Not quite. The brain’s rhythms do drift, and phase matters, but saying tACS “only works” closed-loop is too strong.

• Open-loop can work when you match the stimulation frequency to the endogenous rhythm and get enough field in the right place. You can shift alpha power and behavior with fixed-frequency, individualized tACS, and there’s even direct animal evidence of alpha entrainment without closed-loop. Effects are modest and state-dependent, but they’re real.
• Closed-loop helps a lot when the target oscillation is wobbly or when you need phase-specific effects. Locking stimulation to the ongoing phase boosts reliability and enables bidirectional control (enhance at one phase, suppress at the opposite). That’s been shown cleanly for tremor by phase-locking tACS to the tremor itself.
• Phase really matters to physiology. MEPs and steady-state responses vary with the phase at which the tACS cycle hits cortex. You can get enhancement at one phase, suppression at another, even with open-loop if the endogenous rhythm is stable enough during the block. Closed-loop just keeps you on target as phases drift.
• Practical split:
  • If your rhythm is stable or externally driven (e.g., steady-state sensory), open-loop with individualized frequency often suffices.
  • If your rhythm meanders (most cognition) or you need maximum effect size/precision, go closed-loop phase-locked tACS. Tooling exists to do fast phase tracking in real time.

So yes, ridges and troughs drift. Closed-loop lets you surf the wave; open-loop can still nudge the ocean if you pick the right frequency and montage. You don’t need a sniper rifle every time, but it’s handy when the target won’t sit still.

===

Short answer: not really. The “catch a trough and super-depolarize it” story fits phase-locked stimulation (tACS), not tRNS. tRNS is zero-mean, aperiodic noise, so it isn’t aligned to any particular phase of your ongoing oscillations. It helps mostly because noisy, fast fluctuations nudge a nonlinear, near-threshold system into firing more often and carrying weak signals better. That’s stochastic resonance and temporal summation, not trough hunting

What’s actually doing the work, in brief:

• Stochastic resonance: add the right amount of noise to a nonlinear system and subthreshold inputs cross threshold more often, improving detection and behavior. That’s been shown directly with tRNS over visual cortex. Too little noise does nothing; too much drowns the signal.
• Sodium-channel kinetics: pharmacology shows blocking voltage-gated Na⁺ channels blunts the excitability boost from tRNS, which points to rapid fluctuations biasing channel opening and making spikes easier to trigger. High-frequency tRNS (roughly 100–640 Hz) is the sweet spot for this.
• State dependence, not phase dependence: because the current is random, sometimes it lands on troughs, sometimes on peaks. The net facilitation comes from the system’s nonlinearity and proximity to threshold, not preferentially depolarizing troughs. Some papers see oscillatory changes (e.g., gamma), others don’t; there’s no consistent phase-specific story.
• Lasting excitability changes: brief HF-tRNS over M1 increases corticospinal excitability for ~60 minutes, which is robustly reported, again without any need to invoke trough timing.

So your intuition has the right spirit (timing matters in brainland), but with tRNS the win comes from noise helping a twitchy system do more with faint inputs, not from sniping troughs like a waveform ninja. If you want phase tricks, that’s tACS territory; tRNS is the chaos goblin that boosts SNR when dosed just right.

==

Not quite. The brain’s rhythms do drift, and phase matters, but saying tACS “only works” closed-loop is too strong.

• Open-loop can work when you match the stimulation frequency to the endogenous rhythm and get enough field in the right place. You can shift alpha power and behavior with fixed-frequency, individualized tACS, and there’s even direct animal evidence of alpha entrainment without closed-loop. Effects are modest and state-dependent, but they’re real.
• Closed-loop helps a lot when the target oscillation is wobbly or when you need phase-specific effects. Locking stimulation to the ongoing phase boosts reliability and enables bidirectional control (enhance at one phase, suppress at the opposite). That’s been shown cleanly for tremor by phase-locking tACS to the tremor itself.
• Phase really matters to physiology. MEPs and steady-state responses vary with the phase at which the tACS cycle hits cortex. You can get enhancement at one phase, suppression at another, even with open-loop if the endogenous rhythm is stable enough during the block. Closed-loop just keeps you on target as phases drift.