NMR-baking and Multivariate Prediction of Instrumental Texture Parameters in Bread

Abstract To study the kinetics of the bread baking process, transverse relaxation ( T 2 ) of protons was measured during a baking process performed inside the magnet of a pulsed low field 1 H nuclear magnetic resonance (NMR) instrument. Experimental NMR relaxation data were analysed both by chemometric data analysis and by multi-exponential curve-fitting. Throughout the entire baking process from dough to bread three T 2 -components were determined. During the NMR-baking process significant shifts were observed in the characteristic time constants at c . 55 °C (gelatinisation of starch) and at c . 85 °C. In a second experiment staling of white bread crumb aged 0–8 days was investigated by texture analysis and NMR relaxation. High correlations ( r >0·9) between texture parameters and NMR relaxation data of bread crumb were found by partial least squares regression (PLSR) models. Firmness and elasticity as measured by a Texture Analyser were predicted with an estimated error (RMSECV) of 150 (range 200–2200) and 0·032 (range 0·4–0·7), respectively. Future texture of the bread samples was also predictable by use of NMR relaxation data from the early storage period (day 0 to day 3).