The Tongue as an Excitable Medium

Geographic tongue (GT) is a benign condition affecting approximately 2% of the population, whereby the papillae covering the upper part of the tongue are lost due to a slowly expanding inflammation. The resultant dynamical appearance of the tongue has striking similarities with well known phenomena observed in excitable media, such as forest fires, cardiac dynamics and chemically-driven reaction-diffusion systems. Here we explore the dynamics associated with GT from a dynamical systems perspective, utilizing cellular automata simulations. We emphasize similarities with other excitable systems as well as unique features observed in GT. Our results shed light on the evolution of the inflammation and contribute to the classification of the severity of the condition, based on the characteristic patterns observed in GT patients.

💡 Research Summary

The paper “The Tongue as an Excitable Medium” re‑examines geographic tongue (GT), a benign oral condition affecting roughly two percent of the population, through the lens of excitable‑medium theory. GT is characterized by the gradual loss of filiform papillae on the dorsal tongue surface, producing expanding, irregularly shaped lesions that resemble the patterns seen in forest‑fire spread, cardiac wave propagation, and reaction‑diffusion chemical systems. The authors argue that these visual similarities are not merely superficial; the underlying dynamics of GT can be captured by the same mathematical structures that describe classic excitable media.

To test this hypothesis, the authors construct a two‑dimensional cellular automaton (CA) model. Each lattice site can be in one of three states—resting, excited, or refractory—mirroring the standard three‑state representation of excitable media. Transition rules are defined by a local excitation threshold: a resting cell becomes excited if a sufficient number of its eight nearest neighbours are excited; an excited cell then enters a refractory period of fixed duration before returning to the resting state. The model parameters include excitation propagation speed, refractory duration, and the critical neighbour count. By varying these parameters and by imposing realistic boundary conditions that mimic the tongue’s irregular geometry and heterogeneous mucosal properties, the authors generate a wide spectrum of spatiotemporal patterns.

Simulation outcomes reproduce the main visual motifs observed in clinical photographs of GT patients: circular “target” waves, branching fronts, and concentric rings. When two wave fronts collide, they annihilate or give rise to new fronts, just as in chemical reaction‑diffusion experiments. A crucial divergence from classic excitable media is that, after a wave passes, the affected lattice sites permanently lose their “papilla” state, representing irreversible tissue damage. This irreversible loss is modeled by permanently marking a site as “dead” after a certain number of excitation cycles, thereby coupling the excitable dynamics with a damage accumulation process.

The authors perform a systematic parameter sweep to construct a “severity map” that links model parameters to observable clinical grades. High propagation speed combined with short refractory periods yields large, rapidly expanding lesions that correspond to severe GT, whereas slower propagation and longer refractory periods produce small, localized lesions typical of mild cases. The map provides a quantitative framework for classifying GT severity based on measurable pattern characteristics such as lesion size, curvature, and branching frequency.

In the discussion, the paper acknowledges several limitations. The CA model abstracts away detailed biological processes such as immune cell recruitment, cytokine diffusion, vascular supply, and microbial flora, all of which are known to influence GT progression. Consequently, the model cannot yet explain patient‑specific asymmetries or recurrent flare‑ups. The authors propose extending the framework by coupling the CA with continuous reaction‑diffusion equations (e.g., FitzHugh‑Nagumo or Oregonator formulations) to capture chemical signaling, and by incorporating stochastic immune‑cell dynamics to reflect patient variability. They also suggest using high‑resolution intra‑oral imaging to extract patient‑specific parameters, enabling personalized simulations that could predict the outcome of therapeutic interventions such as topical corticosteroids or anti‑inflammatory mouthwashes.

Overall, the study demonstrates that geographic tongue can be fruitfully interpreted as an excitable medium with an added irreversible damage component. The cellular automaton approach reproduces the hallmark patterns of GT, provides a mechanistic explanation for lesion expansion, and offers a quantitative tool for severity grading. By bridging oral pathology with nonlinear dynamics, the work opens new avenues for predictive modeling, early diagnosis, and individualized treatment planning in a condition that has traditionally been assessed only by visual inspection.