Visualizing the structural and chemical heterogeneity of fruit and vegetables using advanced imaging techniques: fundamentals, instrumental aspects, applications and future perspectives.

Fruit and vegetables (F&V) are vastly complicated products with highly diverse chemical and structural characteristics. Advanced imaging techniques either combine imaging with spectral information or can provide excellent tissue penetration, and enable the possibility to target, visualize and even qualify the chemical and physical (structural) heterogeneity within F&V. In this review, visible and/or near infrared hyperspectral imaging, Fourier transform infrared microspectroscopic imaging, Raman imaging, X-ray and magnetic resonance imaging to reveal chemical and structural information in a spatial context of F&V at the macro- (entire products), meso- (tissues), and micro- (individual cells) scales are comprehensively summarized. In addition, their basic concepts and operational procedures, particularly sample preparation and instrumental parameter adjustments, are addressed. Finally, future challenges and perspectives of these techniques are put forward. These imaging techniques are powerful tools to assess the biochemical and structural heterogeneity of F&V. Cost reduction, sensor fusion and data sharing platforms are future trends. More emphasis on aspects of knowledge and extension at the level of academia and research, especially on how to select techniques, choose operational parameters and prepare samples, are important to overcome barriers for the wider adoption of these techniques to improve the evaluation of F&V quality.

Hyperspectral imaging reveals chemical heterogeneity of fruit and vegetables.Imaging techniques provide spatial insights in fruit and vegetables at multiple scales.Future trends are cost reduction, sensor fusion and data sharing.Instrumental adjustment and sample preparation should receive more attention.

Global and Partial Effect Assessment in Metabolic Syndrome Explored by Metabolomics.

In nutrition and health research, untargeted metabolomics is actually analyzed simultaneously with clinical data to improve prediction and better understand pathological status. This can be modeled using a multiblock supervised model with several input data blocks (metabolomics, clinical data) being potential predictors of the outcome to be explained. Alternatively, this configuration can be represented with a path diagram where the input blocks are each connected by links directed to the outcome-as in multiblock supervised modeling-and are also related to each other, thus allowing one to account for block effects. On the basis of a path model, we show herein how to estimate the effect of an input block, either on its own or conditionally to other(s), on the output response, respectively called "global" and "partial" effects, by percentages of explained variance in dedicated PLS regression models. These effects have been computed in two different path diagrams in a case study relative to metabolic syndrome, involving metabolomics and clinical data from an older men's cohort (NuAge). From the two effects associated with each path, the results highlighted the complementary information provided by metabolomics to clinical data and, reciprocally, in the metabolic syndrome exploration.

Fruit variability impacts puree quality: Assessment on individually processed apples using the visible and near infrared spectroscopy.

This study was designed to have the absolute definition of 'one apple to one puree', which gave a first insight into the impacts of fruit inter-variability (between varieties) and intra-variability (between individual fruits) on the quality of processed purees. Both the inter-variability of apple varieties and the intra-variability of single apples induced intensive changes of appearance, chemical and textural properties of their corresponding microwave-cooked purees. The intra-variability of cooked purees was different according to apple cultivars. Some strong correlations of visible-near infrared (VIS-NIR) spectra were observed between fresh and cooked apples, particularly in the regions 665-685 nm and 1125-1400 nm. These correlations allowed then the indirect predictions of puree color (a* and b*, RPD ≧ 2.1), viscosity (RPD ≧ 2.3), soluble solids content (SSC, RPD = 2.1), titratable acidity (RPD = 2.8), and pH (RPD = 2.5) from the non-destructive acquired VIS-NIR spectra of raw apples.

Mid-infrared technique to forecast cooked puree properties from raw apples: A potential strategy towards sustainability and precision processing.

The potential of MIRS was investigated to: i) differentiate cooked purees issued from different apples and process conditions, and ii) predict the puree quality characteristics from the spectra of homogenized raw apples. Partial least squares (PLS) regression was tested both, on the real spectra of cooked purees and their reconstructed spectra calculated from the spectra of homogenized raw apples by direct standardization. The cooked purees were well-classified according to apple thinning practices and cold storage durations, and to different heating and grinding conditions. PLS models using the spectra of homogenized raw apples can anticipate the titratable acidity (the residual predictive deviation (RPD) = 2.9), soluble solid content (RPD = 2.8), particle averaged size (RPD = 2.6) and viscosity (RPD ≥ 2.5) of cooked purees. MIR technique can provide sustainable evaluations of puree quality, and even forecast texture and taste of purees based on the prior information of raw materials.

Fresh, freeze-dried or cell wall samples: Which is the most appropriate to determine chemical, structural and rheological variations during apple processing using ATR-FTIR spectroscopy?

Attenuated total reflectance Fourier transform spectroscopy (ATR-FTIR) was applied on fresh (NF), freeze-dried (FD) and cell wall materials (AIS) of raw and processed apples. These samples prepared from 36 apple sets and the corresponding 72 purees, issued from different varieties, agricultural practices, storage periods and processing conditions, were used to build models including exploratory analysis, supervised classification and multivariate calibration. Fresh and freeze-dried samples presented similar fingerprint spectral variations due to processing. ATR-FTIR directly on fresh purees satisfactorily predicted textural properties such as particle average size and volume (RPD > 3.0), while freeze-drying improved assessment of chemical (RPD > 3.2) and rheological (RPD > 3.1) parameters using partial least-squares regression. The assessment of texture and macrocomponents of purees can be obtained with a limited sample preparation. For research applications because of a need of sample preparation, changes of cell wall composition during fruit processing could be assessed in relationship with pectin degradation.

Subpixel detection of peanut in wheat flour using a matched subspace detector algorithm and near-infrared hyperspectral imaging.

The detection of adulterations in food powder products represents a high interest especially when it concerns the health of the consumers. The food industry is concerned by peanut adulteration since it is a major food allergen often used in transformed food products. Near-infrared hyperspectral imaging is an emerging technology for food inspection. It was used in this work to detect peanut flour adulteration in wheat flour. The detection of peanut particles was challenging for two reasons: the particle size is smaller than the pixel size leading to impure spectral profiles; peanut and wheat flour exhibit similar spectral signatures and variability. A Matched Subspace Detector (MSD) algorithm was designed to take these difficulties into account and detect peanut adulteration at the pixel scale using the associated spectrum. A set of simulated data was generated to overcome the lack of reference values at the pixel scale and to design appropriate MSD algorithms. The best designs were compared by estimating the detection sensitivity. Defatted peanut flour and wheat flour were mixed in eight different proportions (from 0.02% to 20%) to test the detection performances of the algorithm on real hyperspectral measurements. The number and positions of the detected pixels were investigated to show the relevancy of the results and validate the design of the MSD algorithm. The presented work proved that the use of hyperspectral imaging and a fine-tuned MSD algorithm enables to detect a global adulteration of 0.2% of peanut in wheat flour.

A new application of NIR spectroscopy to describe and predict purees quality from the non-destructive apple measurements.

The potential of NIRS was investigated on both apples and purees to (i) examine factors involving quality variability (variety, agricultural practice, cold storage, puree mechanical refining level) and (ii) establish the link between quality traits before and after processing in order to predict the quality characteristics of purees from spectral information of raw apples. Apples and purees were well-classified at over 82% and 88% according to varieties and storage times respectively. The PLS models showed a good ability to estimate puree characteristics from spectra acquired on corresponding apples such as viscosity (R2 > 0.82), cell wall content (R2 > 0.81) and also dry matter (R2 > 0.83), soluble solids content (R2 > 0.80) and titratable acidity (R2 > 0.80). NIR technique should be a useful tool for industry insofar as it can give a reliable assessment of texture and taste of the final products based on the non-destructive fresh materials evaluation.

Application of direct calibration in multivariate image analysis of heterogeneous materials.

Many scientific instruments produce multivariate images characterized by three-way tables, an element of which represents the intensity value at a spatial location for a given spectral channel. A problem frequently encountered is to attempt estimating the contributions of some compounds at each location of these images. Usual regression methods of calibration, such as PLS, require having a matrix of calibration X (n×p) and the corresponding vector y of the dependent variable (n×1). X can be built up by sampling pixel-vectors in the images, but y is sometimes difficult to obtain, if the surface of the samples is formed by chemically heterogeneous regions. In this case, the quantitative analyses related to y may be difficult, if the pixels represent very small areas (for example on microscopic images) or very large ones (satellite images). This is for example the case when dealing with biological solid samples representing different tissues. Direct Calibration (DC), sometimes referred to as "spectral unmixing", do not require having such a calibration set. However, it is indeed needed to have both a matrix of "perturbing" pixel-vectors (noted K) and a vector of the "pure" component spectrum to be analyzed (p), which are more easily obtainable. For estimating the contribution, the unknown pixel vector x and the pure spectrum p are first projected orthogonally onto K giving the vectors x(⊥) onto p(⊥), respectively. The contribution is then estimated by a second projection of x(⊥) onto p(⊥). A method, based on principal component analysis, for determining the optimal dimensions of K is proposed. DC was applied on a collection of multivariate images of kernel of wheat to estimate the proportion of three tissues, namely out-layers, "waxy"endosperm and normal endosperm. The eventual results are presented as images of wheat kernels in false colors associated to the estimated proportions of the tissues. It is shown that DC is appropriate for estimating contributions in situations in which the more usual methods of calibration cannot be applied.

Molecular size and mass distributions of native starches using complementary separation methods: asymmetrical flow field flow fractionation (A4F) and hydrodynamic and size exclusion chromatography (HDC-SEC).

Starch consists of a mixture of two α-glucans built mainly upon α-(1,4) linkages: amylose, an essentially linear polymer, and amylopectin, a branched polymer containing 5-6% of α-(1,6) linkages. The aim of the present work was to analyze the structural properties of native starches displaying different amylose-to-amylopectin ratios and arising from different botanical sources, using asymmetrical flow field flow fractionation (A4F) and a combination of hydrodynamic and size-exclusion chromatography (HDC-SEC) coupled with multiangle laser light scattering, online quasi-elastic light scattering, and differential refractive index techniques. The procedure, based upon dimethyl sulfoxide pretreatment and then solubilization in water, generates a representative injected sample without altering the initial degree of polymerization. The amylopectin weight-average molar masses and radii of gyration were around 1.0 × 10(8)-4.8 × 10(8) g mol(-1) and 110-267 nm, respectively. For each starch sample, the hydrodynamic radius (R(H)) distributions and the molar mass distributions obtained from the two fractionation systems coupled with light scattering techniques were analyzed. The size determination scales were extended by means of R(H) calibration curves. HDC-SEC and A4F data could be matched. However, A4F enabled a better separation of amylopectins and therefore an enhanced structural characterization of the starches. The two advantages of this experimental approach are (1) it can directly obtain distributions as a function of both molar mass and size, while taking account of sample heterogeneity, and (2) it is possible to compare the results obtained using the different techniques through the direct application of R(H) distributions.