Humans interact with their environment in countless ways and can switch seamlessly between activities. Even for seemingly simple tasks, a variety of sensory inputs and contextual cues are integrated to create a motor plan to complete a task. In our work, we manipulate aspects of task context to study sensorimotor processing during grasping and object exploration tasks. Our experiments are conducted as part of an ongoing effort to develop intracortical-brain computer interface (BCI) devices for people with paralysis. This experimental paradigm allows us to study sensorimotor processing during complex tasks, which will ultimately lead to the development of improved neuroprosthetics. Study participants use the BCI to control a robotic arm and hand, which allows us to precisely control the mapping between neural activity and movement as well as the somatosensory inputs. Previous work supports the idea that neural activity generated during attempted movements in people with spinal cord injury exhibits the same relationship to movement variables as an able-bodied subject. Similarly, stimulation of somatosensory cortex in a person with chronic tetraplegia generates localized sensations that follow the expected spatial organization.
I will provide an overview of our approach to controlling a motor BCI, which is based on an understanding of motor cortical activity measured in non-human primates during overt reaching and grasping behavior. I will also discuss a primary motivation for our current work that is investigating the influence of task context on motor control. We observed that that motor cortical activity is fundamentally different when reaching to and grasping an object, versus reaching with identical kinematics in an empty workspace. In more recent experiments, we have focused on the motor control of grasping. We have observed encoding of implicit grasp force representations associated with familiar objects at particular phases of object grasping and transport movements. Additionally, even when explicit force targets are instructed, we see transient periods of motor cortical activity that are most informative about grasp force control. Overall, our experiments have shown the motor cortical activity is dependent on the context of a task and is not simply communicating a task-invariant motor plan to the upper limb.