We present a new human-computer interface that is based on decoding of attention through pupillometry. Our method builds on the recent finding that covert visual attention affects the pupillary light response: Your pupil constricts when you covertly (without looking at it) attend to a bright, compared to a dark, stimulus. In our method, participants covertly attend to one of several letters with oscillating brightness. Pupil size reflects the brightness of the selected letter, which allows us-with high accuracy and in real time-to determine which letter the participant intends to select. The performance of our method is comparable to the best covert-attention brain-computer interfaces to date, and has several advantages: no movement other than pupil-size change is required; no physical contact is required (i.e. no electrodes); it is easy to use; and it is reliable. Potential applications include: communication with totally locked-in patients, training of sustained attention, and ultra-secure password input. : human-computer interface, brain-computer interface, pupillometry, covert visual attention Manuscript in preparation [v1.1.4; Wed Sep 9 13:22:41 2015; There are many types of BCIs (reviewed in Donoghue, 2008;Nicolas-Alonso & Gomez-Gil, 2012) , which differ in the neural signal that they use (e.g., neural spikes or electroencephalography [EEG]), the way that neural activity is processed (e.g., through a classifier or by measuring overall activity in specific brain areas), and the actions that they perform (e.g. controlling a robotic limb, or writing text). Here we present a new method, which uses pupil size, rather than brain activity, as the controlling signal. Our method is related to two existing methods: the P300 speller (Farwell & Donchin, 1988), which is functionally similar to our method but relies on a different controlling signal; and a recent pupillometry-based method (Stoll et al., 2013) , which is functionally different from our method but relies on the same controlling signal. P300 spellers are among the most successful BCIs (reviewed in Fazel-Rezai et al., 2012) . They exploit the fact that rare visual stimuli elicit positive deflections in the EEG signal about 300 ms after their appearance. This event-related-potential (ERP) component is called the P300, and is largest for stimuli that are overtly (while looking at them) or covertly (without looking at them) attended (e.g., Treder & Blankertz, 2010). In a classic P300 speller, the participant sees a grid of letters. One letter, or sometimes one full column or row of letters, is highlighted at a time. A P300 is elicited each time that a letter is highlighted. The participant selects a letter by attending to it, usually by looking at it directly (i.e. overt attention), which leads to an increased P300 when that letter is highlighted. In its simplest form, the letter that, when highlighted, elicits the highest P300 is selected; however, most P300 spellers now use sophisticated classification techniques, and pool information from mult...