Neural Decoding for Motor and Communication Prostheses

Neural prostheses, which are also termed brain–machine and brain–computer interfaces, offer the potential to substantially increase the quality of life for people suffering from motor disorders, including paralysis and amputation. Such devices translate electrical neural activity from the brain into control signals for guiding paralyzed upper limbs, prosthetic arms, and computer cursors. Several research groups have now demonstrated that monkeys (e.g., Serruya et al., 2002; Taylor et al., 2002; Carmena et al., 2003; Musallam et al., 2004; Santhanam et al., 2006; Mulliken et al., 2008; Velliste et al., 2008) and humans (e.g., Kennedy et al., 2000a; Leuthardt et al., 2004; Wolpaw and McFarland, 2004b; Hochberg et al., 2006; Kim et al., 2008; Schalk et al., 2008) can learn to move computer cursors and robotic arms to various target locations simply by activating neural populations that participate in natural arm movements. Figure 7.1 illustrates the basic operating principle of neural prosthetic systems. Neural activity from various arm movement–related brain regions (e.g., Wise et al., 1997) is electronically processed to create control signals for enacting the desired movement. Non invasive sensors can collect neural signals representing the average activity of many neurons. When invasive permanently implanted arrays of electrodes are employed, as depicted, it is possible to use waveform shape differences to discriminate individual neurons (e.g., Lewicki, 1998; Santhanam et al., 2004). After determining how each neuron responds before and during a movement, which is typically accomplished by correlating arm movements made during a behavioral task with