Introduction

As global life expectancy continues to rise, most people are now expected to live into their sixties and beyond. By 2030, one in six individuals will be 60 or older, with the number of older adults increasing from 1 billion in 2020 to 1.4 billion. By 2050, this number is projected to more than double to 2.1 billion, with the population aged 80 and above tripling to 426 million. While the aging trend initially started in high-income countries, now low- and middle-income nations are experiencing the most rapid growth, with two-thirds of the world’s elderly population expected to reside there by 2050 [1]. This demographic shift, fueled by longer life expectancies and decreasing birth rates, is transforming societal structures and creating new challenges for healthcare systems worldwide. These challenges are particularly evident among older populations, who are increasingly living longer with chronic conditions and reduced physical, cognitive, and social functioning. One key area impacted by this shift is oral health, particularly masticatory function, which plays a vital role in an individual’s overall health and quality of life.

The importance of evaluating masticatory performance extends beyond simple food intake. Dysfunctional mastication can contribute to a range of health issues, including malnutrition, frailty, and cognitive decline, all of which are prevalent among the elderly. Researchers have increasingly recognized the intertwined relationships of oral health, nutritional status, and cognitive function, with studies indicating that poor chewing ability can accelerate the onset of dementia and other age-related disorders [2]. For example, diminished chewing ability may reduce the intake of essential nutrients, leading to deficiencies that contribute to cognitive decline, while poor nutritional status itself has been linked to the increased risk for dementia and decreased ability to perform activities of daily living [3].

Despite advancements in oral care and rehabilitation, age-related factors such as tooth loss, reduced salivary flow, and diminished muscular strength continue to impair chewing ability in older adults. Masticatory performance-the ability to chew food effectively-is a vital indicator of oral health, as it significantly impacts nutritional intake, physical well-being, and social engagement. When masticatory function is inadequate, older adults face difficulties in food consumption, which can lead to malnutrition, weight loss, and other adverse physical health outcomes [4, 5]. Moreover, impaired chewing ability can worsen conditions like frailty, cognitive decline, and even depression, creating a vicious cycle of declining health and well-being [6].

The loss of masticatory function is a prevalent issue among the elderly. Tooth loss remains common, particularly among frail and dependent older adults, despite considerable advancements in dental care. Although the prevalence of edentulism (the condition of being toothless) has decreased in some regions over the past few decades, thanks to improvements in preventive care, it continues to be a major concern in aging populations [7,8,9]. Additionally, individuals with partial dentures or other dental prostheses often experience reduced masticatory efficiency. Consequently, many older adults face challenges in chewing, which impacts their overall health and well-being. Beyond tooth loss, the oral health of the elderly is frequently compromised by dental caries, periodontal disease, dry mouth, and oral cancers-factors that contribute further to the decline in masticatory function [10].

Masticatory performance in the elderly can be objectively assessed using a variety of clinical methods. Masticatory function evaluation has predominantly relied on conventional methodologies since the mid- 20th century. Seminal studies by Prinz [11] and Slagter et al. [12] established quantification protocols through comminution tests (particle size distribution analysis) and mixing ability assessments (e.g., chromatic dispersion scoring of bichromatic chewing gum), which became standardized in controlled laboratory settings, establishing the quantitative framework for objective masticatory assessment. One of the most commonly used methods for evaluating masticatory performance involves sieving food or artificial food particles after a predetermined number of chewing cycles [13]. The particle size distribution is then analyzed, with smaller particle sizes indicating better chewing performance. Additionally, other techniques, such as color-changing chewing gums and optical scanning, have been utilized to assess the efficiency of food breakdown, offering further insights into masticatory function [14].

Despite these advancements, there is no single, universally accepted method for assessing masticatory performance in clinical settings. While sieving and mixing ability tests are commonly used in research, they tend to be time-consuming and impractical for routine clinical use. Moreover, although these methods are valuable for identifying general trends in masticatory efficiency, they may not fully capture the complexities of individual chewing performance. This is particularly true given that chewing ability can vary significantly based on factors such as the type of dental prosthesis, occlusal status, and individual motor skills [15].

And for that, subjective self-assessments of chewing ability continue to be a commonly used tool for evaluating masticatory performance. Questionnaires and interviews enable patients to report their perceived chewing difficulties, offering valuable insights into how masticatory dysfunction affects their daily lives [16]. However, studies have demonstrated mixed correlations between objective and subjective assessments, underscoring the need for more reliable and consistent methods to evaluate masticatory function in both clinical and research settings [16, 17].

Elgestad Stjernfeldt et al. (2019) [18] conducted a systematic review of various methods used to objectively assess masticatory performance and evaluate their measurement properties, including validity, reliability, and responsiveness. The review incorporated 46 studies, with a primary focus on two main assessment methods: comminution and mixing ability. Despite the usefulness of these traditional methods, the review underscores the absence of a single, universally accepted method for assessing chewing ability in clinical practice. It highlights the fact that while these established methods have been effective, they often rely on laboratory-based and labor-intensive equipment, which limits their routine use in clinical settings. As the elderly population grows, the review calls for further research and innovation to develop more accessible, reliable, and clinically feasible methods for evaluating masticatory performance, aiming to better address the oral health challenges faced by aging individuals.

In the early 21st century, with the emergence of digital technologies, researchers began developing more efficient and automated assessment methods. Schimmel et al. (2007) [14] introduced digital image processing for analyzing bichromatic chewing gum, significantly improving the quantification accuracy of mixing ability. However, these approaches still required complex equipment and failed to fully resolve practical challenges in clinical settings.

Recent advancements in artificial intelligence (AI) and wearable technologies have revolutionized masticatory assessment. Since the late 2010 s, computer vision systems [19] and wearable sensors [20] have enabled real-time monitoring of mandibular kinematics and muscular activation. The integration of machine learning algorithms [21] has further automated chewing behavior recognition and analysis. This paradigm shift from manual operation to intelligent systems not only enhances evaluation efficiency but also facilitates personalized healthcare and remote monitoring capabilities.

Fig. 1
Fig. 1
Full size image

Chewing assessment development process

Figure 1 illustrates the evolution of chewing efficiency evaluation methods from 1964 to recent years, transitioning from initially manual and labor-intensive procedures to contemporary intelligent and automated approaches enabled by the continuous advancement of digital technologies and artificial intelligence. However, since Elgestad Stjernfeldt et al. (2019) [18], no scholarly work has systematically summarized the progression from foundational chewing assessments (e.g., Comminution and Mixing Ability) to current AI-integrated methodologies.

This paper presents the first comprehensive review of chewing assessment techniques, tracing their evolution from conventional methodologies to modern AI-driven approaches. It examines the strengths and limitations of different assessment tools, explores their applications in clinical and research settings, and discusses emerging trends in robotic simulation and wearable technology. By identifying gaps in current methodologies, this review aims to highlight future directions for improving masticatory assessment and its role in promoting better oral and overall health, and will subsequently delve into traditional and AI-based methodologies, and Fig. 2 presents the complete set of food specimens employed in the subsequent mastication tests.

Fig. 2
Fig. 2
Full size image

The foods analyzed in the study

Traditional methods

As shown in Fig. 3, conventional chewing assessment techniques encompass subjective and objective methodologies. Subjective approaches involve unaided visual inspection and the Chewing Function Questionnaire (CFQ), while objective methodologies are primarily categorized into Comminution and Mixing Ability analyses. These protocols utilize diverse test foods (e.g., carrot, capsules, gum) for functional evaluation. The following sections provide systematic documentation of these traditional assessment frameworks.

Fig. 3
Fig. 3
Full size image

Taxonomy of Traditional Methods

Subjective

In 1999, Prinz JF [11] introduced a subjective method for assessing chewing efficiency using a custom-designed Bubble Yum™ lime flavour gum (7.5 g, dimensions 25 mm \(\times\) 20 mm \(\times\) 12 mm). The gum consisted of a dual-layer structure: a central rectangular core of yellow lemon-flavoured gum (16 mm \(\times\) 8 mm) enclosed within a green lime-flavoured sheath. Both layers shared similar textural properties. The method involved visual evaluation of the colour mixing homogeneity between the two layers after a predetermined number of chewing cycles, with higher chewing efficiency indicated by more uniform blending of the yellow and green components.

To assess how well observers could distinguish different degrees of mixing, six judges ranked two sets of 10 coded transparent bags containing gum chewed between five and 15 strokes. One set was flattened, while the other kept the gum in its bolus state. An error score was calculated by comparing the judges’ ranks with the actual rank, which was based on the number of chews.

The Chewing Function Questionnaire (CFQ) developed by Peršić et al. [22] in 2013 is another subjective method for assessing chewing ability. The CFQ focuses on assessing an individual’s self-perceived difficulties when chewing various types of food, including items like apples, meat, and gum. The questionnaire includes 10 items, with responses measured using a Likert scale. It was validated with 224 participants and showed a strong correlation with the efficacy of prosthodontic treatments, such as getting dentures fitted, proving to be a reliable and valid tool for assessing chewing function.

Subjective approaches to masticatory evaluation, as outlined in this section, primarily rely on self-reported perceptions and visual judgments to assess chewing efficiency. The two-color chewing gum test, pioneered by Prinz [11] (1999), quantifies masticatory performance through observer-based rankings of color homogeneity after mastication, with error scores calculated against actual chewing cycles. This method emphasizes simplicity and cost-effectiveness but introduces variability due to human interpretation. Complementarily, the Chewing Function Questionnaire (CFQ) by Persic et al. [22] (2013) employs Likert-scale responses to capture individuals’ perceived difficulties in chewing diverse food textures, validated across 224 participants. While subjective methods offer valuable insights into patient-reported functional limitations and align with clinical practicality, their reliability is constrained by inherent biases in self-assessment and inter-observer discrepancies. Studies highlight inconsistent correlations between subjective and objective metrics, underscoring the need for hybrid frameworks that integrate patient perspectives with standardized clinical evaluations. These methods remain indispensable for understanding the psychosocial and functional impacts of masticatory dysfunction, particularly in populations where objective measurements are logistically challenging.

Objective

As shown in Tables 1 and 2, They summarize two traditional methods - Comminution and Mixing Ability with different foods, including capsules, gum, and other special foods to test the chewing ability of the relevant information. We will now analyze masticatory efficiency across standardized food substrates.

Table 1 Comminution methods
Full size table
Table 2 Mixing ability methods
Full size table

Capsule

As shown in Fig. 2b, capsule-based methods for assessing chewing ability are commonly used due to their simplicity, effectiveness, and reliability. Typically, chewy capsules filled with colored granules are chewed, and changes such as color alterations and particle size are measured to evaluate masticatory performance.

In 2006, Escudeiro Santos et al. [29] introduced a colorimetric system to assess masticatory efficiency. Ten volunteers with complete dentition chewed synthetic capsules filled with fuchsine-containing granules for 20 seconds across three evaluation periods, as shown in Fig. 2d. After mastication, the capsule contents were dissolved in 5 mL of water under constant stirring for 30 seconds, filtered to remove residual wrappings and non-crushed granules, and analyzed using a Beckman DU- 7 UV-Visible Spectrophotometer (Beckman Inc., USA) at 546 nm. The released fuchsin concentration (0–12.5 \({\upmu }\)g/mL) was quantified via a pre-calibrated standard curve derived from the absorbance spectrum. This method proved to be quick, straightforward, reproducible, cost-effective, and efficient, making it a useful tool for a variety of applications.

In a study by Cláudia Maria de Felício et al. [28], fuchsin-containing polyvinyl acetate capsules were used to evaluate masticatory efficiency and correlate it with electromyographic (EMG) activities of the anterior temporal and masseter muscles. Nineteen young adults with full dentition and no temporomandibular disorder history participated, chewing fuchsin-filled beads for 20 seconds under varying conditions while synchronized EMG recordings were obtained. Post-mastication, the released fuchsin was quantified using a Beckman DU- 7 UV-Visible Spectrophotometer (Beckman Inc., Palo Alto, CA, USA), establishing a direct metric for chewing efficiency. This objective measurement was then cross-analyzed with EMG signal amplitudes to explore neuromuscular coordination patterns during mastication.

Further research by Augusta Medeiros Ribeiro et al. [27] examined the influence of complete denture (CD) quality and years of use on masticatory efficiency. Special gel beads, dyed with a colored dye, were placed inside sealed capsules, which participants chewed for 20 seconds. Afterward, the beads dissolved in water, and the dye concentration was analyzed to assess masticatory performance.

Rogério Alexandre Modesto et al. [26] later used fuchsin-filled polyvinyl acetate capsules to evaluate both masticatory efficiency and maximum bite force. Their study involved 55 adults with normal occlusion, who chewed four fuchsin capsules (two on each side of the molar region) for 15 cycles, with a 3-minute break between capsules. The fuchsin concentration was measured spectrophotometrically, and bite force was recorded to explore the relationship between these two factors.

Other studies, such as those by Homsi George et al. [25], have integrated capsule-based methods with food comminution and mixing ability tests. These tests assess not only the mechanical breakdown of food (comminution) but also the ability to mix substances in the mouth (mixing ability). For example, the food comminution test records the number of chewing cycles and duration, with the number and size of particles measured to evaluate chewing efficiency. The mixing ability test, on the other hand, involves participants blending two differently colored chewing gums, with the variance of hue (VOH) analyzed after a specified number of cycles.

In conclusion, capsule-based methods provide a reliable, efficient, and cost-effective approach to assessing masticatory performance. By utilizing materials like fuchsin granules or gel beads, researchers can objectively quantify chewing efficiency through colorimetric analysis and particle size measurements. These methods offer valuable insights into oral health, bite force, and denture quality, enabling comparisons across individuals and studies.

Gum

Using chewing gum to evaluate masticatory ability is a simple, effective method that provides quantifiable data on chewing efficiency. This approach typically involves participants chewing a specific type of gum under controlled conditions, with researchers analyzing changes in the gum’s color to assess masticatory performance.

In 2007, Ishikawa et al. [55] developed a novel color-changing chewing gum to evaluate chewing efficiency in complete denture wearers. The chewing performance was assessed by having participants chew the gum for 100 cycles, with color changes (quantified as the a* value in the CIE-Lab color system, representing red-green chromaticity) measured through polyethylene films using a colorimeter (CR- 13; Konica-Minolta, Tokyo, Japan). Comparative analysis of masticatory performance between new and old dentures revealed significant correlations between the gum’s color changes and other evaluation metrics, demonstrating its clinical validity for functional assessment.

In the same year, Schimmel et al. [14] employed a two-color chewing gum test using custom-prepared Hubba-Bubba Tape Gums® (30 \(\times\) 18 \(\times\) 3 mm strips, combining azure Sour Berry and pink Fancy Fruit flavours) to assess chewing efficiency. The method quantified performance by comparing pre- and post-mastication color mixing homogeneity through digital image processing (Epson Scanner® images analyzed via Adobe Photoshop Elements 2.0 on a Windows XP® PC with Intel Pentium 3/2 GHz/256MB). Volunteers chewed the dual-layered gum for varying cycles, with unmixed color pixel proportions calculated to evaluate efficiency. To validate clinical reliability, the digital analysis was cross-referenced with conventional visual assessments. Additionally, maximal voluntary bite force was measured using an Occlusal Force-Meter GM 10®, exploring potential correlations between masticatory force and color-blending metrics.

Then Kamiyama et al. [54] developed a novel color-changing chewing gum to assess chewing efficiency, which undergoes color changes in response to pH variations during mastication. The researchers established a dedicated color scale and validated its effectiveness by comparing results from the scale with a* values (from the CIE-Lab color system) measured via colorimetry. They further evaluated the reliability of the assessment method using intraclass correlation coefficients (ICCs).

Next, Halazonetis et al. [52] employed a novel software named ViewGum to evaluate chewing efficiency. The software quantifies color mixing performance by calculating the standard deviation of hue (SDHue) from images obtained during two-color chewing gum mixing ability tests, using Hubba-Bubba Tape chewing gum. By analyzing color variations across different chewing cycles, ViewGum streamlined the image analysis process for clinical implementation while enabling efficient extraction of actionable clinical insights.

Concurrently within the year, Weijenberg et al. [53] carried out a study to assess the chewing function test based on the mixing ability of two-color gum. They employed multiple digital analysis methods to evaluate the applicability of this test for assessing masticatory function in the elderly. The findings indicated that their approach was sensitive and reliable in detecting changes in the mastication cycle. However, the correlation between this method and the traditional wax block test was not as satisfactory as anticipated. Even though its effectiveness requires further enhancement, it remains a viable option for evaluating masticatory function. This study thus offers a novel perspective and methodology for related research and clinical practice.

In 2014, Hama Yohei et al. [50] introduced a specially developed chewing gum designed to analyze color changes during chewing. The gum’s ingredients included xylitol, citric acid, and red, yellow, and blue dyes. After being chewed, the gum color changed to indicate the chewing efficiency. The red dye is pH-sensitive and changes color in neutral or alkaline conditions. As the chewing progresses, yellow and blue dyes seep into the saliva, and red appears due to citric acid elution. Color measurements were taken using a colorimeter immediately after chewing, with the gum flattened to a thickness of 1.5 mm between polyethylene films.

Subsequent research by the same team [51] used the CIELAB color system to assess changes in gum color, measuring the L*, a*, and b* values. These studies demonstrated that changes in the gum’s color reliably reflected masticatory performance, making it an effective method for evaluating individuals with both natural dentition and complete dentures. The results showed significant color changes across a broad range, providing an efficient measure of chewing ability.

Concurrently, Endo et al. [49] implemented a two-color chewing gum test using standardized Lotte® chewing gum (18.8 \(\times\) 14.2 \(\times\) 3.9 mm, blue/red dye-insoluble strips; Lotte Co., Tokyo, Japan) to evaluate chewing efficiency. Quantitative analysis was performed via Epson ES- 2200® scanner (Seiko Epson, Japan) and Adobe Photoshop CS3® (Adobe Inc., USA) to calculate unmixed color pixel ratios (OCMR-W), following Schimmel et al.’s methodology[14]. Two blinded examiners conducted pixel-level analyses, enabling sex-specific differentiation in masticatory performance through statistically validated color blending metrics.

In 2015, Schimmel et al. [48] employed three customized two-colored chewing gums to evaluate masticatory efficiency: Gum1 (discontinued Hubba-Bubba Tape Gum®, Wrigley Company Ltd., UK; 30 \(\times\) 18 \(\times\) 3 mm azure/pink strips), Gum2 (experimental specimens developed by Lotte™ Co., Japan for the 8020 Promotion Foundation; 18.8 \(\times\) 14.2 \(\times\) 3.9 mm, Shore OO durometer-verified hardness), and Gum3 (commercially available Vivident Fruitswing® by Perfetti van Melle, Turkey; 43 \(\times\) 12 \(\times\) 3 mm green/violet layers). Following standardized chewing cycles, dual assessments were conducted: 1) subjective visual evaluation and 2) objective colorimetric analysis using an Epson Perfection V750 Pro® flatbed scanner (Seiko Epson Corp., Japan; 300 dpi resolution) coupled with ViewGum© software (dHAL Software, Greece; www.dhal.com). The software quantified color-mixing efficiency through hue variance (VOH) calculations in HSI color space after semi-automatic image segmentation. This hybrid methodology demonstrated significant correlations between subjective scores and objective VOH metrics, thereby elucidating condition-dependent variations in masticatory performance across different test materials.

Later researchers have extensively replicated these assessments. This methodological progression manifested through Vaccaro et al.’s MATLAB-based hue blending quantification in 2016 [47], which Wada et al. (2017) [46] expanded by incorporating chromatic aberration mapping with spectrophotometers, culminating in Silva et al.’s (2018) work [45] that standardized colorimetric evaluation framework integrating calibrated reference grids with machine learning-enhanced imaging analytics.

In another study, Đurđa Nedeljković et al. [16] combined the two-color chewing gum test with the ViewGum software, which analyzes color homogeneity and generates a Z-score. This Z-score provides a standardized measure of color mixing, where lower scores indicate better chewing performance and more efficient mixing.

In conclusion, chewing gum offers a non-invasive, easily accessible, and effective method for assessing masticatory ability. Its simplicity, reliability, and ability to generate clear, quantifiable data make it an excellent choice for both clinical and research purposes in evaluating chewing efficiency. With the continued development of standardized assessment protocols, the method’s applicability across various populations is likely to increase.

Specific food

Using specific foods to assess masticatory performance is a widely applied and effective method. This approach typically involves participants chewing a test food, after which the chewed particles are collected, cleaned, dried, sieved, and analyzed. This process allows researchers to measure the degree of food fragmentation, which provides valuable insights into chewing ability.

In 1964, Kapur et al. [24] focused on evaluating diverse food types to assess masticatory efficiency in complete denture wearers. The study systematically examined 33 distinct edible substrates (e.g., peanuts, carrots, ham) to identify optimal test foods for quantifying chewing performance. Selected materials were required to meet four criteria: palatability, structural homogeneity, cost-effectiveness, and provision of graded masticatory challenge levels, enabling precise measurement of performance variations across denture wearer cohorts.

Building upon foundational methodologies, in 1985, Gunne et al. [31] established dual masticatory quantification systems comprising photometric analysis of post-mastication hardened gelatin fragment surface area and sieve-classified almond particulates via masticatory efficiency index (Ci).

Subsequently, in 1994, Mowlana et al. [40] developed an innovative bimodal assessment protocol combining optical scanning and sieving techniques with almond test substrates. Utilizing Seescan Ltd’s optical scanning device (Cambridge, UK), they quantitatively compared post-occlusal particle size distributions between digital image analysis and mechanical sieving methodologies, demonstrating equivalent validity in masticatory efficiency quantification.

In 1999, Al-Ali et al. [37] simplified the chewing assessment process by using microwaved almonds, as shown in Fig. 2h. The almonds were treated to reduce their oil content and prevent clumping with saliva. Participants chewed these almonds in rubber bags to avoid particle loss and saliva interaction. Afterward, the chewed particles were sieved, weighed, and analyzed using optical scanning to determine the number and area of particles. This method streamlined the process by eliminating the need for washing and drying, while still delivering reliable masticatory performance results.

In 1993, Slagter et al. [12] conducted a comparative masticatory analysis of two artificial test foods (Optocal/Optosil) in complete denture wearers versus natural dentition subjects. Participants were instructed to masticate both substrates, with post-occlusal particle size distribution serving as the masticatory efficiency metric. The study characterized Optocal as a low fracture resistance experimental material and Optosil as a standardized silicone-based elastomer, as shown in Fig. 2k, both established in functional oral assessments. Results demonstrated Optocal’s superior suitability for edentulous populations due to its enhanced detection sensitivity in prosthodontic performance evaluation.

Similarly, Lujan-Climent et al. [44] implemented a standardized masticatory protocol using Optosil P Plus (Heraeus Kulzer, Hanau, Germany) as the test substrate. The participants performed 20 chewing cycles, and the subsequent analysis involved particulate sieving (Retsch test sieves) to quantify the masticatory performance. Concurrent measurements of maximum bite force (GM10 occlusal gauge) and static/dynamic occlusal patterns (T-Scan III system) were conducted to establish their biomechanical correlations with chewing efficacy.

Further research by Van der Bilt et al. [56] combined the use of artificial test foods, Optosil and Optocal, with a color-mixing gum test to evaluate chewing performance. Optocal, which was modified to increase brittleness, allowed for more uniform food fragmentation. Participants chewed 17 cubes of each food type before spitting out the particles, which were then processed. The median particle size (X50) was calculated as a key measure of masticatory efficiency. The color-mixing gum tests added another layer of insight by evaluating the degree of color mixing after a set number of chewing cycles.

In the same way, Khoury-Ribas et al. [41] engineered composite test substrates (Optozeta: 50% Optosil/50% Zetalabor) to holistically evaluate masticatory performance, preference, and rate, collectively advancing methodological rigor in oral functional research.

In 1992, Mahmood et al. [23] implemented a novel masticatory evaluation protocol using cored carrot sections as test substrate, as shown in Fig. 2a, with post-mastication particle dimensions quantified through standardized image analysis (ICC= 0.93). Comparative analysis revealed dentate subjects exhibited 38.7% greater masticatory efficiency than both complete denture wearers and immediate prosthesis recipients (\(p<0.01\)). This imaging-based methodology demonstrated superior clinical practicality by eliminating traditional sieving procedures’ procedural discomfort and contamination risks.

Additionally, Ali Alkhalaf et al. [39] used common foods like carrots, apples, and peanuts to estimate chewing ability. By recording and analyzing the masticatory process, researchers measured parameters such as the time to the first swallow, total number of chewing actions, and the number of swallows, which provided valuable insights into overall oral function.

In 2008, Fauzza & Lyons [36] demonstrated the clinical viability of irreversible hydrocolloid impression material (alginate) for assessing masticatory performance in complete denture wearers. The Irreversible Hydrocolloid is shown in Fig. 2g.

Additionally, Eberhard et al. [43] employed dual methodologies for masticatory assessment: conventional sieving and scanning-based image analysis. The study utilized Optocal, a silicone-based artificial test food specifically designed for complete denture wearers. The sieving procedure was carried out using ten sieves (infraTest Prüftechnik GmbH, Brackenheim-Botenheim, Germany) with decreasing apertures, and the scanning method used a flatbed scanner (Epson Expression 1600Pro, Seiko Epson Corporation, Japan, 1,200 dpi). Scanned images underwent image analysis using ImageJ 1.42q (NIH, USA). Comparative analysis of particle size distributions between the two methods validated the efficacy and feasibility of the scanning technique.

In 2016, Sanchez-Ayala et al. [30] validated basic fuchsin-impregnated bead spectrophotometry (UV-Vis) against gold-standard silicone cubes, as shown in Fig. 2c, with multi-sieving methodology,

In 2006, Kobayashi et al. [34] developed a dual-substrate masticatory assessment protocol using 2 g glucose-impregnated gummy jelly and 3 g peanut specimens. Twenty healthy participants conducted standardized masticatory trials across 10/20/30 chewing cycles, with subsequent quantification of post-mastication glucose extraction levels. Comparative analysis against sieving method-derived masticatory indices confirmed glucose release as a valid salivary biomarker for functional chewing assessment.

In a similar approach, Shiga et al. [35] utilized standardized gummy-jelly specimens to evaluate masticatory efficiency. Healthy participants were directed to perform 20-second mastication periods, after which the released glucose was quantitatively analyzed using the Roche Accu-check® comfort portable glucometer and the Hitachi Model 200 - 20 spectrophotometer.

Similarly, Murakami Kazuhiro et al. [33] assessed elderly denture wearers by having them chew gummy jelly (UHA Mikakuto). Participants chewed a 5.5-gram full-size jelly for 30 cycles, with individuals scoring poorly on conventional tests given a smaller half-size jelly. This approach allowed for a tailored and sensitive evaluation of masticatory function, particularly for those with reduced chewing ability.

More recent research by Yokoyama et al. [32] focused on implant-supported denture wearers. In this study, participants chewed gummy jelly containing 5% glucose (GLUCOLUMN, GC, Tokyo, Japan) while their masticatory movements were recorded using a Motion Visi-trainer (MVT V1, GC, Tokyo, Japan). Key parameters such as movement path, opening distance, masticatory width, and cycle time were measured. Masticatory efficiency was further evaluated by quantifying the glucose eluted from the jelly during chewing, offering a comprehensive assessment of chewing performance and denture functionality.

In 2003, Sato et al. [60] developed an evaluation method using two-colored wax as test material. The images of each side of the chewed test food were captured by a CCD camera (xc- 003, Sony Co., Tokyo, Japan). Participants chewed the wax specimens, with masticatory performance quantified through the Mixing Ability Index (MAI) - a composite metric combining color blending degree and post-chewing morphology via discriminant function analysis.

Then Sugiura et al. [59] adopted two-colored wax (red/green) as a masticatory evaluation substrate in 2009. Participants performed 10 chewing cycles, with subsequent color blending and morphological changes quantified through digital image analysis. These parameters generated a Mixing Ability Index (MAI) for functional mastication assessment, demonstrating significant correlation with brittle food fragmentation capacity (e.g., peanuts) but no significant association with glucose dissolution in elastic foods (e.g., gummy-jelly).

In the same year, Speksnijder et al. [58] conducted mastication trials using two-colored wax (red/blue) from Stockmar, a non-toxic Plasticine modeling wax adhering to DIN EN- 71 standards. Post-chewing specimens were flattened and bilaterally scanned using an Epson V750 scanner, with masticatory performance evaluated through chromatic blending analysis. The textural properties of the wax were assessed using a TA-XT Plus texture analyzer from Stable Micro Systems. This wax-based assessment demonstrated enhanced differentiation capability among populations with varying masticatory dysfunction severities.

Subsequently, van der Bilt et al. [57] conducted chromatic mastication assessments with 60 healthy participants using red/blue two-colored wax. Subjects performed 5/10/15/20 chewing cycles, followed by dual-modality evaluation: computerized analysis employing digital image processing algorithms to calculate wax mixing indices, and visual assessment by five trained evaluators using reference scales. The computerized method effectively differentiated masticatory performance across dental status groups, while visual evaluation required 20 cycles to distinguish complete denture wearers from mandibular implant-supported prosthesis users.

For young children, Tournier Carole et al. [38] used a model gel, as shown in Fig. 2i, to assess early masticatory skills. Their study found that masticatory performance improved with age, making the method suitable for evaluating the development of chewing skills in infants who were willing to accept the feeder in their mouth.

Finally, Lorenz Marlene et al. [42] conducted standardized food testing on children with Orofacial Myofunctional Disorders (OMD), using a salted cracker. Researchers measured total bites, chewing cycles, and swallows, offering detailed insights into masticatory and swallowing functions, with video recordings facilitating accurate post-analysis.

In conclusion, the use of specific foods for assessing masticatory performance offers a reliable and practical method for evaluating chewing ability. When combined with color-mixing and food fragmentation tests, this approach provides a comprehensive evaluation of both the mechanical breakdown of food and functional efficiency in the masticatory process.

Artificial intelligence methods

AI-driven masticatory assessment significantly cuts costs versus traditional methods. Conventional approaches such as spectrophotometric analysis of capsules by Escudeiro Santos et al. in 2006 [29] or sieving by Schimmel et al. in 2007 [14] require consumables and labor-intensive workflows. In contrast, AI solutions like wearable sensors described by Hori et al. in 2021 [20], automated video analysis by Kumar et al. in 2024 [21], and robotic simulators by Akarawita et al. in 2024 [61] reduce material waste and human intervention. For instance, AI-based particle analysis by Schmidt et al. in 2023 [62] lowered costs by 40% compared to manual methods. These innovations enhance cost-efficiency through automation, scalability, and minimized consumable dependency.

The digital revolution, while driving the aforementioned cost-efficiency of AI methodologies, has also profoundly enhanced technological capabilities in medical and dental clinical practices [60,61,62]. Historically, AI applications in healthcare-such as medical object detection and disease diagnosis [63, 64]-laid the foundation for its adoption in masticatory science. Recent advancements now extend these principles to revolutionize chewing evaluation through objective, non-invasive, and highly precise tools, addressing both economic and functional limitations of traditional methods. This section explores three AI-driven approaches that epitomize this transformation: computer vision systems, wearable devices, and robotic simulators.

Computer vision systems leverage video analysis, endoscopic imaging, and motion tracking to quantify chewing patterns, bolus formation, and mandibular dynamics. These systems provide granular insights into masticatory efficiency while minimizing physical intrusiveness. Wearable devices, such as inertial sensors and electromyography (EMG) units, enable real-time monitoring of jaw movements and muscle activity in naturalistic settings, enhancing ecological validity. Robotic platforms, designed to replicate human mastication mechanics, standardize food breakdown analysis, eliminating inter-individual variability for controlled experimental studies. The following subsections detail their methodologies and applications.

Table 3 Computer vision system
Full size table
Table 4 Wearable devices and robots
Full size table

As shown in Tables 3 and 4, these tables summarize three modern methods - Computer Vision System, Wearable Devices and Robots with different foods, including capsules, bread, and other special foods to test the chewing ability of the relevant information.Subsequently, we will conduct a comprehensive examination of these three methodologies.

Computer vision system

As shown in Table 3, the table provides a summary of chewing ability based on the visual system test.

Evaluating mastication using a vision system involves leveraging computer vision and image processing technologies to analyze the intricate movements and mechanics of the mouth during chewing. This approach enables researchers and healthcare professionals to gain valuable insights into oral function, diagnose potential issues, and implement targeted interventions to improve both oral health and overall well-being.

Abe Risako et al. [19] conducted a study with ten healthy adults without swallowing difficulties to explore the use of video endoscopy for evaluating chewing functions. Participants were given two types of test foods: two-colored molded rice and a rice cake called ’uirou’, as shown in Fig. 2q & r. They were asked to chew either normally or thoroughly, with some participants having their chewing restricted, while others chewed at their own discretion. A video endoscope was used to observe the formation of the food bolus in the oropharynx, and image analysis software was employed to measure the degree of mixing between the green and white parts of the food. This was used to calculate the bolus formation index (BFI). The results showed a significant increase in BFI with more chewing, demonstrating a strong correlation between the number of chews and the BFI, with correlation coefficients of 0.84 for rice and 0.89 for uirou. Additionally, participants instructed to ’chew well’ had higher BFIs compared to those told to ’chew normally’. This suggests that the conscious effort can improve chewing quality, leading to better bolus formation, which is crucial for digestion.

The study effectively demonstrated that video endoscopy is a valuable tool for quantitatively assessing masticatory function during chewing and swallowing. It also emphasized the importance of proper chewing techniques for optimal oral health and digestion. The typical image from the video endoscopic examination, taken with the endoscope inserted nasally and positioned to capture a clear view of the oropharynx, illustrates the precise and controlled method used to observe the chewing and swallowing process.

While the use of video endoscopes to observe chewing behavior is an intuitive and innovative approach, it is important to note that the procedure requires the insertion of the endoscope through the nasal cavity, which may cause some discomfort for the participant.

Hitos Silvia Fernandes et al. [68] developed a standardized method for assessing mastication in children and teenagers with maxillary atresia, particularly those who primarily breathe through their mouths. Mastication was recognized as a key factor in food preparation and its impact on dentofacial development. The study noted that typical chewing involves incisor cutting, labial occlusion, alternating chewing sides, and coordinated bilateral muscle activity. However, these processes can be disrupted by various oral and facial conditions.

To facilitate the analysis, participants were seated with their feet on the floor and without head support, with a white tag placed on their chin to improve the visibility of mandibular movement during chewing. The test food used was freshly baked French bread sourced from the same supplier, ensuring sufficient quantity for 50 seconds of continuous chewing. Subjects were instructed to chew naturally while minimizing unnecessary movement and maintaining eye contact with the camera lens.

Two expert evaluators, along with the lead author, analyzed the video recordings, focusing on chewing cycles from a frontal perspective. The use of the chin tag, combined with slow-motion video analysis, improved inter-rater reliability. The most common chewing patterns observed were alternating and bilateral, found in 64.7% of the subjects.

The study proposed video recording as a standardized, clinical tool for mastication analysis due to its accessibility, repeatability, and low cost. This method is suitable for daily clinical practice and provides a reliable approach for assessing mastication across different populations, including those with altered breathing patterns, occlusion issues, or temporomandibular joint disorders. The research paves the way for future studies exploring the application of this method across diverse patient groups.

Rovira-Lastra et al. [70] used video recordings to assess mandibular displacement during the closing phase of free-style mastication assays, with the goal of evaluating masticatory performance. The study involved 42 young adults, 23 women and 19 men-aged 21 to 45 years (average age 26.8 years), all with natural dentition. Participants were recruited from volunteer students and staff at the University of Barcelona Faculty of Dentistry. Individuals were excluded if they had fewer than 24 natural teeth, were undergoing active orthodontic treatment, or had orofacial pain.

Each participant completed four types of masticatory assays, each consisting of five trials of 20 chewing cycles using 2 grams of silicon. The tests used two different types of chewing food: unbagged silicon and bagged silicon, the latter sealed in a latex bag using cyanoacrylate adhesive. One of the assays focused on free-style mastication with unbagged silicon, while another used bagged silicon. The other two assays examined unilateral mastication, with participants instructed to use only the right side for one assay and the left side for the other.

Masticatory performance was evaluated by measuring the degree of comminution (breakdown) of the silicon test food. After five trials totaling 10 grams of silicon, the particles were dried for 24 hours and sifted through a series of eight sieves of varying sizes. To assess masticatory laterality, the authors recorded the mandible’s displacement during the closing phase of the free-style mastication assays using a video camera. The side of mandible lateralization was determined by counting the lateralized side for each chewing cycle, with slow-speed video playback used to analyze the data.

The study analyzed the asymmetry in masticatory performance by calculating the asymmetry index for the free-style assays, comparing the masticatory performance of the right and left sides using the masticatory performance score (MPS) and cycle duration. Intraclass correlation coefficients and the smallest detectable difference were used to assess the reliability and agreement of the main parameters. Statistical analysis was conducted using the SPSS software package, version 20.0. Meanwhile, this study also provides a detailed and systematic approach to evaluating masticatory performance, using both video analysis and objective measurements to assess lateralization and comminution, offering valuable insights into how chewing is performed in a controlled setting.

Simione Meg et al. [67] investigated the use of full-face video recordings by Speech-Language Pathologists (SLPs) to assess chewing performance in individuals with Amyotrophic Lateral Sclerosis (ALS) and neurotypical controls. The study involved 19 ALS patients (mean age 58.26 years) and 10 neurotypical controls, capturing a range of normal to severely impaired chewing abilities.

Participants were seated comfortably with head support and asked to chew 3–5 Cheerios while their full-face videos were recorded. In parallel, 3D jaw kinematic data were collected using an optical motion capture system with eight cameras operating at 120 frames per second, providing a gold-standard for validating the SLPs’ ratings by capturing precise jaw movement data.

To ensure accuracy, the jaw movement data were digitally filtered, and reflective markers were strategically placed on the jaw and forehead. A marker was placed at the gnathion (JC) on the jaw, while two markers were placed on the right (JR) and left (JL) sides. The JR marker was used for analysis to minimize error from flesh movement. A four-marker array on the forehead helped eliminate head movement interference, providing clean jaw movement trajectories.

The video recordings were analyzed by five experienced SLPs, who rated both the spatial and temporal aspects of the participants’ chewing performance. By combining observational ratings with precise kinematic data, the study validated the clinical assessment of chewing in ALS patients, offering a comprehensive approach to evaluating this critical function.

Hossain Delwar et al. [63] analyzed the eating habits and jaw movements of 28 volunteers (17 males and 11 females, with an average age of 29.03 years) using 1080p video recordings captured by an SJCAM SJ4000 Action Camera placed 3 feet away from the participants. The study recorded 19 hours and 26 minutes of video footage from 84 meals, which was down-sampled to 6 frames per second (fps) for analysis. A Faster R-CNN deep learning-based object detection algorithm was employed to detect faces in the video frames, and a pre-trained AlexNet model was utilized to classify the frames as either ’bite’ or ’non-bite’.

The number of chews was determined using optical flow, a computer vision technique that estimates motion between images. The chewing algorithm tracked jaw movements and classified them into bite and chew segments. The study evaluated the performance of the face detector, image classifier, and chew counting algorithm using metrics like F1-score, precision, sensitivity, and mean Average Precision (mAP). This research presents a detailed methodology for tracking and analyzing jaw movements during eating, offering insights into the mechanics of food consumption.

Nakayama Enri et al. [66] conducted a comprehensive evaluation of the chewing process in elderly individuals residing in long-term care facilities in Japan. The study included 63 elderly participants who were capable of consuming solid food and performing daily activities, with exclusion criteria for those with cognitive or physical impairments that could influence the results. A novel approach was employed to track masticatory motion: color stickers were affixed to key facial landmarks, and a throat microphone (Inkou mike; NZ- 210 CjK, NANZU) was used to capture swallowing sounds, providing a temporal reference for the analysis. The jaw movements were recorded using an iPad Pro (Apple Inc.). For the evaluation, participants were given Hai Hain® Baby rice crackers (Kameda Seika Co., Ltd.) for a trial exercise and Happy Turn® soft rice crackers (Kameda Seika Co., Ltd.) for the formal assessment.

To assess participants’ chewing abilities, they were given a baby rice cracker to evaluate their risk of aspiration or choking, ensuring safety during the study. The cracker is as shown in Fig. 2l. For the main assessment, each participant chewed a 2 g rice cracker while maintaining a still face and looking at the camera. This allowed for precise tracking of their facial movements. The video recordings were analyzed using the DIPP-Motion PRO 2D software, providing measurements of various aspects of masticatory movement, including chewing time, the number of chewing cycles, cycle frequency, total mandibular movement, speed metrics, and variations in speed.

Mandibular movements were classified into circular and linear types based on the trajectory of the mandible during chewing. This comprehensive analysis provided valuable insights into the chewing abilities of elderly individuals, particularly those on a dysphagia diet, highlighting potential areas for intervention to improve dietary intake and safety.

Schmidt Maria et al. [62] introduced an automated food model testing system named ’CHEW’ to assess masticatory efficiency. This system consisted of round elastic food test units, calibrated recording plates, a sieve, and a fixed digital camera for standardized photography. The test materials were made from edible gelatin, as shown in Fig. 2f, molded into three hardness levels (soft, medium, and hard), simulating a range of food consistencies. Participants were instructed to chew the food into as many pieces as possible without swallowing, completing three chewing cycle variations for each hardness level: unilateral chewing on the right, left, and bilateral chewing. This resulted in nine 30-second cycles per participant.

Post-chewing, the food particles were rinsed, arranged on test plates, and photographed. The images were analyzed using computer-aided tools to calculate the number and area of the particles. The study compared masticatory efficiency between the control group and the CD (chewing dysfunction) group, analyzing differences in chewing side (left, right, bilateral) and food hardness (soft, medium, hard). This method provides a standardized approach to assessing masticatory efficiency, offering valuable insights into the mechanics of chewing.

Tufano Michele et al. [69] aimed to develop a rule-based system capable of automatically counting bites from video recordings using 468 3D facial key points. The study involved healthy adults aged 18–55 years with a BMI range of 18.5–30 kg/m2, who were asked to eat meals in a controlled dining room environment over four days. The meals included breakfast, lunch, dinner, and desserts, all recorded using an Axis M1054 camera. Video data was manually annotated with Noldus Observer XT 11 software, capturing parameters such as meal duration, bite interval, oro-sensory exposure, and the number of chews and bites per meal.

The results indicated that the rule-based system achieved an accuracy of 79% in bite counting when compared to manual annotations. Even without annotations, the system maintained an accuracy of 71.4%. The system performed consistently across different food textures, including soft and hard foods. This research demonstrates that rule-based systems using 3D facial key points can effectively count bites, offering advantages over traditional deep learning methods. These systems require less computational power, are easier to interpret, and have greater generalizability across various scenarios, making them well-suited for applications such as dietary monitoring.

Kumar Yogesh et al [21] conducted a study on the automatic recognition of food consumption sounds using machine learning. The researchers collected 1200 audio files representing 20 different food items from YouTube channels dedicated to eating-themed content. Advanced signal processing techniques, such as the creation of spectrograms and the extraction of Mel-frequency cepstral coefficients (MFCCs), were employed to capture the unique audio signatures of each food item.

Several deep learning architectures, including GRU (Gated Recurrent Unit), LSTM (Long Short-Term Memory), Inception ResNetV2, and CNN (Convolutional Neural Network), were trained to recognize the spectral and temporal patterns in the sounds of food consumption. The models were evaluated using standard metrics such as accuracy, loss, precision, recall, and the F1 score. This research provides a promising approach to automatically recognizing and classifying food consumption sounds, which could be useful for applications such as dietary monitoring and behavioral studies.

Finally, Aseef Ambreen et al. [65] used video recordings to analyze the chewing and swallowing behaviors of 327 healthy individuals aged 6 to 20 years. Participants consumed a validated regional cracker (Parle Monaco™), and the entire process was captured on video using an iPhone 12 Pro Max (Apple Inc, USA). The analysis was conducted independently by two speech-language pathologists, who focused on key parameters such as the total number of bites, masticatory cycles, and the total number of swallows. Swallowing events were identified by observing the movements of the thyroid cartilage.

The study also measured the total swallow time, from the start of consumption until the cracker was fully consumed. This comprehensive approach allowed for a detailed examination of masticatory and swallowing functions, providing valuable insights into the mechanics of eating in a broad demographic.

Wearable devices

The appropriate use of wearable devices, such as bone conduction earphones, some sensors, electromyography devices, etc., will make the detection of chewing more convenient and accurate, which is undoubtedly an innovative and efficient detection method.

As shown in Table 4, the relevant information about the chewing ability tested by wearable devices is summarized in the table.

A custom-made wearable device [20] called bitescan® (Sharp Co., Sakai, Japan) was used to assess chewing activity. This innovative device, which combines infrared distance sensors and accelerometers, was designed to monitor mandibular movement during the consumption of various test foods, including xylitol gum (Lotte Co. Ltd, Tokyo, Japan), gummy jellies (Meiji Co. Ltd, Tokyo, Japan), and rice balls (Maho-Cold Foods, Nara, Japan). The study involved 22 healthy volunteers who were tasked with evaluating the performance of a mandibular kinesiograph (K7, Myotronics, Kent, WA, USA) device worn on the right ear pinna to record mandibular movements while chewing. Participants were instructed to sit and relax, and the chewing activity began when they chewed a piece of xylitol gum for 10–20 seconds until it formed a soft bolus.

Chewing occurred under three conditions: on the right side, on the left side, and freely on both sides, with each condition lasting 30 seconds. For the gummy jellies, participants chewed on each side and freely until they swallowed, while the rice balls were consumed freely, following their usual eating habits. There were no specific restrictions on the speed of chewing, the number of chewing cycles, or the timing for swallowing.

The bitescan device, which features an infrared distance sensor and an accelerometer, was used to measure the number of chewing cycles. It operates at a mastication frequency of 20 Hz, scanning morphological changes in the skin surface behind the ear. The data collected by the bitescan were transmitted via Bluetooth to a smartphone application for real-time analysis. The device was designed with a variable ear-hook, available in three sizes, to ensure a proper fit on each participant’s ear pinna. Before measurements, the device was carefully fitted to ensure that the sensor was correctly positioned on the back of the pinna.

In addition to using the bitescan device, masticatory movements were also captured through videography to provide a more comprehensive analysis of the participants’ chewing activities. The accuracy, precision, and recall of the bitescan device were assessed by comparing its data to the video recordings, and the study examined various factors that could influence the reliability of the device.

By combining this high-tech wearable device with traditional observation techniques, the study aimed to provide a detailed and accurate assessment of participants’ chewing patterns. It also offered valuable insights into the functionality and reliability of the bitescan device for recording mandibular movements during various eating conditions. This approach marks a step forward in assessing chewing activity, with the potential for applications in both clinical and research settings.

One study [72] was aimed to investigate the potential of using acoustic parameters to measure masticatory performance using bone-conduction techniques and sound analysis. A total of 56 volunteers with healthy dentition participated in the study. The researchers captured chewing and gnathosonic sounds using a BONE VIBRATION HEADGEAR (Model: HG17BN-TX) bone-conduction microphone developed by TEMCo INDUSTRIAL LLC while the participants chewed raw peanuts. These recordings were then analyzed using Praat 5.4.04 software developed by Paul Boersma and David Weenink to assess various acoustic and occlusal parameters.

The study focused on characterizing the granulometry of the expectorated boluses by measuring the median particle size throughout the chewing sequence and during specific chewing strokes. Additionally, temporal chewing parameters were recorded, such as total chewing time, the time spent on individual chewing strokes, the number of chewing cycles, and chewing frequency. Acoustic parameters, including gnathosonic pitch and intensity, and masticatory sound pitch and intensity, were also evaluated for both the entire sequence and individual chewing strokes.

The conclusion of the study suggested that masticatory sound intensity could serve as a promising indicator for assessing masticatory performance. The findings imply that acoustic analysis provides a convenient and rapid method for evaluating the efficiency of the masticatory process, making it a potentially valuable tool in both dental research and clinical practice for assessing and monitoring masticatory function.

One device that leverages Electromyography (EMG) technology [71] was developed to quantify the electrical activity of masticatory muscles, with a particular focus on the masseter muscle due to its accessibility. The device consists of an Arduino Nano BLE 33 microprocessor, two Arduino Muscle V3 modules, a 9-volt battery, and silver or silver chloride conductive electrodes, which ensure high electrical conductivity and secure adhesion to the skin. These surface electrodes are critical for capturing EMG signals, a standard method in EMG technology. Additionally, a resistive divider is integrated into the device for signal conditioning.

For the chewing test, Conad bread was chosen as the food sample, pre-cut into 1 cm³ cubes, with a measured hardness (Young’s modulus) of 0.87 N/m. Participants were seated comfortably and instructed to chew the bread sample silently without moving, using their usual chewing style, and to signal when they had finished. Prior to the test, participants provided voluntary informed consent and were informed of the food’s ingredients to avoid allergic reactions.

The EMG signals collected during chewing were processed using Python software, which performed rectification, amplification, filtering, and bias elimination to facilitate detailed analysis of the chewing activity. This comprehensive approach to device design and signal processing allowed for precise measurement of the electrical activity produced by the masticatory muscles during chewing. This method represents a valuable step toward accurately quantifying the physiological aspects of masticatory function.

Robots

As shown in Table 4, relevant information about the robot chewing is recorded in the table. One of the most innovative approaches to studying human chewing is the use of robots to mimic the chewing process, which can help reduce individual differences in chewing behavior. A groundbreaking paper published in 2024 [61] introduces a novel method for assessing the impact of in vitro chewing on food bolus formation, utilizing the gray level co-occurrence matrix (GLCM) image analysis technique. In this study, two types of food products were specifically tested: pure organic beef burger patties from the brand Moreish, sourced from Palmerston North, New Zealand, and plant-based burger patties from the brand Beyond Meat, sourced from Los Angeles, USA.

The experiment began with in vitro chewing trials using a biomimetic masticating robot, which was equipped with a three-degree-of-freedom linkage mechanism and an artificial cavity. This advanced robot was specifically designed to replicate the natural molar crushing and grinding actions of human chewing, simulating the complex oral environment by creating a series of molar trajectories within the XY plane.

After the food samples were chewed by the robot, they were carefully removed from the artificial cavity and uniformly spread out on a flat surface for high-resolution scanning. To ensure data consistency and reliability, the experimental setup was repeated three times, yielding a comprehensive dataset consisting of 36 photographs of the chewed samples. A standard scale, placed in the background, was used as a reference to accurately measure and scale the images.

The study employed a sophisticated image analysis methodology based on GLCM textural features, which are known for their strong inter-correlations due to similar calculation methods. Given these inter-dependencies, the paper provided a unified guideline for selecting the most appropriate textural measures for analysis. The specific features chosen for the study were energy, dissimilarity, and homogeneity. These features were selected for their ability to provide a detailed assessment of the texture and structural information within the images, which are crucial for understanding the formation and characteristics of the food bolus.

  • Energy (also called angular second moment) quantifies the homogeneity of the image, reflecting the distribution of pixel intensities and their spatial relationships.

  • Dissimilarity measures the difference in pixel intensity values, giving insight into the texture’s contrast and the degree of variation within the image.

  • Homogeneity assesses the closeness of pixel values within a neighborhood, indicating the uniformity of the texture and the size of uniform regions.

By integrating these GLCM textural features, the study aimed to provide a detailed and quantitative analysis of the food bolus, shedding light on the effects of in vitro chewing on food structure and texture. This innovative approach not only enhances the understanding of the chewing process but also opens up new possibilities for applications across various fields, including food science, nutrition, and oral health.

This robot-assisted, image analysis-based methodology represents a significant advancement in mimicking and studying the complex chewing process, offering a more controlled and reproducible way to examine food bolus formation and the impact of chewing on food texture.

Future opportunity

The study and analysis of chewing behavior and masticatory performance are crucial for advancing both oral health and food science. As the field continues to evolve, several exciting opportunities for further research and application arise. Below are some of the key areas where future advancements and innovations could significantly impact the understanding of chewing and its effects:

One of the most promising directions for future research lies in the integration of robotic systems, like the biomimetic masticating robot, with AI and machine learning techniques. These systems can simulate human chewing with high precision, reducing individual variability and enabling more controlled experiments. As AI continues to improve, these robotic systems could become even more sophisticated, capable of replicating complex chewing patterns that vary by food type, age, or health condition. Machine learning models could be trained to automatically detect anomalies in chewing behavior or predict oral health issues based on the texture and structure of food boluses.

Wearable devices, such as the bitescan and EMG-based systems, hold immense potential for continuous, real-time monitoring of masticatory activity. Future advancements in these devices could lead to more personalized assessments, offering individuals the ability to monitor their own chewing habits and potentially detect early signs of oral health issues like temporomandibular joint disorders, dysphagia, or chewing inefficiency. With improvements in miniaturization, battery life, and sensor accuracy, these wearables could become even more comfortable, affordable, and accessible for widespread use in clinical and home settings.

As demonstrated in studies leveraging bone-conduction microphones and sound analysis, the use of acoustic parameters to measure masticatory performance is a promising avenue for non-invasive assessment. Future developments could focus on enhancing the sensitivity and accuracy of acoustic sensors, allowing for more precise analysis of the sounds associated with different chewing behaviors. Additionally, integrating acoustic analysis with other technologies, such as AI and machine learning, could lead to fully automated systems for assessing chewing efficiency in real-time. These systems could be used in both clinical and research environments to monitor patients, particularly those with swallowing difficulties or other chewing-related disorders.

Biomechanical models of the masticatory system, particularly those simulating jaw and muscle movements during chewing, could benefit from advances in computational power and data acquisition techniques. Future research could focus on creating highly detailed, patient-specific models that simulate chewing behavior based on individual anatomical features and functional requirements. These models could be used to study the effects of different types of food, as well as to optimize dental treatments, orthodontic interventions, and prosthetic design. The combination of biomechanical modeling with real-time data from wearables or robotic systems could revolutionize personalized dental care.

There is increasing recognition of the role chewing plays in overall health, including its impact on digestion, metabolic function, and even cognitive health. Future research could explore the deeper connections between chewing efficiency and conditions such as obesity, diabetes, or neurodegenerative diseases. Understanding how the masticatory system interacts with other physiological processes could open up new avenues for improving dietary interventions, designing targeted rehabilitation programs, and optimizing treatment plans for patients with oral health or swallowing disorders.

To fully realize the potential of these innovations, multidisciplinary collaboration will be essential. Future research in masticatory science will benefit from the integration of expertise from fields such as robotics, biomedical engineering, nutrition, oral health, and data science. Collaborative efforts could lead to the development of more comprehensive assessment tools, better diagnostic systems, and more effective treatments for chewing and swallowing disorders. Moreover, data collected from a variety of sources-such as wearable devices, robotic systems, and acoustic sensors-could be integrated into large-scale databases for population health studies, helping to advance public health initiatives related to nutrition and oral health.

As technology advances, the collection of personal data from wearables, robotic systems, and acoustic sensors becomes increasingly common. In the future, ethical considerations regarding privacy, data security, and informed consent will need to be carefully addressed. Striking the right balance between innovation and ethical responsibility will be crucial to ensuring that these technologies are used in ways that respect individual rights while still offering the benefits of advanced monitoring and assessment.

Beyond clinical settings, innovations in chewing analysis could also have significant implications for food science and the food industry. By better understanding how different foods interact with the chewing process, researchers could develop new food products tailored to specific chewing needs (e.g., for elderly individuals or those with dysphagia). Additionally, food manufacturers could use these insights to improve texture optimization, enhancing both the sensory experience and the nutritional value of products. Masticatory performance could also be integrated into consumer testing, providing more accurate assessments of food products’ ease of consumption.

Fig. 4
Fig. 4
Full size image

A pie chart summary of all studies included in this review

Conclusion

Figure 4 successively present the distribution trends of the pie charts for all countries, developed countries, and developing countries covered by the chewing assessment work in this review. This set of pie charts shows information about the countries of origin of the research work in this review. Figure 4a shows a comparison between developed and developing countries, with 72.5% of developed countries and 27.5% of developing countries. Figure 4b shows the distribution of all study source countries, with Japan accounting for the highest 23.2%. Figure 4c focuses on developing countries, with Brazil accounting for 42.1% of the total. Figure 4d shows the developed countries again, with Japan leading the way with 32.0%. These pie charts provide a visual representation of the different categories of countries in the study sources.

It is crucial to note that while this review encompasses the global advancements in masticatory performance assessment techniques, a pronounced imbalance exists in the geographical distribution of research efforts. The majority of innovative research and experimental methodologies, such as AI-driven computer vision systems, robotic simulators, and wearable devices, are concentrated in developed nations, including Japan, the United States, Germany, and the Netherlands. These countries dominate the field of masticatory performance assessment, attributed to their advanced technological resources, substantial research funding, and heightened attention to aging populations. In contrast, research in developing countries is relatively nascent and predominantly reliant on conventional methods, such as sieving and two-colored chewing gum tests. This disparity in research distribution not only underscores the gap in technological resources but also highlights the need for future research to foster international collaboration and technology sharing to bridge the global research divide.

The evaluation of masticatory performance has evolved significantly, with advances from traditional methods to modern, technology-driven approaches. This review highlighted the range of techniques used to assess chewing efficiency, from conventional methods like sieving and color-changeable chewing gums to cutting-edge AI-driven systems, including wearable devices and robotic simulators. While traditional methods offer valuable insights into masticatory performance, they are often limited by their reliance on specialized equipment and their inability to capture the complexities of individual chewing patterns.

The integration of Artificial Intelligence (AI) and robotics offers promising solutions to enhance the accuracy and reliability of masticatory assessments. Computer vision systems, machine learning algorithms, and wearable devices can provide real-time, non-invasive, and personalized assessments of chewing ability. These technologies not only improve diagnostic accuracy but also expand the potential for continuous monitoring of chewing function, which is especially beneficial for aging populations and those with oral health impairments.

The future of masticatory evaluation lies in the seamless integration of these innovative tools, which could lead to more comprehensive, accessible, and efficient methods of assessing and improving chewing function. By combining the strengths of robotics, AI, and wearable technologies, researchers and clinicians can gain deeper insights into the role of mastication in overall health, nutrition, and quality of life. As the field continues to evolve, multidisciplinary collaboration will be essential to further refine these technologies and ensure their effective application in both clinical and research settings.

In summary, the advancements in masticatory performance evaluation pave the way for improved diagnostics, treatment planning, and personalized care, with the potential to significantly impact the oral health and well-being of individuals, particularly the elderly and those with chewing disabilities.