neuralink-decoder skill

Overview

This skill simulates a Brain-Computer Interface that decodes synthetic neural spike activity into 2D cursor movement. It generates cosine-tuned spike trains for a population of neurons and applies a linear decoder to reconstruct velocity. The output includes a printed trajectory for quick inspection.

How this skill works

The simulator creates spike trains for a fixed population (default 64 neurons) using cosine tuning relative to intended movement directions and firing-rate noise. Spike counts are converted into rate vectors which a precomputed linear mapping transforms into x/y velocity estimates. The tool runs a time-stepped loop that simulates spikes, decodes velocities, and reports the resulting cursor trajectory.

When to use it

• Prototyping BCI decoding algorithms without real neural data
• Teaching or demonstrating cosine tuning and population decoding concepts
• Comparing simple linear decoders against more complex methods
• Generating synthetic datasets for offline algorithm testing
• Rapidly validating preprocessing and rate-to-velocity pipelines

Best practices

• Normalize spike rates before applying the decoder to reduce bias from high-firing neurons
• Run multiple simulation seeds to assess decoder variability under noise
• Keep simulation time steps and bin sizes consistent with downstream processing
• Visualize both true intended trajectories and decoded trajectories to spot systematic errors
• Tune the number of neurons and tuning width to match your expected SNR

Example use cases

• Simulate cursor-control trials for a research methods class
• Validate a new Kalman filter by comparing its output to the linear decoder baseline
• Create synthetic training data for a machine-learning decoder when neural recordings are limited
• Quickly demonstrate how cosine tuning produces population-level direction encoding
• Archive or reproduce prior decoding behavior for debugging or presentation

FAQ