Journal of Cognitive Neuroscience最新文献

Cognitive fatigue, a key contributor to failures in high-stakes domains, is poorly understood due to imprecise definitions, inconsistent protocols, and neglect of working memory (WM) mechanisms. We propose that active fatigue, arising from sustained cognitive demands, should be studied through WM frameworks, distinguishing it from passive (low arousal) fatigue. Contemporary WM models identify theta/alpha-gamma oscillatory dynamics as fundamental to WM function, plausibly providing testable markers of fatigue-induced breakdown. Conceptualizing active fatigue as specifically a disruption of WM's oscillatory dynamics provides a framework for the precise identification of its core neurophysiological basis. Specifically, tracking destabilized theta/alpha-gamma coupling and frequency synchrony provides a direct link to observed performance declines, enabling targeted rhythm-specific interventions such as frequency-matched brain stimulation. Current induction tasks rarely sustain optimal cognitive difficulty and are confounded by learning effects, prompting us to develop WAND (working-memory adaptive-fatigue with n-back difficulty), an open-source adaptive fatigue induction n-back suite. WAND reduces learning effects, classifies participant performance, and maintains task performance in the "optimal challenge zone"; optional distractor probes and multimodal logging enable robust mechanistic analyses. This approach shifts the field toward mechanistic, intervention-ready insights, enhancing fatigue detection and mitigation through theoretically grounded neural markers and standardized induction protocols.

Although listeners can enhance perception by using prior knowledge to predict the content of degraded speech signals, this process can also elicit "misperceptions." The neurobiological mechanisms responsible for these phenomena remain a topic of debate. There is relatively consistent evidence for involvement of the bilateral posterior superior temporal gyri (pSTG) in speech perception in noise; however, a role for the left premotor cortex (PMC) is debated. In this study, we employed transcranial magnetic stimulation (TMS) and a prime-probe paradigm for the first time to investigate causal roles for the left PMC and pSTG in speech perception and misperception. To produce misperceptions, we created partially mismatched pseudosentence probes via homophonic nonword transformations (e.g., She moved into her apartment soon after signing the lease-Che moffed inso har apachment sool amter siphing tha leals). All probe sentences were then spectrotemporally degraded and preceded by a clear prime sentence. Compared with a control site (vertex), inhibitory stimulation of the left pSTG selectively disrupted priming of real but not pseudosentences. However, inhibitory stimulation of the left PMC did not significantly influence perception of either real sentences or misperceptions of pseudosentences. These results confirm a role for the left pSTG in the perception of degraded speech. However, they do not support a role for the left PMC in either lexical or sublexical processing during perception of degraded speech using ecologically valid sentence stimuli. We discuss the implications of these findings for neurobiological models of speech perception.

Working memory (WM) involves continuous and dynamic processes, including encoding, maintenance, and retrieval. While many studies have focused on the maintenance of WM information, encoding strategies also impact WM performance and can be shaped by the presentation format of stimuli. However, how presentation formats modulate neural responses across WM stages remains unclear. To address this issue, we conducted an EEG study examining the effects of presentation formats (simultaneous, location-sequential, and center-sequential presentation) and WM loads (one and three abstract shapes). Behavioral results showed longer RTs for the location-sequential than for the center-sequential format. Additionally, the recency effects observed in both sequential conditions reflect the influence of ordinal information. EEG results revealed distinct load-dependent alpha activity patterns across presentation formats during WM maintenance. Simultaneous presentations exhibited a persistent decrease in alpha power, whereas both sequential presentations exhibited an initial decrease followed by a subsequent increase. During sequential encoding, alpha power decreased cumulatively with each additional item in the location-sequential format, but not in the center-sequential format. At retrieval, the probe elicited a load-dependent negative potential (i.e., the N3rs) across all formats. The N3rs load modulation was stronger for simultaneous presentations than sequential ones and was more pronounced for earlier positions than for the last position in sequential presentations. In conclusion, our findings demonstrate that the spatial and temporal order information embedded in presentation formats modulates load-dependent neural responses across WM stages. These effects extend beyond maintenance to encoding and retrieval, highlighting the influence of presentation formats on WM neural dynamics.

Harmonic expectation is an important mediator of musical experience. EEG research has identified event-related potential (ERP) components associated with expectation, including the early (right) anterior negativity (E(R)AN), which is theorized to index harmonic surprisal with reference to long-term memory of the statistical structure of music. However, the role of top-down influences on harmonic predictions remains underexplored. One specific influence concerns how a given harmony can be interpreted in different ways, depending on its syntactic role in a musical context. We present data from a novel paradigm that cues listeners to the syntactic structure of the stimuli (but not whether they contain improbable events). Our main result revealed larger E(R)AN amplitudes for improbable chords when listeners knew that additional context would follow a surprising harmony; P3a and P600 amplitudes were also larger in such cases. Using the theoretical framework of predictive coding, we propose that, in such cases, listeners assign higher precision to their predictions, leading to larger prediction errors as indexed by the E(R)AN, P3a, and P600 ERP components, and that prior context alone does not fully explain how unpredictable events are processed. Musical surprisal arises from a dynamic interplay between bottom-up cues and a listener's top-down anticipation within specific syntactic contexts.

In conversation, future speakers often plan speech simultaneously with comprehension, which means that they must divide attentional resources between these processes. In this EEG study, we used responses to linguistic attention probes (i.e., syllable "BA" presented during spoken sentences) to track temporal variations in attention to comprehension. Participants were asked to listen to prerecorded sentences with expected or unexpected sentence-final words. Each sentence was presented twice, once with and once without the attention probe starting 100 msec after the target word onset. Participants saw a picture 50 msec before the target word. Depending on the test block (picture naming or button press), participants either named the picture or pressed the space bar, both after an 850-msec delay. The probes elicited a negative potential approximately 100 msec after probe onset (i.e., an attention probe effect) in all probe conditions. Unexpectedly, neither word expectancy nor speech planning influenced the timing or strength of the attention probe effect. This indicates that expectancy of words in Dutch does not affect the allocation of attention toward these words 100 msec after their onset (i.e., the time of the probe presentation). Interestingly, engaging in speech planning does not seem to divert attentional resources away from comprehension at the moment of probe presentation. These findings imply that listeners are able to effectively distribute their attentional resources between comprehension and speech planning and carry out these processes at the same time. Considering these unexpected findings, using attention probes might not be the best approach to capture variations in temporal attention in dual-task paradigms.