Catherine L. Ojakangas

Brain–Machine and Brain–Computer Interfaces

The idea of connecting the human brain to a computer or machine directly is not novel and its potential has been explored in science fiction. With the rapid advances in the areas of information technology, miniaturization and neurosciences there has been a surge of interest in turning fiction into reality. In this paper the authors review the current state-of-the-art of brain–computer and brain–machine interfaces including neuroprostheses. The general principles and requirements to produce a successful connection between human and artificial intelligence are outlined and the authors’ preliminary experience with a prototype brain–computer interface is reported.

8–12 Hz rhythmic oscillations in human motor cortex during two-dimensional arm movements: evidence for representation of kinematic parameters ☆

Direct cortical recordings were taken from 12 patients with implanted subdural electrode arrays during performance of a 2-dimensional, multi-joint, visually guided arm movement task. Task-related changes in the amplitude of the motor cortex 8-12 Hz surface local field oscillations were evaluated for the encoding of direction and amplitude of movement in the 6 patients in whom no epileptogenic or ECoG background abnormalities were detected over the motor-sensory cortical areas under the recording electrode array. The topography, time of onset and duration of these responses were evaluated in the context of motor cortex somatotopy, as defined by cortical stimulation delivered through the electrode array. Multi-joint arm movements were accompanied by a decrease in the power of the 8-12 Hz frequency components of the ECoG signal. These power changes were spatially distributed over the upper extremity, motor-sensory representation. Movement amplitude influenced the magnitude, duration, and extent of the spatial distribution of ECoG power changes in the 8-12 Hz band. These effects occurred predominantly over cortical areas corresponding to the upper extremity motor-sensory representations. Direction of movement had a weaker influence on the 8-12 Hz frequency components of the ECoG over the upper extremity motor-sensory representations, but influenced the patterns of 8-12 Hz ECoG response on adjacent cortical regions. These results show that the amplitude of surface electrical oscillations generated over the rolandic cortex are correlated with the kinematics of multi-joint arm movements. These changes in the ECoG signal appear to reflect shifts in the functional state of neuronal ensembles involved in the initiation and execution of motor tasks.

Decoding movement intent from human premotor cortex neurons for neural prosthetic applications

Primary motor cortex (M1), a key region for voluntary motor control, has been considered a first choice as the source of neural signals to control prosthetic devices for humans with paralysis. Less is known about the potential for other areas of frontal cortex as prosthesis signal sources. The frontal cortex is widely engaged in voluntary behavior. Single-neuron recordings in monkey frontal cortex beyond M1 have readily identified activity related to planning and initiating movement direction, remembering movement instructions over delays, or mixtures of these features. Human functional imaging and lesion studies also support this role. Intraoperative mapping during deep brain stimulator placement in humans provides a unique opportunity to evaluate potential prosthesis control signals derived from nonprimary areas and to expand our understanding of frontal lobe function and its role in movement disorders. This study shows that recordings from small groups of human prefrontal/premotor cortex neurons can provide information about movement planning, production, and decision-making sufficient to decode the planned direction of movement. Thus, additional frontal areas, beyond M1, may be valuable signal sources for human neuromotor prostheses.

Scaling of the metrics of visually-guided arm movements during motor learning in primates

SummaryHand trajectory, tangential velocity and acceleration, time and distance until peak velocity and reaction time were analyzed during the process of learning a skilled, visually-guided arm movement. Primates were trained to move a cursor with a manipulandum from a start box to target boxes displayed on a horizontal video screen during control conditions and when the relationship (gain) between the cursor and manipulandum was altered. The animals adapted to the altered feedback over 100–200 trials. A subsequent testing phase with randomly interspersed trials using the control gain demonstrated that the animals had modified their movements appropriately for the novel gain. Examination of the kinematics revealed that in adapting to a novel gain, primates scaled movement amplitude, tangential velocity, acceleration, and duration appropriately for the distance the hand needed to travel. Yet time to peak velocity was kept constant. Reaction time also remained unchanged for three of the four animals. Movements were performed in two phases, the first from movement onset to peak velocity and the second from peak velocity until the end of the movement. During the first phase the shape of the trajectory and velocity profile were stereotypic and without evidence of any corrections, consistent with this phase being essentially open loop. However, corrections occurred in the second phase and we propose visual feedback was used to correct for the difference in hand/cursor position. Learning appeared to involve utilizing the errors from previous trials to modify the early feedforward phase of subsequent trials. Peak tangential velocity, total movement duration and distance reached at peak tangential velocity all scaled linearly with the total movement distance required at each gain. Based on regression analyses, for none of these variables were the changes in learning completely adequate to compensate for total distance required. However, distance to peak velocity scaled with peak velocity in relation to the control gain. The results show that non-human primates adopt a consistent strategy when learning to scale a multi-joint movement. The metrics of the movement scaled yet the time to peak velocity remained constant, suggesting independent control of time and amplitude. Keeping time to peak velocity constant as well as the scaling of peak velocity with distance to peak velocity are viewed as ways to simplify the learning process.