Sham Control & Blinding
Methodological safeguards against bias: a sham (placebo) condition that mimics the intervention without its active element, and blinding of participants and/or outcome assessors to which condition applies. In neuromodulation trials a credible sham — inactive stimulation, or in some designs a burr-hole procedure without active therapy — is used to separate specific device effects from placebo and natural history, while blinded assessment reduces expectancy and rater bias.
For recording-and-control BCIs, blinding and sham are often impractical or unethical: a participant plainly knows whether a cursor is moving under their own control, and sham surgery for a purely recording implant is hard to justify. Such trials therefore lean on objective performance metrics, within-participant controls (decoder-on versus decoder-off, or comparison to chance), blinded outcome adjudication, and benchmarking against historical or objective standards instead of a placebo arm.
A closed-loop DBS trial can blind cleanly by turning stimulation on or off invisibly and comparing symptom scores. A cursor BCI cannot: the user always feels whether they are in control. So its efficacy claim rests on decoder-on versus decoder-off blocks and performance far above chance, not on a placebo group.
Stimulation BCIs can be blinded; recording-and-control BCIs mostly cannot, and must argue efficacy differently.
The near-impossibility of clean blinding is a genuine methodological limit of BCI efficacy evidence, not a formality to be waved away.