Efficient Point-Process Modeling of Spiking Neurons for Neuroprosthesis

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computational Neuroscience

Background:

• Neuroprosthetics aim to restore brain function using implantable devices.
• Modeling neural signal transformations between cortical areas is crucial for neuroprosthesis development.
• Spiking Neural Networks (SNNs) are suitable for modeling brain information processing via spike trains.

Purpose of the Study:

• To propose a novel Spiking Neuron Point-Process Model (SNPM) for neuroprosthesis applications.
• To model nonlinear interactions between cortical areas using spike train inputs.
• To explore the potential of SNPM for low-energy neuromorphic computing and clinical use.

Main Methods:

• Developed a Spiking Neuron Point-Process Model (SNPM) that accepts spike times as input.
• Investigated the model's capability to capture nonlinear interactions between neural populations.
• Evaluated the model's performance in reconstructing functional relationships between specific cortical areas (PMd to M1).

Main Results:

• The proposed SNPM accurately reconstructs functional relationships between the dorsal premotor cortex (PMd) and primary motor cortex (M1).
• The model demonstrates the ability to handle nonlinear interactions inherent in neural signal processing.
• The SNPM is suitable for implementation on neuromorphic chips, suggesting potential for efficient, low-power neuroprosthetic devices.

Conclusions:

• The SNPM offers a viable approach for modeling neural signal transformations in neuroprosthetics.
• This model has significant potential for clinical applications due to its accuracy and compatibility with neuromorphic hardware.
• Further research can leverage SNPM for developing sophisticated brain-computer interfaces and neural prostheses.