Frequency modulation of large oscillatory neural networks

Area of Science:

• Dynamical systems and control theory
• Computational neuroscience
• Robotics

Background:

• Periodic signal generation is crucial for biological central pattern generators and robotic applications.
• Frequency modulation is a fundamental requirement for controlling these dynamical systems.
• Generic mechanisms for controlling neural oscillator frequency remain an open challenge.

Purpose of the Study:

• To investigate the challenges in controlling the frequency of neural oscillators compared to other oscillation characteristics.
• To propose a generic mechanism for achieving frequency modulation in neural oscillators.
• To establish a link between oscillation frequency and the geometric properties of the neural oscillator's phase portrait.

Main Methods:

• Analysis of dynamical systems exhibiting periodic behavior.
• Development of a simple linear controller for frequency modulation.
• Investigation of the relationship between phase portrait geometry and oscillation frequency.

Main Results:

• Frequency modulation is inherently more complex than amplitude or shape modulation in neural oscillators.
• A generic method using a linear controller was proposed to achieve frequency control.
• A bidirectional dependency between oscillation frequency and phase portrait geometry was identified.

Conclusions:

• Controlling the geometry of neural state orbits offers a viable pathway for frequency modulation.
• The proposed method enables frequency control by shaping the state space for easy manipulation.
• This approach provides a generic solution applicable to both biological modeling and robotics.