Identifying Functional Components of the Motor Cortical Gamma Broadband

Restoring movement to individuals with paralysis is the primary goal of motor brain-machine interfaces (BMIs). BMIs use movement intention signals typically recorded from the primary motor cortex (M1) to drive assistive devices. One of the major limitations of BMIs however, is the duration and stability of the signals recorded from M1. Currently signals can only be recorded for months to a few years, which falls far short of the decades required for successful clinical translation. Two particularly promising signals to overcome this limitation are multi-unit spikes (MSPs) and local field potentials (LFPs), i.e. spatially summed extracellular potentials, recorded from the motor cortex, which have been shown to reliably decode arm movement in order to intuitively control a computer cursor. However, to date there has not been a direct comparison of these two signal sources. In Chapter 1, monkeys used either their hand or fixed decoders of LFP and MSP activity to control a computer cursor over the course of several years and 200 days respectively. We compared the stability of these signals with respect to movement and found they are both very stable over the course of 6 months to several years during brain and hand control.;Looking closer at LFPs, one of the most movement-informative components of LFPs is the gamma frequency band (30-300 Hz). There is, however, a gap in knowledge about the mechanisms that generate the gamma band in motor cortex. This limits our ability to understand motor cortical physiology and also limits the optimal design of a motor BMI utilizing the gamma LFP band. There are two ongoing debates about 1) the relationship between spiking activity and high gamma rhythms and 2) whether the gamma band contains numerous distinct sub-bands or is one monolithic broadband. These features of the gamma band have been largely unexplored in M1. In Chapter 2, I show that the high gamma rhythm is distinct from spiking activity in motor cortex, and in Chapter 3, I show that the gamma band contains two distinct sub-bands. In Chapter 2 and 3 monkeys used a BMI to independently control M1 spiking activity and high gamma rhythms or two M1 gamma sub-bands (high and low gamma), to determine whether spiking is distinct from high gamma rhythms or whether multiple gamma bands exist. The data I show in Chapter 2 and 3 suggests that subjects can differentially modulate spiking activity from high gamma power as well as the power in various motor cortical gamma sub-bands. Spiking activity can be modulated independently from the high gamma 2 band (200-300 Hz) and the high gamma band (130-200 Hz) power can be modulated independently of low gamma band power (30-50 Hz).