Artificial tongue

Chemometric Brains for Artificial Tongues Paolo Oliver / M. Chiara Casolino, and ... 

The last years showed a significant trend toward the exploitation of rapid and economic analytical devices able to provide multiple information about samples. Among these, the so-called artificial tongues represent effective tools which allow a global sample... 

In this review, a critical overview of artificial tongue applications over the last decade is outlined. In particular, the focus is centered on the chemometric techniques, which allow the extraction of valuable information from nonspecific data. The basic steps of signal processing and pattern recognition are discussed and the principal chemometric techniques are described in detail, highlighting benefits and drawbacks of each one. Furthermore, some novel methods recently introduced and particularly suitable for artificial tongue data are presented. 

The term artificial tongue is used in two main branches of science. The first one concerns the neurophysiological studies aimed at developing perceptual supplementation devices, with biomedical engineering applications to human disabilities. The second utilization of the term artificial tongue concerns, instead, the laboratory analytical instruments used in combination with chemometric techniques to obtain complex information (often sensory-like, but not only) on samples. As for this latter meaning, also the synonymous electronic tongue is frequently used, particularly for electroanalytical devices. 

A wide number of sensor types have been described in the literature, from optical to mass spectrometry-based devices, but the sensors most commonly used in artificial tongues are electrochemical. 

The first utilization of the term artificial tongue dates to 1978, when H.W. Harper and M. Rossetto presented an apparatus based on conductance measurements able to mimic the taste stimulus delivery systems (Harper and Rossetto, 1978). This pioneering work has represented an isolated study for several years. The first example of a factual taste sensor was developed, in fact, by Toko and coworkers in 1990 (Hayashi et ah, 1990 Toko et ah, 1990). It was based on ion-sensitive lipid membranes and it was claimed to be able to respond to the basic tastes of the human tongue sour, sweet, bitter, salt, and umami. 

From examination of almost 230 papers published in ISI journals over the period 1996-2009, it emerges that the many sectors in which artificial tongues have found applications go from the industrial plant processmonitoring to biomedical and clinical studies (see Fig. 2.4). As for this latter field, a number of sensors for the determination of several clinical... 

An)nvay, the principal use of artificial tongues is within the food sciences. The applications concern almost exclusively liquid food mainly wine (about 18% of the studies examined), fruit juices (almost 15%), mineral water (about 13%), followed by infusions like tea and coffee, soft drinks, milk, beer, and other alcoholic beverages. All these liquid foods are characterized by both low-viscosity and high-polarity values. 

In some interesting studies, artificial tongues are employed to evaluate the effect of a number of process factors on the quality of the finished products (Esbensen et al., 2004 Rollm de Moura et al., 2007 Rudnitskaya et ah, 2009b) or the effects of storage time and conditions (Apetrei et ah, 2007 Cosio et ah, 2007 Kantor et ah, 2008 Parra et ah, 2006a Rodriguez-Mendez et ah, 2007). 

Figure 2.6 shows that almost one-half of the artificial tongues described in the scientific literature of the last decade are based on potentiometric devices. 

In the case of the complex analytical signals arising from artificial tongue instruments, a number of preprocessing tools may be employed for three main purposes, namely ... 

Nevertheless, in many electronic tongue studies, such constraints are ignored and ANNs are used as the default choice. This choice is also made in cases with very poor data sets and without performing a proper validation. This may be due to the fact that the related computational software is easily available and that many people have a propensity to follow the predominant trends and to use the most potent instruments available, without critical considerations. Furthermore, perhaps, there is a fashionable association of ideas coimecting the concepts of artificial tongue and artificial intelligence. 

Regression techniques provide models for quantitative predictions. The ordinary least squares (OLS) method is probably the most used and studied historically. Nevertheless, it presents a number of restrictions which often limit its applicability in the case of artificial tongue data. 

Oliveri et al. (2009) presented the development of an artificial tongue based on cyclic voltammetry at Pt microdisk electrodes for the classification of olive oils according to their geographical origin the measurements are made directly in the oil samples, previously mixed with a proper quantity of a RTIL (room temperature ionic liquid). The pattern recognition techniques applied were PCA for data exploration and fc-NN for classification, validating the results by means of a cross-validation procedure with five cancellation groups. 

An interesting issue, in the field of nonspecific analysis, is the fusion of data coming from different analytical techniques. As for artificial tongues, potentiometric and voltammetric data have been employed... 

Artificial tongues represent effective analytical tools able to characterize samples by means of a nonspecific approach. They may provide information useful for many purposes, allowing both qualitative and quantitative applications. 

---