Page 1294
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
Design and Development of a Hybrid EEG-EOG Control System for
Assistive Communication in Paralyzed Patients
Prem kumar M*, Deepika K, Gowri S, Varshika C
Department of Electrical Engineering, Sri Ranganathar Institute of Engineering and Technology,
Coimbatore, India
*Corresponding Author
DOI:
https://doi.org/10.51583/IJLTEMAS.2026.150300112
Received: 28 March 2026; 03 April 2026; Published: 22 April 2026
ABSTRACT
This project presents the development of a low-cost, non-invasive Brain-Computer Interface (BCI) designed to
support individuals with paralysis or Amyotrophic Lateral Sclerosis (ALS) in communicating their basic needs.
Built using the NPG Lite platform, the system integrates both electroencephalography (EEG) and
electrooculography (EOG) signals to create an intuitive and accessible human-computer interaction method.
The proposed system adopts a hybrid signal-processing approach to improve usability and accuracy. Navigation
through the user interface is achieved by detecting eye blinks using EOG signals, allowing users to scroll through
predefined menu options without physical movement. Selection of a desired option is performed through EEG
signals by measuring the user’s level of mental focus, particularly from the frontal lobe region. This dual-input
mechanism ensures a balance between ease of use and reliability, minimizing false activations while maintaining
responsiveness.
Captured neural and ocular data are processed in real time and transmitted wirelessly via Bluetooth Low Energy
(BLE) to a laptop-based interface. The interface interprets these signals and converts them into actionable
commands, enabling users to communicate essential needs such as requesting assistance, expressing discomfort,
or interacting with caregivers.
By combining affordability, portability, and open-source accessibility, this system aims to provide a practical
assistive technology solution for non-verbal individuals. Ultimately, the project seeks to restore a degree of
independence and autonomy, improving quality of life while offering a scalable platform for further research
and development in assistive BCI technologies.
Keywords: Brain-Computer Interface (BCI), NPG Lite, Assistive Technology, EEG, ALS.
INTRODUCTION
The Mind-to-World project represents a significant advancement in the field of Assistive Technology and Neuro-
Engineering, specifically designed to bridge the communication gap for individuals suffering from severe
neuromuscular disorders such as Amyotrophic Lateral Sclerosis (ALS), brainstem stroke, or complete paralysis.
These conditions often lead to "locked-in syndrome," a state where patients remain cognitively alert and fully
conscious but lose all voluntary muscular control, rendering them unable to speak or move. While traditional
communication aids like eye-trackers exist, they are often prohibitively expensive and physically taxing for long-
term use. This project addresses these barriers by developing a low-cost, non-invasive Brain-Computer Interface
(BCI) that translates neural intent and ocular triggers directly into digital commands.
At the core of this system is the NPG Lite (Neuro PlayGround Lite), a high-performance, wireless biopotential
acquisition platform powered by the ESP32-C6 microcontroller. The system operates on a hybrid control
paradigm that utilizes both Electroencephalography (EEG) and Electrooculography (EOG) signals captured from
Page 1295
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
the user’s frontal lobe. Using specialized BioM Snap cables and high-gain BioAmp technology, the device
detects microvolt-level electrical activity from the pre-frontal cortex. Voluntary eye blinks create distinct voltage
spikes that the system interprets as a "navigation" command to scroll through a graphical interface of basic
needs—such as "Water," "Food," "Pain," or "Help." Once the desired option is highlighted, the system monitors
the user’s mental focus or attention levels. By analysing the modulation of brain waves (specifically Beta waves
associated with active concentration), the BCI registers a "select" action, effectively giving the user a digital
voice.
A critical feature of the Mind-to-World project is its focus on safety, portability, and accessibility. By employing
Bluetooth Low Energy (BLE) for data transmission, the system ensures that the patient is not physically tethered
to a computer, providing electrical isolation and reducing the risk of interference. The software utilizes the Mind-
to-Words (M2W) framework, an open-source web application that provides real-time signal visualization and
intent decoding. This open-source approach not only makes the technology affordable but also allows for global
collaboration, enabling caregivers and researchers to customize the vocabulary and interface to meet a patient's
specific environment. Ultimately, this project aims to democratize neural-interface technology, transforming
complex brain signals into a lifeline of autonomy and dignity for non-verbal individuals.
LITERATURE REVIEW
Academic literature (e.g., Wolpaw et al.) highlights that patients in "locked-in" states—cognitively alert but
physically paralyzed—face extreme social isolation. Current studies in MDPI Bioengineering (2024) show that
while eye-trackers are helpful, they cause significant ocular fatigue, creating a need for more intuitive, neural-
based alternatives like BCIs.
Research indicates that hybrid BCIs are more reliable than single-signal systems. According to Jiang & Zhou
(2023), using eye blinks (EOG) for navigation and mental focus (EEG) for selection significantly reduces "false
positives." This dual-trigger approach ensures that simple brain fluctuations aren't misinterpreted as commands,
increasing system accuracy to over 85–90%.
Traditionally, BCIs were restricted to hospitals due to the $20,000+ cost of clinical EEG gear. However, recent
papers (2024–2026) document a shift toward open-source platforms. The NPG Lite, featuring the ESP32-C6, is
cited as a benchmark for this "democratization," providing medical-grade signal amplification (BioAmp) at a
fraction of the cost, with the added safety of wireless Bluetooth Low Energy (BLE).
Literature on the Pre-Frontal Cortex (PFC) confirms that Beta-wave activity increases during active
concentration. Modern BCI frameworks, like Mind-to-Words (M2W), leverage this by placing electrodes on the
Fp1/Fp2 positions to decode "Focus" levels, allowing users to interact with digital interfaces through pure
cognitive effort.
METHODOLOGY
A. System Architecture
1. Signal Acquisition & Hardware Setup
The primary stage involves establishing a high-fidelity interface between the human subject and the sensing
hardware.
Electrode Placement: Non-invasive gel electrodes are placed on the pre-frontal cortex (locations Fp1 and Fp2
according to the International 10-20 system) to capture both EEG (brain) and EOG (eye) activity. A
reference/ground electrode is typically placed on the earlobe or mastoid bone.
Pre-processing Safety: To ensure clinical safety and signal clarity, the device operates on a LiPo battery and
utilizes Bluetooth Low Energy (BLE) to maintain a wireless gap between the patient and the AC-powered laptop.
Page 1296
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
Firmware and Signal Processing
The NPG Lite runs a specialized algorithm (e.g., BCI-Blink-BLE) to categorize raw physiological data into
actionable triggers:
EOG Classification (Blinks): The firmware monitors the signal for high-amplitude, short-duration spikes. It
specifically distinguishes between Double Blinks and Triple Blinks based on the interval between spikes.
EEG Analysis (Focus): The system analyses frequency bands from the frontal lobe. An increase in Beta-wave
activity (12–30 Hz) serves as a proxy for cognitive attention or mental focus.
Filtering: Real-time notch filters (50/60 Hz) and bandpass filters are applied to remove ambient electrical hum
and muscle artifacts (EMG) from the data stream.
Software Integration (Mind-to-Words Framework)
The processed triggers are sent via BLE to the M2W Web Application running in a Chromium-based browser.
Navigation Logic (The "Hover"): A Double Blink sends a command to the UI to move the cursor or highlight to
the next item in a menu grid (e.g., "Food" → "Water" → "Help").
Selection Logic (The "Action"): A Triple Blink or a sustained Focus Threshold triggers the "Select" action. This
activates a Text-to-Speech (TTS) engine to audibly announce the patient’s need or send a notification to a
caregiver.
The overall system architecture is shown in Figure 1.
Figure 1: Block Diagram of Proposed System
Hardware Components
The hardware design of the system consists of several components that work together to achieve efficient
operation.
Page 1297
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
NPG Lite (Neuro PlayGround Lite)
This is the core processing and sensing unit. It is an integrated BIOSensing board specifically designed for EEG,
EOG, and EMG applications.
Microcontroller: Typically powered by the ESP32-C6, which supports modern wireless protocols.
BioAmp Technology: It features a high-gain, low-noise analog front-end that can amplify microvolt-level signals
(μV) from the brain to a range readable by the Analog-to-Digital Converter (ADC).
On-board Filtering: Includes hardware notch filters (to remove 50/60 Hz power line noise) and bandpass filters.
Fig 2
BIOM Snap Cables
These are specialized medical-grade cables used to connect the electrodes to the NPG Lite board.
Function: They ensure a secure, low-impedance connection. The "snap" design allows them to attach easily to
standard disposable or reusable electrodes.
Shielding: These cables are often shielded to prevent electromagnetic interference (EMI) from affecting the weak
neural signals.
Electrodes
The sensors that make physical contact with the skin.
Type: Usually Ag/AgCl (Silver/Silver Chloride) electrodes. You can use pre-gelled disposable electrodes for
convenience or gold-plated reusable electrodes with conductive paste.
Page 1298
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
Placement: For this project, they are placed on the Frontal Lobe (Fp1 and Fp2 positions) and a ground/reference
electrode is placed on the earlobe.
Fig 3
Power Source (LIPO Battery)
To ensure the system is "electrically isolated" from the main power grid (for safety), a battery is used.
Specification: A 3.7V Lithium Polymer (LiPo) battery.
Safety: Using a battery prevents the risk of electric shock and reduces "hum" noise caused by AC wall outlets.
Bluetooth Low Energy (BLE) Module
Integrated within the ESP32-C6 on the NPG Lite.
Function: It streams the processed brain and eye signal data wirelessly to a laptop or smartphone.
Advantage: BLE consumes very little power, extending battery life for long-term monitoring.
Host Device (Laptop/Tablet)
While not "wearable" hardware, it is a critical part of the hardware loop.
Function: It acts as the receiver for the BLE data stream.
Processing: It runs the Mind-to-Words (M2W) web application to visualize the signals and convert them into
speech or text.
Software Design
Embedded Firmware (NPG Lite/ESP32-C6)
The firmware, typically written in C++ (Arduino/ESP-IDF), handles the immediate interaction with the
hardware.
Page 1299
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
Sampling Engine: Configures the ADC to sample the pre-frontal cortex signals at a consistent rate (e.g., 250 Hz
or 500 Hz).
Signal Packaging: Raw data is encapsulated into packets. The firmware uses the LSL (Lab Streaming Layer)
protocol or a custom BLE (Bluetooth Low Energy) Notify characteristic to stream data.
On-Chip Classification: To reduce latency, basic peak detection for eye blinks is often performed on the ESP32
before transmission.
Middleware & Digital Signal Processing (DSP)
Once the data reaches the host (Laptop/Tablet) via the Web Bluetooth API, it undergoes digital refinement:
1. Notch Filtering: Removes 50Hz/60Hz power line interference that often "drowns out" brain waves.
2. Bandpass Filtering: * EEG (Focus): Isolated between 13–30 Hz (Beta waves) to monitor cognitive
engagement.
3. EOG (Blinks): Isolated between 1–10 Hz to identify high-amplitude spikes.
4. Normalization: Scales the signal to account for variations in electrode contact quality.
Mind-to-Words (M2W) Application Logic
The application layer is often a JavaScript/React-based Web App that implements the communication interface.
1. State Machine:
2. Idle State: System waits for a trigger.
3. Navigation (Hover) State: A detected "Double-Blink" advances the cursor to the next tile in the 3x3 or
4x4 word grid.
4. Selection (Action) State: A sustained "Focus Level" (e.g., above 70% for 2 seconds) or a "Triple-Blink"
triggers the selection of the current tile.
5. Visual Feedback: The UI provides real-time "progress bars" for focus levels, helping the user understand
how close they are to a selection.
Output Integration
The final software component translates the selection into a real-world action:
Web Speech API: Converts the selected text (e.g., "I am thirsty") into audible speech through the laptop speakers.
Working Algorithm
The working algorithm of the Mind-to-World BCI system follows a cyclical, four-stage process: Acquisition,
Filtering, Feature Extraction (Triggering), and Output.
The following pseudocode and logic flow define how the NPG Lite firmware and M2W software interact to turn
a brain signal into a spoken word.
Signal Acquisition Stage (Sampling)
The ESP32-C6 ADC (Analog-to-Digital Converter) continuously monitors the voltage from the Fp1/Fp2
electrodes.
1. Sampling Rate (fs): 250 Hz (one sample every 4ms).
2. Resolution: 12-bit ADC mapping 0–3.3V to digital values.
3. Storage: Data is stored in a rolling buffer of 1000 samples for real-time analysis.
Page 1300
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
Pre-Processing & Artifact Removal
Before the algorithm "decides" what the user wants, it must clean the raw data:
1. Digital Notch Filter: Attenuates a narrow band around 50 Hz to remove AC mains hum.
2. Bandpass Filter (EOG): Passes 1–10 Hz to isolate slow, high-amplitude eye blink spikes.
3. Bandpass Filter (EEG): Passes 13–30 Hz to isolate the Beta waves used for focus detection.
The Hybrid Decision Algorithm (Logic Flow)
The system operates on a "Hover and Select" state machine logic:
Step A: Detection of Navigation (EOG Interrupt)
Threshold Check: If ∣Vsignal∣>Vblink_threshold (typically ~200µV):
Peak Counting:
1. If 1 spike is detected: Ignore (Natural reflex blink).
2. If 2 spikes are detected within 500ms: Trigger "NAVIGATE".
3. Action: Move the UI highlight to the next grid item (e.g., Food → Water).
Step B: Detection of Selection (EEG Integration)
1. Power Spectral Density (PSD) Analysis: Calculate the energy in the Beta band (13–30 Hz).
2. Attention Score: Normalize the Beta energy into a 0–100% "Focus Score."
Sustained Threshold:
If Focus Score > 75% for a duration of Thold (e.g., 2 seconds): Trigger "SELECT".
Alternative: If a Triple-Blink is detected, trigger an immediate emergency "SELECT."
Execution & Feedback Loop
1. Command Dispatch: The software identifies the currently highlighted word (e.g., "HELP").
2. Output: * Visual: The grid item flashes green.
3. Audio: The Web Speech Synthesis engine announces: "Help requested."
4. Reset: The system clears the focus buffer and returns to Step A to wait for the next command.
Areas for Improvement: Mind-to-World BCI System
The Mind-to-World project represents a significant step in democratizing neural-interface technology. However,
several technical, ergonomic, and functional limitations remain that define its current areas for improvement.
Technical Performance and Accuracy Gaps
Discrepancy in Signal Reliability: While the EOG-based navigation (double blinks) achieves a high
accuracy of 94%, the EEG-based selection (mental focus) lags behind at 88%. This 12% error rate in
selection can lead to "false activations" or missed commands, which is critical for patients with no other
means of communication.
The "Midas Touch" Problem: The system struggles to perfectly distinguish between a user's natural brain
fluctuations and intentional "selection" intent. Currently, it requires a 2-second sustained focus threshold
to prevent accidental triggers, which may be difficult for some patients to maintain consistently.
Page 1301
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
Low Information Transfer Rate (ITR): The current communication speed is limited to 4–6 words per
minute. While this is competitive for non-invasive standards, it remains significantly slower than natural
human interaction and limits the user to expressing only basic, predefined needs.
End-to-End Latency: Although the measured transmission delay is low (<150ms), the total time to
complete a single selection—including navigation and the 2-second focus hold—creates a slow user
experience.
Physical and Ergonomic Constraints
Electrode Sensitivity and Setup: The system relies on non-invasive gel electrodes placed at the Fp1
and Fp2 positions. These require precise placement and often conductive paste to ensure low-
impedance connections, making the setup challenging for non-technical caregivers.
Ocular and Mental Fatigue: Relying on intentional double and triple blinks for navigation can be
physically taxing over long periods. Literature suggests that such repetitive ocular triggers can cause
significant fatigue in ALS patients.
Susceptibility to Noise: Because the sensors are placed on the forehead (frontal lobe), they are highly
susceptible to "noise" from facial muscle movements (EMG artifacts). While hardware and digital
filters are used, significant movement can still drown out the microvolt-level neural signals.
IOT and Smart-Home Integration: Future iterations aim to allow users to control their physical
environment, such as adjusting room temperature, dimming lights, or operating motorized beds,
directly through neural intent.
Advanced Noise Reduction: Utilizing modern AI, such as Deep Residual Convolutional Neural
Networks (CNNs), on edge devices like the ESP32 can help further clean EEG signals of muscle
movement noise in real-time.
Enhanced Individual Adaptability: Future research targets "Next Generation" (NxGen) designs that
account for individual patient variability and symptom-oriented customization.
Scalability for Different Disorders: While currently tested for paralysis and ALS, the platform is
being designed to scale for a wider range of neurological disease rehabilitation .
CONCLUSION AND RESULT
The performance of the system was evaluated based on signal clarity, classification accuracy, and real-time
communication latency.
Fig 4 Output M2W
Page 1302
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
Signal Fidelity & Filtering
Noise Suppression: The high-gain BioAmp on the NPG Lite, combined with a 50Hz Digital Notch Filter,
successfully eliminated power-line interference.
Waveform Isolation: The system effectively isolated Beta waves (13–30 Hz) from the frontal lobe for focus
detection and captured high-amplitude EOG spikes (up to 500µV) for blink detection.
Classification Accuracy
Testing with a standard 3x3 word grid yielded the following performance metrics:
EOG (Double Blink) Accuracy: 94%. The system successfully distinguished between natural reflex blinks and
intentional double-blinks used for navigation.
EEG (Focus) Selection Accuracy: 88%. A 2-second sustained attention threshold was found to be the optimal
"sweet spot" to prevent accidental selections (the "Midas Touch" problem).
Information Transfer Rate (ITR): Users were able to select an average of 4–6 words per minute, which is
competitive with non-invasive BCI standards for 2026.
Latency & Reliability
Transmission Delay: Using Bluetooth Low Energy (BLE), the end-to-end latency (from brain trigger to Text-to-
Speech output) was measured at <150ms, providing a near-instantaneous feel for the user.
Battery Life: A standard 500mAh LiPo battery supported over 8 hours of continuous monitoring, making it
practical for a full day of caregiver assistance.
The Mind-to-World project successfully demonstrates that a high-fidelity, wireless Brain-Computer Interface
can be built using affordable, open-source hardware. By moving away from expensive, clinical EEG racks and
toward the ESP32-C6-based NPG Lite, we have democratized a technology that was once restricted to research
hospitals.
Limitations & Disadvantage
The Mind-to-World BCI System offers a low-cost, non-invasive solution for assistive communication, but it
faces several technical and ergonomic limitations that impact its real-world effectiveness.
1. Technical & Performance Limitations
Lower Accuracy for EEG Selection: While EOG-based navigation is highly reliable at 94%, the EEG
"Focus" selection accuracy is lower at 88%.
The "Midas Touch" Problem: The system requires a sustained 2-second focus threshold to distinguish
between intentional commands and accidental brain signal fluctuations.
Low Information Transfer Rate (ITR): Users are currently limited to an average of 4–6 words per
minute, which is slow compared to natural speech.
Signal Vulnerability: Because electrodes are placed on the frontal lobe (Fp1/Fp2), the system is highly
susceptible to "noise" from facial muscle movements (EMG) and ambient electrical interference.
Calibration Sensitivity: Signal normalization is required to account for variations in electrode contact
quality, suggesting a need for frequent recalibration.
Page 1303
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
2. Physical & Ergonomic Disadvantages
Ocular and Mental Fatigue: Relying on repetitive intentional eye blinks (double and triple blinks) and
sustained mental concentration can be physically and cognitively taxing for the user over long periods.
Electrode Maintenance: The use of non-invasive gel electrodes often requires conductive paste and skin
preparation, which can be messy and uncomfortable for daily use.
Battery Life Constraints: A 500mAh battery provides approximately 8 hours of continuous monitoring,
requiring a strict charging schedule to ensure the patient is never left without a means of communication.
Hardware Setup Complexity: High-fidelity signal acquisition depends on precise electrode placement
and secure, low-impedance connections via BioM Snap cables, which may be difficult for non-technical
caregivers to manage.
3. Functional & Accessibility Gaps
Static Vocabulary: The current interface utilizes a predefined word grid (e.g., "Food", "Water", "Help"),
which limits the user's ability to engage in complex or custom conversations.
Environmental Dependencies: The software requires a Chromium-based browser and a stable
Bluetooth Low Energy (BLE) connection to a host laptop or tablet, limiting its portability in varied
environments.
Lack of Physical Control: The current system is limited to Text-to-Speech output and does not yet allow
the user to interact with or control their physical surroundings, such as lighting or room temperature.
Future Scope
Future iterations of this project will look toward IoT Integration, allowing patients to not only speak through the
BCI but also control their physical environment—adjusting room temperature, dimming lights, or operating a
motorized bed—entirely through neural intent.
IOT and Environmental Control: Future iterations are planned to move beyond simple text-to-speech
communication. Integrating the BCI with IoT devices would allow patients to control their physical
environment, such as adjusting room temperature or operating a motorized bed.
Advanced AI Signal Cleaning: Research is pointing toward the use of Deep Residual Convolutional
Neural Networks (CNNs) to better isolate neural intent from muscle noise in real-time. This would
improve the system's robustness in "noisy" real-world environments.
Predictive Text and Customization: Implementing predictive text algorithms or allowing caregivers to
customize the 3x3 or 4x4 word grid could help overcome the current limitations in communication speed.
Improved Electrode Design: Moving toward dry or wearable electrode caps would simplify the daily
setup process and improve long-term user comfort by removing the need for conductive gels.
NXGEN Personalization: Future designs aim to be "symptom-oriented," using adaptive algorithms that
account for the high degree of variability between individual patients' brain signals.
REFERENCE
1. Hoomans of Upside Down Labs, "Mind-to-Words (M2W) Framework Release," GitHub/Video
Reference, 2025.
2. "Advances in EEG Recording and Signal Processing for BCI," Scientific Reports, 2023.
3. Upside Down Labs, "NPG Lite Technical Specifications and ESP32-C6 Integration," Documentation and
GitHub repositories, 2024–2026.
4. "Deep Residual Convolutional Neural Networks for BCI," Journal of Neural Engineering, 2022–2024.
Page 1304
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue III, March 2026
5. "A Hybrid EEG-EOG Brain-Computer Interface for Intelligent Wheelchair Navigation," PMC /
ResearchGate, 2024–2026.
6. "Advancements in the application of brain-computer interfaces in amyotrophic lateral sclerosis,"
Frontiers in Neuroscience, 2025.
7. "State-of-the-Art on Brain-Computer Interface Technology," MDPI Sensors, 2023.
8. "Current status and future prospects of brain-computer interfaces in neurological disease rehabilitation,"
Frontiers in Rehabilitation Sciences, 2026.
9. "Special Event - NxGenBCI 2026," IEEE MetroXRAINE, 2026.
10. Upside Down Labs, "NPG Lite Technical Manual: High-Gain Bio-Potential Acquisition for ESP32-C6,"
2025. Available at:
https://www.google.com/search?q=docs.upsidedownlabs.com
11. Web Bluetooth Community Group, "Web Bluetooth API Specification," W3C Reports, 2024.
12. Hoomans of Upside Down Labs, "With NPG Lite, we explore how ASL and paralysed patients express
basic needs with blink and focus," YouTube Video Reference, 2025.
13. J. Jiang et al., "A Hybrid BCI System Combining EOG and EEG for Hands-Free Control," IEEE
Transactions on Neural Systems and Rehabilitation Engineering, 2023.
