Towards hybrid brain-computer interfaces

A promising future direction for neural engineering will be the
development of implantable electronic circuitry that continuously
interacts with the brain. Today, brain-computer interfaces and
brain-machine interfaces allow neural activity to control computer
cursors or external devices. In contrast, a recurrent brain-computer
interface (R-BCI) can generate activity-dependent output that is
directly fed back into the nervous system or muscles. Such R-BCIs
implement an artificial connection that the adaptive brain could learn
to incorporate into normal function. We are investigating an
implantable R-BCI called “Neurochip” that includes amplifiers, a
programmable computer chip and stimulator, and is connected to
electrodes in motor cortex and/or muscles. The neural activity can be
converted in real-time to deliver activity-contingent electrical
stimuli back to the cortex, spinal cord or muscles. This autonomous
recurrent paradigm opens new experimental directions and has promise
for several clinical applications.

First, the artificial feed-back loop could implement a long-term
recurrent connection that the adaptive brain could learn to utilize.
Clinically, such R-BCIs could bridge impaired biological connections
and allow the subject to use this prosthetic connection to compensate
for lost pathways. We are currently testing monkeys’ ability to use
motor cortical cell activity to control functional electrical
stimulation of muscles paralyzed by peripheral nerve block.
Surprisingly, neurons could control goal-directed stimulation equally
well regardless of whether they were originally modulated with
movement or not. These results demonstrate that direct artificial
connections between single cortical cells and muscles can be used to
compensate for interrupted physiological pathways and restore
volitional control of movement to paralyzed limbs. So far this
paradigm has been implemented primarily with conventional rack-mounted
instrumentation, but autonomous operation with the Neurochip will
allow longer practice sessions and improved accuracy. The degree to
which prolonged implementation of these artificial connections will
allow the brain to integrate them into normal movements remains to be
tested.

A second consequence of the R-BCI paradigm is the long-term
modification of physiological connections through Hebbian mechanisms.
By delivering stimuli synchronized with cell activity, continuous
operation of the Neurochip has produced significant long-lasting
changes in neural connections in motor cortex. Remarkably, these
changes, as measured by cortical output, remained stable for at least
10 days following the end of conditioning, indicating a potent
long-lasting effect. Such R-BCI-induced plasticity has potential
therapeutic application to strengthen weak biological connections and
facilitate recovery from stroke or traumatic brain injury.

Clearly, the R-BCI paradigm has numerous possible applications,
depending on the input signals (neuronal action potentials, field
potentials, ECoG, EMG), and the output sites (cortex, spinal cord,
cerebellum, muscles, reinforcement sites), and the transform between
them (direct conversion to proportional stimulation, computed
functions of detected activity, simulated neural networks). Looking
ahead, the possibility of transforming spatiotemporal patterns of
multichannel neural recordings to patterns of multichannel stimulation
with custom VLSI may make it possible to implement recurrent
computations in higher-order cognitive areas of the brain, conceivably
producing a “cognitive prosthesis”. Ultimately, successful
incorporation of these hybrid BCIs into brain function will depend on
the brain’s proven ability to adapt to consistent
contingencies. These numerous possibilities, combined with inevitable
technical advances, promise to make the R-BCI a productive paradigm
for neural engineering, basic research and clinical applications.

Biography

Eberhard Fetz received a B.S. in Physics from Rensselaer Polytechnic
Institute (1961) and a Ph.D. in Physics from the Massachusetts
Institute of Technology (1967). Since 1968 he has been at the
University of Washington, most recently as Professor of Physiology and
Biophysics and Core Staff member of the Washington National Primate
Research Center. His research has investigated the neural control of
limb movement in primates. This has included studies of monkeys’
ability to volitionally control the activity of brain cells and
muscles with biofeedback; the functional organization of motor cortex
cells controlling forearm muscles, particularly output cells
identified by correlation techniques; interactions between cortical
neurons identified by cross-correlation and in vivo intracellular
recording; the response properties and correlational linkages of
cervical spinal cord interneurons during volitional hand movement;
artificial neural network modeling; and most recently, the development
of an implantable recurrent brain-computer interface.