Embedded Files
Longitudinal Motor Imagery EEG Data Collection (2026-Current)
Longitudinal Motor Imagery EEG Data Collection (2026-Current)
Summary: This project investigates hemispheric dominance using EEG signals, examines differences in motor imagery (MI) behavior between ALS and non-ALS participants, and analyzes session-to-session neural variability to identify potential underlying causes.
Summary: This project investigates hemispheric dominance using EEG signals, examines differences in motor imagery (MI) behavior between ALS and non-ALS participants, and analyzes session-to-session neural variability to identify potential underlying causes.
PhD Dissertation (2022-2025)
PhD Dissertation (2022-2025)
Mapping Resting-State Electroencephalography (EEG) Onto Motor Imagery for Cross-Subject Brain-Computer Interface (BCI) Applications
Mapping Resting-State Electroencephalography (EEG) Onto Motor Imagery for Cross-Subject Brain-Computer Interface (BCI) Applications
Summary: Motor Imagery (MI) Brain–Computer Interfaces still rely heavily on subject-specific training because EEG signals vary widely across individuals. This dissertation examines whether resting-state EEG contains subject-dependent information that can reduce or replace the need for per-user MI training. The core idea is that resting activity acts as a neural signature that shapes how MI patterns later appear. The work shows that Alpha-band resting power systematically predicts hemispheric activation during MI, allowing subjects to be grouped into three meaningful neurophysiological profiles. These rest-based clusters consistently improve cross-subject decoding across datasets and models, supporting the feasibility of plug-and-play MI-BCIs.
Summary: Motor Imagery (MI) Brain–Computer Interfaces still rely heavily on subject-specific training because EEG signals vary widely across individuals. This dissertation examines whether resting-state EEG contains subject-dependent information that can reduce or replace the need for per-user MI training. The core idea is that resting activity acts as a neural signature that shapes how MI patterns later appear. The work shows that Alpha-band resting power systematically predicts hemispheric activation during MI, allowing subjects to be grouped into three meaningful neurophysiological profiles. These rest-based clusters consistently improve cross-subject decoding across datasets and models, supporting the feasibility of plug-and-play MI-BCIs.
Publications:
· M. Y. M. Naser and S. Bhattacharya, "Resting-State EEG’s Influence on Ipsilateral Motor Imagery ERS Could Enhance BCI Generalizability," 12th IEEE EMBS International Conference on Neural Engineering, San Diego, CA, USA, 2025 (accepted, access from here).
· M. Y. M. Naser and S. Bhattacharya, “Mapping resting-state EEG onto motor imagery EEG signals via data clustering for reduced classifier training requirements,” U.S. Patent Application 18/629,156, Oct. 10, 2024.
· M. Y. M. Naser and S. Bhattacharya, "Resting EEG State: An Insight into Motor Imagery Signal Characteristics," 2023 15th Biomedical Engineering International Conference (BMEiCON), Tokyo, Japan, 2023, pp. 1-5, doi: 10.1109/BMEiCON60347.2023.10321821.
· M. Y. M. Naser and S. Bhattacharya, "Towards Practical BCI-Driven Wheelchairs: A Systematic Review Study," IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 31, pp. 1030-1044, 2023, doi: 10.1109/TNSRE.2023.3236251.
U.S. A.R.L. Project (2023-2026)
U.S. A.R.L. Project (2023-2026)
Overcoming Confusion in AI: Integrating Human-like “gist” into Complex Dynamically Evolving Human-AI Systems
Overcoming Confusion in AI: Integrating Human-like “gist” into Complex Dynamically Evolving Human-AI Systems
Summary: This project addresses the growing gap between the rapid rollout of AI technologies and the limited real-world experience needed to reveal rare but high-impact failures. Because many AI systems are pushed into public use before accumulating enough operational data, they suffer from “AI blind spots” that can lead to serious errors, especially in high-risk contexts. We argue that safe and reliable deployment requires stronger human-AI partnerships, moving beyond simple “Human-OR-AI” task allocation toward systems that leverage human adaptive decision-making. The goal is to understand how humans selectively process information under uncertainty and use that knowledge to design multimodal machine-learning systems capable of robust, risk-aware performance. Ultimately, the project seeks to build AI agents that operate safely, adaptively, and in alignment with human priorities in complex environments.
Summary: This project addresses the growing gap between the rapid rollout of AI technologies and the limited real-world experience needed to reveal rare but high-impact failures. Because many AI systems are pushed into public use before accumulating enough operational data, they suffer from “AI blind spots” that can lead to serious errors, especially in high-risk contexts. We argue that safe and reliable deployment requires stronger human-AI partnerships, moving beyond simple “Human-OR-AI” task allocation toward systems that leverage human adaptive decision-making. The goal is to understand how humans selectively process information under uncertainty and use that knowledge to design multimodal machine-learning systems capable of robust, risk-aware performance. Ultimately, the project seeks to build AI agents that operate safely, adaptively, and in alignment with human priorities in complex environments.
Publications:
· Z. Chen, S. B. Kelil, M. Y. M. Naser, J. S. Metcalfe and S. Bhattacharya, "Human-Guided Artificial Intelligence (HGAI), a Framework for Overcoming AI's Blind Spots," SoutheastCon 2025, Concord, NC, USA, 2025, pp. 1417-1423, doi: 10.1109/SoutheastCon56624.2025.10971607.
U.S. A.R.L. Project (2022-2024)
U.S. A.R.L. Project (2022-2024)
Multimodal Inference of Human State: Tracking Cognition in a Risky Environment
Multimodal Inference of Human State: Tracking Cognition in a Risky Environment
Summary: The goal of this project was to develop a reliable multimodal sensing framework that integrates physiological, behavioral, and environmental data to infer human cognitive and emotional states during interactions with vehicle-based autonomy. The first phase focused on analyzing human-only data from ARL’s real-world driving experiment. The first aim was to establish and validate multimodal methods capable of characterizing passenger states as they relate to vehicle performance and environmental conditions. The second aim was to develop machine-learning models that can systematically classify relevant passenger states in relation to the driver’s actions and the ongoing driving context.
Summary: The goal of this project was to develop a reliable multimodal sensing framework that integrates physiological, behavioral, and environmental data to infer human cognitive and emotional states during interactions with vehicle-based autonomy. The first phase focused on analyzing human-only data from ARL’s real-world driving experiment. The first aim was to establish and validate multimodal methods capable of characterizing passenger states as they relate to vehicle performance and environmental conditions. The second aim was to develop machine-learning models that can systematically classify relevant passenger states in relation to the driver’s actions and the ongoing driving context.
Publications:
· J. S. Metcalfe, W. D. Hairston, N. Ishmakhametov, M. Y. M. Naser, and S. Bhattacharya, " Final Report on Multimodal Inference of Human State: Tracking Cognition in a Risky Environment" DEVCOM Army Research Laboratory, report number: ARL-TR-10048, Jan 2025 (access from here).
· M. Y. M. Naser and S. Bhattacharya, "Empowering Human-AI Teams via Intentional Behavioral Synchrony," Frontiers in Neuroergonomics, vol. 4, 2023, doi: 10.3389/fnrgo.2023.1181827.
· M. Y. M. Naser, S. Bhattacharya, K. Alame, and W. Hairston, "Modeling of Passenger Postures for Predicting Driving Events Using a Support Vector Machine," International Association of Journals and Conferences, Orlando, FL, USA, 2023, doi: 10.5281/zenodo.8212504.
Digital Signal Processing Internship at Bose Corporation (Jun 2024-Aug 2024)
Digital Signal Processing Internship at Bose Corporation (Jun 2024-Aug 2024)
My work during this internship focused on developing machine-learning algorithms for in-cabin noise estimation from microphone signals, targeting the isolation of real driving noise, and benchmarking them against adaptive filtering methods while accounting for real-time constraints in model design, mathematical formulation, and code efficiency.
My work during this internship focused on developing machine-learning algorithms for in-cabin noise estimation from microphone signals, targeting the isolation of real driving noise, and benchmarking them against adaptive filtering methods while accounting for real-time constraints in model design, mathematical formulation, and code efficiency.
Physiological signal acquisition project (Jan 2025-May 2025)
Physiological signal acquisition project (Jan 2025-May 2025)
DEJA-VU: A Multimodal Physiological Dataset (EEG, ECG, EMG, GSR) for Dynamic Emotion Transitions in Virtual Reality
DEJA-VU: A Multimodal Physiological Dataset (EEG, ECG, EMG, GSR) for Dynamic Emotion Transitions in Virtual Reality
Summary: Emotion recognition from physiological signals typically treats emotions as discrete, static states rather than dynamic processes, creating limitations for real-world affective computing applications. This dataset contains multimodal physiological recordings from 28 participants experiencing systematically designed emotional transitions in virtual reality environments. Participants viewed validated emotion-eliciting video stimuli (2:13-3:40 duration) across three emotional quadrants with 69-second neutral reset periods between stimuli. Four physiological modalities were recorded simultaneously: EEG (7 channels, 300 Hz), ECG (4 leads, 512 Hz), EMG (2 channels, 512 Hz), and GSR (3 channels, 10 Hz). The experimental protocol employed balanced incomplete block design across six possible emotional sequences. Statistical validation demonstrates quadrant differentiation with 65-78% emotion induction success rates. Individual journey analysis reveals emotional mobility variation ranging 5.9%-70.2% of theoretical maximum with systematic sequence effects. The dataset comprises 1.85GB including 840 video-aligned physiological segments, comprehensive self-assessment ratings, and extracted features.
Summary: Emotion recognition from physiological signals typically treats emotions as discrete, static states rather than dynamic processes, creating limitations for real-world affective computing applications. This dataset contains multimodal physiological recordings from 28 participants experiencing systematically designed emotional transitions in virtual reality environments. Participants viewed validated emotion-eliciting video stimuli (2:13-3:40 duration) across three emotional quadrants with 69-second neutral reset periods between stimuli. Four physiological modalities were recorded simultaneously: EEG (7 channels, 300 Hz), ECG (4 leads, 512 Hz), EMG (2 channels, 512 Hz), and GSR (3 channels, 10 Hz). The experimental protocol employed balanced incomplete block design across six possible emotional sequences. Statistical validation demonstrates quadrant differentiation with 65-78% emotion induction success rates. Individual journey analysis reveals emotional mobility variation ranging 5.9%-70.2% of theoretical maximum with systematic sequence effects. The dataset comprises 1.85GB including 840 video-aligned physiological segments, comprehensive self-assessment ratings, and extracted features.
Publications:
· N. Ishmakhametov, M. Y. M. Naser , S. Kelil, C. McClary, J. Metcalfe and S. Bhattacharya, " DEJA-VU: A Multimodal Physiological Dataset (EEG, ECG, EMG, GSR) for Dynamic Emotion Transitions in Virtual Reality", Nature Scientific Data (under 2nd round of review, access from here).
Physiological signal acquisition project (Aug 2024-Nov 2024)
Physiological signal acquisition project (Aug 2024-Nov 2024)
Dataset of Multi-Modal Physiological Signals (fNIRS, EEG, ECG, EMG) Recorded Across Different Emotional States
Dataset of Multi-Modal Physiological Signals (fNIRS, EEG, ECG, EMG) Recorded Across Different Emotional States
Summary: This dataset comprises synchronized multi-modal physiological recordings—functional Near-Infrared Spectroscopy (fNIRS), Electroencephalography (EEG), Electrocardiography (ECG), and Electromyography (EMG)—collected from 16 participants exposed to emotion-eliciting video stimuli. It includes raw signals, event markers, and Python scripts for data import and preprocessing. Special emphasis is placed on fNIRS, which, though less common in affective computing, provides valuable hemodynamic insights that complement electrical signals from EEG, ECG, and EMG. The dataset is structured to facilitate reproducibility and ease of integration across platforms. It aims to support research in emotion recognition, multimodal data fusion, and machine learning applications in emotion-aware and human-centered systems.
Publications:
· C. McClary, M. Y. M. Naser , S. Repole, B. McKinney, S. Bhattacharya, "Dataset of Multi-Modal Physiological Signals (fNIRS, EEG, ECG, EMG) Recorded Across Different Emotional States", IEEE Dataport, April 25, 2025, doi:10.21227/tm30-9744.
Master's thesis (2018-2021)
Master's thesis (2018-2021)
Computer Modeling of Solar Thermal System with Underground Storage Tank for Space Heating
Computer Modeling of Solar Thermal System with Underground Storage Tank for Space Heating
Summary: Space heating is required in almost every dwelling across the country for different periods of time. The thermal energy needed to meet a heating demand can be supplied using different conventional and/or renewable technologies. Solar energy is one example of a renewable resource that can be used for supplying heating needs. It can be utilized either by using photovoltaic panels to generate electricity, that in turn can be used to operate heaters, or by using solar thermal panels. Solar thermal panels obtain higher operating efficiencies than photovoltaic panels, but solar energy for heating purposes suffers from a mismatch between supply and demand. This problem can be solved by employing large tanks that serve as seasonal thermal energy storage units. This work focuses on assessing the performance of solar thermal panels in supplying the space heating needs of a single-family dwelling in two different cities in the United States. These panels are coupled to a cylindrical tank buried in the ground for seasonal and daily energy storage, and a heat exchanger to transfer heat to the home. Storage tank size, collector area, and the working fluid mass flow rate are investigated to determine adequate values for these parameters to enhance the overall system performance. In addition, the simulation timestep for the program and the spatial grid sizes used in the CFD (computational fluid dynamics simulation) model of the storage unit have been examined to determine values of each to keep the simulation time reasonable without substantial loss in accuracy. These results are obtained by mathematically modeling the system components, the solar thermal panels, the heat exchanger, and the storage tank; then programming this mathematical model in MATLAB. Parts of the developed computer code were obtained from previous work and have been modified to suit the purpose of this project. Flat plate solar panels have been chosen for use as the solar collectors. An unmixed, crossflow heat exchanger is considered for transferring the thermal energy from the glycol-water mixture in the collectors to the air in the house. The storage unit CFD simulation is composed of the tank, insulation, and ground surrounding the tank and is based on the SIMPLE algorithm developed by Patankar.
Summary: Space heating is required in almost every dwelling across the country for different periods of time. The thermal energy needed to meet a heating demand can be supplied using different conventional and/or renewable technologies. Solar energy is one example of a renewable resource that can be used for supplying heating needs. It can be utilized either by using photovoltaic panels to generate electricity, that in turn can be used to operate heaters, or by using solar thermal panels. Solar thermal panels obtain higher operating efficiencies than photovoltaic panels, but solar energy for heating purposes suffers from a mismatch between supply and demand. This problem can be solved by employing large tanks that serve as seasonal thermal energy storage units. This work focuses on assessing the performance of solar thermal panels in supplying the space heating needs of a single-family dwelling in two different cities in the United States. These panels are coupled to a cylindrical tank buried in the ground for seasonal and daily energy storage, and a heat exchanger to transfer heat to the home. Storage tank size, collector area, and the working fluid mass flow rate are investigated to determine adequate values for these parameters to enhance the overall system performance. In addition, the simulation timestep for the program and the spatial grid sizes used in the CFD (computational fluid dynamics simulation) model of the storage unit have been examined to determine values of each to keep the simulation time reasonable without substantial loss in accuracy. These results are obtained by mathematically modeling the system components, the solar thermal panels, the heat exchanger, and the storage tank; then programming this mathematical model in MATLAB. Parts of the developed computer code were obtained from previous work and have been modified to suit the purpose of this project. Flat plate solar panels have been chosen for use as the solar collectors. An unmixed, crossflow heat exchanger is considered for transferring the thermal energy from the glycol-water mixture in the collectors to the air in the house. The storage unit CFD simulation is composed of the tank, insulation, and ground surrounding the tank and is based on the SIMPLE algorithm developed by Patankar.
Publications:
· M. Y. M. Naser, Computer Modeling of Solar Thermal System with Underground Storage Tank for Space Heating, 2021.
Mechanical Engineering Intern at the National Renewable Energy Lab (May 2020-Dec 2020)
Mechanical Engineering Intern at the National Renewable Energy Lab (May 2020-Dec 2020)
The main duties of this internship were to develop and validate a simulation-based thermal energy storage model for integration into the REopt software, including benchmarking its performance against high-fidelity models. I also worked on analyzing how demand charges affect electric bills and used that insight to identify optimal periods where renewable-energy availability aligns with peak demand to reduce costs for end users.
The main duties of this internship were to develop and validate a simulation-based thermal energy storage model for integration into the REopt software, including benchmarking its performance against high-fidelity models. I also worked on analyzing how demand charges affect electric bills and used that insight to identify optimal periods where renewable-energy availability aligns with peak demand to reduce costs for end users.
Senior design project (Jan 2017-Jan 2018)
Senior design project (Jan 2017-Jan 2018)
Automated Solar-Powered Street lighting System
Automated Solar-Powered Street lighting System
Summary: With the global shift toward solar energy and automation—and the essential role of street lighting in maintaining public safety—this project was developed to modernize the lighting system on Wasfi Al-Tal Street. The existing system was analyzed and redesigned in Dialux to meet international street-lighting standards, after which a PV solar system was engineered in PVsyst to supply the required energy for the upgraded lighting layout. An automated control system was then built using Arduino, capable of identifying the location of faulty luminaires and operating in three distinct modes: an OFF mode that disables the lighting system, an AUTO mode that uses an LDR sensor to automatically turn the lights on or off, and a SAVING mode that also relies on LDR-based logic to enable dimming for improved energy efficiency.
Summary: With the global shift toward solar energy and automation—and the essential role of street lighting in maintaining public safety—this project was developed to modernize the lighting system on Wasfi Al-Tal Street. The existing system was analyzed and redesigned in Dialux to meet international street-lighting standards, after which a PV solar system was engineered in PVsyst to supply the required energy for the upgraded lighting layout. An automated control system was then built using Arduino, capable of identifying the location of faulty luminaires and operating in three distinct modes: an OFF mode that disables the lighting system, an AUTO mode that uses an LDR sensor to automatically turn the lights on or off, and a SAVING mode that also relies on LDR-based logic to enable dimming for improved energy efficiency.
Page updated
Google Sites
Report abuse