Introduction
High resolution nuclear magnetic resonance (NMR) spectroscopy is a
powerful and attractive approach in biochemistry (e.g., protein
analysis[1],
metabolomics[2–4]) and inorganic
chemistry[5]. In the recent years however, with
the rapid advances in NMR engineering (e.g., IC-based
spectrometer[6–12], microfluidic-based
chip[13–17], artificial
intelligence[18,19]) utilizing small foot-print
permanent magnet, the time-domain NMR instrumentations have seen a
myriad of interesting applications from point-of-care testing (PoCT)
medical diagnosis[7,20–23], industrial food
science [24,25], and in-situ oil-gas
exploration[26,27].
Biochemical information is typically detected and encoded in the
frequency domain (´chemical shift´) in the high-field NMR. In contra,
the low-field NMR, information is encoded in the time domain, with the
dephasing of the spin-spin relaxation (T2 relaxation) of
the water-proton of the observed sample used as diagnostic
criterion[20,21]. Time domain NMR however, suffers
from inherently long relaxation time of the (liquid) sample, (and hence
low signal-to-noise ratio (SNR)) causes unnecessarily long repetition
time[28,29]. Furthermore, the
T2-relaxation measurement (in one-dimensional) which is
frequently reported in NMR relaxometry experiments has limited number of
dimensionality (e.g., healthy/non-healthy)[20,21].
In this work, a new class of methodology is presented for rapid and
accurate object classification using PoCT NMR relaxometry with the aid
of machine learning (Fig. 1). It is demonstrated (using various edible
oils as proof-of-concept) that the sensitivity (´true positive rate´)
and specificity (´true negative rate´) of the classification is
substantially improved using higher order of (pseudo)-dimensionality
(e.g., 2D or multidimensional). Further, by leveraging on the advances
in machine learning techniques (e.g., pre-trained dataset) the detection
time was sped up (in minutes) as compared to conventional 2D or
multidimensional NMR (>hours), without resorting to using
Ultrafast NMR[30]. This methodology (termed as
Clustering NMR) is extremely useful for rapid and accurate
classification of objects (in less than a minute) using the low-field
NMR at point-of-need.