Measuring biological activities in human organic structure is based on the electrobiological measurings which are caused by the human bio-potentials. Electrobiological measurings include electrocardiography ( ECG, bosom ) , electroencephalography ( EEG, encephalon ) , electromyography ( EMG, encephalon and muscular system ) , magnetoencephalography ( MEG, encephalon ) , electrogastrography ( EGG, tummy ) , functional Magnetic Resonance Imaging ( functional magnetic resonance imaging, encephalon ) and electrooptigraphy ( EOG, oculus dipole field ) .
The chief three used techniques to mensurate encephalon signals are MEG, functional magnetic resonance imaging and EEG. Each method has advantages and disadvantages leting it to be utile for BCI research. A brief description for each method is in the followers:
There's a specialist from your university waiting to help you with that essay.
Tell us what you need to have done now!
order now
MEG is mapping encephalon activities through entering magnetic field produced by electrical currents inside the encephalon cells. MEG needs expensive and intensive devices moreover the measurings are sensitive to ferromagnetism hence MEG devises should be isolated inside Magnetically Shielded Room -MSR- where this room will insulate all external magnetic field even Earth ‘s magnetic field. MEG is considered as extremely accurate technique and normally it is utile for activities which take less than 10 msecs.
fMRI utilizations atomic magnetisation of the hydro atoms in the fluids -mainly the blood- to aline a powerful magnetic field. Because fMRI depends on the fluids motions inside the organic structure tissues it will be more utile for slow events -relatively- several hundred msecs.
EEG is entering of the electrical activities of the encephalon from the scalp. The electrical activities occurred by fire of the nervous in the encephalon. Even EEG is less accurate ( low Signal-to-Noise Ratio SNR ) than MEG and functional magnetic resonance imaging, EEG is the most used techniques in Brain Computer Interface which represents more than 80 % of BCI published work ( 1 ) . EEG has many advantages over MEG and fMRI where EEG needs inexpensive apparatus cost and non-intensive interface comparing with other methods.
Brain Anatomy:
The nerve cell is the basic processing unit in the cardinal nervous system ( CNS ) . Figure 1.1 shows basic construction of the nerve cell.
Figure ?1. Neuron Structure [ from Attwood and MacKey 1 ]
The chief three parts of nerve cell are Cell organic structure, Axon and Dendrites. Neuron cell is different from other cell types by the Axon carries the information through the nervous web. Information carried and transmitted by nerve cells is called action potency ( AP ) . Information is carried by the Axon through the alteration in the electrical potency in the membrane of the Axon itself. To bring forth the possible alteration in the membrane, Axon use particular channels to pump the positive ions ( K+ , Na+ and Ca++ ) and negative ions ( Cl- ) between membranes ‘ inside and outside. For human encephalon the AP ranges between -60mV and 10mV ( 2 ) .
When the nerve cell becomes active it do current within the dendrites, this current generate magnetic Fieldss which can be measured by EMG technique and secondary electrical field over the scalp which will be mensurable by EEG technique. The secondary electrical current weakens by the caput beds ( scalp, skull and encephalon ) . The opposition of the skull ( difficult tissue ) is higher than the opposition of scalp and encephalon ( soft tissue ) doing more perturbation to the secondary current. Another job is the noise which is generated by the encephalon itself and the scalp, internal and external noise severally.
The complete and to the full functional human encephalon contains around 1011 nerve cells. All nerve cells are connected with each other through synapses doing a big nervous web. Adults have about 5 ten 1014 synapses but this figure increases with age while the figure of nerve cells lessenings with age ( 3 ) .
Neuroscientists refer to all witting mental activity, e.g. , thought, retrieving, voluntary moves and others, by utilizing higher cortical maps term. Constructing a structural theoretical account for the encephalon utilizing higher cortical maps allows us to pull out the encephalon activates by mensurating the degree of stimulation in a certain country based on that theoretical account. Figure 1.2 shows the functional theoretical account for the human encephalon.
Figure ?1. functional countries of the encephalon ( 4 )
Primary sensory and motor countries are responsible for feeling and bring forthing traveling orders to different portion of the human organic structure. Figure 1.3 shows the parts sensed and controlled by sensory and motor countries ( 5 ) .
Figure ?1. Primary Motor and Sensory countries ( 6 )
Through supervising the primary motor and centripetal countries activities, we can calculate out if the encephalon is executing any mental undertaking or having any information from any portion of the organic structure.
The EEG recordings include different frequences, e.g. , ? , ? , ? , ? and ? , and each frequence correspond to certain encephalon activities. Table 1.1 shows the frequence ranges and the mental activities stand behind them.
Name
Frequency Range ( Hz )
Metallic element activity
alpha ( ? )
8-13
Relaxation and concentration
theta ( ? )
4-7.5
Unconscious, inspiration and deep speculation
beta ( ? )
14-26
Thinking, attending and work outing concrete jobs
delta ( ? )
0.5-4
Deep slumber
gamma ( ? )
30-45
Finger motion, toes and lingua
Table ?1. Frequency ranges of the encephalon beat
Brain Computer Interface -BCI-
BCI is a communicating and control channel which depends on ideas merely and does n’t affect any musculus or peripheral nervousnesss ( 7 ) . BCI is considered as united research among neuroscience, technology and machine acquisition.
The chief two types of BCI utilizing EEG are direct ( invasive ) BCI and non-invasive BCI. In the following a brief elucidation for each one:
Direct BCI: utilizing implant electrodes inside the encephalon to maximise the signal to resound ratio ( SNR ) by capturing EEG signals from closer country. The chief jobs with this technique that it is hard to implement, there will be ethical issues behind it and unexpected side effects for the deep-rooted electrodes. Though these jobs, direct BCI was tested on animate beings ( monkeys, rates and others ) . Implanted electrodes were tested on human topic ( Mattew Nagle ) in 2005 ( 8 ) ( 9 ) .
Non-Invasive BCI: measurement encephalon beat through EEG signal from the surface of the scalp utilizing external electrodes ( dry or wet electrodes ) . This method is noisier and the signal is weaker than direct BCI, but it is more practical than the direct BCI since it is portable and no side consequence of using this technique on any topic.
BCI uses signal processing and machine larning algorithms to acknowledge the encephalon beat and link them to certain mental undertakings. In this thesis we will cast the visible radiation on signal processing and machine acquisition algorithms for BCI. Figure 1.4 shows a general block diagram for BCI.
BCI algorithms contain normally four procedures which are preprocessing, characteristic extraction, categorization and post-processing.
Preprocessing procedure may include one or more of the undermentioned procedures:
Temporal Filtering is pull outing the frequences which related to targeted mental undertakings.
Common Spatial Patterns ( CSP ) is taking to divide between categories by increasing the discrepancy of certain category and diminishing it the other category.
Normalization procedure is seeking to pull strings the EEG signal to hold normalized one which might assist in the classifier subsequently.
Feature extraction considered the nucleus of the BCI algorithm where this procedure will seek to happen the most inhibitory characteristics of the mark mental undertakings ( labels ) . Many characteristics have been used in BCI research, e.g. , the discrepancy of the signal, the power denseness of the signal, ripple transform and others.
Categorization algorithms are good known in machine acquisition field. In BCI, categorization algorithm is seeking to observe the mental undertaking ( label ) behind the characteristics for period of clip of EEG signal based on a preparation datasets provided for the classifier before the categorization procedure. Some of the common categorization algorithms in BCI are Support Vector Machine ( SVM ) , Artificial Neural Network ( ANN ) , k-Nearest Neighbor ( kNN ) and other classifiers.
Post-Processing procedure depends on the job itself where post-processing can smooth the end product of the classifier to do the end product consistent with the job ‘s constrains.
Finally the interlingual rendition procedure where in this procedure the computing machine performs a set of actions based on the predicted mental undertaking, e.g. , operate certain devices or travel an electrical wheelchair into a certain way based on the predicted mental undertaking.
Figure ?1. Basic BCI block diagram
