EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Hybrid Electroencephalogram-genomic Deep Learning for Personalised Depth of Anaesthesia Monitoring: A Transformer-based Depth of Anaesthesia Index Calculator with Real-time Pharmacogenomic Adaptation

Turki Khalid Alsayed, Rawan Faris Almalki, Maria Suliman Aljumah, Faisal Mahmoud Habib, Ibrahim Abdrabalreda Alhumaidan, Tahani Marzoog Almuteri, Ayman Mohammed Kharaba

Journal of Advanced Trends in Medical Research · 2025

DOI: 10.4103/atmr.atmr_115_25 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Advanced Trends in Medical Research, 2025. Verfasst von Turki Khalid Alsayed und 6 weiteren Autor:innen. Die Arbeit nennt 43 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Traditional depth of anaesthesia (DoA) monitoring systems, such as Bispectral Index (BIS), overlook individual genetic differences in anaesthetic drug metabolism, leading to suboptimal dosing. These limitations highlight the need for personalised monitoring tools that integrate both neurophysiological signals and genetic information.

Methods

We introduce a hybrid deep learning system that combines electroencephalogram (EEG) signal processing with genomic data to enhance real-time DoA assessment. EEG features are extracted using wavelet-based time–frequency analysis and integrated with single-nucleotide polymorphism profiles through a transformer-based neural network. The model includes a pharmacokinetic feedback mechanism to dynamically recalibrate predictions in response to changing drug concentrations. Real-time implementation is achieved using field-programmable gate array (FPGA) acceleration for EEG processing and transformer inference.

Results

Evaluation on a dataset of 1872 patients revealed that the hybrid model achieved a concordance correlation coefficient of 0.91 and a mean absolute error of 4.2, outperforming conventional EEG-only and BIS-based systems. Personalised performance improvements were especially notable in pharmacogenetically distinct groups such as CYP2B6 poor and ultra-rapid metabolisers. The model also demonstrated reliable sub-10 ms processing latency, suitable for clinical deployment.

Conclusion

The proposed EEG-genomic framework represents a significant advancement in personalised anaesthesia monitoring. By combining real-time neurophysiological and genetic data, it enables more accurate and adaptive DoA estimation, with the potential to improve patient safety and dosing precision in diverse surgical populations.

Zitieren

APA

Alsayed, T. K., Almalki, R. F., Aljumah, M. S., Habib, F. M., Alhumaidan, I. A., Almuteri, T. M., & Kharaba, A. M. (2025). Hybrid Electroencephalogram-genomic Deep Learning for Personalised Depth of Anaesthesia Monitoring: A Transformer-based Depth of Anaesthesia Index Calculator with Real-time Pharmacogenomic Adaptation. Journal of Advanced Trends in Medical Research. https://doi.org/10.4103/atmr.atmr_115_25

Vancouver

Alsayed TK, Almalki RF, Aljumah MS, Habib FM, Alhumaidan IA, Almuteri TM, et al. Hybrid Electroencephalogram-genomic Deep Learning for Personalised Depth of Anaesthesia Monitoring: A Transformer-based Depth of Anaesthesia Index Calculator with Real-time Pharmacogenomic Adaptation. Journal of Advanced Trends in Medical Research. 2025. doi:10.4103/atmr.atmr_115_25

BibTeX
@article{Alsayed2025,
  author = {Alsayed, Turki Khalid and Almalki, Rawan Faris and Aljumah, Maria Suliman and Habib, Faisal Mahmoud and Alhumaidan, Ibrahim Abdrabalreda and Almuteri, Tahani Marzoog and Kharaba, Ayman Mohammed},
  title = {Hybrid Electroencephalogram-genomic Deep Learning for Personalised Depth of Anaesthesia Monitoring: A Transformer-based Depth of Anaesthesia Index Calculator with Real-time Pharmacogenomic Adaptation},
  journal = {Journal of Advanced Trends in Medical Research},
  year = {2025},
  doi = {10.4103/atmr.atmr_115_25}
}
Kostenlos starten →