Despite the growing interest in the study of attentional refreshing, the functioning of this working memory maintenance mechanism, including its cerebral underpinnings, is still debated. In particular, it remains unclear whether refreshing promotes long-term memory and whether it, in return, depends on long-term memory content to operate. Here, we used direct maintenance instructions and measured brain activity to investigate working memory maintenance with two objectives: (1) test if different behavioral and oscillatory patterns could be observed when participants were instructed to use attentional refreshing versus verbal rehearsal, and (2) observe whether and how refreshing is modulated when maintaining novel (pseudowords) versus familiar (words) memoranda. We conducted an EEG experiment using a modified Brown-Peterson task, in which we manipulated the type of maintenance engaged through explicit instructions (verbal rehearsal vs. refreshing), the type of memoranda (words vs. pseudowords), and the memory load (2 vs. 6). Using scalp EEG, we measured both neural oscillations during working memory maintenance and ERPs during the concurrent parity judgment task. For words, we showed that verbal rehearsal benefited more short-term recall whereas refreshing benefited more delayed recall. In keeping with these behavioral differences between maintenance instructions, frontal–midline theta power increased with memory load only when using verbal rehearsal, whereas occipito-parietal alpha desynchronization was larger with refreshing than verbal rehearsal. When maintaining pseudowords, verbal rehearsal also benefitted short-term recall more than refreshing. However, no long-term memory benefit of refreshing was observed for pseudowords, and oscillatory activity was not different under the two maintenance instructions. Our results provide new evidence supporting the independence between attentional refreshing and verbal rehearsal, and bring new insight into refreshing functioning. We discuss the possible interpretations of these results and the implications for the attentional refreshing literature.

Congenital amusia is a neurodevelopmental disorder characterized by difficulties in the perception and production of music, including the perception of consonance and dissonance, or the judgment of certain combinations of pitches as more pleasant than others. Two perceptual cues for dissonance are inharmonicity (the lack of a common fundamental frequency between components) and beating (amplitude fluctuations produced by close, interacting frequency components). Amusic individuals have previously been reported to be insensitive to inharmonicity, but to exhibit normal sensitivity to beats. In the present study, we measured adaptive discrimination thresholds in amusic participants and found elevated thresholds for both cues. We recorded EEG and measured the MMN in evoked potentials to consonance and dissonance deviants in an oddball paradigm. The amplitude of the MMN response was similar overall for amusic and control participants; however, in controls, there was a tendency toward larger MMNs for inharmonicity than for beating cues, whereas the opposite tendency was observed for the amusic participants. These findings suggest that initial encoding of consonance cues may be intact in amusia despite impaired behavioral performance, but that the relative weight of nonspectral (beating) cues may be increased for amusic individuals.

How are we able to easily and accurately recognize speech sounds despite the lack of acoustic invariance? One proposed solution is the existence of a neural representation of speech syllable perception that transcends its sensory properties. In the present fMRI study, we used two different audiovisual speech contexts both intended to identify brain areas whose levels of activation would be conditioned by the speech percept independent from its sensory source information. We exploited McGurk audiovisual fusion to obtain short oddball sequences of syllables that were either (a) acoustically different but perceived as similar or (b) acoustically identical but perceived as different. We reasoned that, if there is a single network of brain areas representing abstract speech perception, this network would show a reduction of activity when presented with syllables that are acoustically different but perceived as similar and an increase in activity when presented with syllables that are acoustically similar but perceived as distinct. Consistent with the long-standing idea that speech production areas may be involved in speech perception, we found that frontal areas were part of the neural network that showed reduced activity for sequences of perceptually similar syllables. Another network was revealed, however, when focusing on areas that exhibited increased activity for perceptually different but acoustically identical syllables. This alternative network included auditory areas but no left frontal activations. In addition, our findings point to the importance of subcortical structures much less often considered when addressing issues pertaining to perceptual representations.

How does the sleeping brain process external stimuli, and in particular, up to which extent does the sleeping brain detect and process modifications in its sensory environment? In order to address this issue, we investigated brain reactivity to simple auditory stimulations during sleep in young healthy subjects. Electroencephalogram signal was acquired continuously during a whole night of sleep while a classical oddball paradigm with duration deviance was applied. In all sleep stages, except Sleep Stage 4, a mismatch negativity (MMN) was unquestionably found in response to deviant tones, revealing for the first time preserved sensory memory processing during almost the whole night. Surprisingly, during Sleep Stage 2 and paradoxical sleep, both P3a-like and P3b-like components were identified after the MMN, whereas a P3a alone followed the MMN in wakefulness and in Sleep Stage 1. This totally new result suggests elaborated processing of external stimulation during sleep. We propose that the P3b-like response could be associated to an active processing of the deviant tone in the dream's consciousness.

Timbre characterizes the identity of a sound source. On psychoacoustic grounds, it has been described as a multidimensional perceptual attribute of complex sounds. Using Garner's interference paradigm, we found in a previous behavioral study that three timbral dimensions exhibited interactive processing. These timbral dimensions acoustically corresponded to attack time, spectral centroid, and spectrum fine structure. Here, using event-related potentials (ERPs), we sought neurophysiological correlates of the interactive processing of these dimensions of timbre. ERPs allowed us to dissociate several levels of interaction, at both early perceptual and late stimulus identification stages of processing. The cost of filtering out an irrelevant timbral dimension was accompanied by a late negative-going activity, whereas congruency effects between timbre dimensions were associated with interactions in both early sensory and late processing stages. ERPs also helped to determine the similarities and differences in the interactions displayed by the different pairs of timbre dimensions, revealing in particular variations in the latencies at which temporal and spectral timbre dimensions can interfere with the processing of another spectral timbre dimension.

Timbre is a multidimensional perceptual attribute of complex tones that characterizes the identity of a sound source. Our study explores the representation in auditory sensory memory of three timbre dimensions (acoustically related to attack time, spectral centroid, and spectrum fine structure), using the mismatch negativity (MMN) component of the auditory event-related potential. MMN is elicited by a discriminable change in a sound sequence and reflects the detection of the discrepancy between the current stimulus and traces in auditory sensory memory. The stimuli used in the present study were carefully controlled synthetic tones. MMNs were recorded after changes along each of the three timbre dimensions and their combinations. Additivity of unidimensional MMNs and dipole modeling results suggest partially separate MMN generators for different timbre dimensions, reflecting their mainly separate processing in auditory sensory memory. The results expand to timbre dimensions a property of separation of the representation in sensory memory that has already been reported between basic perceptual attributes (pitch, loudness, duration, and location) of sound sources.