Progress in Neurobiology

Value-based decision-making is a dynamic and idiosyncratic process which requires appraising the value of options and comparing the values to select the best choice. An emerging view of orbitofrontal cortex (OFC) is that it achieves this by representing values serially during deliberation, alternating back-and-forth between transient states that encode the value of different options. While the time spent in each value state is known to reflect choice behavior, the source of the alternating dynamics remains unclear. One possibility is that fluctuations in value may be driven by attentional shifts between the choice options. Conversely, value dynamics in OFC may be generated locally, enabling OFC value signals to influence decision-making independently from attention. To test these hypotheses, we recorded from OFC and lateral prefrontal cortex (LPFC), a major attentional area in prefrontal cortex, to determine whether their population-level activity correlated in a manner consistent with crosstalk between neuronal systems involved in value and spatial attention. We found that OFC and LPFC selectively encoded option values and spatial locations, respectively, reflecting their specialized roles in cognition. Despite this functional dissociation, both OFC and LPFC dynamics were strongly affected by overt attention: which caused the value and spatial location of the fixated option to be represented at the same time. Additionally, fluctuations in the encoding strength of value in OFC and space in LPFC were temporally correlated above and beyond the effect of gaze, reflecting the effect of covert attention.

Working memory (WM) enables us to maintain and manipulate information over time, but how the brain organizes sequential information locally and across networks remains unclear. Recent work suggests that slow and fast theta oscillations serve different roles in memory, yet their distinct contributions to sequential WM are unknown. Based on evidence that the hippocampus (HC) and orbitofrontal cortex (OFC) support sequential WM and that slower theta cycles provide optimal temporal windows for organizing items in WM, we predicted that these regions would coordinate via slow theta dynamics. We analyzed intracranial EEG from the HC, OFC, and amygdala (AMY) in 21 neurosurgical patients (7 female, 13-54 years of age; M {+/-} SD, 30 {+/-} 11.2 years) performing a delayed match-to-sample WM task. We assessed phase locking between regions, phase-amplitude coupling within regions, and neuronal phase coding for slow (~1-4.5 Hz) and fast (~4.5-8 Hz) theta oscillations. We found significant slow and fast theta synchrony between all regions, but identical anatomical pathways produced opposing behavioral effects depending on oscillatory frequency, particularly during higher cognitive demand. Slow theta synchrony was associated with faster response times (RTs), while fast theta synchrony between HC and OFC hindered both accuracy and RTs. Unexpectedly, AMY modulated RT through demand-dependent slow theta synchrony, where AMY-OFC synchrony predicted faster RTs during maintenance and HC-AMY synchrony predicted faster RTs during higher cognitive demand. Sustained coupling between slow theta oscillations and high-frequency broadband activity within each region suggests that local organization coincides with beneficial network behavioral effects. These results establish a frequency-opponent mechanism in which theta oscillation frequencies determine whether HC-OFC circuits facilitate or impair sequential WM.

Given the limited capacity of working memory (WM), prioritization is essential for efficient information processing. Whether prioritization acts primarily at encoding, or dynamically shapes representations during maintenance, is currently unclear. Here, we employed a two-item delayed-match-to-sample task and compared prioritization conditions in which the testing order of items was either known in advance or not. Behaviorally, prioritization selectively reduced guess rates, without affecting precision. Using multivariate pattern analysis, we decoded stimulus information from EEG voltage and indexed internal attention using alpha-band patterns. Prioritization did not alter decodable representations during encoding. During maintenance, however, prioritization enhanced both voltage-based decodability and alpha power-based decodability for the currently prioritized item. Mediation analyses further indicated that alpha-based attentional signals influenced behavior indirectly, via voltage-based representational strength, which is consistent with the idea that internal attention supports performance by strengthening prioritized representations during memory maintenance. Significance StatementWM is capacity-limited, requiring the prioritization of information most relevant to current task demands. Whether prioritization is established at encoding or emerges during maintenance, and how it improves working memory performance, remains unclear. Comparing conditions with and without advance priority knowledge, we found that prioritization occurred primarily during maintenance rather than encoding. We also found that prioritization improved performance by directing internal attention to prioritized items, strengthening their neural representations and increasing their accessibility. This finding provides insight into the flexibility of working memory in the updating of already-encoded information.

The hippocampus has been proposed to support visual processing and perception, challenging longstanding accounts that emphasize navigation or declarative memory. A key prediction of visual-processing accounts is that the hippocampus should exhibit similar visuospatial coding properties to those of higher-order visual neocortical areas, such as sensitivity to the size of visual stimuli and contralateral visual field biases. We tested for these properties using intracranial EEG to measure hippocampal neural activity during a retinotopic mapping task. The hippocampus exhibited characteristic slow ([~]2 Hz) and fast ([~]8 Hz) theta oscillations throughout the task. Fast theta was responsive to the presence but not the amount of visual stimulation. In contrast, slow theta did not generally respond to stimulus presence but scaled with the size of the visual stimulus, consistent with larger receptive fields. Slow theta also showed a contralateral bias, an effect that was specific to the right hippocampus. None of these effects were attributable to microsaccades or performance of the concurrent vigilance task. These findings provide electrophysiological evidence for visual field coding by human hippocampus, supporting accounts of hippocampal function that emphasize its role atop the visual hierarchy. Visual processing of this kind may combine with self-motion, memory, and other signals to support the broader spatial and mnemonic functions with which hippocampal theta oscillations have long been associated.

Sudden and isolated sensory stimuli (SISS) engage the extralemniscal system and elicit widespread electrocortical responses in the brain. These responses, consisting of both time-domain transients and spectral changes, reflect a switch of the global brain state that likely prepares the organism for subsequent urgent behaviours. Crucially, SISS also elicit a short-latency phasic response in a key component of the extralemniscal system in the brainstem, the noradrenergic Locus Coeruleus (LC) nucleus. Such stimulus-evoked LC firing is associated with the electrocortical markers of extralemniscal activation. LC neurons also display burst-like firing spontaneously, i.e., without imposed sensory stimuli, for example, during quiet wakefulness, sleep, or anaesthesia. However, this phenomenon remains underexplored. We therefore measured, in freely behaving rats, the prefrontal electrocorticogram (ECoG) responses following spontaneous LC bursts. In addition, we compared these ECoG responses to those triggered by electrical LC stimulation or auditory SISS. We found that ECoG responses were proportional to the magnitude of the spontaneous LC bursts or microstimulation, and remarkably similar to those elicited by SISS. Finally, suppression of noradrenergic transmission with systemic clonidine administration attenuated the auditory-evoked ECoG response. These results suggest that LC plays a role in generating the transient brain state changes elicited by SISS.

When humans learn under conditions of uncertainty, they dynamically adjust how they prepare for and respond to feedback. In navigating uncertain environments, the brain minimizes error by continuously refining internal models via memory updating (MU). Feedback is critical for MU, and anticipatory neural mechanisms shape how feedback is processed, likely reflecting learned environmental certainty. However, the literature has largely focused on post-feedback activity, leaving pre-feedback certainty-related mechanisms less understood. The present study aims to address this gap by examining how certainty modulates anticipatory states, preceding feedback and subsequent MU. We examined oscillatory activity prior to performance feedback in a reanalysis of EEG data previously published by Hassall and colleagues (2023). Twenty-one participants (16 female, Mage = 25.81 years) predicted the strength of cartoon characters with varying predictability levels which were learned through exposure. Feedback on prediction accuracy was presented via an animated rising bar. Results revealed that theta power is modulated by accumulative feedback. Linear mixed-effects models revealed an interaction between predictability-related certainty and learning stage: in late learning, higher performance was associated with increased pre-feedback alpha and beta power for low-certainty trials, whereas in early learning, higher performance was associated with decreased beta power. These learning-related modulations in alpha and beta power suggest that initial learning is marked by adaptable exploratory processing. Subsequent learning exhibited increased alpha-mediated inhibition and beta-related anticipatory activity for lower certainty trials, indicative of dynamic strategy refinement and selective engagement of task-relevant information. These results demonstrate that certainty shapes preparatory oscillatory activity associated with MU.

The medial temporal lobe (MTL) houses specialised cell types supporting spatial representation in the human brain. Among these are grid cells that encode the location of the navigator, displaying a geometrically structured anchoring to the environment. While macroscopic grid-cell-like coding in humans is typically investigated using functional magnetic resonance imaging or invasive electrophysiological methods in clinical populations, we demonstrate the feasibility of using non-invasive scalp electroencephalography (EEG) to capture the characteristic six-fold modulation of high-frequency activity source-localised to the MTL. We found hexadirectional modulations of the low- and high-gamma band activity by movement direction in virtual reality. Furthermore, individual preference to use an allocentric reference frame was linked to higher hexadirectional modulation in the low gamma band and encoding a location along the putative grid axes of high gamma band predicted faster retrieval of the remembered location. The use of scalp EEG to capture grid-cell-like activity and its functional relevance sets the foundation for investigating the MTL activity in richer experimental contexts.

The ability of the brain to encode information and control behavior depends on the coordinated activity of large and distributed neuronal populations. Correlations in neuronal spiking activity across trials of the same condition, or noise correlations (NCs), have been interpreted as a reflection of shared synaptic connectivity and as a contributing factor to the information capacity of neuronal populations. The impact of NCs on coding is most often considered in populations of neurons exhibiting robust condition-dependent information in their spike counts (SCs). However, theoretical work suggests that NCs could provide a source of condition-dependent information separate from SCs. We examined the activity of large neuronal populations in prefrontal cortex of macaques while they performed a spatial delayed response task composed of visual, memory, and motor epochs. We found that pairs of neurons that displayed visual, memory, and motor selectivity in their SCs often exhibited selectivity in their NCs, independent of spike count. However, we also found that pairs of neurons without SC selectivity during the different task epochs nonetheless exhibited condition-dependent NCs. Moreover, we found that the magnitude of condition-dependent NCs were largely comparable across neuronal pairs with or without SC selectivity. These results demonstrate that correlated variability in spiking activity can be condition-dependent even in the absence of condition-dependent SCs.

Value-based decision making emerges from coordinated neural dynamics across distributed brain networks. Recent studies using noninvasive whole-brain measurements in humans have highlighted the importance of neural activity in the 2-10 Hz frequency band for value-based decision making. Using magnetoencephalography and hidden Markov model (HMM) analysis, we examined whether and how whole-brain neural dynamics in this frequency band, evolving on a timescale of a few hundred milliseconds, reflect value-based decision processes. Thirty-five healthy adults (females and males) made binary choices between risky and sure options. Trial-wise subjective values were estimated using behavioral economic modeling based on prospect theory. We found that HMM-derived trial-by-trial whole-brain neural dynamics (defined by 2-10 Hz amplitude envelopes in distributed brain regions and their interregional coupling) were associated with the subjective values of choice options in a manner distinct from simple perceptual- or motor-evoked activity. Notably, these trial-by-trial whole-brain dynamics covaried with the difference in subjective values between the chosen and unchosen options when the neural data were time-locked to participants responses, but not when time-locked to option onset. These findings revealed a crucial link between subsecond whole-brain neural dynamics and trial-by-trial decision variables, providing insights into how value-based decision processes unfold over time in the human brain.

Homeostatic plasticity of intrinsic excitability (IE) in the visual system has been essentially shown at the cortical level but whether thalamic nuclei also express homeostatic plasticity of IE is unknown. We show here that 4 days of monocular deprivation (MD) at eye opening induces a homeostatic change in IE in dorsal lateral geniculate nucleus (dLGN) neurons. Neurons recorded in the dLGN region activated by the deprived eye are more excitable than neurons recorded in the dLGN region activated by the open eye. No significant changes were observed following 7 days of MD, however. Enhanced excitability in neurons from the deprived side after 4 days of MD was associated with a reduced Kv1-dependent LTP-IE, a smaller voltage ramp, and a reduced inter-spike interval, suggesting that Kv1 channels are down-regulated in deprived dLGN neurons. Furthermore, the ankyrin G signal of the axon initial segment was larger in deprived dLGN neurons compared with open ones, indicating that Nav1 channel number also undergoes homeostatic regulation, and Kv1.1 channel signals were lower in deprived neurons compared to open ones. In addition, electrical coupling was found to be strengthened in neurons displaying enhanced IE following either brief (4 days) or long (10 days) MD. These results suggest that homeostatic and Hebbian plasticity in the dLGN share common expression mechanisms involving the regulation of Kv1 channels, Nav1 channels and electrical coupling between relay neurons.

CNS oligodendrocytes generate myelin, an RNA-containing proteolipid substance that enhances axonal transmission. In multiple sclerosis (MS), myelin debris is phagocytosed by microglia (MG), and prior studies have detected myelin-derived mRNA in MG nuclei, suggesting a retrograde transport pathway. We report myelin basic protein (MBP) is a nucleic acid-binding and trafficking protein. We found that retro-transport of myelin RNA into the MG nucleus was phagocytosis and importin-dependent. Transcriptomic and proteomic analyses of MG nuclei revealed enrichment of myelin mRNAs and proteins, with MBP singularly detected in soluble and chromatin-associated fractions. MBP bound mRNA with high affinity (Kd {approx} 0.30 nM) and was sufficient to facilitate MG RNA nuclear import in vitro and in vivo. Functionally, MBP mediated the delivery of small interfering RNAs for targeted knockdown of toll-like receptor 4. These findings indicate MBP as an RNA-binding protein capable of MG nuclear import, providing insight into neuroinflammatory pathology of MS.

Although remyelination, a central nervous system (CNS) regenerative process mediated by oligodendrocyte progenitor cells (OPCs), takes place in an inflammatory environment the long-term impact of inflammation on OPC remyelination capacity remains unclear. Here, we studied the short- and long-term impact of systemic inflammation on adult OPCs to assess whether transient inflammation triggers enduring chromatin remodelling indicative of inflammatory memory in OPCs. We observed long-lasting epigenetic modifications in response to both lipopolyssaccharide (LPS) and polyinosinic:polycytidylic acid (Poly(I:C)), but only LPS induced a tolerance-like memory. LPS-mediated tolerance-like memory enhanced OPC differentiation after demyelination in aged mice, reducing axonal damage. Our findings reveal OPC epigenetic memory of inflammation as a mechanism by which adult OPCs adapt to inflammatory challenges, which could be harnessed to reduce neuroinflammation and enhance remyelination efficiency in ageing and neurodegenerative diseases.

Dopaminergic neurons in the substantia nigra depend on iron for dopamine synthesis but are vulnerable to iron-induced oxidative stress. Many of these neurons synthesize neuromelanin, an iron-chelating pigment that accumulates across the lifespan and makes them vulnerable in Parkinsons disease. It remains unclear whether their selective vulnerability arises from neuromelanin overload or from the release of toxic labile iron from the oversaturated pigment. We quantified iron and neuromelanin at the single-cell level across the lifespan of chimpanzees, a species closely resembling humans in pigment and iron accumulation. Combining quantitative MRI, X-ray fluorescence imaging, and microscopic colorimetry, we found that the iron-to-neuromelanin ratio remains stable with age across large neuronal populations. Chemical equilibrium modeling of the iron binding in neuromelanin indicated that cytosolic labile iron concentrations remain low throughout adulthood. We have found no evidence for neuromelanin saturation or increased iron-mediated toxicity with age. This finding challenges the hypothesis that neuromelanin saturation drives age-related dopaminergic vulnerability. The presented method provides a quantitative framework for studying iron homeostasis in these neurons.

Dendritic arbor morphology is shaped in part by interactions with neighboring dendrites, and its geometry strongly influences the spatial distribution and strength of synapses. These observations raise the possibility that local dendritic contacts help determine where synapses accumulate and strengthen. Previous work in cultured hippocampal neurons showed that dendrite-dendrite contact sites are non-random and associated with local synaptic clustering. Here we asked whether a different type of dendritic contact, formed between a dendrite and the soma of a neighboring neuron, behaves similarly. Using dissociated hippocampal cultures, immunofluorescence imaging, time-lapse microscopy, quantitative image analysis, stochastic spatial simulations, and minimal quantitative modeling, we identified three recurrent classes of dendrite-soma interactions (DSIs): dendrites crossing directly over a neighboring soma, growing tangentially along the soma perimeter, or contacting the proximal region where a neighboring dendrite emerges from the soma. These interactions were abundant, occurred exclusively between different neurons, and showed substantial structural persistence over several days. Their overall frequency exceeded stochastic predictions across culture densities, and two configurations - proximal and tangential contacts - were selectively enriched above random expectation, whereas soma-crossing contacts were largely consistent with stochastic overlap. DSI composition also changed over development, with proximal contacts becoming progressively more prevalent. At DSI sites, synaptophysin-positive puncta were significantly denser and more intense than on non-interacting dendritic segments, consistent with local enrichment and strengthening of presynaptic specializations. Minimal modeling further indicated that biased formation together with developmental stabilization explains the observed organization better than stochastic geometry alone. These findings identify DSIs as non-random structural motifs in cultured hippocampal networks and suggest that dendrite contact geometry can contribute to synaptic distribution and strengthening.

BackgroundMicroglial heterogeneity is a fundamental feature of brain homeostasis and pathology. The purpose of this study was to investigate the complexity of microglial plasticity by characterizing specialized oligodendrocyte-like microglial subsets. MethodsThe study was performed utilizing single-cell transcriptomics analyses and immunofluorescence staining to identify and profile microglial subpopulations. Additionally, spatial transferring and morphological analyses were conducted to determine the anatomical distribution and structural features of these specific cells. ResultsWe identified a distinct microglial subset termed dual-phenotype microglia (DPM), which co-expresses microglial and oligodendrocyte markers. DPM consisted of two subtypes with distinct functions: myelin-associated DPM (mDPM) and neuron-associated DPM (nDPM). Spatial and morphological evaluations revealed that mDPMs were sparsely distributed across the whole brain and exhibited a highly ramified architecture, whereas nDPMs were enriched in the hippocampal dentate gyrus. Mechanistically, we found that mDPM function was driven by the Sox10 regulon to modulate myelin maintenance and axonal ensheathment, while nDPM was orchestrated by Glis2, facilitating essential neuron-glia crosstalk and synaptic regulation. Furthermore, we demonstrated that nDPM and mDPM were predicted to undergo significant alterations in multiple sclerosis and Alzheimers disease. Notably, mDPMs were selectively enriched in active multiple sclerosis lesions, revealing that DPM were closely related to neuropsychiatric disorders. ConclusionsBy comprehensively characterizing the morphology, molecular signatures, and spatial logic of these oligodendrocyte-like microglial subsets, our study elucidated the complexity of microglial plasticity. These findings provided new insights into their diverse roles in central nervous system health and disease. Graphical abstractIdentification, Molecular Profiling, and Functional Modeling of Dual-Phenotype Microglia (DPM). (1) Discovery: Identification of the dual-phenotype microglia (DPM) population through single-cell transcriptomics. (2) Molecular Signatures: The transcriptomic identity of DPM subtypes is governed by specific regulatory networks. (3) Distribution & Pathology: Spatial mapping reveals divergent anatomical logic and disease relations for DPM subtypes. (4) Mechanism/Theory: A proposed functional model of mDPMs as "metabolic relay" and support units. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=113 SRC="FIGDIR/small/724239v2_ufig1.gif" ALT="Figure 1"> View larger version (39K): org.highwire.dtl.DTLVardef@b7db1dorg.highwire.dtl.DTLVardef@9265e7org.highwire.dtl.DTLVardef@1605d82org.highwire.dtl.DTLVardef@19b048f_HPS_FORMAT_FIGEXP M_FIG C_FIG

Focusing on an organisms task at hand is instrumental for intelligent and goal-driven behavior. However, humans and other animals often fail to pay sustained attention across long time intervals. Failing to stay on-task may cause one to miss crucial task-relevant signals, leading to impaired performance, which can have serious consequences. Therefore, it is important to understand the neural basis of attentional lapses. One promising neural marker of attentional lapses is the frontal P3 (fP3) EEG component, which has been suggested to reflect the susceptibility to incoming sensory input. Following this, we hypothesized that the fP3 1) predicts imminent lapses of attention, and 2) that it should predict upcoming lapses of attention across modalities. In two experiments, we found that the fP3 reliably tracked lapses of attention of sustained attention already seconds preceding the crucial visual signal. We further extended this to the auditory domain: Already 1.5s ahead of the incoming auditory target, the fP3 revealed whether that target was detected or not. Detailed topographic analyses did, however, reveal a slight dissociation between modalities in underlying intracranial source configurations. In sum, this work revealed a supramodal neural signature of susceptibility, which tracks lapses of sustained attention seconds ahead of the critical incoming sensory input.

Strikingly, highly trained athletes engaged in vertiginous activities (e.g., dance and slacklining) and patients with bilateral vestibular loss show a similar pattern of neural plasticity, likely resulting from reduced vestibular sensory processes. However, unlike patients, these athletes show no balance impairments, quite the opposite. This suggests that the attenuation of vestibular processing represents an adaptive recalibration to excessive vestibular stimulation rather than a sign of dysfunction. Concurrently, tactile processing increases as vestibular processing attenuates. Our findings indicate that effective adaptation extends beyond simple tactile compensation: it involves a strengthened tactile-brain pathway. Indeed, following unexpected base-of-support translations, the coupling between plantar shear forces (i.e., a proxy of plantar sole tactile afferents) and cortical responses over the somatosensory areas was markedly enhanced in Athletes. Cross-correlation analysis revealed stronger (r = 0.71) and faster (36 ms) tactile-brain coupling in Athletes (n = 25) compared with age- and gender-matched Controls (n = 18). This enhancement occurred within the first 180 ms following translation, that is, during the critical early phase of skin-surface interaction. Notably, artistic swimmers, who undergo intense vestibular stimulation in a weightless underwater environment without balance equilibrium constraints, also exhibit enhanced tactile-brain coupling. This suggests that strengthening the tactile-brain coupling is not merely a byproduct of balance expertise, but rather a broader adaptive response to sustained vestibular stimulation. Multimodal neurons integrating vestibular and somatosensory inputs, such as those in the somatosensory cortex and thalamus, may increase their responsiveness to foot tactile afferents when vestibular inputs become excessive. In such contexts, the somatosensory system may assume a dominant role in providing gravity-related information for balance control.

Rhythmic temporal structure improves working memory, but how this benefit emerges from recurrent dynamics remains unclear. Here, we trained excitatory-inhibitory recurrent neural networks with short-term synaptic plasticity to perform a sequential delayed match-to-sample task with either regular or jittered sample timing. Rhythmic input produced a small but reliable improvement in task accuracy and was associated with more differentiated population trajectories during encoding. This behavioral advantage was accompanied by an organization of population dynamics around the dominant input frequency: temporal regularity progressively brought stimulus arrivals closer to preferred encoding phases, modulated phase advancement during stimulus presentation, and reduced the deviation of inter-stimulus phase-progression frequency from the dominant input rhythm. As a result, internal oscillations increasingly tracked the temporal structure of the input across the sequence, providing a phase-based scaffold for encoding ordered information. This scaffold preferentially supported temporal-order representations rather than uniformly enhancing all stimulus features. Decoding analyses further showed that stronger temporal regularity increased the fidelity and persistence of stimulus information in both neuronal activity and synaptic efficacy, whereas perturbing synaptic efficacy produced the largest impairment during delay-period maintenance. These findings suggest that rhythmic input supports sequential working memory by imposing a reliable temporal structure on recurrent dynamics and stabilizing synaptic-state representations.

Interlimb skill generalization, defined as the transfer of a newly learned skill from the trained to the untrained limb, represents a fundamental aspect of human motor behavior with significant implications for rehabilitation and athletic training. Skill generalization is influenced by processes that drive learning and interact with the newly acquired memory. For instance, in our recent work, we reported that performing a secondary, cognitively demanding task immediately after a short skill-training session impaired skill generalization when the untrained arm was tested 24-hour later. This suggests that working memory (WM) interacts with the early stage of skill memory consolidation processes and thereby impacts skill generalization. Motivated by this finding, in the current study, we investigate how WM interacts with reactivated skill memory and its subsequent impact on skill generalization, tested 24 or 48-hour post skill training. We recruited right-handed young participants (n=95) who performed a fast, accurate reaching task with their dominant right arm during a short training session (50 trials) on Day-1. After 24-hour on Day-2, depending on the group type, participants had a brief skill reactivation session (10 trials or no reactivation) and then performed the WM task (or a control task) with their right arm. Interlimb generalization to the untrained left arm was assessed either immediately after the WM/control task on Day-2 or after a 24-hour gap on Day-3. We found that, engaging in the WM task (compared to the control task) after skill reactivation on Day-2 enhanced immediate generalization. Conversely, when generalization was tested 24-hour later on Day-3, the same WM engagement impaired skill generalization. These findings demonstrate that WM engagement during the post-reactivation phase has a time-dependent influence on interlimb generalization. WM can facilitate immediate generalization, possibly by sustaining neural processes that promote skill memory generalization across effectors. However, when a 24-hour time gap is introduced, generalization is disrupted following WM engagement, possibly because of interference between underlying neural processes involved in WM and reactivation-induced (re)consolidation of the skill memory. This study highlights the delicate interplay among WM, motor memory reactivation dynamics, and skill generalization and suggests a time-dependent interplay of neural processes critical for optimizing outcomes in motor learning and clinical rehabilitation protocols.

Cortical brain regions integrate information across different timescales, ranging from fast sensory processing to longer integration windows, allowing cognitive functions like working memory. At the large-scale, brain regions organize into transient network states that rapidly switch over time and similarly contribute to cognition. Both cortical timescales and large-scale network dynamics are proposed to be determined by the balance between recurrent synaptic excitation and GABAergic inhibition. Here, we pharmacologically manipulated synaptic transmission at GABAA and NMDA receptors in 60 healthy male participants and acquired resting-state magnetoencephalography. Neuronal timescales followed a hierarchical gradient with shorter timescales in early sensory regions. Increasing GABAergic activity prolonged neuronal timescales across cortical regions. This effect was most prominent in the frontal default mode and in the dorsal attention network. Notably, dynamic network analyses revealed that the occurrence probability of the frontal default mode network increased, whereas the occurrence of the dorsal attention network was reduced. NMDA receptor modulation resulted in no significant changes. Together, these findings provide causal evidence that GABAergic inhibition is a key regulator of cortical temporal organization, linking microscale synaptic mechanisms to neuronal timescales and network dynamics that support diverse cognitive function.