Homeostatic learning rules for gap junction coupling

Goel and Mehta observed that the voltage-gated gap junction appears to follow a homeostatic principle with respect to transjunctional current. When that current is small, as in synchronous bursting, junctional conductance is large. When it is large, as in anti-synchrony, conductance is small.¹ The authors argued that firing patterns change conductance in a way that stabilises the current. From this they built a rule for slow modification of gap junctions. They then enumerated firing combinations of coupled cells, with their probabilities, to write an equation for how a junction's strength evolves.²

Baluška and Levin placed the study within work on cognition in non-neural systems. They noted that Goel and Mehta examined insulin secretion from the islets of Langerhans. Past measurements of junctional conductance could not explain systemic properties, such as reduced coupling in type-2 diabetes.³ Goel and Mehta instead treated the process top-down, as a learning-like adaptation, rather than following the prevailing focus on molecules and their interactions.⁴ Baluška and Levin report that, in this model, reduced coupling in diabetes is required for blood insulin to rise after hyperglycemia.

Later modelling work in the Levin group and colleagues treats junction conductance as a key parameter of multicellular voltage patterns. Pietak and Levin reported simulations in which junction conductivity, boundary restriction and cluster geometry shape the resting voltage of a cell cluster.⁵ Pai and colleagues represented each cell's voltage by two opposing voltage-gated channel groups, with the state of a cell modulated by its neighbours.⁶ Riol and colleagues analysed how connexin types with different voltage-gated conductances can hold regions at distinct potentials.⁷ They show conductance curves for Cx43 and Cx45 homotypic and heterotypic channels.⁸

Acharya and colleagues, in 2026, drew on simulations by Cervera and colleagues. In those simulations junction conductance decides whether single-cell voltage perturbations are buffered by neighbours or persist as stable local regions.⁹ The Acharya paper applies this to oral squamous cell carcinoma, where fragmented junctional communication is expected to turn transient voltage noise into lasting regional change.⁹ This differs in aim from the 2013 rule. The 2013 work asked how firing patterns set coupling. The later work asks what a given coupling level does to voltage patterns.