Figure 1.

Sparse-sampling design. A, Timeline of events within a trial of overt singing. After audio presentation of two-tone intervals (2 s), participants matched the target tones with their singing voice (3 s). B, Timeline of events within an active baseline trial, in which participants only listened to the same two-tone intervals without subsequent overt reproduction. All trials were presented in blocks of six consecutive trials of the same trial type (sing, listen-only), pseudorandomized throughout the session, and interspersed with three silent (i.e., no stimulation) baseline blocks across the run. Trial onsets were randomly jittered over 1 s. A total of 36 singing trials, 36 listening-only trials, and 18 silent baseline trials were measured per condition (normal singing and anesthesia).