Methods

Participants

One healthy adult will be recruited and will provide written informed consent before participation. The participant will report normal hearing and no neurological or psychiatric diagnoses. Vision will be fully occluded for the duration of the task by preventing any view of the stimulus monitor (see Apparatus); no other sensory impairments are expected.

Apparatus and Stimuli

Stimuli are two geometric targets: a star and a square. Targets are rendered in identical size and luminance on a computer display (black background; white foreground) controlled by a custom script that also handles randomization and logging. To preclude visual access, the display is placed inside an opaque shroud (monitor hood with a blackout curtain) oriented away from the participant; no reflections or secondary screens are visible from the participant’s position.

Auditory events are delivered over closed-back, over-ear headphones. The only task-relevant sound is a brief auditory cue (“beep”) indicating the start of the response window; no other auditory information about the stimulus is presented. A calibrated USB microphone records verbal responses and timestamps them for trial-level alignment. The experimenter’s console shows trial timing only (no stimulus identity) to maintain blinding. All stimulus identities, randomization seeds, and timestamps are written to an append-only log.

Procedure

Sessions are conducted in a quiet room with constant lighting and temperature. The participant is seated ~60–80 cm from the shrouded display. Each trial proceeds as follows:

1. Stimulus selection and presentation. The program samples the target (star vs. square) using computer-generated randomization and displays it on the covered screen for a fixed duration (not visible to the participant).
2. Response cue. After the display period, a brief tone is played over headphones to signal “guess now.”
3. Verbal report. The participant states the target aloud (“star” or “square”). No button presses are used. The microphone captures the utterance; an experimenter, blind to the stimulus, codes the response in real time (or from the audio recording immediately after the session).
4. Inter-trial interval. A short pause minimizes carryover effects.

Trials are run in blocks of ~25 to limit fatigue, with short breaks between blocks. A single session comprises 70 trials, with independent randomization on each trial. Missed or ambiguous responses (e.g., no speech within the response window) are logged and treated as incorrect unless pre-specified otherwise.

Experimental Design and Statistical Analysis

An a priori power analysis (one-tailed binomial test, α = 0.05) indicated that 70 trials provide ~80% power to detect accuracy of 65% versus chance (50%). Each trial is modeled as a Bernoulli outcome (correct/incorrect) under H₀: p = 0.5 (random guessing) and H₁: p > 0.5. After the session, the total number of correct identifications is evaluated with a one-tailed binomial test. For this design, performance of ≥45/70 correct is statistically significant at p < 0.05.

To document blinding and integrity, we save the randomization seed, trial-by-trial targets, onset times, and coded responses. Any deviations (e.g., discarded trials) are reported with reasons.

Replication and Quality Control

If accuracy exceeds the significance threshold, the full protocol is repeated on a separate day under identical conditions. Replication uses a new randomization seed and preserves blinding. All software (versioned), randomization logs, and audio recordings are archived to support audit and re-analysis. This configuration eliminates conventional visual access to the display and minimizes experimenter cueing; therefore, consistent above-chance performance would constitute evidence for an ability exceeding chance inference given the information available to the participant.