Brain Computer Interface Mina Mikhail minamohebn@gmail.com
Introduction Ways for controlling computers Keyboard Mouse Voice Gestures Ways for communicating with people Talking Writing Gestures
Problem Shortage of the current ways of interaction Require muscle movements Disabled people Totally paralyzed people are estimated to be 2 cases per 100,000 each year Amyotrohic lateral sclerosis (ALS) This raises the need of a new way of communication
Brain Computer Interface Direct Neural Interface or Brain-Machine interface An interface between the human brain and computers A New communication Channel
BCI Misconceptions Cannot read thoughts Cannot write to the brain Cannot repair injured areas Cannot operate without your will
Human Brain The Ultimate Parallel Machine! Billions of neurons require a lot of energy. 15% of the cardiac output 20% of total body oxygen consumption 25% of total body glucose utilization. Energy consumption for the brain to simply survive is 0.1 calories per minute and 1.5 calories per minute during crossword puzzle
Brain Regions Frontal Lobe Primary motor cortex, Frontal Eye, information processing, Parietal Sensory information, taste, pressure, sound, temp.. Occipital Visual processing center Temporal Auditory processing Frontal lobe Temporal Parietal Occipital
Human Brain Whenever a neuron is active, its voltage changes
Human Brain Million of neurons fire together Each mental state produces a distinct pattern of electrical activity
Measuring Brain Activity
Measuring Brain Activity Positron emission tomography A radio Active isotope is injected into the subject s blood Isotopes emits positrons Advantages High spatial resolution Disadvantages Expensive Low time resolution Not portable
Functional Magnetic Resonance FMRI depends on blood flow It measures the magnetic properties of the hoemoglobin Active neurons consume oxygen carried by hemoglobin Advantages High spatial resolution Disadvantages Expensive Low time resolution Not portable
EEG Measures the electrical activity of the neurons. Advantages High time resolution Cheaper portable Disadvantages Low spatial resolution Still not user friendly
EEG Montage 10-20 system An international system that describes and applies the location of the electrodes
Rhythmic Activity Delta Band < 3 Hz Deep sleep Theta Band 4-7 Hz Drowsiness and meditation Alpha Band 8-12 Hz Awake Beta Band 13-30 Hz Concentration and thinking
BCI Categories
BCI Categories
BCI Categories
Differences Electrode placements Number of electrodes Number of trial before taking a decision
General Approach
Signal Acquisition EEG CAP Bioamplifier Electrodes Active electrodes Conductive gel Impedance Checker
Signal Preprocessing Artifacts Technical Artifacts Line noise Electrode Artifacts
Signal Preprocessing Physiological Artifacts Eye Blinking artifacts Eye movement Artifacts Muscle Activity artifact
Signal Preprocessing
Methods for Artifact Rejection Filters Artifact Rejection Artifact Subtraction (using EMG sensors) Blind Source Separation Independent Component Analysis (ICA)
Feature Extraction (frequency Frequency Domain Features domain) FFT, wavelets, finite impulse response EEG Frequency Band Power most of the times a measure of event related desynchronization (ERD) is used
Feature Extraction (Time domain) Spatial Domain Feautres Hjorth parameters Three parameters are used to characterize the EEG Activity (mean power) Mobility (mean frequency) Complexity
Classification Bayes Classifiers Support Vector Machines Artifical Neural Networks
Application Wheel Chair Controlling Cursor Controlling OS Word Processing
Research Labs Graz Brain Computer Interface BCI Research at Alberta University BCI Research at Oxford University Berlin Brain Computer Interface Computer Vision and Multimedia Laboratory Geneva University
BCI Systems