Influence of movement speed on plan activity in monkey pre-motor cortex and implications for high-performance neural prosthetic system design

The success of human motor prosthetics will largely depend on increasing system performance by maximizing the movement-related information that can be recorded from cortical neurons. This will in turn depend on successfully combining information from different neurons, and across times for the same neuron. Here we investigate neural activity that occurs prior to reaching movements, and attempt to understand how this activity varies with the speed of the upcoming movement. We recorded pre-motor cortex neural activity from a rhesus monkey trained to perform delayed-reaches to targets at two different speeds. We found that movement speed information is present in action potential emission rates and 20-40 Hz local field potential power during a "plan" period preceding movement. We then further analyzed the action potential data to determine the accuracy at which trial conditions, including the speed of the upcoming movement, can be estimated from neural data. System performance is expected to increase when plan-period speed information is included in prosthetic estimation algorithms.