Neuroprosthetics

1
Neuroprosthetics Brain-Machine
Interfaces An Integration of Technology and
Neuroscience

• Jessica Blose

2
Basic Motor Areas

• Frontal Lobe
• - Prefrontal Cortex- Plans movements
• - Premotor Cortex- Organizes sequences
• - Motor Cortex- Produces specific movements
• Parietal Lobe
• - Processes and integrates somatosensory and
  visual info for guiding movement
• Spinal Cord and Corresponding Motor and Sensory
  Neurons

3
http//www.sirinet.net/jgjohnso/brainpics.html
4
Loss of Motor Control

• Spinal Cord and Neuron Damage
• Brainstem Stroke
• Neurodegenerative Disease
• Amputation

5
Early Neuroprosthetic Developments

• Freehand System
• - implanted electrodes w/in muscles can be
  stimulated by movement of other muscles

http//remote-ability.com/unique/freehand.htm
6
X-rays of inserted electrodes
http//remote-ability.com/unique/freehand.htm
7
Recent Neuroprosthetic Developments

• Brain-Machine Interfaces (BMIs)
• - an interface in which the brain accepts and
  controls a mechanical device as a natural
  representation of the body
• Types
• - Surface Recordings
• - Intra-cortical recording

8
BMIs

• Recordings of extracellular activity of frontal
  and parietal cortical areas
• Activity of neuronal populations is processed by
  simple mathematical models designed to extract
  motor-control parameters from the raw brain
  signals.
• The outputs of these models are used to control
  the movements of a robot arm.

9
Surface Recordings

• Electroencephalographs (EEGs)
• Detect electrical signals of underlying neurons
• - Use of specific components of EEG signals
• ex. µ-rhythm
• Technical Limitations
• - Low info rate of 20-30 bits/min

10
http//www.cs.colostate.edu/eeg/Data
EEG Equipment includes Computer program, surface
electrodes, and amplifier/recorder.
11
http//www.cs.colostate.edu/eeg/Data
12
Schematic representation of a cortical
neuroprosthetic device.   Nature Reviews
Neuroscience 4 417-422 (2003)
13
BMIc

• Visual or haptic feedback produces a closed loop
  BMI.
• This is known as BMIc

1999 SPIE - The International Society for Optical
Engineering
14
Monkey vs. Rats

• Evolutionarily similar
• Similarities in brain structure and function

15
Carmena et al. Study

• Two adult female Mocaca mulatta monkeys
• Implanted areas included dorsal premotor
  cortices, supplementary motor areas, and primary
  motor cortices in both hemispheres
• Monkey 1 additional implantation in primary
  somatosensory cortex
• Monkey 2 additional implantation in posterior
  parietal cortex

16
Insertion of Electrodes
17
Electromyograms

• EMGs
• Recording of electrical activity of the muscles
  and electrical response of the peripheral nerves
• Recordings of wrist flexors extensors, biceps

18
Motor Parameters

• Hand position
• Velocity of Movement
• Gripping Force
• Direction of Movement/ Hand Trajectory

19
Pole Control Mode

• Tasks w/ joystick and computer screen
• 1) Reaching - moving cursor to target
• 2) Hand-gripping- force required to increase
  cursor size to match target size
• - provides visual feedback
• 3) Reach-and-grasp- combo of 1) 2)
• Incorporation of robotic movement

20
Closed-loop control brainmachine interface for
motor control monkeys.

• Artificial visual feedback signals inform animal
  about performance of robot arm controlled by
  brain-derived and mechanical signals.
• Moving cursor on a video screen to inform animal
  about the position of robot arm in space.
• ? size of cursor to inform animal about robotic
  gripping force.

21
Pole Cont.

• Computer tracked bioelectrical activity patterns
• Specific patterns corresponded with specific
  commands
• Computer learned to read monkeys minds

22
Brain Control Mode

• Researchers disconnected joystick
• Robotic arm movement depended on brain activity
• Computer decoded brain signals and sent
  appropriate instructions to mechanical device

23
Brain Cont.

• Initially relied on joystick for movements
• Learned that thoughts could control movements
• IMPORTANT 1st time gt 1 motor parameter could be
  activated by single ensembles

24
Questions Answered

• How many neurons from which to record?
• - larger neuronal ensemble is better
• Which cortical areas to sample?
• - any part of frontalparietal area
• Can motor parameters be predicted in real time
  with less accuracy in brain signal measurement?
• - Multiunit signals can be used instead of the
  more specific single unit

25
Additional Results

• Both long- and short-term cortical changes
  occurred
• - of single neurons contributing increased
  across motor areas
• ? in neuronal direction tuning from pole
  control to brain control

26
Experimental design used to test a closed-loop
control brainmachine interface for motor control
in macaque monkeys.   Nature Reviews Neuroscience
4 417-422 (2003)
27
Implications and Complications

• Brain Plasticity
• - changing/adapting to incorporate new devices
• Limited Neuronal Input
• - How many neurons can and need to be recruited
  to perform smooth, complicated movements?

28
Current Future Research

• Total control of prostheses/devices with brain
  implants
• Wireless or Portable BMIs
• Direct transmission of brain commands to muscles
  in latent limbs
• Electrode implant incorporation of larger neural
  populations

29
(No Transcript)
30
A Big Thank You to Matt!