Abstract Cooperative or joint actions involve two or more agents coordinating their behavior in space and time to perform a particular task. Joint actions are an important component of interaction and cooperation in both the social and motor domains however surprisingly little is known about the planning and control of these actions, particularly their neural correlates. The work proposed here is aimed at providing preliminary data for a federal grant application focused on characterizing the neural correlates of one of the most fundamental and commonly performed joint actions, object handovers, in the sensorimotor cortices of awake, behaving monkeys. In addition to expanding our knowledge of joint actions and the sensorimotor systems of the brain, the proposed research has significant potential to impact state and national needs in the consumer, healthcare, military, and industrial settings by advancing the fundamental engineering and neuroscience knowledge necessary to create the next generation of neural prosthetic and rehabilitative systems, specifically those employing robots as cooperating agents. In the healthcare domain, such systems are expected to be in widespread usage in the next few decades, and will be critical for improving the quality of life of Arizonans with disabilities resulting from stroke, Parkinsons disease and other conditions. The research proposed here will provide preliminary data for the resubmission of a NSF Collaborative Research in Computational Neuroscience (CRCNS) proposal, originally submitted in the fall of 2012. The reviewers of that proposal acknowledged that joint action is a novel and important topic, and noted that our research plan was innovative and ambitious and possessed the potential for transformative advances in the field of brain-machine interfaces (BMI). They also noted that the already established collaboration between the PIs suggested a high likelihood of success. However, the reviewers also expressed concern that the declared main goal of the research (characterizing the mechanisms of joint action) would be difficult to achieve with our original paradigm. As a result, we have substantially revised our approach to address this criticism and others. In our revised federal application we will first train animals to perform an object handover task with a human subject (Aim 1). We will then model the kinematics of the arms during this task and use these models to develop control laws for robots that will interact with the animals in subsequent aims. These models and control laws will facilitate characterization of the neural representation of an object handover task performed between the robot and the animals (Aim 2). Lastly, we will use neural activity recorded from the animals to implement a neurally-controlled handover task (Aim 3). Here the neural activity characterized in Aim 2 will be mapped directly to the kinematic models developed in Aim 1 and transformed to control laws that will be applied in real time to direct motion of a robotic arm during performance of a handover task with the animals. The aims of our revised federal proposal are described in detail below, with Aims 1 and 2a being the focus of any efforts funded by ABRC.
|Effective start/end date||10/23/14 → 4/22/16|
- Arizona Department of Health Services: $100,000.00