Modeling gravity-dependent plasticity of the angular vestibuloocular reflex with a physiologically based neural...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Vestibular System Research

Background:

• The angular vestibuloocular reflex (aVOR) stabilizes gaze during head rotations.
• Adaptation of the aVOR exhibits gravity-dependent properties, varying with head orientation.
• Understanding this adaptation is crucial for diagnosing and treating vestibular disorders.

Purpose of the Study:

• To develop a neural network model explaining gravity-dependent gain adaptation of the aVOR.
• To investigate the three-dimensional mechanisms underlying aVOR adaptation.
• To test the hypothesis of direct otolith input influencing canal-otolith neurons.

Main Methods:

• A three-dimensional, physiologically based neural network model of canal-otolith-convergent neurons was developed.
• Model parameters were trained using experimental vertical aVOR gain values.
• A learning rule minimized the error between experimental and model-derived eye velocities.

Main Results:

• The model successfully predicted experimental aVOR gain data across all head positions, not just trained positions.
• Model predictions compared favorably with a standard double-sinusoid function fit.
• Altering relative learning rates of weights and bias improved model-data accuracy.

Conclusions:

• The model supports a three-dimensional adaptation mechanism for the aVOR, driven by direct otolith input to canal-otolith neurons.
• Visual-vestibular mismatch appears to adapt canal sensitivities.
• Adaptation tuning is influenced by otolith input weights to canal-otolith-convergent neurons.