Summary: nces in Cognitive Neural Prosthesis 177
11.3 Advances in Cognitive Neural Prosthesis
The motor-based approach, although predominantly used, is certainly not the only way
of using brain data for neuroprosthetic applications. Shenoy et al. (2003) argue that neural
activity presentbeforeor even without naturalarm movementprovides an importantsource
of control signals. In nonhuman primates, these types of neural signals can be found,
among other areas, in parietal reach region (PRR) of the posterior parietal cortex (PPC).
PPC is an area located at an early stage in the sensory-motor pathway (Andersen et al.
(1997)), and is involved in transforming sensory inputs into plans for actions, so-called
"sensory-motorintegration."In particular,PRR was shown toexhibitdirectionalselectivity
with respect to planned reaching movements (Snyder et al. (1997)). Moreover, these
plans are encoded in visual coordinates (also called retinal or eye-centered coordinates)
relative to the current direction of gaze (Batista et al. (1999)), thus providing extrinsic
spatial information and underscoring the cognitive nature of these signals. We refer to this
approach to neural prostheses as "cognitive-based." The human homologue of PRR has
recently been identified in functional-magnetic-resonanceimaging experiments (Connolly
et al. (2003)).
The cognitive-based approach to neural prostheses does not require the execution of arm
movements; its true potential lies in assisting paralyzed individuals who are unable to