Understanding reward-to-reward transformations in the prefrontal cortex
In social interactions, an action value often critically depends on potential reward delivered to self as well as potential reward delivered to other. In such situations, the brain must distinguish reward outcomes delivered to self and others to compute a relative reward value. Metaphorically to spatial and temporal representations in the brain, this mechanism could be approached from the framework of “reference frame transformation” in individual reward-processing neurons.
Reference frame transformations in the macaque parietal reach region (PRR)
In order to make a voluntary movement to a visual location, the brain needs to solve a non-trivial problem of converting target location represented under one coding scheme (e.g., visual) to another (e.g., motor). In case for visually-guided reaching, the brain must transform a representation of target location relative to the retina into a new representation relative to a particular limb, and then ultimately relative to the joint angles and muscular contractions of the limb. This conversion process (taking place between “sensory” and “motor” coordinates) is known as the sensorimotor transformation. This transformation involves multiple visuomotor areas and many different computations in a distributed cortical reach network. In particular, converging evidence suggests a critical role of the primate posterior parietal cortex in transforming target representation from eye-centered to arm-centered coordinates.
With Larry Snyder, I have studied the neuronal mechanisms underlying the sensorimotor transformation in the parietal reach region (PRR), located in the posterior parietal cortex. We explored whether PRR neurons represent target location in a limb-specific fashion (Chang et al., 2008, Journal of Neuroscience). We recorded the activity of neurons while monkeys planned and executed reaching movements using either the contralateral or ipsilateral limb. Some PRR neurons selectively represented target location for the contralateral limb but not the ipsilateral limb, whereas others represented target location indistinguishably for either the contralateral or ipsilateral limb. Only few cells selectively represented target location for the ipsilateral limb. The activity of PRR neurons were inversely correlated with reaction times of the contralateral limb, but not the ipsilateral limb. Surprisingly, this contralateral limb-specific reaction time correlation was also present in the cells that indistinguishably represented target location for both limbs. Based on these findings, we suggest that PRR operates as a limb-dependent stage of the coordinate transformation. We are currently investigating reference frames employed by PRR neurons, and how various information represented in PRR is used to transform an eye-centered target information into an arm-centered motor error.