Pre-processing and transfer entropy measures in motor neurons controlling limb movements

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Directed information transfer measures are valuable for modeling neural systems due to their model-free nature and applicability to complex signals.
• Intracellular recordings of synaptic potentials present challenges like non-stationary behavior, necessitating advanced analytical techniques.

Purpose of the Study:

• To develop and validate a preprocessing method for analyzing intracellular recordings of neural activity.
• To assess the consistency and predictive power of directed information transfer measures in neural connectivity.

Main Methods:

• Singular Spectrum Analysis (SSA) was employed as a preprocessing step to remove nonlinear trends and discontinuities from intracellular recordings.
• Normalized delayed transfer entropy was applied to analyze neural responses in locust motor neurons evoked by proprioceptor stimulation.

Main Results:

• The SSA preprocessing effectively handled nonlinear trends in synaptic potential recordings.
• Delayed transfer entropy measures demonstrated consistency within and between animals for identified neurons.
• These measures successfully predicted neural connectivity, specifically for the fast extensor tibiae motor neuron.

Conclusions:

• The proposed SSA-based preprocessing method enhances the analysis of challenging intracellular neural recordings.
• Directed information transfer measures, when applied to preprocessed data, are reliable indicators of neural connectivity in biological systems.