EXPRESS: Adulteration Detection for Poria Cocos Based on Near-Infrared Two-Dimensional Correlation Spectroscopy and...

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Machine Learning

Background:

• Poria cocos (PC) is a vital traditional Chinese medicine, but market adulteration compromises its efficacy and safety.
• Current identification methods are often destructive, slow, and inadequate for complex adulteration scenarios.

Purpose of the Study:

• To develop a rapid, non-destructive method for qualitative and quantitative analysis of PC adulteration.
• To integrate near-infrared two-dimensional correlation spectroscopy (NIR 2DCOS) with deep learning for enhanced detection.

Main Methods:

• Collected NIR spectra (960-1600 nm) from pure and adulterated PC samples with varying concentrations.
• Generated a systematic 2DCOS dataset using diverse spectral preprocessing and 2DCOS techniques.
• Developed a Bidirectional Long Short-Term Memory network with an attention mechanism (BiLSTM-Attn) for spectral analysis.

Main Results:

• The BiLSTM-Attn model achieved 100% classification accuracy for PC adulteration.
• Superior regression performance was observed with R² = 0.9920, RMSE = 2.41, and RPD = 12.85.
• The integrated approach demonstrated robust and accurate detection of PC adulteration.

Conclusions:

• The combination of NIR 2DCOS and deep learning provides an effective framework for PC quality control.
• This non-destructive strategy offers significant advantages over conventional methods for traditional Chinese medicine analysis.
• The proposed method has broad applicability for ensuring the authenticity and safety of medicinal materials.