Learning induces activation-mechanism–dependent neural plasticity in an intracortical microstimulation task

Abstract

Electrical microstimulation provides high-resolution control of neural circuits for causal studies and restoration of impaired functions, yet how responses to artificial activation evolve with learning remains unclear. Here, we deploy a detection task and pair ultraflexible electrodes for stable intracortical microstimulation (ICMS) with longitudinal imaging and recordings to track single-cell and population responses across weeks of learning. Detection thresholds decreased with learning, indicating plasticity. Chronic imaging showed that stimulus-evoked recruitment expanded at a fixed current, while a consistent number of neurons continued to underlie behavioral responses. A subset of learning-sensitive cells enhanced modulation and reduced latency. Electrophysiological recordings further distinguished two forms of adaptation: Directly activated, pulse-locked neurons strengthened their excitability, whereas polysynaptically recruited neurons expanded in number and were predictive of behavioral outcomes. These results show that learning in an ICMS task reshapes cortical circuits through activation-mechanism–dependent plasticity, underscoring the need for stimulation paradigms that adapt to both cell-intrinsic and network dynamics.