting bidirectionally with external devices.
• To assess the dynamical dimension of neural populations as a measure of their complexity and predictability.
• To demonstrate a simple experimental approach for measuring neural dynamical dimension through closed-loop interactions.

Main Methods:

• Utilized in vitro lamprey brainstem preparations.
• Engaged preparations in a closed-loop interaction with simulated dynamical devices of varying degrees of freedom.
• Analyzed the composite system's behavior to determine the number of independent parameters (state variables) governing neural output.

Main Results:

• Successfully assessed the dynamical dimension of neural populations in real-time.
• Demonstrated that the dynamical dimension provides insight into the complexity of neural population dynamics.
• Achieved stable and reliable measurements of the dynamical dimension using the proposed experimental technique.

Conclusions:

• The developed technique offers a practical method for quantifying neural population dynamics.
• Measuring dynamical dimension is a valuable computational property for understanding neural computations in bidirectional BMI.
• This approach facilitates the prediction of future neural behaviors based on current states and external inputs.