JOVANA
Explore Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Timing the Brain: State- and Phase-Dependent Stimulation

Excitability is a moving target. This guide shows why the moment of stimulation matters as much as the fact of it, and assembles the full stroke-rehabilitation loop.

Excitability is a moving target

The cortex is not equally plastic from one moment to the next; its excitability rides on ongoing oscillations. Brain-state-dependent stimulation delivers the pulse only when the brain is in a receptive state — for instance during the desynchronization of the sensorimotor mu rhythm that marks movement preparation. Same pulse, better moment, larger effect.

This is the natural marriage of the two halves of a BCI: you use the decoder not to move a cursor, but to detect the plastic state and time the write. The reading side exists to schedule the writing side.

Phase-dependent stimulation

Finer still, phase-dependent stimulation locks the pulse to a specific phase of the ongoing rhythm. The remarkable finding is that the same stimulus can potentiate at one phase and do nothing — or even depress — at the opposite phase. Excitability, and with it plasticity, is phase-organized.

p_{\text{LTP}} \;\propto\; \tfrac{1}{2}\big(1+\cos(\phi-\phi_{0})\big)

A schematic of phase-gated plasticity: the probability of potentiation is modulated by the phase phi of the sensorimotor rhythm at which the pulse lands, peaking at a preferred phase phi-nought. The engineering catch is that you must fire slightly ahead of phi-nought to hit it after the loop delay.

The rehabilitation loop, assembled

Put the pieces together for stroke. Detect movement intention through motor imagery and ERD; then, inside the plastic window, deliver functional electrical stimulation or an orthosis that completes the movement, together with contingent proprioceptive feedback. The reafferent volley now arrives paired with the cortical command — exactly the Hebbian coincidence recovery needs — closing the rehabilitation loop. This is the design behind the field's landmark contingent-feedback trials.

P_{\text{eff}} \;=\; \Pr\!\big(\,|t_{\text{stim}}-t_{\text{intent}}| < \tau_{\text{STDP}}\,\big)

The effective dose of a closed-loop session is the fraction of paired events whose stimulus lands inside the plasticity window. Loop latency and jitter shrink P-eff directly — a slow or noisy loop wastes repetitions no matter how many you run.

Why contingency and effort matter

The patient is not a passive recipient. User learning and BMI skill learning mean the loop trains two learners at once — the decoder and the brain. Agency and attention amplify plasticity; a bored patient watching non-contingent stimulation learns little, and reward can gate the effect further. The therapeutic loop works best when the patient is genuinely trying.