Brain Computer Interface (BCI) enables collaboration between brain and machine. This interface was developed to allow lost sensory information to be transmitted to the brain or to stimulate the brain by artificially generating electrical signals. But vice versa, too. that the signals of an active brain can be used to make a machine do what it wants. BCI can not be used to read thoughts. But it can dramatically improve the lives of people with physical limitations.

 

The brain is very complex. It consists of about 100 billion neurons that constantly receive and/or transmit signals. This creates one of the most complex networks we know. There are also various chemical processes that we can understand only gradually.

At SNAP, we measure these signals using noninvasive electroencephalographs (EEG). These are electrodes that measure the potential on our heads. The tension differences on the head thus obtained, provide information about whether one is currently looking, hearing, feeling, active or sleeping.
Being able to use non-invasive BCIs for which surgery is not required makes the use of this technology possible and interesting beyond purely medical purposes. In this case there are two main technologies: fMRI and EEG. The first requires a huge machine, the second with handy headsets is usable for a more general audience.

 

BCI can therefore also be a promising interactive tool for healthy people, with several potential multimedia, VR or video game applications, among many other potential uses.
 

The EEG hardware is absolutely safe for the user, but the received signals are often overlaid by disturbances.

This sometimes results in problems in the interpretation of EEG data, since the data contains many noise, which can be caused by teeth grinding, eye movements or network noise. These interfering signals must be detected and filtered out.

Thereafter, these data can be used to evaluate the actual signals.
An example of this brain-generated signal is the P300 wave, a so-called event-related potential that manifests itself when a less than odten occuring stimulus is generated. This wave is displayed as a large peak in the data history and various machine learning techniques can be used to detect such peaks.

After the interesting signals were filtered out of the data, they should also be used as efficiently as possible. The possible application scenarios can only be guessed at the moment. However, with increasing digitization, the number of different control applications is also growing.

 

How a Brain-Computer-Interface works

 

Simplified, the brain is divided into two main sections:
The limbic system and the neocortex.

The limbic system is responsible for our basic drives as well as for survival, such as eating and reproduction.

Our neocortex is the most advanced and responsible for logical functions that enables us to use language, technology, business, and philosophy.

The human brain contains approximately 86 billion nerve cells, called neurons, which are individually linked to other neurons through connecting elements called axons and dendrites. Every time we think, move or feel, neurons are active. In fact, the brain generates a great deal of neuronal activity. Basically, small electrical signals that move from neuron to neuron are responsible for the cognitive processes.

There are many signals that can be used for BCI's. These signals can be divided into two categories: spikes and field potentials.
These can be recognized, interpreted and used to interact with a device.

Artificial intelligence or machine learning has received a great deal of attention in developing BCI applications to solve difficult problems in various fields, particularly in medical and robotics.
AI / ML has since become the most efficient tool for BCI systems.

Gefördert durch:

SNAP GmbH - Universitätsstr. 136 - 44799 Bochum

© 2018 by SNAP GmbH