A recent study demonstrates that updated predictive models based on NIR spectra can outperform traditional nitrogen-based prescreening methods in identifying samples suitable for radiocarbon dating.
According to a research team from the University of Colorado Boulder, near-infrared (NIR) spectroscopy can facilitate rapid, fully non-destructive assessments of collagen preservation in archaeological bone. This study, which was published in the Journal of Archaeological Science, demonstrates that updated predictive models based on NIR spectra can outperform traditional nitrogen-based prescreening methods in identifying samples suitable for radiocarbon dating and other biomolecular analyses.1
Collagen degradation is a natural process.2 As a major protein found in bone and tissue, collagen is primarily responsible for supporting teeth, internal organs, skin, and bones.2,3 As people age, their body begins to break down and decay, and collagen degradation is a huge part of this process. Apart from aging, exposure to UV radiation and vitamin C deficiency can also lead to disruptions in collagen production and result in degradation.3
Typical collagen degradation leads to complications to archaeological science because archaeologists rely on this protein to acquire better samples of old bones and tissues from human remains. What complicates this further is that collagen degradation because of diagenesis is often invisible, yet it directly affects the success of techniques such as radiocarbon dating, stable isotope analysis, and zooarchaeology by mass spectrometry (ZooMS).1 Although there are some prescreening methods that have been used, including percent nitrogen (%N) measurements and C:N ratios, these methods are far from perfect. They typically require destructive sampling and laboratory infrastructure, limiting their use in field settings and on rare collections.1
In their study, the research team investigated how single-point NIR spectroscopy can probe collagen preservation beneath the bone surface and enable representative assessment of bulk collagen content without preparation or sampling. By training partial least squares regression (PLSR) and random forest (RF) models on whole bones with known collagen yields and validating them on an independent archaeological assemblage, the team identified a spectral window (2030–2060 nm) that delivers strong predictive performance while avoiding interference from consolidants commonly applied in conservation.1
Both the PLSR and RF models showed comparable accuracy in predicting collagen yield. They surpassed %N-based screening in selecting bones likely to produce datable collagen.1 The authors emphasized that restricting models to the 2030–2060 nm range reduces false positives linked to conservation materials in reference libraries, improving the reliability in curated collections.1
The researchers demonstrated in their study the practical applicability of NIR spectroscopy in studying bones. Because NIR instruments are portable and require only seconds per measurement, these tools have practical advantages.1 They allow for high-throughput screening of large faunal collections before destructive analyses. The method can also guide sampling strategy: the study recommends averaging predictions from two to three non-overlapping scan locations per bone (more for larger specimens) to ensure overall preservation quality, and aligning scan sites closely with intended sampling points to minimize spatial variability.
The research builds on earlier explorations of Raman and Fourier transform infrared (FT-IR) spectroscopy for collagen assessment but argues that those methods remain constrained by shallow penetration depth, sample preparation requirements, and limited generalizability across bone types and burial conditions.
“NIR spectroscopy combines high predictive accuracy with deep penetration and rapid, non-destructive analysis, making it particularly well-suited for large-scale prescreening of archaeological bone,” the authors wrote in their study.1 “The greater penetration depth of NIR light into bone allows for more representative sampling of bulk collagen content, improving the correlation between spectral data and actual collagen yields.”
The findings of this study indicate where the field of archaeology may head in the future and how spectroscopic tools would be deployed. By reducing unnecessary destructive testing, NIR prescreening supports conservation goals and ethical stewardship of archaeological materials, particularly in museum and heritage contexts where specimen integrity is important.1 Improved screening accuracy also increases the success rate of costly radiocarbon dating campaigns and biomolecular studies, enabling more reliable reconstructions of past diets, environments, and human–animal interactions.1
Although acknowledging that predictive models cannot fully replace extraction-based confirmation, the research team concludes that single-point NIR spectroscopy provides a robust complementary tool for archaeological bone analysis.
“Overall, single-point NIR spectroscopy is a valuable, complementary tool for archaeological bone analysis, offering actionable insights that complement established methods,” the authors conclude in their study.1
