BBCI Workshop 2009 - Advances in Neurotechnologies

About Learning, Predictions and Adaptivity of Brains and Machines

author:Eilon Vaadia, Department of Physiology, Faculty of Medicine, The Hebrew University of Jerusalem
published: Aug. 10, 2009, recorded: July 2009, views: 167

Turn off the lights

You might be experiencing some problems with Your Video player.

Slides

Slides
0:00	Learning and control of sensorimotor Functions in Motor Cortical Fields
0:41	Acknowledgments
0:51	The Edmond and Lily Center for Brain Sciences
1:52	Outlines
3:42	René Descartes 1569 - 1650
5:44	The Sensory-motor Loop
6:22	Quote
7:08	Action is essential for learning and perception
8:48	Modeling the Action-Perception Loop: “Internal Models”
11:21	The Significance of Previous Knowledge
12:07	Action-perception loop
12:56	Kalman Filter
14:23	Outlines
14:39	Billions of nerve cells in many brain regions orchestrate behavior (1)
15:59	Billions of nerve cells in many brain regions orchestrate behavior (2)
17:04	Single Units Recordings
18:09	The Voice of a single neuron in motor cortex
18:48	The “Center Out” Task
19:16	Directional Tuning of Single cells in motor cortex
19:34	Hypothesis: 1.Each neurons has a “preferred direction”
20:04	2. A Population of neurons encodes accurately movement direction
20:47	Simultaneous Recordings allow on-line inference of movements from Neuronal Activity: Example : OLE* / Linear Regression
21:44	Multi channel recordings of brain signals
22:39	Simultaneous Recordings and Spikes Sorting
23:19	Simultaneous Recordings of ~200 Spikes Channels
23:52	Local Field Potentials - LFP
24:08	Outlines
24:33	Sensorimotor Learning
24:46	The hypothesis: The population code is shaped by learning
25:18	Three learning tasks
25:58	In All Experiments: Delayed Reaching Tasks
26:43	1. Learning Reaching: More tuned cells as learning
28:13	2. Learning Visuomotor Rotation
28:58	After learning Visuomotor rotation of 45°
29:31	Increased firing rate in selected population during learning Single units Activity changes only in the preparatory period
30:48	Single units Trial-to-Trial Variability : Modulation only in the preparatory period!
31:39	Modulation of Local Field Potentials Gamma Excess – Only During movement execution!
32:43	LFP and Spikes reflect different aspects of Learning?
33:22	After Learning: memory traces?
33:24	After Learning: Better Representation of Learned movements
33:35	Intermediate Conclusions
34:18	Outlines
34:29	Neuronal signals
36:12	BMI action-perception loop
36:45	Model Requirments: Choosing a BMI algorithm (1)
37:47	Model Requirments: Choosing a BMI algorithm (2)
40:52	Kernel Auto Regressive Moving Average (KARMA):
42:17	KARMA: General scheme
42:22	Making KARMA Adaptive
43:55	Experiments
44:07	Outlines
44:12	3D instructed-delay with continuous target-totarget reaching
44:59	Arm Control
45:46	Real Time tracking of hand movements by KARMA
48:46	Adaptivity model is helpful in tracking
50:53	Adaptive Models are Useful for These Brains
50:55	Brain Control – Reaching Targets in 3-D
51:42	First days of BMI adaptation
54:28	Switching from Arm control to Brain Control is fast!
54:55	Adaptive model performs better in Brain control
56:27	Learning novel visuomotor task using a BMI
58:28	Target Rotation Task
59:37	Learning anew in BMI - Target Rotation Task
61:02	Target Rotation Task
62:45	Gradual emergence of Population responses
64:25	Reach and Rotate
65:00	Summary and conclusions
67:52	More investment in funds and manpower will put us on the yellow brick road
68:07	Acknowledgments
71:18	- Questions

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.

Lecture popularity: You need to login to cast your vote.

Description

Using useful signals from the brain, and useful computer algorithms to improve brain machine interfaces. I will talk about the physiology of motor cortex and the nature of activity of population of neurons in motor cortex during Sensorimotor learning, movement preparation and execution. I will present the approach of internal models of the brain as the basis for learning and perception and use all of the above to show how our current knowledge can facilitate approaches to adaptive brain machine interfaces.