Iyad Obeid
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Iyad Obeid has a total of 101 co-authors. Here are the top co-authors by number of shared publications:
J. Picone
Neural Engineering Data Consortium, Temple University, Philadelphia, Pennsylvania, USA
13 shared publications- Alessandro Napoli
Raphael Center for Neurorestoration, Thomas Jefferson University Hospital, Philadelphia, PA 19107, USA
12 shared publications - Joseph Picone
Temple University
9 shared publications M. Golmohammadi
Neural Engineering Data Consortium, Temple University, Philadelphia, Pennsylvania, USA
7 shared publications- Christian R. Ward
Electrical Engineering Department at Temple University Philadelphia, PA 19121 USA
6 shared publications Meysam Golmohammadi
Neural Engineering Data Consortium Temple University Philadelphia Pennsylvania USA
5 shared publications
Latest Publications
- 9 July 2026
- Published by IOP Publishing in Journal of Neural Engineering
- Vol. 23 (4) , 046008
Developing a binary communication protocol between biological neural networks using virtual white matter
Objective. Biological neural networks (BNNs) are composed of interconnected neurons, where a single neuron may connect to thousands of other neurons, and this interregional connectivity enables distributed computation across multiple nodes. Biocomputing exploits these inherent processing capabilities of neural tissue for computational tasks. However, most in vitro biocomputing systems rely on isolated neural cultures that may not be suitable for complex multi-network processing tasks. Building scalable biocomputing architectures requires reliable communication frameworks between physically separated biological processing units. Structured information exchange across distinct BNNs remains largely unexplored. Approach. virtual white matter (VWM) is a previously established platform that enables real-time functional connectivity between neural cultures in separate microelectrode array (MEA) dishes. In this study, we expanded the VWM platform to enable structured binary communication between dissociated cortical cells cultured in separate MEAs implemented on MED64 MEAs with Intan RHS stimulation/recording hardware and a custom real-time decoding pipeline. The system encodes 3‐bit data words using spatiotemporal electrical stimulation patterns, with designated electrodes representing binary states; evoked responses from selected output electrodes are decoded in real time using machine learning, and parity-based error correction is applied to enhance transmission fidelity. We validated the framework through unidirectional and bidirectional communication experiments across multiple independent neural preparation pairs. Main Results. The VWM system successfully transmitted 3-bit data packets between physically separate neural cultures, achieving individual bit decoding accuracies of 75%–90% and aggregate word accuracy exceeding 52%, well above the 12.5% chance level and an optimal-threshold spike-count baseline. Parity-based error correction with confidence-based bit toggling enhanced transmission fidelity, enabling reliable word-level communication. In bidirectional experiments, end-to-end round-trip accuracy reached approximately 20%, remaining above chance. Significance. This work demonstrates that physically separated neural culturescan reliably exchange structured information through engineered communication frameworks. By establishing reliable structured communication between living neural networks, VWM provides a foundational framework for scalable, distributed biocomputing architectures that may eventually support multi-network in vitro models of neural computation.
- 9 July 2026
- Published by IOP Publishing in Journal of Neural Engineering
- Vol. 23 (4) , 046008
Developing a binary communication protocol between biological neural networks using virtual white matter
Objective. Biological neural networks (BNNs) are composed of interconnected neurons, where a single neuron may connect to thousands of other neurons, and this interregional connectivity enables distributed computation across multiple nodes. Biocomputing exploits these inherent processing capabilities of neural tissue for computational tasks. However, most in vitro biocomputing systems rely on isolated neural cultures that may not be suitable for complex multi-network processing tasks. Building scalable biocomputing architectures requires reliable communication frameworks between physically separated biological processing units. Structured information exchange across distinct BNNs remains largely unexplored. Approach. virtual white matter (VWM) is a previously established platform that enables real-time functional connectivity between neural cultures in separate microelectrode array (MEA) dishes. In this study, we expanded the VWM platform to enable structured binary communication between dissociated cortical cells cultured in separate MEAs implemented on MED64 MEAs with Intan RHS stimulation/recording hardware and a custom real-time decoding pipeline. The system encodes 3‐bit data words using spatiotemporal electrical stimulation patterns, with designated electrodes representing binary states; evoked responses from selected output electrodes are decoded in real time using machine learning, and parity-based error correction is applied to enhance transmission fidelity. We validated the framework through unidirectional and bidirectional communication experiments across multiple independent neural preparation pairs. Main Results. The VWM system successfully transmitted 3-bit data packets between physically separate neural cultures, achieving individual bit decoding accuracies of 75%–90% and aggregate word accuracy exceeding 52%, well above the 12.5% chance level and an optimal-threshold spike-count baseline. Parity-based error correction with confidence-based bit toggling enhanced transmission fidelity, enabling reliable word-level communication. In bidirectional experiments, end-to-end round-trip accuracy reached approximately 20%, remaining above chance. Significance. This work demonstrates that physically separated neural culturescan reliably exchange structured information through engineered communication frameworks. By establishing reliable structured communication between living neural networks, VWM provides a foundational framework for scalable, distributed biocomputing architectures that may eventually support multi-network in vitro models of neural computation.
- 2 July 2026
- Published by Springer Nature
- p. 191-232
Enabling Microsegmentation: Digital Pathology Corpora for Advanced Model Development
- 2 July 2026
- Published by Springer Nature
- p. 191-232
Enabling Microsegmentation: Digital Pathology Corpora for Advanced Model Development
- Open Access
- 14 May 2025
- Published by IOP Publishing in Journal of Neural Engineering
- Vol. 22 (3) , 036013
Virtual white matter: a novel system for cross-dish neural interaction and modulation
Objective. Biological neural networks (BNNs) are characterized by complex interregional connectivity, allowing for seamless communication between different brain regions. In vitro models traditionally consist of single-dish neural cultures that cannot recapitulate the dynamics of interregional interactions and little effort has been made to interconnect multiple BNNs to process information through a hybrid interconnection of the biological and digital systems. Approach. We introduce virtual white matter (VWM), a novel platform enabling real-time functional digital connectivity between neural cultures in separate microelectrode array dishes. By detecting neural activity in one dish and providing precisely timed electrical stimulation to another, VWM recreates bidirectional inter-regional neural communication. The study introduces the conceptual framework, technical implementation, and proof-of-concept validation of the VWM system. Main Results. VWM represents a significant advancement in vitro modeling and data processing by enabling controlled interactions between heterogeneous neural cultures, such as different brain regions or cell types. The platform successfully enables the investigation of dynamic network behaviors and integration with both biological and artificial neural networks. Significance. VWM will push forward biocomputing, wetware computing, and organic intelligence by establishing a reliable form of interconnection between these systems. Furthermore, VWM has the potential to be applied in fields like therapeutic interventions that use directed neural plasticity to promote recovery from brain injury or disease responses. VWM enables complex in vitro models to be built with the same neural connectivity as in the human brain. VWM is versatile, placing it at the core of a transformational tool for experimental neuroscience, biocomputing, and translational research to bridge biological and digital systems.
- 14 May 2025
- Published by IOP Publishing in Journal of Neural Engineering
- Vol. 22 (3) , 036013
Open AccessVirtual white matter: a novel system for cross-dish neural interaction and modulation
Objective. Biological neural networks (BNNs) are characterized by complex interregional connectivity, allowing for seamless communication between different brain regions. In vitro models traditionally consist of single-dish neural cultures that cannot recapitulate the dynamics of interregional interactions and little effort has been made to interconnect multiple BNNs to process information through a hybrid interconnection of the biological and digital systems. Approach. We introduce virtual white matter (VWM), a novel platform enabling real-time functional digital connectivity between neural cultures in separate microelectrode array dishes. By detecting neural activity in one dish and providing precisely timed electrical stimulation to another, VWM recreates bidirectional inter-regional neural communication. The study introduces the conceptual framework, technical implementation, and proof-of-concept validation of the VWM system. Main Results. VWM represents a significant advancement in vitro modeling and data processing by enabling controlled interactions between heterogeneous neural cultures, such as different brain regions or cell types. The platform successfully enables the investigation of dynamic network behaviors and integration with both biological and artificial neural networks. Significance. VWM will push forward biocomputing, wetware computing, and organic intelligence by establishing a reliable form of interconnection between these systems. Furthermore, VWM has the potential to be applied in fields like therapeutic interventions that use directed neural plasticity to promote recovery from brain injury or disease responses. VWM enables complex in vitro models to be built with the same neural connectivity as in the human brain. VWM is versatile, placing it at the core of a transformational tool for experimental neuroscience, biocomputing, and translational research to bridge biological and digital systems.
- 22 July 2023
- Published by Springer Nature in Technology, Knowledge and Learning
- Vol. 28 (4) , 1937-1943
Multimodal Learning Analytics and Neurofeedback for Optimizing Online Learners’ Self-Regulation
- 22 July 2023
- Published by Springer Nature in Technology, Knowledge and Learning
- Vol. 28 (4) , 1937-1943
Multimodal Learning Analytics and Neurofeedback for Optimizing Online Learners’ Self-Regulation
- Open Access
Improved EEG Event Classification Using Differential Energy
Feature extraction for automatic classification of EEG signals typically relies on time frequency representations of the signal. Techniques such as cepstral-based filter banks or wavelets are popular analysis techniques in many signal processing applications including EEG classification. In this paper, we present a comparison of a variety of approaches to estimating and postprocessing features. To further aid in discrimination of periodic signals from aperiodic signals, we add a differential energy term. We evaluate our approaches on the TUH EEG Corpus, which is the largest publicly available EEG corpus and an exceedingly challenging task due to the clinical nature of the data. We demonstrate that a variant of a standard filter bank-based approach, coupled with first and second derivatives, provides a substantial reduction in the overall error rate. The combination of differential energy and derivatives produces a 24% absolute reduction in the error rate and improves our ability to discriminate between signal events and background noise. This relatively simple approach proves to be comparable to other popular feature extraction approaches such as wavelets, but is much more computationally efficient.
Open AccessImproved EEG Event Classification Using Differential Energy
Feature extraction for automatic classification of EEG signals typically relies on time frequency representations of the signal. Techniques such as cepstral-based filter banks or wavelets are popular analysis techniques in many signal processing applications including EEG classification. In this paper, we present a comparison of a variety of approaches to estimating and postprocessing features. To further aid in discrimination of periodic signals from aperiodic signals, we add a differential energy term. We evaluate our approaches on the TUH EEG Corpus, which is the largest publicly available EEG corpus and an exceedingly challenging task due to the clinical nature of the data. We demonstrate that a variant of a standard filter bank-based approach, coupled with first and second derivatives, provides a substantial reduction in the overall error rate. The combination of differential energy and derivatives produces a 24% absolute reduction in the error rate and improves our ability to discriminate between signal events and background noise. This relatively simple approach proves to be comparable to other popular feature extraction approaches such as wavelets, but is much more computationally efficient.
- Open Access
Semi-automated Annotation of Signal Events in Clinical EEG Data
To be effective, state of the art machine learning technology needs large amounts of annotated data. There are numerous compelling applications in healthcare that can benefit from high performance automated decision support systems provided by deep learning technology, but they lack the comprehensive data resources required to apply sophisticated machine learning models. Further, for economic reasons, it is very difficult to justify the creation of large annotated corpora for these applications. Hence, automated annotation techniques become increasingly important. In this study, we investigated the effectiveness of using an active learning algorithm to automatically annotate a large EEG corpus. The algorithm is designed to annotate six types of EEG events. Two model training schemes, namely threshold-based and volume-based, are evaluated. In the threshold-based scheme the threshold of confidence scores is optimized in the initial training iteration, whereas for the volume-based scheme only a certain amount of data is preserved after each iteration. Recognition performance is improved 2% absolute and the system is capable of automatically annotating previously unlabeled data. Given that the interpretation of clinical EEG data is an exceedingly difficult task, this study provides some evidence that the proposed method is a viable alternative to expensive manual annotation.
Open AccessSemi-automated Annotation of Signal Events in Clinical EEG Data
To be effective, state of the art machine learning technology needs large amounts of annotated data. There are numerous compelling applications in healthcare that can benefit from high performance automated decision support systems provided by deep learning technology, but they lack the comprehensive data resources required to apply sophisticated machine learning models. Further, for economic reasons, it is very difficult to justify the creation of large annotated corpora for these applications. Hence, automated annotation techniques become increasingly important. In this study, we investigated the effectiveness of using an active learning algorithm to automatically annotate a large EEG corpus. The algorithm is designed to annotate six types of EEG events. Two model training schemes, namely threshold-based and volume-based, are evaluated. In the threshold-based scheme the threshold of confidence scores is optimized in the initial training iteration, whereas for the volume-based scheme only a certain amount of data is preserved after each iteration. Recognition performance is improved 2% absolute and the system is capable of automatically annotating previously unlabeled data. Given that the interpretation of clinical EEG data is an exceedingly difficult task, this study provides some evidence that the proposed method is a viable alternative to expensive manual annotation.
- Open Access
An Analysis of Two Common Reference Points for EEGs
Clinical electroencephalographic (EEG) data varies significantly depending on a number of operational conditions (e.g., the type and placement of electrodes, the type of electrical grounding used). This investigation explores the statistical differences present in two different referential montages: Linked Ear (LE) and Averaged Reference (AR). Each of these accounts for approximately 45% of the data in the TUH EEG Corpus. In this study, we explore the impact this variability has on machine learning performance. We compare the statistical properties of features generated using these two montages, and explore the impact of performance on our standard Hidden Markov Model (HMM) based classification system. We show that a system trained on LE data significantly outperforms one trained only on AR data (77.2% vs. 61.4%). We also demonstrate that performance of a system trained on both data sets is somewhat compromised (71.4% vs. 77.2%). A statistical analysis of the data suggests that mean, variance and channel normalization should be considered. However, cepstral mean subtraction failed to produce an improvement in performance, suggesting that the impact of these statistical differences is subtler.
Open AccessAn Analysis of Two Common Reference Points for EEGs
Clinical electroencephalographic (EEG) data varies significantly depending on a number of operational conditions (e.g., the type and placement of electrodes, the type of electrical grounding used). This investigation explores the statistical differences present in two different referential montages: Linked Ear (LE) and Averaged Reference (AR). Each of these accounts for approximately 45% of the data in the TUH EEG Corpus. In this study, we explore the impact this variability has on machine learning performance. We compare the statistical properties of features generated using these two montages, and explore the impact of performance on our standard Hidden Markov Model (HMM) based classification system. We show that a system trained on LE data significantly outperforms one trained only on AR data (77.2% vs. 61.4%). We also demonstrate that performance of a system trained on both data sets is somewhat compromised (71.4% vs. 77.2%). A statistical analysis of the data suggests that mean, variance and channel normalization should be considered. However, cepstral mean subtraction failed to produce an improvement in performance, suggesting that the impact of these statistical differences is subtler.
- Open Access
Optimizing Channel Selection for Seizure Detection
Interpretation of electroencephalogram (EEG) signals can be complicated by obfuscating artifacts. Artifact detection plays an important role in the observation and analysis of EEG signals. Spatial information contained in the placement of the electrodes can be exploited to accurately detect artifacts. However, when fewer electrodes are used, less spatial information is available, making it harder to detect artifacts. In this study, we investigate the performance of a deep learning algorithm, CNN-LSTM, on several channel configurations. Each configuration was designed to minimize the amount of spatial information lost compared to a standard 22-channel EEG. Systems using a reduced number of channels ranging from 8 to 20 achieved sensitivities between 33% and 37% with false alarms in the range of [38, 50] per 24 hours. False alarms increased dramatically (e.g., over 300 per 24 hours) when the number of channels was further reduced. Baseline performance of a system that used all 22 channels was 39% sensitivity with 23 false alarms. Since the 22-channel system was the only system that included referential channels, the rapid increase in the false alarm rate as the number of channels was reduced underscores the importance of retaining referential channels for artifact reduction. This cautionary result is important because one of the biggest differences between various types of EEGs administered is the type of referential channel used.
Open AccessOptimizing Channel Selection for Seizure Detection
Interpretation of electroencephalogram (EEG) signals can be complicated by obfuscating artifacts. Artifact detection plays an important role in the observation and analysis of EEG signals. Spatial information contained in the placement of the electrodes can be exploited to accurately detect artifacts. However, when fewer electrodes are used, less spatial information is available, making it harder to detect artifacts. In this study, we investigate the performance of a deep learning algorithm, CNN-LSTM, on several channel configurations. Each configuration was designed to minimize the amount of spatial information lost compared to a standard 22-channel EEG. Systems using a reduced number of channels ranging from 8 to 20 achieved sensitivities between 33% and 37% with false alarms in the range of [38, 50] per 24 hours. False alarms increased dramatically (e.g., over 300 per 24 hours) when the number of channels was further reduced. Baseline performance of a system that used all 22 channels was 39% sensitivity with 23 false alarms. Since the 22-channel system was the only system that included referential channels, the rapid increase in the false alarm rate as the number of channels was reduced underscores the importance of retaining referential channels for artifact reduction. This cautionary result is important because one of the biggest differences between various types of EEGs administered is the type of referential channel used.
- Open Access
Gated Recurrent Networks for Seizure Detection
Recurrent Neural Networks (RNNs) with sophisticated units that implement a gating mechanism have emerged as powerful technique for modeling sequential signals such as speech or electroencephalography (EEG). The latter is the focus on this paper. A significant big data resource, known as the TUH EEG Corpus (TUEEG), has recently become available for EEG research, creating a unique opportunity to evaluate these recurrent units on the task of seizure detection. In this study, we compare two types of recurrent units: long short-term memory units (LSTM) and gated recurrent units (GRU). These are evaluated using a state of the art hybrid architecture that integrates Convolutional Neural Networks (CNNs) with RNNs. We also investigate a variety of initialization methods and show that initialization is crucial since poorly initialized networks cannot be trained. Furthermore, we explore regularization of these convolutional gated recurrent networks to address the problem of overfitting. Our experiments revealed that convolutional LSTM networks can achieve significantly better performance than convolutional GRU networks. The convolutional LSTM architecture with proper initialization and regularization delivers 30% sensitivity at 6 false alarms per 24 hours.
Open AccessGated Recurrent Networks for Seizure Detection
Recurrent Neural Networks (RNNs) with sophisticated units that implement a gating mechanism have emerged as powerful technique for modeling sequential signals such as speech or electroencephalography (EEG). The latter is the focus on this paper. A significant big data resource, known as the TUH EEG Corpus (TUEEG), has recently become available for EEG research, creating a unique opportunity to evaluate these recurrent units on the task of seizure detection. In this study, we compare two types of recurrent units: long short-term memory units (LSTM) and gated recurrent units (GRU). These are evaluated using a state of the art hybrid architecture that integrates Convolutional Neural Networks (CNNs) with RNNs. We also investigate a variety of initialization methods and show that initialization is crucial since poorly initialized networks cannot be trained. Furthermore, we explore regularization of these convolutional gated recurrent networks to address the problem of overfitting. Our experiments revealed that convolutional LSTM networks can achieve significantly better performance than convolutional GRU networks. The convolutional LSTM architecture with proper initialization and regularization delivers 30% sensitivity at 6 false alarms per 24 hours.
- 1 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
The Neuronix HPC cluster: Cluster management using free and open source software tools
One of the most notable impacts of computing advancements over the last few decades has been the decentralization of resources. As the cost of computer hardware continues to decrease, significant computational power continues to become more accessible to the consumer market. Similar to personal computing, this phenomenon has also enabled the growth of low-cost high-performance computing (HPC) (e.g., desktop supercomputers). Combined with advances in computational statistics and machine learning, HPC systems can now accommodate computationally expensive research using consumer-grade hardware.
- 1 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
The Neuronix HPC cluster: Cluster management using free and open source software tools
One of the most notable impacts of computing advancements over the last few decades has been the decentralization of resources. As the cost of computer hardware continues to decrease, significant computational power continues to become more accessible to the consumer market. Similar to personal computing, this phenomenon has also enabled the growth of low-cost high-performance computing (HPC) (e.g., desktop supercomputers). Combined with advances in computational statistics and machine learning, HPC systems can now accommodate computationally expensive research using consumer-grade hardware.