The Insula Mediates Access to Awareness of Visual Stimuli Presented Synchronously to the Heartbeat

Salomon et al. (2016) — The insula mediates access to awareness of visual stimuli presented synchronously to the heartbeat

A nine-experiment series (seven psychophysical, two 7T fMRI) from Olaf Blanke’s Lausanne lab arguing that the timing of a visual stimulus relative to your own heartbeat changes whether and how fast you become conscious of it — and that the anterior insula is where the two signals meet. It is the wiki’s first source in which interoceptive timing gates exteroceptive (visual) awareness, and one of the few that puts a cardiac-interoceptive manipulation and a high-resolution insula measurement in the same study.

The core result

Using continuous flash suppression — high-contrast dynamic masks shown to the dominant eye render a target to the other eye invisible until it slowly “breaks through” into awareness — the authors flashed a target (a yellow octagon) either synchronously with the participant’s heartbeat (locked to the ECG R-peak, ~10 Hz mask) or asynchronously (at 80% or 120% of the heart rate). Targets flashing in time with the heart took reliably longer to reach awareness (Exp 1, n=31; replicated within-subject in Exp 2, n=15). Cardiac-synchronous visual events are suppressed from consciousness.

The interpretation the authors favour: the heartbeat is a lifelong self-generated signal with sensory consequences across touch, proprioception, audition and vision (it moves the eyes and retina; de Kinkelder et al. 2011). A predictive system should attenuate the sensory consequences of self-generated cardiac activity — the same logic by which self-produced touch is cancelled (Blakemore et al. 1998). A visual stimulus that happens to arrive at the cardiac frequency is treated as one of those predictable self-consequences and suppressed. See interoceptive-inference.

Why the control experiments matter

The claim is fragile in exactly the way cardiac-timing studies are fragile — a “cardiac effect” can be a heart-rate or attention artefact in disguise — so the paper’s spine is its controls, and they are unusually complete:

• Not explicit heartbeat perception (Exp 3): the same participants were at chance judging whether the flashes were synchronous with their heartbeat (BF favours the null), so the suppression does not require perceiving the cardio-visual relationship.
• Not a response/detection bias (Exp 4): a classic CFS control with the target blended into both eyes’ masks (no interocular suppression) showed no synchronous/asynchronous difference (BF=0.34), so the effect is in conscious access, not decision.
• Not the visual stimuli themselves (Exp 5): replaying another participant’s heartbeat-locked sequence, decoupled from the viewer’s own heart, abolished the effect — the difference depends on coupling to the viewer’s own heartbeat, not on any visual property of synchronous vs asynchronous sequences.
• Frequency, not phase (Exp 6): phase-shifting the target by half a cardiac cycle while holding frequency constant left the suppression unchanged (BF=0.25). This is the key dissociation from the pulsed-inhibition tradition: the effect is not about where in the cardiac cycle the stimulus lands (systole vs diastole) but about matching the cardiac frequency.
• Not CFS-specific (Exp 7): visual crowding — a wholly different paradigm using discrimination accuracy, not breaking-time — reproduced the suppression (synchronous accuracy 56% vs 69%), ruling out artefacts specific to the time-to-emergence measure.

The neural claim

Two 7T fMRI experiments localized the effect. In both a visible condition (Exp 8, unmasked octagon) and an invisible condition (Exp 9, target suppressed by CFS, analyzed only on trials where participants were fully unaware of it), the bilateral anterior insula showed lower BOLD to synchronous than asynchronous cardio-visual stimuli. Control regions from an independent interoceptive-attention localizer (ACC, right STG, occipital cortex — see Fig 6) showed no such difference, and regressing out the cardiac signal (RETROICOR) left the insular result intact, arguing against a vascular pulsation confound.

That the effect survives full invisibility (Exp 9) is the strongest part: the insula tracks cardio-visual synchrony for stimuli the participant never consciously sees, consistent with a low-level, pre-conscious suppression rather than a report-driven one.

Where it sits in the wiki

• It converges with the wiki’s reading of the anterior insula as a multimodal integration / salience hub (Craig 2009; Seth 2013) — here integrating interoceptive (cardiac) and exteroceptive (visual) signals — and extends its remit beyond self-awareness to the perception of the external world.
• It is a distinct kind of evidence for interoceptive predictive coding: not the usual interoceptive-report or emotion paradigm, but a case where the prediction of the heartbeat’s sensory consequences is claimed to shape exteroceptive consciousness. The authors are careful that a non-predictive temporal-disparity account also fits.
• It sharpens the cardiac-cycle-effects page’s note that vision shows mostly null cardiac-phase effects: Salomon et al. find a robust cardiac-frequency effect on vision precisely by dissociating it from phase — a different axis of cardio-visual coupling from the systole/diastole one.
• It complements Haruki & Ogawa (2023) and the domain-generality question: where Haruki narrow the right anterior insula toward cardiac interoception, Salomon show the (bilateral) anterior insula also arbitrating cardiac-vs-visual synchrony — the same cortex handling cross-domain temporal relations, not body-signals alone.