Scribd
Upload
72 views3 pages

College of Engineering Pune: Marathwada Mitra Mandal's

This document is a student project report on brain-computer interfaces (BCI). It summarizes that a BCI system acquires brain activity signals, interprets them, and uses them to interact with computers. The report outlines the four main steps of a BCI system: signal acquisition, processing, classification, and application. It then discusses applications in medical rehabilitation and treatment, as well as potential entertainment uses. The report will cover topics of data acquisition and analysis, feature extraction and classification, and translating features for control applications. It presents analysis of EEG data from multiple test subjects during motor imagery tasks. Neural networks are used to classify signals and generate control signals to applications like prosthetics.

Uploaded by

Tejas Talele
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
72 views3 pages

College of Engineering Pune: Marathwada Mitra Mandal's

This document is a student project report on brain-computer interfaces (BCI). It summarizes that a BCI system acquires brain activity signals, interprets them, and uses them to interact with computers. The report outlines the four main steps of a BCI system: signal acquisition, processing, classification, and application. It then discusses applications in medical rehabilitation and treatment, as well as potential entertainment uses. The report will cover topics of data acquisition and analysis, feature extraction and classification, and translating features for control applications. It presents analysis of EEG data from multiple test subjects during motor imagery tasks. Neural networks are used to classify signals and generate control signals to applications like prosthetics.

Uploaded by

Tejas Talele
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 3

Marathwada Mitra Mandal’s

College of Engineering Pune

Academic Year: 2010-2011

DEPARTMENT : Computer Engineering

Roll No : - TC 260 Class :- TE COMP II


Name of the Student :- Tejas Talele

1. TITLE – Brain Computer Interface (BCI)

2. ABSTRACT

BCI stands for Brain Computer Interface, a system capable of acquiring brain activity’s signals,
interpret them and eventually use them for any kind of purpose. This technology focuses on specific
interaction between an individual and any computer. Its prime objective is to allow the user to
communicate with a digital environment resorting exclusively to “thought stimuli” [1; 7]. The
functional system of BCI could be summarized in four steps: signal acquisition; signal processing
(which includes: pre-processing and feature extraction); classification (of the signal); application
interface, which can be biofeedback (the interface itself gives the user information about his brain
activity) or output devices like neuroprosthesis and wheeling chairs) [7]. There are two major fields of
BCI applications: medical field (rehabilitation and treatment) and entertainment field, where it is
thought that, in a distant future, this technology will achieve an enormous success. Concerning
rehabilitation, this technology’s main contribution is based upon prostheses, but it can also have an
important role in: functional control of electric stimuli; virtual keyboards; pointing device equipment;
vehicle control; etc. Some significant applications of a BCI system in treatment are, for example, the
restitution of a communication signal for patients with locked-in syndrome or the control of
neuroprosthesis by patients affected with Spinal Cord injuries. Even though it is still a low performance
system, BCI attracts people who only seek entertainment as well. Some companies have begun to
invest in this area, since it will undoubtedly revolutionize the way people carry out simple tasks (in
terms of commodity and amusement).

3. CONTENTS

Much of the progress on this project has been dependent upon technical developments in
advanced pattern recognition. These have included developing measures of signal complexity,
calculating error bars on neural network predictions, assessing model selection methods, developing
methods for combining network predictions and investigating the use of dynamic models. It’s not only
a new step towards automation but it can also be adapted to turn the miserably painful lives of the
persons lying with severe motor disabilities into heavens. Further, this work can be extended to various
scenarios where security and safety is a big issue.

The following topics are intended to be covered during the course of the seminar:
 Data Acquisition and Analysis
 Feature Extraction and Classification
 Feature Translation for Control Application

Data Acquisition and Analysis:


Data has been acquired by applying non-invasive technique i.e. scalp EEG. Dataset for various
subjects under different test cases and scenarios has been analyzed and waveforms of two subjects
(Laraib and houkat) have been collected during motor imageries (through training sessions). Analysis
of similar tests on different subjects helped reaching the conclusion that the brain wave patterns of
adults are not much different from each other corresponding to a particular activity. However, since
infants have delta waves superimposed on the scalp EEG patterns which cause their patterns to vary
from those one of normal adults. Thus, this encourages the development of a generalized system that
can be applicable to a number of persons. In this paper the work presented is limited to the
implementation of few motor controls because of inherent complexity of brain waves as multiple
processes are running at a time and it’s tough to determine the desired signal out of the highly damped
signal received over the scalp. The results have been studied and variance/changes have been analyzed
as slightly different stimulus was presented to the subjects. Figure 4, 5 and 6 show various mental
states of the subject under test.

Feature Extraction and Classification:


After data acquisition and analysis, the next task is to identify the feature presented to the
system and generate the control signals accordingly. Since, machines don’t understand analog
waveforms, so this data has been converted to ASCII format and imported into the MATLAB, which
read the files. Since motor imageries are concerned with the frontal part of brain, so 6 electrodes
peculiarly associated with frontal regions contained the signals of the concerned activities, a routine has
been defined to access the entire dataset and pick the required data. Next routine has been defined to set
the target values for the classification system which are used to train the system. Multiple neural
networks (with different layers and neurons) have been created to analyze the response.

Feature Translation for Control Application:


Finally, all the routines and scripts have been integrated into one file which displays the
information in a GUI window which allows the user to select or import the data. Then, after
classification, it generates the control signals which are being written to the parallel port of the PC.

4. REFERENCES

[1] Wikipedia [Online]. Available: http://en.wikipedia.org/wiki/Brain-computer_interface


[2] Controlling robots with the mind(2002,Sept 16) Scientific American, article on Miguel Nicolelis

Tejas Talele Prof. Ram Joshi


(Internal Guide)

You might also like