![An Investigation on Non-Invasive Brain-Computer Interfaces:
Emotiv Epoc+ Neuroheadset and Its Effectiveness
Md Jobair Hossain Faruk
Faculty Mentor: Professors Maria Valero and Hossain Shahriar
INTRODUCTION
▪ Neurotechnology focuses on nervous system
with goals to directly “wire up” human brains
to machines.
▪ Brain-Computer Interfaces (BCI) is one of the
forefronts of many neurotechnological
discoveries.
▪ BCI research was introduced in the 1970s and
primary attempts are to decoding human speech
from brain signals, implementing creativity by
imagination, and controlling neuro-
psychological patterns.
▪ BCI utilizes billion of neural activities that
would significantly benefit people suffering
from neurological disorders.
STUDY OBJECTIVES
▪ Study the progress of BCI research and
Illustrate scores of unveiled contemporary
approaches.
▪ Investigate the mind machine-based Epoc+
neuroheadset in identifying emotional
parameters among human subjects and
effectiveness of neurotechnological research.
MATERIALS AND METHODS
▪ Three non-KSU human subjects for Epoc+
neuroheadset experiments focus on emotional
parameters.
▪ Naïve Bayes and Linear Regression classifiers
are applied to identify, and Kappa metric
optimizes the accuracy of the performance
metrics.
▪ Comparative study between Experimental,
External, and Survey Data.
RESULTS
▪ 69% and 62% improved accuracy for the
Naïve Bayes and Linear Regression
classifiers, respectively in reading the
performance matrices of the participants.
▪ A comparative representation between the
subjects and their survey report on
performance metrics draw a parallel
implication.
CONCLUSIONS
▪ Current Essence: We envision that non-
invasive, insertable, and low-cost BCI
approaches shall be the focal point for not
only an alternative for patients with physical
paralysis but also brain understanding that
would pave us to access and control the
memories and brain somewhere very near.
▪ Future Direction: We aim to continue our
effort explicitly towards significant
accomplishments around the domain of BCIs;
following the vision 2050 when BCI could
become a magic wand for developing men
control objects with the mind.
REFERENCE CITED
[1] H. Jobair, M. Valero, H. Shahriar. “An
Investigation on Non-Invasive Brain Computer
Interfaces: Emotiv Epoc+ Neuroheadset and Its
Effectiveness”. To appear in IEEE COMPSAC
2021, Full paper.
Literature Review
August 2020
Study Timeline
Internal Experiments
November 2020
Data Classifying
December 2020
Result Presentation
February 2021
FINDINGS
DATASETS
▪ External Datasets from Emotiv,
consist of about 2,170 events
(Emotional Parameters).
▪ Experimental Datasets (from 3
different subjects) consist of
6,086 events.
▪ Datasets from the Preliminary
Survey.
DISCUSSION
▪ The BCI’s approaches, both
Neuralink and FLR revealed their
progressional promises in BCIs
research towards a technological
wonderland.
▪ Experimental findings appear a
promising research tool around
the field of neurotechnology.
Naïve Bayes
Initial
Accuracy
Improved
Accuracy
N=1,242
57% 69% Training: 70%
32% (Kappa) 41% (Kappa) Testing: 30%
Linear Regression
Initial/Improv
ed Accuracy
Standard
Error/t-Stat
N=1,242
45% 15% Training: 70%
62% 12% Testing: 30%
ACKNOWLEDGEMENT
Fig. 1: Represents the contact quality of Epoc+ Neuroheadset (left) and illustrate the probability chart of performance
metrics (right)
Fig. 2: Graphical illustration of performance metrics using EEG dataset
Fig. 3: Analogy Between Naïve Bayes And Linear Regression Using Confusion Matrix](https://image.slidesharecdn.com/jobairsposterforsymposiumspring21-210502051713/85/Poster-Presentation-An-Investigation-on-Non-Invasive-Brain-Computer-Interfaces-Emotiv-Epoc-Neuroheadset-and-Its-Effectiveness-1-320.jpg)
Poster Presentation: An Investigation on Non-Invasive Brain-Computer Interfaces: Emotiv Epoc+ Neuroheadset and Its Effectiveness
- 1. An Investigation on Non-Invasive Brain-Computer Interfaces: Emotiv Epoc+ Neuroheadset and Its Effectiveness Md Jobair Hossain Faruk Faculty Mentor: Professors Maria Valero and Hossain Shahriar INTRODUCTION ▪ Neurotechnology focuses on nervous system with goals to directly “wire up” human brains to machines. ▪ Brain-Computer Interfaces (BCI) is one of the forefronts of many neurotechnological discoveries. ▪ BCI research was introduced in the 1970s and primary attempts are to decoding human speech from brain signals, implementing creativity by imagination, and controlling neuro- psychological patterns. ▪ BCI utilizes billion of neural activities that would significantly benefit people suffering from neurological disorders. STUDY OBJECTIVES ▪ Study the progress of BCI research and Illustrate scores of unveiled contemporary approaches. ▪ Investigate the mind machine-based Epoc+ neuroheadset in identifying emotional parameters among human subjects and effectiveness of neurotechnological research. MATERIALS AND METHODS ▪ Three non-KSU human subjects for Epoc+ neuroheadset experiments focus on emotional parameters. ▪ Naïve Bayes and Linear Regression classifiers are applied to identify, and Kappa metric optimizes the accuracy of the performance metrics. ▪ Comparative study between Experimental, External, and Survey Data. RESULTS ▪ 69% and 62% improved accuracy for the Naïve Bayes and Linear Regression classifiers, respectively in reading the performance matrices of the participants. ▪ A comparative representation between the subjects and their survey report on performance metrics draw a parallel implication. CONCLUSIONS ▪ Current Essence: We envision that non- invasive, insertable, and low-cost BCI approaches shall be the focal point for not only an alternative for patients with physical paralysis but also brain understanding that would pave us to access and control the memories and brain somewhere very near. ▪ Future Direction: We aim to continue our effort explicitly towards significant accomplishments around the domain of BCIs; following the vision 2050 when BCI could become a magic wand for developing men control objects with the mind. REFERENCE CITED [1] H. Jobair, M. Valero, H. Shahriar. “An Investigation on Non-Invasive Brain Computer Interfaces: Emotiv Epoc+ Neuroheadset and Its Effectiveness”. To appear in IEEE COMPSAC 2021, Full paper. Literature Review August 2020 Study Timeline Internal Experiments November 2020 Data Classifying December 2020 Result Presentation February 2021 FINDINGS DATASETS ▪ External Datasets from Emotiv, consist of about 2,170 events (Emotional Parameters). ▪ Experimental Datasets (from 3 different subjects) consist of 6,086 events. ▪ Datasets from the Preliminary Survey. DISCUSSION ▪ The BCI’s approaches, both Neuralink and FLR revealed their progressional promises in BCIs research towards a technological wonderland. ▪ Experimental findings appear a promising research tool around the field of neurotechnology. Naïve Bayes Initial Accuracy Improved Accuracy N=1,242 57% 69% Training: 70% 32% (Kappa) 41% (Kappa) Testing: 30% Linear Regression Initial/Improv ed Accuracy Standard Error/t-Stat N=1,242 45% 15% Training: 70% 62% 12% Testing: 30% ACKNOWLEDGEMENT Fig. 1: Represents the contact quality of Epoc+ Neuroheadset (left) and illustrate the probability chart of performance metrics (right) Fig. 2: Graphical illustration of performance metrics using EEG dataset Fig. 3: Analogy Between Naïve Bayes And Linear Regression Using Confusion Matrix