Neural signatures of automatic repetition detection in temporally regular and jittered acoustic sequences

Abstract

Detection of repeating patterns within continuous sound streams is crucial for efficient auditory perception. Previous studies demonstrated a remarkable sensitivity of the human auditory system to periodic repetitions in unfamiliar, meaningless sounds. Automatic repetition detection was reflected in different EEG markers, including sustained activity, neural synchronisation, and event-related responses to pattern occurrences. The current study investigated how listeners' attention and the temporal regularity of a sound modulate repetition perception, and how this influence is reflected in different EEG markers that were previously suggested to subserve dissociable functions. We reanalysed data of a previous study in which listeners were presented with sequences of unfamiliar artificial sounds that either contained repetitions of a certain sound segment or not. Repeating patterns occurred either regularly or with a temporal jitter within the sequences, and participants' attention was directed either towards the pattern repetitions or away from the auditory stimulation. Across both regular and jittered sequences during both attention and in-attention, pattern repetitions led to increased sustained activity throughout the sequence, evoked a characteristic positivity-negativity complex in the event-related potential, and enhanced inter-trial phase coherence of low-frequency oscillatory activity time-locked to repeating pattern onsets. While regularity only had a minor (if any) influence, attention significantly strengthened pattern repetition perception, which was consistently reflected in all three EEG markers. These findings suggest that the detection of pattern repetitions within continuous sounds relies on a flexible mechanism that is robust against in-attention and temporal irregularity, both of which typically occur in naturalistic listening situations. Yet, attention to the auditory input can enhance processing of repeating patterns and improve repetition detection.

Fig 1. Illustration of the study design.

A: Illustration of regular and jittered acoustic stimulus sequences with and without pattern repetitions. B: Experimental design. Participants took part in two EEG sessions. In the first session, their attention was directed away from the auditory stimulation, and in the second session, they were instructed to focus on repetitions in the sounds. Both sessions consisted of ten blocks in a random order, each of which contained 30 trials.

Fig 2. Sustained activity throughout the sequence.

Global field power (GFP) relative to the onset of the first pattern occurrence per sequence (0 ms) for rep and no-rep sequences in each of the four Attention x Regularity conditions. Bar plots display mean amplitudes in the time window of interest (500 to 3000 ms; marked by the light-grey box). Shaded areas and error bars indicate ± 1 SEM.

Fig 3. Event-related responses to repeating pattern onsets.

Event-related potentials (ERPs) relative to the onset of repeating patterns at position 2 to 6 within the sequence (0 ms) for rep and no-rep sequences in each of the four Attention x Regularity conditions. Bar plots display mean amplitudes in the time windows of interest (early positivity: 0 to 160 ms; later negativity: 190 to 380 ms; marked by light-grey boxes) at a fronto-central electrode cluster. Topographies show the rep-minus-no-rep difference potential in the respective time window of interest. Shaded areas and error bars indicate ± 1 SEM.

Fig 4. Phase coherence of neural oscillations.

Inter-trial phase coherence (ITPC) over frequencies and time relative to the onset of repeating patterns at position 2 to 6 within the sequence (0 ms) at a fronto-central electrode cluster for rep and no-rep sequences in each of the four Attention x Regularity conditions. Bar plots display mean ITPC between 1 and 4 Hz in a time window from 0 to 500 ms relative to pattern onset (marked by dotted lines). Error bars indicate ± 1 SEM.