electronic tongue (taste sensing system)
concept / concept-electronic-tongue
電子舌
Electronic tongue (taste sensing system) is an instrument combining an array of nonspecific, low-selectivity chemical sensors with overlapping signals and chemometric data processing. It is used as a nonhuman alternative for taste evaluation.
On this page (5 sections)
Reading the bitterness
An electronic tongue arranges sensors that each have low specificity and respond to many overlapping chemical signals, then analyzes their outputs by chemometrics (multivariate data processing). It is used, especially to evaluate pharmaceutical bitterness, as an alternative to human sensory assessment. A review notes that electronic-tongue output does correlate with conventional human taste panels, but not always consistently.
Its predictive power is disputed. For chemically similar drugs, such as quinine hydrochloride solutions anchored to 8 charged H1-receptor antagonists, correlation between electronic-tongue signals and human scores was high (R2 = 0.910; R2 = 0.9621 in a model using 2 other sensors). In contrast, a study ranking chemically diverse drugs by electronic-tongue Euclidean distance from least bitter said the ranking could not be compared with previously published human taste-threshold data. Ranitidine hydrochloride ranked most bitter by electronic tongue but received a bitterness probability of only 0.737 from a bitterness-prediction web server; conversely, sucrose octaacetate, a known bitter acetylated sugar, was predicted sweet with probability 0.752.
Of 55 electronic-tongue bitterness projects published in the preceding 10 years, only 26 validated their results with human or animal data. Validation was described as working particularly well for charged drugs, especially ones resembling H1 antihistamines, but was difficult for uncharged substances. Thus an electronic tongue is suitable for taste assessment in at least some projects and particular drugs, but fully modeling a biological taste system with artificial sensors is complex and requires validation by sample and condition.
This configuration also appears as an IUPAC definition. A 2026 review quotes it as an instrument combining an array of low-selectivity sensors with multivariate analysis or pattern recognition.
Another study examined bitterness masking in 5 herbal mother tinctures—white horehound, propolis, thyme, plantain, and eucalyptus—using an electronic tongue with 7 potentiometric sensors and PCA with caffeine as a standard bitter-direction reference. It says electronic-tongue analysis cannot quantitatively assess bitterness or taste in general. Only the abstract has been checked.
Sources for this overview
Checked for this overview: Checked the review's statement that correlation with human panels is inconsistent.
Checked for this overview: Checked R2 values of 0.910 and 0.9621 with H1-antagonist anchors. / Checked ranking disagreement in a 7-drug set and probabilities 0.737/0.752. / Checked 26 validated projects out of 55 and performance for ionized drugs.
Checked for this overview: Quotation of the official IUPAC definition of electronic tongue.
Checked for this overview: Statement that electronic tongue cannot quantitatively assess bitterness.
Structured claims — open a section to read each claim's confidence, evidence and limits
Sensory (3)
Scope: Oral drug formulation development. The evidence base reviewed in Pharmaceutics 2024 (55 projects over the past 10 years) and Frontiers in Chemistry 2024
Note: Both the supporting observation and the contradicting observation come from the same 2024 review, which presents both side by side. The reviewers' own summary is that the electronic tongue can succeed as an evaluation of bitterness rather than a prediction of it, but that correlation with human data is still needed even so.
Show data
concept-electronic-tongue —can_predict_human_bitterness_ratings_of→ "Drugs in solution. The reported agreement is high for chemically similar, charged (ionized) drugs (R2 = 0.910 with the Insent SB2BT0 sensor, R2 = 0.9621 with the ASTREE PLS model, both against 11 trained volunteers who rated 8 H1-receptor antagonists using quinine hydrochloride as the anchor), but ranking agreement breaks down across a chemically heterogeneous set of drugs"Scope: Layer: literature. A definition quoted as IUPAC's in a 2026 review.
Show data
concept-electronic-tongue —is_defined_as→ "In the IUPAC (International Union of Pure and Applied Chemistry) definition, an electronic tongue is described as an instrument combining an array of low-selectivity sensors with multivariate analysis or pattern recognition techniques."Scope: Layer: solution. Electronic-tongue data analyzed by PCA (principal component analysis), using caffeine as a standard reference for the bitterness direction.
Show data
concept-electronic-tongue —has_methodological_limitation→ "A bitterness-masking study of 5 herbal tinctures (white horehound, propolis, thyme, plantain, eucalyptus) using an Alpha Astree II electronic tongue with 7 potentiometric sensors states that electronic-tongue analysis cannot quantitatively assess bitterness or taste in general."