Pooyan Mobtahej
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
Pooyan Mobtahej has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Seyed Ahmad Sajjadi
Department of Neurology, University of California, Irvine, CA, USA
3 shared publications Jing Zhang
Lamar University
2 shared publications- Sam T. Gouron
Department of Neurology, University of California, Irvine, CA, USA
2 shared publications - Thiri Htet Naing
Department of Neurology, University of California, Irvine, CA, USA
2 shared publications Xulong Zhang
Lamar University
2 shared publications- Anne-Marie Leiby
Department of Neurology, UCI School of Medicine, United States
1 shared publication
Latest Publications
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30 (3) , 2034-2048
Transformer-Based Deep Learning Approaches for Speech-Based Dementia Detection: A Systematic Review
As the population of older adults continues growing, so will the need for cost-effective approaches to early dementia detection. Deep learning approaches using patient speech samples show promising results. This systematic review examines studies utilizing speech-based deep learning for dementia diagnosis with the objective of identifying best practices for future data-driven dementia research. Studies researching speech-based deep learning for dementia were obtained from PubMed, Wiley Library, Science Direct, IEEE, Web of Science, Google Scholar, and arXiv. 80 studies were reviewed. Studies were analyzed in terms of model architecture and performance, speech features employed, and databases used. We observed that transformer-based approaches were most frequent, achieving an average accuracy of 85.71%, and that linguistic features outperform acoustic features. Our review identified several limitations, including a lack of dataset diversity, inconsistent classification of dementia severity levels across studies, and variability in how sample sizes and model performance metrics (e.g., accuracy, sensitivity, specificity) are reported. These inconsistencies hinder direct comparisons between studies and limit the reproducibility of findings. Still, our findings suggest that incorporation of transformers into current speech-based deep learning models can further improve detection of cognitive impairment. Consideration of our observations in future data-driven dementia research will lead to advancements in the development of diagnostic decision support systems for clinical practice.
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30 (3) , 2034-2048
Transformer-Based Deep Learning Approaches for Speech-Based Dementia Detection: A Systematic Review
As the population of older adults continues growing, so will the need for cost-effective approaches to early dementia detection. Deep learning approaches using patient speech samples show promising results. This systematic review examines studies utilizing speech-based deep learning for dementia diagnosis with the objective of identifying best practices for future data-driven dementia research. Studies researching speech-based deep learning for dementia were obtained from PubMed, Wiley Library, Science Direct, IEEE, Web of Science, Google Scholar, and arXiv. 80 studies were reviewed. Studies were analyzed in terms of model architecture and performance, speech features employed, and databases used. We observed that transformer-based approaches were most frequent, achieving an average accuracy of 85.71%, and that linguistic features outperform acoustic features. Our review identified several limitations, including a lack of dataset diversity, inconsistent classification of dementia severity levels across studies, and variability in how sample sizes and model performance metrics (e.g., accuracy, sensitivity, specificity) are reported. These inconsistencies hinder direct comparisons between studies and limit the reproducibility of findings. Still, our findings suggest that incorporation of transformers into current speech-based deep learning models can further improve detection of cognitive impairment. Consideration of our observations in future data-driven dementia research will lead to advancements in the development of diagnostic decision support systems for clinical practice.
- Open Access
- 24 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 486-487
Poster: Multimodal Deep Learning Approach to Detect Cognitive Impairment in the Oldest-Old with Digital Speech Markers
- 24 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 486-487
Open AccessPoster: Multimodal Deep Learning Approach to Detect Cognitive Impairment in the Oldest-Old with Digital Speech Markers
- Open Access
- 31 January 2024
- Published by Wiley
An LSTM-Autoencoder Architecture for Anomaly Detection Applied on Compressors Audio Data
Anomaly detection for the compressor systems is essential for the midstream industry. In this paper, anomaly classification and detections method based on neural network hybrid model named as Long Short-Term Memory (LSTM)-Autoencoder (AE) is proposed to detect anomalies on sequence pattern of audio data, collected by multiple sound sensors deployed at different components of each compressor system. To create a baseline for model evaluation this paper has also conducted experiments on different RNN architectures such as GRU, LSTM, Stacked LSTM and Stacked GRU with different functions. Each architecture used audio signals dataset received from the compressor system for experiments to consider the performance in each neural network model. All the network architectures and experiment results have considered with various model configuration and layers with different functions. According to performance results, optimal model for anomaly detection with best performance scores has proposed in this research. In conducted experiments, combined one-dimensional raw audio signals features using SC and Mel spectrogram features were fed to deep learning models to evaluate performance. Using SC and Mel spectrogram features achieved the best performance for anomaly detection in audio data based on the LSTM-AE in all evaluation metrics. Hence such hybrid methods can detect normal and anomaly audio signals collected from a compressor system effectively, which can increase the compressor reliability and the gas production line sustainability.
- 31 January 2024
- Published by Wiley
Open AccessAn LSTM-Autoencoder Architecture for Anomaly Detection Applied on Compressors Audio Data
Anomaly detection for the compressor systems is essential for the midstream industry. In this paper, anomaly classification and detections method based on neural network hybrid model named as Long Short-Term Memory (LSTM)-Autoencoder (AE) is proposed to detect anomalies on sequence pattern of audio data, collected by multiple sound sensors deployed at different components of each compressor system. To create a baseline for model evaluation this paper has also conducted experiments on different RNN architectures such as GRU, LSTM, Stacked LSTM and Stacked GRU with different functions. Each architecture used audio signals dataset received from the compressor system for experiments to consider the performance in each neural network model. All the network architectures and experiment results have considered with various model configuration and layers with different functions. According to performance results, optimal model for anomaly detection with best performance scores has proposed in this research. In conducted experiments, combined one-dimensional raw audio signals features using SC and Mel spectrogram features were fed to deep learning models to evaluate performance. Using SC and Mel spectrogram features achieved the best performance for anomaly detection in audio data based on the LSTM-AE in all evaluation metrics. Hence such hybrid methods can detect normal and anomaly audio signals collected from a compressor system effectively, which can increase the compressor reliability and the gas production line sustainability.
- 14 June 2023
- Published by H1 Connect
Faculty Opinions recommendation of Dementia Detection from Speech Using Machine Learning and Deep Learning Architectures.
This study focuses on using speech features to recognize dementia in patients. Previous studies have used speech features along with other linguistic features obtained from transcriptions, but this study explores using standalone speech features. The study utilizes speech samples from the Pitt corpus in the Dementia Bank and proposes a compact set of prosodic, voice quality, and cepstral speech features for the recognition task. The study uses machine learning (ML) and deep learning (DL) models and found that ML models performed better (87.6%) compared to DL models (85%). The results show that the proposed approach of using standalone speech features and ML models is promising for recognizing dementia with lower time and memory consumption compared to existing works.
- 14 June 2023
- Published by H1 Connect
Faculty Opinions recommendation of Dementia Detection from Speech Using Machine Learning and Deep Learning Architectures.
This study focuses on using speech features to recognize dementia in patients. Previous studies have used speech features along with other linguistic features obtained from transcriptions, but this study explores using standalone speech features. The study utilizes speech samples from the Pitt corpus in the Dementia Bank and proposes a compact set of prosodic, voice quality, and cepstral speech features for the recognition task. The study uses machine learning (ML) and deep learning (DL) models and found that ML models performed better (87.6%) compared to DL models (85%). The results show that the proposed approach of using standalone speech features and ML models is promising for recognizing dementia with lower time and memory consumption compared to existing works.
- Open Access
- 29 September 2022
- Published by Wiley in Computational and Mathematical Methods
- Vol. 2022, 1-22
An LSTM-Autoencoder Architecture for Anomaly Detection Applied on Compressors Audio Data
The compressors used in today’s natural gas production industry have an essential role in maintaining the production line operational. Each of the compressors’ components has routine maintenance tasks to avoid sudden failures. Hence, the significant advantages and benefits of performing preventative maintenance tasks in time are decreasing equipment downtime, saving additional costs, and improving the safety and reliability of the whole system. In this paper, anomaly classification and detection methods based on a neural network hybrid model named Long Short-Term Memory (LSTM)-Autoencoder (AE) is proposed to detect anomalies in sequence pattern of audio data, collected by multiple sound sensors deployed at different components of each compressor system for predictive maintenance. In research methodology, this paper has conducted experiments that employed different RNN architectures such as GRU, LSTM, Stacked LSTM, and Stacked GRU with various functions to create a baseline for model evaluation. Each architecture used audio signals dataset received from the compressor system for experiments to consider each neural network model’s performance. According to performance results, an optimal model for anomaly detection with the best performance scores has been proposed in this research. Experiments combined one-dimensional raw audio signal features using SC and Mel spectrogram features were fed to deep learning models to evaluate performance. Hence, such hybrid methods can effectively detect normal and anomaly audio signals collected from a compressor system, increasing the compressor system’s reliability and the sustainability of the gas production line. The combination of multiple-resource features in the proposed hybrid model showed a 100% score in all four-evaluation metrics such as accuracy, precision, recall, and F1 in LSTM-based autoencoder in both test and train results.
- 29 September 2022
- Published by Wiley in Computational and Mathematical Methods
- Vol. 2022, 1-22
Open AccessAn LSTM-Autoencoder Architecture for Anomaly Detection Applied on Compressors Audio Data
The compressors used in today’s natural gas production industry have an essential role in maintaining the production line operational. Each of the compressors’ components has routine maintenance tasks to avoid sudden failures. Hence, the significant advantages and benefits of performing preventative maintenance tasks in time are decreasing equipment downtime, saving additional costs, and improving the safety and reliability of the whole system. In this paper, anomaly classification and detection methods based on a neural network hybrid model named Long Short-Term Memory (LSTM)-Autoencoder (AE) is proposed to detect anomalies in sequence pattern of audio data, collected by multiple sound sensors deployed at different components of each compressor system for predictive maintenance. In research methodology, this paper has conducted experiments that employed different RNN architectures such as GRU, LSTM, Stacked LSTM, and Stacked GRU with various functions to create a baseline for model evaluation. Each architecture used audio signals dataset received from the compressor system for experiments to consider each neural network model’s performance. According to performance results, an optimal model for anomaly detection with the best performance scores has been proposed in this research. Experiments combined one-dimensional raw audio signal features using SC and Mel spectrogram features were fed to deep learning models to evaluate performance. Hence, such hybrid methods can effectively detect normal and anomaly audio signals collected from a compressor system, increasing the compressor system’s reliability and the sustainability of the gas production line. The combination of multiple-resource features in the proposed hybrid model showed a 100% score in all four-evaluation metrics such as accuracy, precision, recall, and F1 in LSTM-based autoencoder in both test and train results.