ISI

Experiments

ISI (inter-stimulus interval) does two things at once: it SPLITS a trial’s stimuli into sequential screens, and it sets the blank gap, in milliseconds, between consecutive screens. A trial with more than one screen (sequential presentation) exists either because ISI says so or because an explicit stimScreens grouping does; when stimScreens resolves, it takes over the split and ISI keeps only its gap role (see Works with). Use it for prime/probe sequences, RSVP-like streams, target-then-choices layouts, and gaps between sequential audio/video files.

Syntax

Put a single number in the cell, in milliseconds, for the blank gap between two stimuli: ISI 100 shows the first stimulus, blanks the screen for 100 ms, then shows the rest.

For longer sequences, give one value per gap, separated by semicolons. The first value is the gap after the first stimulus, the second value the gap after the second, and so on. Each value can be:

• a plain number of milliseconds, such as 250
• min_max (for example 200_400), a random gap drawn once when the experiment loads
• a|b|c (for example 100|200|400), one of the listed gaps picked at random when the experiment loads
• V*N (for example 250*3), the value V repeated for N gaps
• left empty, as in 10;;100, meaning no screen break at that point, so the two stimuli on either side of it appear together on one screen

The screens follow from the breaks: the first stimulus opens the first screen, and every value that is a real gap (including 0) starts a new one. If you write fewer values than there are breaks to make, all the remaining stimuli land together on the last screen.

Options

Value	Behavior
empty (or absent)	No screen breaks: all stimuli simultaneous, single screen. The ISI results columns are left empty.
0	A real screen break with a 0 ms gap: stimuli are on separate sequential screens that follow each other immediately.
positive integer	Blank gap of that many ms between the two screens, starting when the previous screen’s presTime has elapsed.
list g1;g2;...	Per-gap durations in order.
empty entry, as in ;;	Keeps the neighboring stimuli on the same screen.
min_max, a|b|c	Random gap, drawn once per trial row when the experiment loads.
V*N	N identical gaps.
Nf	Accepted as a frames value but currently behaves as N ms (see Tips & gotchas on the presTime page).

Defaults & missing values

Leave the cell empty, or leave the column out, and the trial has no screen breaks: every stimulus appears at once on a single screen, with no gaps. The results columns ISI_ms, ISI_f and ISI_fDuration are left empty for that trial.

When you do use ISI, each gap shows the mask if the trial has one, plus the ISI trial text and the ISI fixation if you asked for those.

Works with

• presTime: the mandatory partner. Every screen except the last needs a positive presTime, because a gap only starts once the screen before it has hidden.
• stimScreens: on trials without a target, a stimScreens grouping that matches the stimulus count decides how the stimuli are split into screens, and ISI then only supplies the gap durations. If ISI is left empty on such a multi-screen trial, every gap is 500 ms. With no stimScreens value, one that does not match the stimulus count, or on any trial with a target, ISI splits the screens itself as described above.
• target: the target occupies the first screen. When the target is alone on its screen, it is hidden after its presTime and the rest of the stimuli follow after a wait, see Tips & gotchas for which value that wait actually uses.
• mask: shown during each ISI gap in sequential trials.
• fixation value ISI: a fixation cross is shown during the gaps.
• trialTextOptions value ISI: per-gap text segments from trialText.
• Audio and video trials: the same per-gap values set the silent gaps between consecutive media files.
• Word trials: a numeric ISI splits the trial’s words across sequential screens in the same way.
• stimOptions loop: the last positive ISI value is reused as the blank period between presentation loops.
• Results: the gaps actually measured during the trial are saved for test trials, one per gap and semicolon-joined, in ISI_ms, alongside ISI_f and ISI_fDuration. Trials without an ISI value leave those columns empty.

Caution

With presTime set to 0 or left empty, the sequence stops after the first screen: the gap never starts and the later screens never appear.

Examples

1. Prime then probe with a 100 ms blank (presTime list times each screen):

type	stimFormat	stim1	stim2	presTime	ISI	keyboard	key
test	.jpg	prime	probe	200;500	100	f j	f

2. Grouping with an empty entry: stim1 alone (10 ms gap), stim2 and stim3 together (100 ms gap), stim4 alone:

type	stimFormat	stim1	stim2	stim3	stim4	presTime	ISI
test	.png	a	b	c	d	300;300;300	10;;100

3. RSVP-style stream, 4 stimuli, identical 250 ms gaps via multiply syntax, last screen stays up for a response:

type	stimFormat	stim1	stim2	stim3	stim4	presTime	ISI	keyboard	key
test	.png	w1	w2	w3	w4	150*3	250*3	f j	j

Tips & gotchas

Caution

Avoid trailing semicolons. A trailing empty or invalid value still counts as a value, so 100; on a three-stimulus trial gives two screens of one stimulus each and silently drops the third, where plain 100 shows it.

• A single ISI value on a trial with many stimuli lumps everything after the first break onto one screen: ISI 100 with four stimuli gives screens of one and three stimuli.
• On trials where the target is alone on its screen, the wait between the target disappearing and the test screen appearing is the target’s presTime, not the ISI value.
• Gaps that measure as 0 ms are not reported in ISI_ms.
• ISI random values are drawn once per trial row when the experiment loads, so a row that runs several times reuses the same gap. ITI re-rolls on every presentation instead.
• A trial with only one stimulus always produces a single screen, so ISI does nothing without at least two stimuli (or a target plus one stimulus).

Common combinations

ISI appears in these worked recipes:

• Two-interval forced choice: Present two stimuli one after the other and ask which interval contained the target.
• Masked priming: Flash a prime briefly, mask it, then time a choice response to a probe.
• Match to sample: Show a sample stimulus, then an array of choices in shuffled positions, scoring the matching choice correctly wherever it appears.
• AX-CPT: Present a cue and then a probe as two screens of one trial, with the correct response depending on the cue-probe pair.
• Paired presentation: Present items in fixed pairs whose order shuffles for every participant while the members of each pair always stay together on screen.