Cerebral capillary computation

This article synthesizes recent evidence to propose a framework for control of energy delivery in the brain in which the capillary bed functions as an active, distributed signal-processing network that senses neuronal activity and metabolic state and converts these inputs into electrical commands that regulate upstream diameter to control blood flow. Capillary endothelial cells (ECs) form an electrically coupled syncytium via gap junctions, whereas pericytes are vertically integrated into this network at peg-socket junctions, enabling bidirectional electrical communication. It is proposed that thin-strand pericytes and their associated underlying ECs constitute a "capillary computational unit" (CCU): a local transformer-like module in which pericytes act as rich multimodal sensors and signal generators, whereas ECs are optimized for signal amplification and long-range transmission. Emphasis is placed on the ion channel toolkit that implements CCU computations, with discussion of how different conductances shape membrane voltage to encode local energetic demand and propagate signals over long distances. Kir2.1 channels emerge as a keystone conductor and regenerative carrier of hyperpolarizing signals; K_ATP channels couple energy status and adenosine levels/glucose availability with electrical output; small- and intermediate-conductance Ca²⁺-activated K⁺ channels in the arteriole-capillary transition zone provide amplification; transient receptor potential and Piezo1 channels impose depolarizing and mechanosensory feedback constraints; and chloride channels (notably TMEM16A) act as voltage tethers that clamp or reset local membrane potential. Framing these elements computationally suggests that addition and subtraction, gain control, shunting, and veto-like logic may arise naturally from network architecture and channel biophysics.