Independent Mobility Achieved through a Wireless Brain-Machine Interface

Abstract

Individuals with tetraplegia lack independent mobility, making them highly dependent on others to move from one place to another. Here, we describe how two macaques were able to use a wireless integrated system to control a robotic platform, over which they were sitting, to achieve independent mobility using the neuronal activity in their motor cortices. The activity of populations of single neurons was recorded using multiple electrode arrays implanted in the arm region of primary motor cortex, and decoded to achieve brain control of the platform. We found that free-running brain control of the platform (which was not equipped with any machine intelligence) was fast and accurate, resembling the performance achieved using joystick control. The decoding algorithms can be trained in the absence of joystick movements, as would be required for use by tetraplegic individuals, demonstrating that the non-human primate model is a good pre-clinical model for developing such a cortically-controlled movement prosthetic. Interestingly, we found that the response properties of some neurons differed greatly depending on the mode of control (joystick or brain control), suggesting different roles for these neurons in encoding movement intention and movement execution. These results demonstrate that independent mobility can be achieved without first training on prescribed motor movements, opening the door for the implementation of this technology in persons with tetraplegia.

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1. Setup Description and Electrode Locations.

Animals were trained to move a robotic platform (right) using a joystick. The joystick was spring-loaded, returning to the center position when released. Movement of the joystick was restricted to left, right, and forward movements, while not allowing for diagonal movements. The robotic platform rotated in place in the counter-clockwise and clockwise directions with left and right joystick movements, respectively, and moved forward with forward joystick movements (bottom-middle inset). Movement commands from the joystick reached the platform serially, so at any point in time only one command was executed. Joystick movements were translated to platform movements in discrete states, such that if the joystick moved past a threshold, the platform would move with a fixed speed after initial acceleration. Multiple microelectrode arrays were implanted in the arm and hand areas of primary motor cortex (top middle).

Fig 2. Performance under different modes of control.

(A) Accuracy of decoder, defined as the proportion of decoded directions that matched the target location, in the single-movement task (chance performance 25%). (B) Success rate, defined as the percentage of trials in which the animals reached the reward location within 15 seconds, in the single-movement task (chance performance ~0%). (C) Average time that the animals took to reach the target during correct trials in the single-movement task. Error bars represent the standard error of the mean, and asterisks denote results that were significantly different from those of the Joystick Control task (blue bars, t-test, pRecalibrated Decoder (red lines) in the free-movement task that required the monkeys to move sequentially through a series of targets. The gray circles represent target locations. Animals controlled the platform continuously from the start until the end point. Trajectories were collected during a single experimental session.

Fig 3. Linear discriminant space for a sample session.

Projection of the firing rates onto two largest components of the linear space for the population of neurons collected during a sample BMI Control session (see detailed Methods).

Fig 4. Neurons exhibited selectivity for movement direction and mode of control.

(Pie chart) percentage of cells with different response profiles. Cells were categorized as selective based on the activity 500–1500 ms after the trial-start cue (left, forward, right, or stop; one-way ANOVA p<0.01). In our sample, 25% of cells showed no selectivity (red), 22% showed selectivity during BMI Control only (orange), and 9% showed selectivity during Joystick (motor) Control only (green). The rest of the cells (45%) showed selectivity during both modes of control. This last group was further subdivided into cells where activity during Joystick and BMI Control showed no significant differences (24%, purple) and cells where at least one category of movement was significantly different between Joystick and BMI Control (21%, light blue). (Bar plots) (A–E): Mean firing rates of example cells with different response profiles. Colored bars represent the activity during Joystick Control and gray bars during BMI Control. Error bars represent the standard error of the mean across trials.