Skip to main content
. Author manuscript; available in PMC: 2022 Aug 24.
Published in final edited form as: Nat Hum Behav. 2022 Feb 24;6(5):732–742. doi: 10.1038/s41562-022-01294-x

Figure 1. Experimental design, example trials, alpha frequency predictions and behavioural results.

Figure 1

a, Study design and example trials. The study design manipulates i. the sensory context ‘# Sounds’: one or two flashes were presented with none, one or two sounds and ii. experimental paradigm: ‘yes-no SOA’, ‘yes-no threshold’ and ‘2IFC’ design. Example trials: In the ‘yes-no SOA’ and ‘yes-no threshold’ experiments, observers were presented within a single interval with either one or two flashes (together with zero, one or two sounds). They reported whether they perceived one or two flashes (followed by a confidence judgement that is not included in this report).

In the ‘yes-no SOA’ experiment, the SOAs between the two flashes (or two sounds) varied (i.e. 25, 42, 50, 58, 75, 108, 158, 225 ms) to enable the estimation of psychometric functions. In the ‘yes-no threshold’ experiment, the SOAs between the two flashes (or sounds) were titrated individually for each participant to yield approximately 50% “One flash” and 50% “two flash” responses for each of the following stimulus combinations: ‘2 flash + 0 sound’; ‘2 flash + 1 sound’; ‘1 flash + 2 sound’.

In the ‘2IFC’ experiment, observers were presented on each trial in one interval with a probe (e.g. ‘2 flash + x sound’) and in the other interval with a standard (e.g. ‘1 flash + x sound’). They reported which interval contained the probe (i.e. ‘2 flash’ stimulus). b, Alpha temporal resolution hypothesis16,18,30,48. Two sensory inputs are bound into one event if they fall within the same alpha cycle, but as two separate events, if they fall into different cycles. As a result, observers should be less sensitive at discriminating between one and two flashes for lower alpha frequency in purely visual perception. In audiovisual perception, audiovisual interactions or more specifically the influence of an incongruent number of sounds on the perceived number of flashes should be greater for lower alpha frequency leading to stronger crossmodal biases. Observers should be more likely to experience fusion illusions (i.e. perceive two flashes as one flash) when two flashes are paired with one sound and more fission illusions (i.e. perceive one flash as two flashes) when one flash is paired with two sounds. These audiovisual perceptual illusions introduce audiovisual biases and may reduce observers’ sensitivity to discriminate between one and two flashes for lower alpha frequency. c, Behavioural results. Perceptual sensitivity (d’ = z(HR) – z(FAR)) and bias (Biascentre= -0.5(z(HR) + z(FAR))) for ‘yes-no SOA’ and ‘yes-no threshold’ experiments for each auditory context (irrespective of alpha frequency). In the ‘yes-no threshold’ experiment differences in perceptual sensitivity and bias between sensory contexts cannot be interpreted, because the SOAs differed between no sound, one sound and two sound contexts. Error bars represent 1 SEM. SEM, Standard error of the mean after Bonferroni-Holm correction (per task for d’ or Biascentre). Detailed statistics and effect sizes are in Supplementary Table 3.