The world of the food industry search to ensure satisfactory multisensory experiences for consumers through the consolidation of quality standards for food products ( Blissett & Fogel, 2013; Tuorila & Hartmann, 2020). The first approach to each food matrix allows the consumer to identify attributes related to size, shape, color, and brightness. A second approach allows more direct interactions related to the perception of smell, aroma, taste, temperature, and texture of the product ( Fine & Riera, 2019; Isogai & Wise, 2016; Moding et al., 2020; Nederkoorn et al., 2018). Recognizing these sensory characteristics determines the acceptance or rejection of the food ( Costell et al., 2009; Torres Gonzalez et al., 2015; Wadhera & Capaldi-Phillips, 2014). One of the disciplines that study the sensory characteristics of food is sensory analysis. This term became a field of study in the 17th century when Jean Anthelme Brillat-Savarin, in 1825, wrote his first book entitled Philosophy of Taste, in which he established the basis for the analysis of food and how it is perceived ( Chong, 2012). The constant evolution of the concept and applicability of sensory analysis has consolidated its study using trained panelists or instrumental methods. Although the analyses carried out by these panelists constitute an essential source of information for the acceptance or rejection of a food product, this can be subjective due to biological, social, and other external factors surrounding the subject ( Buratti et al., 2018; Loutfi et al., 2015; Tan & Xu, 2020).

One of the main limitations when implementing sensory tests is the number of required panelists, ranging from 7 to 100 depending on the test type ( Lawless & Heymann, 2010; O’Mahony, 2017). This implies an investment of human and economic resources, raw materials, and/or time. This limitation has motivated researchers to generate technologies to identify and quantify some sensory characteristics of foods with greater precision ( Akimoto et al., 2017; Kusumi et al., 2020; Pascual et al., 2018).

Such developments search to mimic the functioning of the five senses, such is the case of electronic noses (e-noses) and tongues (e-tongues), which upon contact with food, generate an electronic response from a chemical interaction, which is interpreted by a digital information processing system ( Banerjee et al., 2019; Bonah et al., 2020). Similarly, image analysis through devices such as cameras seek to simulate the sense of eyesight ( Ansari et al., 2021; Barbon et al., 2017; Kakani et al., 2020; Khojastehnazhand & Ramezani, 2020); concerning touch and hearing, some reports show various technological tools that measure force and sound, seeking to imitate the behavior of these senses ( Akimoto et al., 2019; Kato et al., 2017; Kusumi et al., 2020).

Each of the technological tools mentioned above contributes a description of the primary sensory characteristics of the food matrix to be evaluated. Few works show the use of more than one technological tool, despite the fact that the combination of these tools allows for better management of different types of resources (scientific personnel, economic resources, time, raw materials). However, Huang et al. (2023), Chen et al. (2023), Gao et al. (2022) and Martínez-Velasco et al. (2022) made use of more than one technological tool. This article consolidates information on some technological tools reported in the literature for sensory analysis in various food matrices.

Odor is one of the most representative attributes of food. This can be expressed as one of the qualities of Volatile Organic Compounds (VOCs), so unique and distinctive that they are considered fingerprints ( Bonah et al., 2020; Tan & Xu, 2020).

Generally, the sensory analysis method to identify such components is performed by panelists who rate and classify on different scales the odor perceived in the sample ( Barbieri et al., 2021; Giungato et al., 2018; Niu et al., 2019; Świąder & Marczewska, 2021). On the other hand, different methods have been developed for the identification of VOCs, such as: 1) Gas Chromatography-Olfactometry (GC-O): this methodology is to assess the odor impact of volatile compounds present in a sample extract and assign a degree of significance to each individual compound. GC-olfactometry, or GC-O, encompasses a range of techniques that rely on human assessors to detect and assess the volatile compounds released during a gas chromatography separation ( Delahunty et al., 2006). 2) Gas Chromatography-Mass Spectrometry (GC-MS): is an analytical technique that integrates the capabilities of both gas chromatography and mass spectrometry to detect and identify various substances present in a given test sample. This technique is utilized within flavor research to identify the aroma-contributing compounds in various food products. These methodologies encompass approaches such as dilution analysis, detection frequency techniques, posterior intensity assessments, and time-intensity evaluations ( Van Ruth, 2001). 3) Headspace Solid Phase Microextraction (HS-SPME), is a modern and highly sensitive sample preparation technique that does not require solvents. HS-SPME has emerged as a potent sample preparation method that efficiently enables the isolation and concentration of analytes from intricate matrices. It utilizes a coated fiber to concentrate volatile and semi-volatile compounds from a sample, operating on the principles of adsorption/absorption and subsequent desorption ( Lancioni et al., 2022).

These methods (GC-O, GC-MS, HS-SPME) are characterized by high accuracy and reliability, as some of the most used methods ( Attchelouwa et al., 2020; Chen et al., 2021). However, these methods usually require sample conditioning, which involves investing many different types of resources ( Shi et al., 2018). Considering the above, devices such as the e-nose have been developed, consisting of an array of electrochemical sensors articulated with a pattern recognition system that identifies, groups, and discriminates the VOCs ( Gliszczyńska-Świgło & Chmielewski, 2017; Loutfi et al., 2015). This has become an alternative to generating fast and reliable results in the food industry ( Barbosa-Pereira et al., 2019; Conti et al., 2021; Wasilewski et al., 2019).

The human tongue can identify five basic tastes: sour, salty, sweet, bitter, and umami ( Beauchamp, 2019). Usually, the evaluation and classification of the basic flavors of a product are done through trained panelists and sometimes consumers ( Jiang et al., 2018). However, these measurements can be subjective, which can be reduced by using technological tools such as the e-tongue, thus ensuring repeatability and reproducibility of the results ( Schlossareck & Ross, 2019). Ross (2021) showed that combining different electrodes makes it possible to identify different flavors, such as fatty, metallic, and others. Different investigations have shown that by using the e-tongue, it is possible to determine the quality, adulteration, classification, or origin of food ( de Morais et al., 2019; Elamine et al., 2019; Jiang et al., 2018; Sobrino-Gregorio et al., 2018). The previously mentioned characteristics have allowed the e-tongue to become a fast, economical and impartial detection alternative ( Titova & Nachev, 2018); this is because it allows the characterization of the flavor of the food matrix ( di Rosa et al., 2017). Additionally, the e-tongue has a matrix of electrodes that, according to their combination and characteristics, produce potentiometric, voltametric, and impedimetric signals ( Jiang et al., 2018).

The texture of a food is perceived through the response to the contact between the body part and the food. It is a determining characteristic in the acceptance of the product by the consumer ( Civille, 2011; Liu et al., 2019; Muthukumarappan & Karunanithy, 2021). Texture is a quality attribute used in the food industry ( Torres Gonzalez et al., 2015), allowing the parameterization and standardization of food products ( Liu et al., 2019). For example, freshness, a determining characteristic in selecting a vegetable or fruit, can be described by its hardness ( Liu & Zhang, 2021). The latter is one of the primary properties of texture, as well as cohesiveness, viscosity, elasticity, and adhesiveness ( Foegeding et al., 2011).

To determine some of the main textural characteristics mentioned, Friedman in 1963 established a method called Texture Profile Testing (TPA) ( Nishinari et al., 2019). This method generates characteristic curves from the force measurement performed by the jaw to realize a change in the geometrical property of the product, generating deformation or fracture ( Kohyama, 2020; Peleg, 2019). The study of these curves allows for establishing and quantifying texture characteristics such as: brittleness, hardness, adhesiveness, cohesiveness, elasticity, gumminess, and chewiness ( Nishinari et al., 2019).

For the measurement of texture characteristics, different methodologies and instruments have been developed, the most widely used technology is centered on texture analyzers or texturometers ( Torres Gonzalez et al., 2015), which are based on the TPA principle, this device simulates the bite of the jaw in two cycles (compression and decompression), through a controlled mechanism that vertically displaces a uniaxial compression cell ( Peleg, 2019). When the cell comes into contact with the product, it generates an electrical signal conditioned by a transducer and sent to a computer to be read by operating software ( Taniwaki & Kohyama, 2012). The displacement is carried out until it reaches either a distance threshold or a force level defined by the operator. When this limit is exceeded, the cell moves back and repeats the cycle ( Liu et al., 2019), simulating the chewing process. Chewing is the first step in the digestion process, and this seeks to prepare food for swallowing. During this process, saliva moistens the chewed food, generating a bolus, which is reduced in size so it can be swallowed. Additionally, saliva helps release flavors and perceive the texture of the food. To achieve this, the intervention of teeth, tongue, saliva, cheeks, and palate is required ( Pereira et al., 2007).

Although TPA is a method that simulates the chewing process, its shear rate is low compared to that of the human bite ( Nishinari & Fang, 2018). Therefore, some researchers have focused on finding other mechanisms that allow an understanding of the bite processes of people in a real environment. One of the alternatives is the study of Electromyographic (EMG) signals, which are produced by the nervous system so that the muscles involved during the chewing process react in a certain way producing electrical signals that can be measured ( Besomi et al., 2020; Pereira de Caxias et al., 2021). These signals are captured with an electromyograph, which integrates an instrumentation amplifier that captures and amplifies the EMG signal with the help of three reference electrodes ( Fang et al., 2020), measuring the activity of the jaw muscles and the coordination between them, as well as the movement of the jaw. This signal is sent through a data acquisition board (DAQ), to a processing system where it is processed and sent to a data acquisition system (DAS) ( Gohel & Mehendale, 2020) to a processing system where it is subjected to extraction methods that perform the analysis of the signal ( Ahsan et al., 2009; Zabala et al., 2019). Sodhi et al. (2019) correlated bite EMG signals with texture variables (instrumental and sensory) of seven Indian sweets, identifying EMG parameters that distinguish the different textured foods. In addition, the PCA determined the significant correlation between hardness (instrumental and sensory) and sensory stickiness. Similarly, Shimada et al. (2012) established intraoral force recordings to analyze the mechanics of human chewing by measuring the force (using strain gauges located on the molars) and the EMG signals (using electrodes located on the masseter muscle) during the biting process of five different products (rice, bread, almonds, banana, and apple). Other relevant studies where the effectiveness of the analysis of EMG signals to determine the texture of a food matrix is sought to be validated are shown in Table 6.

Food products have the characteristic that when consumed they generate sounds that allow identifying or relating some textural properties such as hardness, crispness, and crunchiness to it ( Dias-Faceto et al., 2020). These sensory properties are related to the freshness of the food. When food is ingested, sound waves are generated that can be perceived through the ears by air conduction or through the jaw by bone conduction. Crispy foods, having a more fragile structure, generate high-frequency sounds; while crunchy foods produce low frequency sounds ( Tunick et al., 2013). Some of the equipment to perform these measurements use devices such as microphones connected to computers o texture analyzers integrated with microphones ( Dias-Faceto & Conti-Silva, 2022), and alternative designs with oscillating tips and piezoelectric sensors ( Taniwaki et al., 2006). All these devices allow capturing the acoustic waves produced by the deformation of the product. These sound waves are captured by microphones which are made up of a diaphragm, a grille, and a transducer. The transducer is responsible for converting the movement generated in the membrane (by detecting a sound wave) into electrical signals. Additionally, it has electronic circuits that help manipulate and improve the electrical signal, among these elements are Light Emitting Diodes (LEDs) whose function is to indicate an on/off state, buttons to manipulate the volume amplitude, filters, among others. This electrical signal is sent to a computer where software analyzes and graphs it ( Dias-Faceto et al., 2020). In humans, the sound wave reaches the outer ear, where the sound is collected and transmitted to the middle ear through the ear canal. Between the outer ear and the middle ear is the tympanic membrane or eardrum, which vibrates when it detects sound (behavior imitated by the microphone diaphragm). The middle ear, made up of the tympanic cavity, houses the auditory ossicles (malleus, incus, and stapes), which transform high-amplitude, low-intensity sound waves into low-amplitude, high-intensity vibrations (behavior similar to that of amplifiers). In this way, the ossicles are the intermediaries in the transmission of vibrations from the eardrum to the inner ear. The inner ear detects and transmits auditory impulses to the brain (function carried out by the software housed in the computer) through the vestibulocochlear nerve (function carried out by a transmission medium such as cables) ( Duizer, 2001). It is known that soft tissues have a damping effect (a function performed by attenuating circuits) on the sound produced when chewing food. To match what is heard during the consumption of a product, bone-borne noise must be attenuated at a frequency of 160 Hz, while air-borne sound must be attenuated at 160 Hz and amplified at 3.5 kHz ( Dacremont Colas & Sauvageot, 1991). Due to these differences in sound contribution, the two sounds must be combined and equalized to fully quantify the acoustic sensations perceived during the consumption of crunchy or crispy products ( Vickers & Bourne, 1976).

Researchers such as Błońska et al. (2014), showed that adding inulin with reduced fat content significantly affected the acoustic parameters of Short-Dough Biscuits. Eight Short-Dough Biscuits with different percentages of inulin addition were compared, determining the impact on the acoustic properties and the decrease in the breaking workforce. For example, the biscuit with 74.1% fat and 18.5% inulin, showed a low acoustic energy level of 1.134 a. u. this compared to a biscuit with 55.6% fat and 9.3% inulin, in which a high acoustic energy level of 17.373 a. u. was found, the former being less brittle and hard compared to the latter. This was achieved using a Zwick 1445 measuring system (Zwick GmbH & Co. KG, Ulm, Germany). Separately, Jakubczyk et al. (2017) studied the acoustic signals generated during puncture tests on some coextruded cereal products with different fillings (toffee, milk, fruit jelly, coconut, and chocolate creams), to perform the analysis of hardness, crunchiness, and texture sound attributes for each product. The results showed that the snacks with jelly filling were perceived as less crunchy and soft, compared to the snack with milk cream filling, which showed high acoustic and mechanical values that link it to crunchiness. The variables were measured with a BC45 cooking extruder (Clextral, Firminy, France). Other relevant investigation, such as those mentioned above, where acoustic analysis was performed to determine some textural properties of certain foods, can be seen in Table 7.

Other considerations to take into account when acquiring a technological tool for the study of food matrices are: •

Costs: The acquisition of technological tools, such as those presented in this article, can be expensive. For example, the cost of e-nose ranges from USD 100 to USD 1000; the texturometer at an approximate value of USD 23900; a colorimeter at an approximate value of USD 12000 and an electromyograph (for medical use) at an approximate value of USD 10560. Regarding accessories and/or additional components, these can vary between USD 10 and USD 2300. This aspect must be considered when purchasing any of these technological tools to keep them operational.

Finally, it must be considered that the study of various food matrices is characterized by the variability of the samples, this being biological material. Thus, some e-tongues and e-noses are designed to identify specific compounds with high precision, thus limiting the analysis of food matrices for which they were not designed. Regarding the use of the colorimeter, lighting conditions and texture of the food matrix can affect the results. All of the above shows the economic and technical limitations that these technological tools may have.

As evidenced in this review, some technological tools have been developed to emulate the functioning of the five senses (smell, taste, sight, touch, and hearing), seeking to quantify and characterize some sensory properties of different food matrices, to compare, parameterize and standardize a product. These investigations show that the use of technological tools guarantees the repeatability and reproducibility of the process, compared to the results obtained when working with trained panelists. Therefore, the use of this type of device reduces the number of samples required to perform the analysis, in addition to dispensing with the need for a team of trained panelists, which generates a reduction in costs. In addition, another advantage of these tools is the wider measurement capacity compared to that of human beings. However, most of the tools analyzed only have the property of measuring a single characteristic in a food matrix, this becomes an inconvenience when it comes to characterizing an entire product, for which many tools must be available, samples required and therefore an increase in the time of the analysis and availability of personnel to carry out the process. This is why both the scientific community and the industry, increasing the development of research that seeks to create new technological tools that allow the measurement of two or more sensory characteristics in a food matrix. All the above, seeking to develop new food products and improve existing ones to satisfy the sensory experiences of the consumer, driving growth in the food sector.

No data are associated with this article.