Non-causal spike filtering improves decoding of movement intention for intracortical BCIs

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interfaces (BCIs) utilize neural signals for device control.
• Intracortical spiking neural signals offer high-fidelity information for BCIs.
• Current methods extract information using causal filtering and spike detection.

Purpose of the Study:

• To investigate the impact of non-causal filtering on neural signal information content.
• To improve the decoding accuracy of intended movement direction in BCIs.
• To assess the real-time applicability of non-causal filtering for BCIs.

Main Methods:

• Replaced causal filters with equivalent non-causal filters for neural signal processing.
• Applied a 4ms lag for real-time processing in brain-computer interface applications.
• Extracted threshold crossing events from filtered neural signals.

Main Results:

• Non-causal filtering significantly increased information content compared to causal filtering.
• Decoding of intended cursor direction improved, with reduced angular error.
• Demonstrated effectiveness across multiple sessions and participants with tetraplegia.

Conclusions:

• Non-causal filtering is a simple yet effective method for processing intracortical neural signals.
• This approach enhances neural signal conditioning for direct control of external devices via BCIs.
• The findings suggest a promising advancement for assistive technology development.