Random matrix theory for analysing the brain functional network in lower limb motor imagery

Area of Science:

• Neuroscience
• Brain-Computer Interfaces
• Statistical Physics

Background:

• Understanding brain functional networks is crucial for decoding motor intentions.
• Motor imagery, particularly of lower limbs, presents unique challenges in neural signal analysis.
• Random matrix theory (RMT) offers a framework for analyzing complex systems like brain networks.

Purpose of the Study:

• To investigate statistical properties of brain functional networks during lower limb motor imagery using RMT.
• To explore the applicability of RMT in analyzing electroencephalogram (EEG) data for motor imagery tasks.
• To identify potential neural markers for differentiating between left and right lower limb movements.

Main Methods:

• Electroencephalogram (EEG) signals were recorded from subjects performing lower limb motor imagery.
• Functional connectivity was quantified using Pearson correlation coefficient (PCC), mutual information (MTI), and phase locking value (PLV).
• Statistical properties of the functional networks were analyzed using random matrix theory (RMT).

Main Results:

• Deviations from RMT predictions were observed in spectral density and level spacings during lower limb motor imagery.
• A significant difference in the maximum eigenvalue of the PCC-derived network was found between left and right foot imagination.
• These findings suggest distinct statistical network characteristics associated with unilateral lower limb movements.

Conclusions:

• RMT provides valuable insights into the statistical structure of brain functional networks during motor imagery.
• The observed differences in network properties between left and right foot imagery offer a potential basis for classification.
• This study contributes to the development of more sophisticated brain-computer interfaces for motor control and rehabilitation.