Browse over 10,000 Electronics Projects

How Brain-Controlled Devices Actually Work: Decoding EEG Signals

How Brain-Controlled Devices Actually Work: Decoding EEG Signals

3.) How the EEG Signals Are Recorded and Processed

Brain-computer interfaces (BCIs) are a combination of software and hardware systems that sample EEG signals from electrodes positioned on the scalp and identify patterns in these signals that indicate the brain function being carried out by the subject. The output of the brain–computer interface or brain–machine interface (BMI) can be utilized to control external devices, or to perform specific actions, as indicated by the EEG signals.

The EEG signals may be recorded from electrode sites within or outside the brain. Various BCI technologies that have been developed over the years are mentioned below, along with their respective pros and cons.

Invasive BCI devices are directly implanted in the grey matter, with the help of a surgical procedure. Invasive BCI produces the best quality of signals with minimal noise. However, such implants may be susceptible to the development of scar tissue. Despite this risk, invasive electrodes are quite useful in the fields of neuroprosthetics and vision correction.

Partially invasive BCI devices are implanted within the skull but outside the brain, instead of within the grey matter. They provide a better resolution of EEG signals than non-invasive BCIs because, in the case of non-invasive BCI’s, the skull tissue tends to deform and deflect the waves. Partially invasive BCI’s are also associated with a lower risk of scar tissue formation in the brain than the fully invasive variants.

Non-invasive BCIs do not require tissue-penetrating surgical procedures and instead rely on recording the brain signals using surface electrodes or sensors. Although they are easy to use, non-invasive implants provide a poor signal resolution because the skull causes the dampening of the signals and disperses and blurs the electromagnetic waves generated by the neurons.

The different types of electrodes used for non-invasive purposes include reusable disks, EEG caps with disks, subdermal needles, and adhesive gel electrodes. In most cases, brain-controlled devices rely on non-invasive EEG headgear or caps for the recording of the signals. Some of the most common uses of such brain-controlled interfaces include-



Advertisement1


  • Monitoring of EEG waves for sleep disorders, neurological diseases etc.
  • Development of thought-driven interfaces between humans and computers.
  • Development of devices to assist disabled people, such as brain-controlled prosthetics.

The fundamental steps involved in the development of BCI-based technology are-

1) Signal Acquisition: The EEG signals are recorded with the help of invasive or non-invasive techniques. These signals measure a few millivolts, and therefore they are amplified prior to sampling.

2) Signal Pre-Processing: Noise and artefacts are removed from the signal with the help of appropriate filters.

3) Signal Classification:  The pre-processed signal is then classified, in order to identify the task that it corresponds to.

4) Device Interaction: The classified output is then fed to the brain-controlled device/computer, which performs the action accordingly.

To summarize, the electrodes take care of signal acquisition but hardware or software filters are required for the pre-processing. Furthermore, signal classification involves the creation of machine learning algorithms whose output can then be fed to the device to obtain the desired results. These steps often form a closed loop in real-time, as depicted in the diagram below.


Brain Computer Interface Cycle

Now that we have the EEG recording in the desired format, let us discuss how to utilize the same for training the brain-controlled device.

UP NEXT:

4.) How to Train a Device Using EEG Signals

Pages: 1 2 3 4 5

 


Top