Source
Szczepanski, Crone, Kuperman, Auguste, Parvizi and Knight (2014), LIT-1982, pp. 2–12; read in NOTE-1528.
What was actually shown
Eight patients with subdural grids held covert attention on a cued hemifield for 1000–2000 ms. In 123 of 273 high-gamma-responsive electrodes, 80–250 Hz amplitude was locked to the 2–5 Hz phase (surrogate test, FDR q < 0.01), and locked to its trough in 117 (pp. 3–4). In 55 the coupling was stronger for the contralateral field, with no matching power difference in 53 of them (p. 4). Dropping the first 500 ms, where the evoked potential lies, left 107 electrodes significant (p. 4). In 28 electrodes, stronger coupling in the 1000 ms before the target went with faster reaction times, contralateral trials only; neither band’s amplitude alone predicted RT (pp. 4–5).
Each of these could have failed: coupling could have been absent, equal across hemifields, confined to the evoked window, or unrelated to RT.
What this does not say
- That the coupling improves attention or causes faster responses. Nothing was manipulated. The title’s “facilitate” and the Discussion’s “serves as a mechanism” (p. 9) are interpretation; the closing reading is that it indexes “the overall level of engagement” (p. 10).
- That slow waves constrain fast ones in any sense beyond a statistical dependence. The paper’s mechanism is excitability modulation through membrane potentials (p. 2), entirely micro-level; it does not bear on whether a macro pattern has a power its parts lack (THEORY-1314).
- That the coupling is free of waveform-shape artefact. No shape control is reported (see THEORY-1312).
- Anything about choice, value or stalemates. The task is instructed spatial attention.