Related Experiment Video
Updated: Jul 11, 2025
Brain Source Imaging in Preclinical Rat Models of Focal Epilepsy using High-Resolution EEG Recordings
Published on: June 6, 2015
Antiepileptic drug concentration detection based on Raman spectroscopy and an improved snake
Xinghu Fu1, Xiqing Cao1, Zizhen Fu1
1School of Information Science and Engineering, The Key Laboratory for Special Fiber and Fiber Sensor of Hebei Province, Yanshan University, Qinhuangdao, China. fuxinghu@ysu.edu.cn.
This study introduces a novel method combining Raman spectroscopy with an Improved Snake Optimization-Convolutional Neural Network (ISO-CNN) algorithm for precise antiepileptic drug concentration measurement. The ISO-CNN model demonstrated high accuracy in predicting drug levels, offering a promising tool for pharmaceutical analysis.
Area of Science:
- Analytical Chemistry
- Spectroscopy
- Machine Learning
Background:
- Accurate measurement of antiepileptic drug concentrations is crucial for therapeutic efficacy.
- Raman spectroscopy offers non-destructive, rapid analysis of pharmaceutical compounds.
- Existing methods for spectral data analysis may require complex pre-processing or lack precision.
Purpose of the Study:
- To develop and evaluate a novel method for measuring mixed antiepileptic drug concentrations using Raman spectroscopy.
- To investigate the performance of an Improved Snake Optimization-Convolutional Neural Network (ISO-CNN) algorithm for spectral data analysis.
- To compare the proposed ISO-CNN algorithm with other optimization algorithms for drug concentration prediction.
Main Methods:
- Utilized Raman spectroscopy to measure 360 mixtures of oxcarbazepine, carbamazepine, and lamotrigine.
- Developed and applied an Improved Snake Optimization (ISO) algorithm integrated with a Convolutional Neural Network (CNN) for spectral data analysis.
- Pre-processed raw spectral data and trained/evaluated the ISO-CNN model.
Main Results:
- The ISO-CNN algorithm achieved excellent performance in predicting oxcarbazepine concentration, with high determination coefficients (R²) of 0.99378 for validation and 0.99627 for testing.
- The root mean square errors (RMSE) for oxcarbazepine prediction were low (0.0295 for validation, 0.0278 for testing), indicating high precision.
- Comparative analysis showed the superiority of the ISO-CNN algorithm over other tested optimization algorithms.
Conclusions:
- Raman spectroscopy combined with the ISO-CNN machine learning algorithm provides a powerful and accurate approach for predicting antiepileptic drug concentrations.
- The developed method demonstrates potential for routine pharmaceutical analysis and therapeutic drug monitoring.
- The improved optimization strategy of the ISO algorithm enhances the speed and accuracy of spectral data analysis.
