Lecture Seven;Brain Machine interface part 2

Question 1

What was the focus of the Nicolelis study?

Answer 1

Electrophysiological paradigm capable of simultaneously recording the EC activity of large populations of single neurons

i.e a method to study multiple neurons at once not individual ones

Question 2

What were the results of the nicolelis study;

Answer 2

100 individual neurons were able to be recorded from 48 microwires implanted in the brainstem, thalamus and somatosensory cortex in rats

Question 3

What were the conclusions of the nicolelis study?

Answer 3

Every neuron has their own Baseline, modulation depth and preferred direction

Different neurons have different wave forms and it is therefore possible to differentiate neurons.

From movement direction, firing rates, we can get tuning parameters.

Question 4

Describe the purpose of the taylor study;

Answer 4

3D movement of neuroprosthetic devices can be controlled by the activity of cortical neurons when appropriate algorithms are used to decode intended movement in real time.

Question 5

Describe the set up of the taylor study;

Answer 5

Subjects had real time visual feedback of their brain controlled trajectories

They had to move a prosthesis to the 8 corners of the cube.

Question 6

What were the conclusions of the taylor study;

Answer 6

Cell tuning properties changed when used for brain controlled movements. By using control algorithms to track these changes, subjects made long sequences of 3D movements using far fewer cortical neurons than expected. Daily practice improved movement accuracy and the directional tuning of these units.

Question 7

How were the algorthims calibrated for a tetraplegic? (typically would use functioning arm)

Answer 7

Co-adaptation

Random parameters -> no movement of mouse -> record subject thinking of movement -> Use this to improve the parameters -> Repeat this each time giving more free movement of the cursor until no restrictions and the subject can accurately move the cursor.

Question 8

Describe what the taylor study found when comparing HC (hand controlled) and BC (brain controlled) movements of a mouse over time.

Answer 8

With practice the preferred direction of neurons for HC movements changed with practice to a greater extent than BC neurons.

i.e neurons are adapting to the control scheme (cursor movement)

BC neurons become more accurate, performance improved

Question 9

What did the taylor study show regarding cosine tuning?

Answer 9

Neurons arent exactly cosine tuned

Question 10

What did the jones study show?

Answer 10

How new electrodes are silicon based with 100 needle electrodes

UTAH electrode, low resistance very good

Question 11

Describe the Hochberg 2006 study set up;

Answer 11

Aim ; To translate preclinical results from intact animals to a clinically useful NMP, movement signals must persist in cortex after spinal cord injury and be engaged by movement intent when sensory inputs and limb movement are long absent.

i.e checking that a tetraplegic still encodes normally for movement

Question 12

What was found in the hochberg 2006 study?

Answer 12

• Demonstrated that intended hand motion modulates cortical spiking patterns three years after spinal cord injury.
• These neurons also showed directional tuning
• They could record a single neuron
• He could control a cursor
• Consistant in response
• Also managed to control a tv while conversing and a multi jointed robotic arm.

Question 13

Describe the velliste study;

Answer 13

Application of a prosthesis to a monkey.

A system that permits embodied prosthetic control , showing that monkeys use their motor cortical activity to control a mechanised arm replica in self feeding task

The arm was 4d in that it also gripped too.

Question 14

What did the vistille study found?

Answer 14

The subjects cortical area controlled arm movement and proportionally controlled gripper strength

i.e monkey could control movement and grip in real time, big step towards human prosthesis

Question 15

What are the two possible neural decoding technologies?

Answer 15

Population Vector Algorithm (PVA)

Optimal Line Estimator (OLE)

Question 16

Describe the fundamentals of decoding technology;

Answer 16

• Dozens of simultaneously recorded neurons can be used to control movement of a robotic arm.
• Therefore algorithms are needed by which motor intention can be inferred

Question 17

Describe PVA;

Answer 17

• Activity of each cell is used to weight a vector pointing in the neurons preferred direction.

Question 18

What is the problem with PVA?

Answer 18

Offline it is possible to yeild more accurate hand movement reconstruction using OLE than PVA

i.e PVA is biased / not uniform in its predictions

This is called open loop performance

This difference is less present on-line closed loop.

Question 19

What did the chase study show regarding OLE vs PVA;

Answer 19

1. The performance gain with certain decoders can be far less extreme than predicted off line results
2. Subjects are not able to compensate for certain types of bias in decoders
3. Care must be taken to ensure that estimation error does not degrade the performance of the theoretical optimal recorders

OLE has less error the PVA

Question 20

Describe the holchberg 2012 study;

Answer 20

The use of a neural interface system in two tetraplegic people to perform 2d movements including grasping using a robotic arm.

No training

Question 21

Describe the set up of the holchberg 2012 study;

Answer 21

• NO training
• Singals decoded from a 96 channel microelectrode array in Motor cortex.

7D = translation (XYZ), rotation, aperature

Question 22

What were the conclusions of the holchberg 2012 study?

Answer 22

Robotic grasp and actions were not as fast or accurate as an abled bodied person, the results demonstrate a feasibility for people with tetraplegia , years after injury to the CNS to recreate useful multidimensional control of complex devices directly from a small sample of neuronal signals

Question 23

Describe the collinger study set up;

Answer 23

Tested whether and individual with tetraplegia could rapidly achieve neurological control with a high performance prosthetic limb using a brain machine interface

Question 24

Describe the method of the collinger study;

Answer 24

Implanted 2 96 intracortical micro electrodes in the motor cortex, 13 weeks of brain machine interface training with the goal of controlling an antrhopomorphic prosthetic limb with seven degrees of freedom (3d translation, 3d orientation, 1d grasping).

Question 25

What were the findings of the collinger study;

Answer 25

• Patient could freely perform 3d movements on day 2 of training
• After 13 weeks, robust movements were performed routinely.
• Patients efficiency improved over time with 91% accuracy.

Clinically significant gains on the tests of upper limb function

Question 26

What is the next step in prosthetic development?

Answer 26

The use of intracortical microstimulation (ICMS) with input magnitude graded according to the force exerted on the prosthetic finger

i.e creating somesthetic feedback (tactile stimulation) to reduce concentration to perform tasks

Question 27

Describe the tests used to encode sensory substitution devices’ input into the somatosensory cortex;

Answer 27

• Performed mechanical stimulation on a real finger
• Recorded the stimulation in primary somatosensory cortex and correlated the two
• Performed psychometric testing

Then created a model with the same preciseness as human fingers. Simulating the brain thus prosthesis can have feeling.

Question 28

Describe the tests used to encode sensory substitution devices’ input into the somatosensory cortex;

Answer 28

• Performed mechanical stimulation on a real finger
• Recorded the stimulation in primary somatosensory cortex and correlated the two
• Performed psychometric testing

Then created a model with the same preciseness as human fingers. Simulating the brain thus prosthesis can have feeling.