Page 27 - BH-2-2

Brain & Heart Predictive modeling using electroencephalogram

A B

Figure 1. Standard electrode positions. (A) Left side view and (B) Top view.

users struggle to master this method. The use of mobile 0.5–10.0 s. As a rule, negative SCPs are linked to motor and
phones, even when hands-free or activated by voice, is just other cortically activated functions, while positive SCPs
as risky as driving while intoxicated, according to recent are associated with deactivated cortical regions. 12
studies that analyzed brain imaging results of cell phone The biopotentials from the scalp’s surface are recorded
use while driving. The purpose of this study is to examine to create an EEG. These recordings capture electrochemical
11
better classifier models to work with EEG data. The main potentials gathered from neurons in the human brain’s
contributions of this study are twofold: (i) to develop an cerebrum. Initial inspection of EEG data may reveal
approach for choosing a better predictive model to address a chaotic, non-stationary, and noisy signal. To isolate
EEG data for use in BCI systems, and (ii) to validate this
approach through the analysis of human driving behavior. individual brain wave components, cutting-edge signal
processing methods are required. After being broken down
2. Related studies into their constituent parts, these pieces can be linked to
specific regions and functions of the brain. Acquiring
Data collected and processed in the time domain, the potential from a single neuron is extremely inefficient.
spatial domain, or both can serve as the basis for an Instead, the final electrical potential recorded at a single
EEG-based conversation. Users can manipulate the sensor node is sufficient for subsequent signal processing
characteristics of electrophysiological signals on demand stages. Studies of brain topology have shown that neurons
with adequate training and practice, making EEG signals in the same general area will all fire in response to any
widely employed in BCI research. Most current BCIs are mental or physical activity that has a localized basis. 13
built to detect specific patterns of brain activity that can
then be translated into physical actions. Many different In addition, choosing between a non-invasive and
signal patterns have been studied, and it has been claimed an invasive approach to obtaining the signals is crucial.
that some of them are particularly easy to recognize and There is no strict rule against using both invasive and non-
control. There are essentially two types of signals here: invasive techniques, as well as both evoked and spontaneous
(i) visual-evoked potentials and (ii) slow cortical potentials inputs. Electrode signals are collected, amplified, and
(SCPs). The brain’s electrical potentials that are elicited by digitized to make them usable in the subsequent steps.
visual stimuli, such as light, are known as visual evoked However, eye (EOG: electrooculogram) or muscle (EMG:
potentials. Gaze control is crucial for the effective operation electromyogram) electrical activity can introduce noise
of such signals. Applications currently utilizing these into the acquired EEG data, which can be mitigated by
signals aim to generate motor output in robots in response maintaining ideal conditions during the signal acquisition,
to the user’s gaze direction. When an EEG is recorded over such as staying in a comfortable, motionless position.
the scalp, the lowest frequency components are typically However, in real-world scenarios, BCIs cannot mimic
attributed to slow voltage fluctuations in the cortex. SCPs laboratory conditions, and such systems are not trusted to
refer to changes in cortical potential that take place over control embedded applications like wheelchairs in outdoor

Volume 2 Issue 2 (2024) 3 doi: 10.36922/bh.2819