SOMOTE has four settings that actually change how the words reach you — how long each phrase appears, how visible it is, where it lands on screen, and how often it returns. None of the defaults are aesthetic guesses. Each one sits where fifty years of exposure research suggests it should.
Flash length. Around seventeen milliseconds — one frame on a 60 Hz display — a word slips past conscious report; ask someone what they saw and they will tell you nothing appeared. By roughly thirty to fifty milliseconds, a masked word reliably shapes the response that follows even when it still cannot be named, and past about eighty milliseconds it becomes plainly readable. SOMOTE's flash dial spans that entire curve deliberately, and Deep Mode parks at the seventeen-millisecond threshold, repeating every seven seconds so that even the faintest registration compounds across a working day.
Opacity. The counterintuitive finding in this literature is that quieter often means stronger. Across Bornstein's meta-analysis of 208 exposure experiments, stimuli participants never consciously registered produced substantially larger liking effects than stimuli they noticed; in head-to-head comparisons, five-millisecond flashes outperformed five-hundred-millisecond views. That is the case for the three-to-eight-percent opacity zone: present in every glance, beneath the threshold of noticing, and therefore never something you start arguing with.
Position. Priming studies place the word at the point of fixation — the fovea, the eye's reading center — because that is where a brief word is processed most fully. Words presented off to the side register weakly and lopsidedly. So SOMOTE centers phrases mid-screen, in the path of your gaze, rather than tucking them into a corner you have already trained yourself to ignore.
Frequency and rotation. Liking climbs steeply across the first ten to twenty repetitions of a stimulus and then flattens; Berlyne's two-factor account explains the plateau as familiarity gains giving way to satiation. Rotating several messages reliably beats hammering one. Hence playlists, shuffle, and a frequency dial rather than a single phrase on an endless loop — a handful of lines taking turns, each arriving fresh enough to keep working.
A necessary caveat: none of these studies used SOMOTE. What they establish is the mechanism SOMOTE automates. The honest claim is not that this software has been clinically validated, but that the settings were chosen by reading the literature on repetition and exposure closely, and then building the thing those papers imply.