Abstract.
Technological developments in the past decade have accelerated the pace
of research in brain computer interfaces. Multiple research groups across
the country are pursuing this area of research as a possible solution
to spinal cord injury. The Andersen lab at Caltech specializes in studying
brain areas in the parietal cortex, which is associated with vision
and motor planning, and in particular the Parietal Reach Region (PRR)
which encodes the plan for the next intended reach movement, which is
markedly different than the approach taken by other research groups
which are using the motor cortex as the source of control signal. The
Cortical Prosthetic Project at the Andersen lab has multiple research
areas, including the development of an implantable chip to read signals
from the parietal cortex, development of computational models for the
neural signals involved, development of an online decoding algorithm
for the intended movements, and finally the implementation of the real
time control of a robotic arm through a brain computer interface.
This project seeks to investigate the encoding of depth by PRR neurons
by carrying out experiments that in essence characterize the system.
The first experiment will involve training non-human primates to maintain
fixed eye positions while reaching to targets at various locations in
three dimensional space. The second experiment will have the primates
vary eye positions, however maintain fixed reach locations. Subsequently,
we will investigate the neural mechanism by which PRR neurons encode
the intended three dimensional reach location and develop a computational
model to simulate the process. Lastly, we will augment the online decoding
algorithm that is under development to decode PRR signals from implanted
arrays in non human primates to control a robotic arm in real time to
make reaches to locations in three dimensional space.
Parietal cortex is generally associated with processing visual information
to plan motor actions. In particular, the posterior parietal cortex
is well known for its role in visuo-motor coordinate transformations.
This region of the brain contains specialized areas that perform these
coordinate transformations in order to plan motor actions in a common
reference frame. Studies conducted in this lab show that the Parietal
Reach Region (PRR) is involved in planning reaches, and integrates head
and eye position signals (using gain fields) to encode the next intended
reach in eye centered (or retinal) coordinates, which is advantageous
for a number of reasons including the facilitation of hand eye coordination.
Currently, a model for 2D reach target location has been developed,
and an online Bayesian decoding algorithm for the next intended reach
has been implemented. The model uses the firing rates of PRR neurons
with different receptive fields to predict the location of the intended
reach by using Bayes rule to find the most likely target location (currently,
only discrete locations are used) from the observed firing rates (see
Figure 1). The next step in this project seeks to assess how a) reach
depth with constant fixation depth, and b) fixation depth with constant
reach depth modulate the PRR signal. In a sense, through the two types
of modulation above, we hope to characterize the system and develop
a computational model of how 3D target locations for the next intended
reach are coded by Parietal Reach Region neurons.
It is unknown how depth vision and reaching in depth modulates PRR neurons.
When a subject proceeds to fixate an object of interest in the visual
scene, the brain uses various cues to position the eyes correctly and
modify of the lens for focus. For objects within a close of distance
(6ft, and certainly within reach distance), eye position can fully specify
where the subject is fixating, and in particular, the amount the eyes
rotate inward, called the vergence angle, specify where in depth the
subject of fixating (accommodation, or the focusing of the lenses in
the eyes is very rarely out of register with vergence angle). Other
objects at different depths from the fixation depth in the visual scene
have binocular disparity (a difference in position of the image of the
object cast on the two retinas); binocular disparity provides the information
that the brain uses to calculate the depth or distance to these objects
relative to the fixation depth. By presenting fixation stimuli at different
depths, measuring eye position (vergence angle), and presenting reach
targets in depth (with known binocular disparities), we seek to understand
how the PRR signals are modulated by these parameters (See Figure 2).
Eye position is measured with the scleral search coil technique, and
neural recordings will be obtained by the implantation of micromachined
silicon or tungsten electrode arrays. Our experimental setup consists
of a track of LEDs, which will serve as fixation targets, and a 3 axis
Cartesian robot arm with an attached touch sensor as the reach target.
The implantation of the arrays will be performed stereotaxically, with
the aid of an MRI to localize the region of implantation before surgery.
A 32 channel Plexon data acquisition system is in place for recording
from the arrays, and all stimuli (LEDs and robot arm) are administered
by a real time operating system through LabView (National Instruments).
The outlined experiments will provide the data necessary to develop
a computational model of the PRR neurons. The phasic activity of the
PRR neurons are understood and broken into various epochs, wherein the
firing rate during the memory/planning period determines the location
of the next intended reach. The form of the model will be to express
firing rate of a PRR neuron as a function of fixation distance and reach
depth, and will most likely be nonlinear. The general expected result
is that reach depth will be represented as a gain field (Andersen et
al. 1985), wherein the tuning curve of the neuron, or in this case,
two dimensional receptive field in the frontoparallel plane, is simply
multiplied by a constant proportional to the reach depth. This is the
result found in a close by brain area also located in the intraparietal
sulcus, LIP, which signals the next intended eye movement (Gnadt and
Mays, 1995). Currently, it is unknown how fixation depth will be represented
in the activity of the PRR neurons, however it is likely that it results
in gain modulation as well; we expect that the firing rate function
will be nonlinear, and that the gain modulation will be nonlinear as
well. Additionally, we will attempt white noise analysis, or the reverse
correlation technique to plot a three dimensional receptive field for
a PRR neuron, however this can only be done by holding one parameter
constant (such as fixation distance, so [reach] depth becomes relative
to this).
Once a computational model is verified, the final goal of this research
is to augment the online decoding algorithm to decode three dimensional
intended reach location. This will then be used to control a prosthetic
device to physically move to the location of the intended reach. In
order to accomplish this, two robot arms will be in place, one to present
the stimulus and the other for the primate to control with its thoughts
(PRR online decoding controls the arm). The success of these experiments
will then allow the implantation of electrode arrays in patients suffering
from spinal cord injury in an attempt to control prosthetic devices.
This project advances neuroscience and the understanding of the Parietal
Reach Region as well as integrating math and engineering techniques
for the development of the computational model. In addition, the research
involved in developing an implantable chip for recording signals furthers
the ultimate goal of the robust and dependable brain computer interface.
Most importantly, the efforts in truly understanding the neural signal
to develop a real time decoding algorithm is the most important aspect
in this project because acceptable reliability and accuracy has not
been achieved by other methods in this area of research, and is critical
to the success of any cortical prosthetic project.
References
Andersen RA and Buneo CA. Intentional Maps in the Posterior Parietal
Cortex. Annual Review of Neuroscience, 2002.
Batista AP and Andersen RA. The parietal reach region codes the next
planned movement in a sequential reach task. Journal of Neurophysiology
85: 539:544, 2001.
Buneo CA, Jarvis MR, Batista AP and Andernes RA. Direct visuomotor transformations
fo reaching. Nature 416:632:636, 2002.
Cumming BG and DeAngelis GC. The physiology of stereopsis. Annual Review
of Neuroscience 24: 203-238, 2001.
Gnadt JW and Mays LE. Neurons in mokey parietal area LIP are tuned for
eye movement parameters in three dimensional space. Journal of Neurophysiology
73:280-297, 1995
Meeker D, Shenoy KV, Cao S, Pesaren B, Scherberger H, Jarvis M, Buneo
CA, Batista AP, Kureshi SA, Mitra PP, Burdick JW, and Andersen RA. Cognitive
control signals for prosthetic systems. Society of Neuroscience Abstracts
27, 2001.
Rosenbluth D and Allman JM. The effect of gaze angle and fixation distance
on the responses of neurons in V1, V2, and V4. Neuron 33:143:149, 2002.
