Neural function is maintained through homeostatic mechanisms that are engaged following perturbations to the nervous system. Homeostatic plasticity is thought to be critical for establishing and stabilizing appropriate levels of network function. Neurons are proposed to detect deviations in activity through intracellular calcium signaling, such that changes in calcium levels initiate compensatory mechanisms that restore activity and calcium to baseline. This sensing process is generally assumed to occur in the cytoplasm, however, recent work suggests that it may reside inside mitochondria. We test this in the chick embryo (either sex) spinal cord. We show that perturbations known to induce homeostatic plasticity preferentially alter the mitochondrial proteome, including components of the tricarboxylic acid (TCA) cycle, a pathway sensitive to calcium entry into mitochondria. We then tested whether calcium influx into the mitochondrial matrix contributes to the induction of homeostatic plasticity in motoneurons. Pharmacological blockade of the mitochondrial calcium uniporter (MCU), which mediates calcium entry into the matrix, produced a robust and sustained increase in spontaneous network activity (SNA). Using Ru265 to inhibit MCU function, we confirmed a reduction in mitochondrial calcium, while cytoplasmic calcium levels were largely unchanged or slightly elevated. MCU blockade was accompanied by an increase in excitatory synaptic strength consistent with homeostatic synaptic plasticity. The underlying mechanisms overlapped with those previously described in this preparation following activity or neurotransmitter blockade. Together, these findings support a model in which mitochondria contribute to the initiation of homeostatic synaptic plasticity, potentially by sensing changes in calcium transients within the mitochondrial matrix.Significance Statement Homeostatic plasticity is thought to play a critical role in maintaining circuit function. Although substantial progress has been made in identifying the mechanisms underlying the expression of homeostatic plasticity, the upstream triggers remain poorly understood. Cytoplasmic calcium has been proposed as a key signal in the detection of perturbations in neural circuit activity and in initiating compensatory responses. Here, we present findings consistent with the idea that the sensor for network activity and homeostatic synaptic plasticity tracks mitochondrial calcium. Identifying the sensor that initiates homeostatic mechanisms will be essential for understanding the functional objectives of this form of plasticity and may provide a foundation for pharmacologically targeting this pathway in conditions characterized by altered network activity.
During audiovisual perception, spatial information from vision and audition is combined, often producing biases such as the ventriloquist effect. While these interactions are well documented behaviourally, it remains unclear when cross-modal information begins to alter modality-specific spatial representations in the brain. Here we used cross-generalised inverted encoding modelling of electroencephalography (EEG) data to track the temporal evolution of spatial representations during a spatial ventriloquist task. Human participants (both sexes) localised audiovisual stimuli with horizontally offset auditory and visual components. Decoders trained on unisensory EEG responses were applied to audiovisual trials to estimate whether and when spatial representations of one modality were biased toward the other. Behaviourally, participants integrated cues but showed slight visual over-weighting. Neural decoding revealed robust spatial representations for both modalities, with early unisensory encoding remaining unaffected by cross-modal input. Cross-modal biases emerged only later (from ∼200 ms onwards), within a generalised representational window, and occurred sequentially: auditory representations were biased earlier than visual. These findings indicate that multisensory spatial integration arises via recurrent feedback during later processing stages, while early sensory-specific representations remain independent, providing a neural basis for the temporal dynamics underlying the ventriloquist effect. Significance statement Understanding how the brain integrates spatial information across the senses is central to theories of perception, yet the timing of cross-modal influences on modality-specific neural representations remains unclear. Using EEG and cross-generalised inverted encoding models, we tracked auditory and visual spatial representations with millisecond precision. Early spatial representations remained unisensory despite synchronous audiovisual stimulation. A shared, generalised spatial representation emerged at ∼200 ms, after which cross-modal biases appeared-first in auditory and later in visual codes. These results show that spatial integration is implemented through late, feedback-driven processes operating on initially independent sensory estimates, resolving a long-standing discrepancy between early multisensory interactions and late spatial ventriloquism.
Although we increasingly understand how the brain processes phrase-level meaning, the role of semantic content remains underexplored. Recent models grounded in componential experiential features offer a principled framework for linking conceptual representation to the neurobiology of semantic composition. Here we used MEG to test whether experiential semantic features are retrieved and composed during two-word phrase comprehension, and whether lexical and phrase-level representations show temporally overlapping or strictly serial activation. Participants (16 women, 12 men) read two-word phrases and their component words in a rapid parallel visual presentation (RPVP) paradigm, equating the temporal dimension of the two types of stimuli. The phrases spanned six linguistic relations. All words and phrases were independently rated on 65 experiential features across 14 experiential domains and representational similarity analysis (RSA) was used to relate neural and semantic structure. Rather than emerging after or in parallel with single-word meanings, phrase-level meanings arose earlier, outpacing single words in isolation. Phrasal contexts also elicited earlier activation of word-level meaning: a word such as ‘cake’ showed semantic activation at ∼312 ms in isolation, but at ∼236 ms in ‘green cake.’ Experiential feature representations of both component words were detectable during phrase processing, with partially overlapping temporal windows. The magnitude of composition-related effects varied systematically across phrasal relation types and experiential domains. Together, these findings show that experiential semantic features are integrated during composition in ways that vary across phrasal relations and that combining words into phrases speeds semantic access, accelerating both phrase- and word-level meaning relative to isolated words. Significance Statement This study provides temporally resolved evidence that the brain simultaneously represents multiple levels of meaning during language comprehension. Using MEG and representational similarity analysis grounded in experiential semantic features, we show that compared to single words, two-word phrases elicit earlier and more sustained alignment between neural activity patterns and semantic feature representations. Further, the meanings of both constituent words remain detectable in overlapping time windows during phrase processing. Exploratory analyses further suggest that residual phrase-level semantic alignment varies with phrasal relations and semantic content. This work advances understanding of how distributed experiential knowledge is recruited during real-time language processing and how relational structure and semantic content jointly shape meaning composition.
Multiple sclerosis (MS) is a chronic inflammatory and neurodegenerative disease. While onset usually occurs in young adulthood, older age worsens MS prognosis with accelerated accumulation of disability and higher susceptibility to the progressive form of MS, characterized by continuous symptoms without periods of remission. Treatment responsiveness also declines with age, changing the landscape of available therapies for older patients. Recent work has begun to uncover an accelerated aging and senescence-like phenotype arising in patients with MS and its preclinical animal models. This review explores the current evidence for inflammatory injury driving age-related changes in brain cells in patients with MS and preclinical animal models, its interactions with sex and hormones, and the possible future for anti-aging therapeutic strategies to address this evolving neurodegenerative phenotype of MS during aging.
Motor sequence learning recruits a distributed sensorimotor network. Within this network, the premotor cortex (PMC) encodes and represents sequence information while the primary motor cortex (M1) executes movement sequences. Like M1, PMC neuronal populations exhibit mu (8-13 Hz) rhythms, which are typically non-sinusoidal. Previous studies showed that mu rhythm phase gates corticospinal transmission, sensitivity to LTP-like plasticity, and learning-related corticospinal plasticity. However, these studies treated mu peak and trough phases as discrete functional states and focused exclusively on M1. Here, we characterized mu phase-dependent mechanisms of motor sequence learning by measuring waveform shape, which is a more holistic measure that treats peak and trough phases as continuous components of the same oscillatory cycle. Using cycle-by-cycle analysis of resting EEG, we quantified the peak-trough symmetry of mu rhythms recorded over premotor regions during, before, and after healthy adults (25 females, 11 males) practiced a serial reaction time task (SRTT) containing a repeating, embedded sequence (sequence group) or no sequence (no-sequence group). As learning progressed, premotor mu rhythms became more symmetric in the sequence than the no-sequence group; these symmetry increases were also more positively correlated with skill acquisition in the sequence than the no-sequence group. Further, premotor mu asymmetry before task exposure predicted sequence acquisition, such that participants with longer baseline trough phases acquired greater skill and showed larger learning-related increases in mu peak-trough symmetry. Overall, these findings provide first evidence that motor sequence learning re-shapes premotor mu rhythms, consistent with learning-related redistribution of cortical excitability across the mu cycle. Significance statement Motor sequence learning requires coordinated activity between premotor and sensorimotor cortices. Although sensorimotor mu rhythm phase gates LTP-like corticospinal plasticity, the effects of TMS on motor learning, and learning-related plasticity, the role of premotor mu waveform shape in motor sequence learning is unknown. Using cycle-by-cycle analysis, we characterized mu peak-trough symmetry recorded over premotor regions before, during, and after healthy adults learned an implicit motor sequence or completed a control task. We found that premotor mu peak-trough symmetry progressively increased during learning, with baseline asymmetries and learning-related symmetry increases both predicting sequence learning. Our findings suggest that sequence learning redistributes excitability across the mu cycle and identify premotor mu rhythm waveform shape as a novel marker of motor sequence learning.
A habit develops as a motivated behavior that is strengthened by extended experience and feedback. While the performance component of a habit is relatively well studied, being linked to action-related neural dynamics in the basal ganglia and beyond, neural mechanisms for outcome feedback during habit formation, the reinforcement component, remain unknown. One candidate for this feedback mechanism is in the central nucleus of the amygdala (CeA). Here, we identify CeA neural firing dynamics in male and female rats that serve to promote habit formation. We used a novel maze task with rewards of differing identity and value to show that habits arise with task overtraining. After showing that overtraining engages CeA as shown through elevated cFos expression, we recorded in-vivo CeA activity across learning, and after outcome devaluation when habitual behavior is most identifiable. Neuronal activity changes tracked with habit formation. During learning, a group of recorded cells were significantly responsive at the choice point of maze trials while others encoded outcome consumption. By late training, neural activity exhibited a rapid depression at the choice point of the maze and excitation during reward receipt. In both cases, outcome magnitude, but not identity, was a significant modulator of neural activity. Additionally, a population of neurons tracked instantaneous changes in animal run speed. These speed cells dramatically decreased in number as habits formed. Together, these findings support a role for the CeA in providing reinforcement for habitual behavior, offering a signal that marks successful performance, while identifying speed representations in the CeA.Significance Statement Habits enable efficient behavior but can also become inflexibly maladaptive. Although the neural circuits underlying habitual action execution have been studied extensively, the mechanisms by which outcome feedback reinforces habit formation remain poorly understood. Here, we show that neural activity in the central nucleus of the amygdala evolves with habit development, shifting from representations of choice and reward consumption to reinforcement-related signals during overtraining. Activity reflected outcome magnitude rather than identity, consistent with a role for the central amygdala in reinforcement. We also identify a previously unrecognized population of CeA neurons that track animal running speed. Together, these findings implicate the CeA as a source of reinforcement signals that promote habit formation.
The formation of the glial sheath is essential in nervous system development to insulate and protect peripheral nerves and axons, yet the regulation of adhesion junctions in non-myelinating glia has not been clearly established. Many components of adhering junctions contain PDZ domains or are recruited by PDZ binding motifs. To identify PDZ domain proteins with roles in peripheral glial sheath formation, we carried out an RNAi screen using Drosophila melanogaster to knockdown each of the 66 predicted PDZ proteins in larvae of either sex. We identified six PDZ genes with potential roles in glial morphology, and we further characterized the role of Dlg5, a scaffolding protein with no previously known function in any glia. To further our investigation of Dlg5, we focused on cadherins and found both N-Cadherin and E-Cadherin are expressed throughout peripheral glia. Knockdown of E-Cadherin (ECad) phenocopied the loss of Dlg5 leading to gaps in the subperineurial glia and septate junctions, while only simultaneous loss of both N-Cadherins (NCad and CadN2) had the same effect. The loss of all three Cadherins enhanced these phenotypes as did loss of Dlg5 when paired with cadherin knockdown yet Dlg5 does not colocalize with Cadherins. The Hippo signaling pathway represents a common convergence point for both Dlg5 and cadherins and we found that expression of Yorkie is able to rescue the loss of Dlg5 and ECad in the SPG. This leads to a model where Dlg5 and cadherins function in the same pathway and play a role in glial membrane stabilization and septate junction formation.Significance Statement Glial ensheathment of peripheral nerves is critical to generate the insulating sheath and the blood-nerve barrier. We found that the PDZ domain protein Discs-large 5 (Dlg5) plays a key role in the developmental of the insulating glial sheath and does so in cooperation with Cadherins, both E- and N-Cadherins. Dlg5 and Cadherins converge to regulate the Hippo pathway and thus glial growth and blood-brain barrier formation.
Cholesterol-rich lipid rafts are organizing platforms for many excitatory ion channels and receptors in neurons and for inflammatory receptors like toll-like receptor 4 (TLR4) in immune cells. Human dorsal root ganglion (hDRG) neurons that express nociceptor markers also express high levels of TLR4. Apolipoprotein A-I binding protein (AIBP) binds to TLR4 and promotes cholesterol depletion and lipid raft disruption in cells where TLR4 is expressed. The current study conducted with male and female hDRG demonstrates that TLR4-lipid rafts are expressed with higher density in pain- versus non-pain-associated tissues and localized to neurons expressing transient receptor potential vanilloid 1 (TRPV1) typically expressed by nociceptors. This increased density of TLR4-lipid rafts results in increased excitability of hDRG TRPV1 receptors. Additionally, whole cell recordings show that disruption of TLR4-rafts with AIBP inhibited pain-associated ectopic action potential generation and evoked neuronal excitation in hDRG neurons. AIBP treatment impacted several action potential dynamics specifically in cells with on-going ectopic action potentials rather than non-spontaneously active neurons. Finally, AIBP is shown to be localized with satellite cells in human DRG and its expression increases along with glutamine synthetase in pain-associated hDRG. Given that TLR4-raft expression in DRG neurons is largely confined to nociceptors, a TLR4-raft specific therapeutic such as AIBP may have potential as a novel pain therapeutic with a low-side effect profile.Significance statement Lipid rafts are critical regulators of numerous neurological processes and disorders and are shown here to increase in density in pain-associated human dorsal root ganglia. Selective disruption of lipid rafts in toll-like receptor 4-expressing nociceptors significantly reduced the excitability of these neurons and suppressed ectopic, pain-associated action potential discharges. The agent used to disrupt lipid rafts, apolipoprotein A-I binding protein, localizes to macrophages and satellite glial cells and engages an endogenous pain relief mechanism with an expected limited side effect profile. This work reveals a novel mechanism underlying the generation of chronic pain in patients and identifies a potential therapeutic strategy for its management.
The cochlear amplifier, mediated by outer hair cell-driven nonlinear mechanics, enables the remarkable sensitivity and dynamic range of human hearing. Cochlear compression is a defining feature of this active process and is known to deteriorate with hearing loss. Among individuals with normal audiograms, cochlear compression is generally presumed to be invariant, an assumption that has not been systematically examined. If incipient cochlear changes precede measurable threshold shifts, variability in compressive nonlinearity may represent an early peripheral correlate of emerging dysfunction and a potential mechanistic contributor to auditory processing differences. Whether cochlear compression varies meaningfully among individuals with normal audiograms, and what such variability reveals about the earliest stages of cochlear dysfunction, remains unknown. Here, we quantified compressive nonlinearity using distortion-product otoacoustic emission (DPOAE) input/output functions with stringent signal-to-noise criteria, applying covariate-controlled mixed-effects models within a dimensional framework that treated compression as a continuous outcome to isolate structured individual differences (n=144 ears, male=46, female=98). Compression metrics demonstrated substantial interindividual variability despite normal hearing thresholds. Subtle differences in standard-frequency thresholds were associated with variations in cochlear gain. Extended high-frequency sensitivity, an index of basal cochlear integrity, showed additional associations with DPOAE input/output function, implicating basal vulnerability in shaping cochlear nonlinearity. Age-related effects were detectable even within a restricted young-adult cohort, indicating that cochlear operating characteristics may begin to diverge early in adulthood. These findings challenge the assumption of invariant compression in normal hearing and suggest that alterations in cochlear nonlinear processing may reflect incipient cochlear dysfunction that precedes overt hearing loss.Significance Statement Hearing has traditionally been defined by audiometric thresholds, with normal results interpreted as evidence of an intact auditory periphery. This study challenges that assumption by demonstrating meaningful physiological variation within the clinically normal-hearing range. Specifically, cochlear nonlinear amplification, the core mechanism regulating sensitivity and dynamic range, differs across individuals and is associated with subtle biological factors, including age-related variation detectable in early adulthood. These findings indicate that cochlear function reflects a continuum of underlying biological states, some of which may represent incipient dysfunction preceding measurable hearing loss.
Evidence implicates the cerebellum in motor learning and performance, and relatively simple sensorimotor associations such as classical conditioning. However, little is known about how Purkinje cells encode more complex associations, such as arbitrary stimulus-response mappings. To address this, we recorded from Crus I/II Purkinje cells while two male macaques learned to associate novel fractal cues with left- or right-hand bar release. Here we show that as learning progresses, the simple spike activity of individual neurons becomes more selective for stimulus-response associations, with selectivity for left or right association developing closer to the appearance of visual stimuli. Initially, most neurons respond to both associations, irrespective of the identity of the stimulus and the associated movement, but as learning advances, more neurons distinguish between specific stimulus-hand associations. Moreover, complex spikes, which respond after the appearance of the stimuli, decrease their activity as the monkeys learn the task, suggesting that climbing-fiber activity encodes higher-order sensorimotor associations beyond mere error signaling. Using a linear decoder, we found that in early learning stages, the visual stimulus can be decoded only when the choice can also be decoded. As learning progresses, the visual stimulus is decoded earlier than the choice. A simple model can replicate the observed Purkinje cell signals in both the early and late learning stages.Significant Statement The cerebellum is traditionally associated with motor learning, yet its contribution to complex stimulus-response associations remains unclear. Recording Purkinje cells in Crus I/II as monkeys learned visuomotor associations, we found that activity evolves from reactive to predictive with learning. Simple spikes encode the upcoming choice progressively earlier, while complex spikes shift from a reactive response to the stimulus during early learning to a predictive signal before movement. Thus, cerebellar populations integrate sensory and motor inputs to generate anticipatory, decision-related activity.
Neural pruning optimizes the structure of neural circuits by selectively eliminating redundant or weakened synaptic connections, thereby enhancing the precision and efficiency of information transmission. Dendrites of class IV dendritic arborization (C4da) sensory neurons of Drosophila undergo a large-scale pruning process during development. However, it is currently ambiguous whether peripheral glia non-cell-autonomously modulate dendritic pruning in C4da neurons by transmitting signals. Here, we show that Unpaired1 (Upd1), a secreted glycoprotein, acts as a glial-originated signaling molecule to promote dendritic pruning during development in both male and female Drosophila Our further investigation reveals that downregulation of Domeless (Dome), the receptor of Upd1, in C4da neurons also triggers dendritic pruning defects. Hopscotch (Hop) kinase and transcription factor Stat92E, as downstream factors activated by Upd1 and Dome, similarly facilitate dendritic pruning. Notably, the Upd1 deficiency in glia produces a significant diminishment in JAK/STAT activity, and reactivation of this signaling in C4da neurons alleviates the dendritic pruning phenotype caused by glial Upd1 deficits. Furthermore, casein kinase II alpha (CK2α), the catalytic subunit of the serine/threonine protein kinase CK2, acts as a downstream target of the JAK/STAT and contributes to dendritic pruning. Collectively, we reveal a novel mechanism by which glia-derived Upd1 promotes dendritic pruning in C4da neurons by interacting with neuronal receptor Dome and activating the JAK/STAT signaling and downstream CK2α.Significance Statement Neural circuits are refined during development by selectively removing unnecessary branches and connections, but whether neighboring glia control this process via intercellular signaling remains not fully understood. Using Drosophila sensory neurons, we show that glia release the signaling protein Unpaired1 (Upd1), which activates neuronal JAK/STAT signaling to promote dendritic pruning. Restoring this pathway in neurons alleviates pruning defects caused by glial Upd1 deficiency, and we identify casein kinase CK2α as a downstream mediator. By revealing a glia-to-neuron signaling mechanism that coordinates developmental dendrite remodeling, this study expands our understanding of how neural circuits are refined during development and provides a conceptual basis for investigating how disrupted glial signaling and aberrant pruning may contribute to neurodevelopmental disorders.
Human vision efficiently navigates information-dense environments by extracting summary statistics—the average properties of item groups—to circumvent capacity limits. However, the neural transition from location-specific sensory registration to location-invariant abstract representation remains poorly understood. We recorded high-density EEG while participants of either sex performed an ensemble size discrimination task, using time-resolved multivariate pattern analysis (MVPA) and cross-visual-field generalization to dissociate these two levels of representation. Our results reveal a clear temporal hierarchy: location-specific ensemble size information emerged as early as ∼40 ms post-stimulus, significantly preceding the onset of location-generalized, abstract representation at ∼89 ms. During the early location-specific decoding phase, we observed a distinct left-visual-field advantage, with higher neural decoding accuracy predicting superior behavioral precision. Critically, error trials were characterized by premature neural generalization, suggesting an inherent trade-off: while abstraction is essential for efficient summarization, sacrificing sensory fidelity too early impairs perceptual accuracy. Furthermore, distinct oscillatory mechanisms supported this transformation—low-frequency (delta/theta) activity underpinned location-specific encoding, whereas mid-frequency (alpha/beta) oscillations robustly sustained location-invariant abstraction. Individual differences in the strength of these late-stage location-specific and location-invariant representations were predictable from intrinsic resting-state occipito-parietal gamma power. Together, these findings provide a comprehensive neural model of ensemble perception, demonstrating that the brain constructs abstract summary statistics through a temporally ordered, spectrally specific transformation that balances sensory fidelity with representational abstraction Significance Statement How the brain transforms complex sensory input into simplified summary statistics—like the average size of a group—is fundamental to efficient vision. This study reveals a critical temporal and spectral hierarchy in this process. Using high-density EEG and multivariate analysis, we show that the brain first registers ensemble information at specific locations before transforming it into a location-invariant, abstract representation. We demonstrate that premature abstraction leads to perceptual errors, suggesting a vital balance between sensory detail and abstract summary. Furthermore, we identify distinct neural oscillations that coordinate this transformation. These findings provide a new model for how the human brain constructs stable, abstract representations from a dynamic and cluttered visual world.
Anhedonia reflects a transdiagnostic deficit in a range of processes that impact reward and motivation. While human neuroimaging has mainly focused on striatal-related alterations in anhedonia, animal models suggest hippocampal [HPC] novelty processing regulates mesolimbic dopamine activity, implicating mesolimbic-HPC alterations in anhedonia. Childhood trauma, which disproportionately impacts HPC structure and function, may exacerbate this vulnerability. The present study of 37 males and 55 females examined whether HPC alterations interact with childhood trauma to predict anhedonia in humans. Using fMRI in a sample enriched for anhedonia, we assessed three HPC-related processes: resting-state connectivity with mesolimbic targets in the ventral tegmental area [VTA] and nucleus accumbens [NAc], task-based HPC novelty response, and task-based HPC modulation of VTA activation reflecting novelty-evoked facilitation of target detection. Significant interactions emerged for anticipatory anhedonia: reduced HPC-NAc connectivity, reduced novelty response, and weaker HPC→VTA modulation were each associated with greater anticipatory anhedonia among individuals with high childhood trauma. Moreover, LASSO regression confirmed these interactions as unique predictors. These findings suggest that early life adversity interacts with alterations in HPC-mesolimbic signaling to contribute to individual differences in anhedonia, highlighting the HPC as a potential target of motivation-related deficits in striatum. Significance Statement The ability to process rewarding stimuli and initiate goal-directed behavior is critical for human functioning. While human neuroimaging has highlighted the role of striatal function in these processes, animal evidence suggests that hippocampal modulation of mesolimbic dopamine signaling is critical for motivated behavior. The present study translates this framework to humans, demonstrating that hippocampal-mesolimbic functional integrity interacts with childhood trauma to predict individual differences in anticipatory anhedonia (i.e., impairments in one's ability to anticipate and pursue rewards). These findings implicate the hippocampus as a potential upstream contributor to motivational deficits and highlight early life stress as a key context in which mesolimbic circuit dysfunction becomes behaviorally relevant.
A central question in motor neuroscience is how the brain represents the state of the limbs to guide volitional movements. While the primate motor cortex is known to encode movement kinematics, such as velocity and direction, whether it also maintains a direct and explicit representation of hand position in 3D space remains debated. To address this, we recorded the activity of single neurons in the primary motor cortex (M1) and dorsal premotor cortex (PMd) of two male rhesus macaques performing a naturalistic, self-paced 3D reach-and-grasp task. We found significant populations of neurons in both M1 (36.2%) and PMd (21.3%) that are robustly tuned to the instantaneous 3D position of the hand. In these neurons, the tuning for hand position-characterized by localized, elongated fields-coexists with tunings for other kinematic variables, reflecting the principle of mixed selectivity. Critically, the spatial organization of these representations differs between the two areas: M1 fields are systematically oriented along cardinal axes and exhibit multi-scale spatial clustering, whereas PMd fields are more randomly organized. Furthermore, a small subset of these hand position-tuned cells is sufficient to decode the hand's 3D trajectory with high fidelity. Our findings demonstrate that an explicit and functionally organized representation of 3D hand position is a fundamental component of primate motor cortex, complementing dynamic motor signals to support high-fidelity motor control.Significance Statement To guide skilled actions, the brain must track the hand's location. While the motor cortex is known for controlling movement commands, we reveal it also creates an explicit, highly organized 3D map of hand position. This neural representation is systematically structured, differing between primary and premotor areas. This discovery reshapes our understanding of motor control, showing the brain merges spatial information ("where") with motor commands ("how") of the hand in the same areas. These insights are crucial for creating more effective brain-computer interfaces for individuals with paralysis.
Animals must continually evaluate the relative value of social and nonsocial rewards to guide adaptive behavior. While both food and social stimuli engage overlapping reward networks, how these reward types are represented within specific nodes of the circuit remains unresolved. The basolateral amygdala (BLA), a region critical for valence and motivational processing, has been implicated in encoding both shared and distinct representations of social and food stimuli. Using in vivo calcium imaging in freely behaving male and female mice performing a two-choice social-sucrose operant task, we examined how individual BLA neurons encode these two different rewards within the same behavioral framework. We found that largely non-overlapping populations of BLA neurons respond to social and sucrose rewards, revealing distinct representational subspaces for each reward type. Under baseline conditions both sexes robustly encoded social and sucrose rewards, but water deprivation revealed a pronounced sex difference, shifting representations toward sucrose reward in females. Populations of sucrose reward responsive neurons in the BLA were selectively sensitive to reward omission. Specifically, sucrose-excited neurons were more strongly modulated by the absence of expected reward than sucrose-inhibited neurons, suggesting a key role for these neurons in signaling reward prediction violations. Together, these findings demonstrate that the BLA encodes social and sucrose rewards through distinct, state-dependent, and functionally heterogeneous neuronal populations, highlighting region-specific strategies for representing reward identity and internal motivational context.Significance Statement The brain's capacity to weigh different types of rewards, like sucrose versus social interaction, is fundamental to survival yet remains poorly understood. By imaging neuronal activity in the basolateral amygdala (BLA) as mice freely choose between social and sucrose rewards, we found that largely distinct populations of BLA neurons encode each reward type. Internal state selectively altered the recruitment of reward-responsive BLA neurons in a sex-dependent manner. We further show that individual BLA neurons are differentially sensitive to the omission of expected rewards based on baseline responses. Together, these findings highlight the functional diversity through which the brain assigns value across reward modalities and dynamically adapts to changing motivational states.
To investigate how altered neural signal processing influences fMRI-BOLD responses in the hippocampus, we performed simultaneous in vivo electrophysiology and BOLD-fMRI in male Wistar rats during electrical stimulation of the perforant pathway. By defining input activity via applied pulses and measuring output activity through population spikes, we were able to identify qualitative and quantitative changes in signal processing once the relationship between input and output changed. An initial series of three low-intensity stimulations (LIS) induced clear, consistent BOLD responses. However, following a high-intensity stimulation (HIS) that triggered brief neuronal after-discharges, subsequent series of three identical LIS resulted in significantly attenuated BOLD responses. Electrophysiological data revealed that while total neuronal activity remained stable across all LIS, only the initial LIS induced long-lasting changes in signal processing (persisting beyond 1 minute) and transiently increased gamma band activity. By contrast, after HIS, these changes were reversed and could no longer be re-induced, coinciding with the absence of further increases in gamma band activity. Pharmacological experiments using MK801 and isoflurane further demonstrated that the mechanisms underlying long-lasting changes in signal processing also enhance LIS-induced BOLD responses. These findings suggest that positive fMRI-BOLD responses reflect functionally relevant changes in neural network properties-such as long-lasting modifications in signal processing-rather than simple increases in total neuronal output.Significance Statement An increase in BOLD fMRI signal in a specific brain region is generally interpreted as reflecting increased neuronal activity, with larger increases taken to indicate higher levels of activity. Here, we show that identical inputs can elicit stronger BOLD responses when they also induce long-lasting changes in local network properties, even when the overall activity of principal neurons remains similar. This suggests that the BOLD response is better understood as an indicator of functionally relevant changes in local network processing rather than a purely quantitative measure of neuronal activity.
Schizophrenia is associated with prefrontal cortex dysfunction, including neural disinhibition (reduced GABAergic inhibition) and hypoactivation ('hypofrontality'), alongside impaired reversal learning. However, evidence implicating prefrontal regions, including the rodent medial PFC (mPFC), in reversal learning is mixed. mPFC involvement may scale with demand for mPFC-dependent attention and cognitive control to overcome prepotent responses and to learn that reward contingencies can reverse, which is highest during early reversals. We therefore hypothesized that (1) the mPFC is required for early reversal learning but less important when reversal proficiency is high during late reversals. Furthermore, mPFC disinhibition may impair cognitive performance by causing circuit-level disruption outside the mPFC. Therefore, we hypothesized that (2) even when the mPFC is not required, mPFC disinhibition may impair reversal performance. To test hypotheses (1) and (2), we first examined the effect of mPFC functional inhibition and disinhibition, by microinfusion of the GABA-A receptor agonist muscimol and antagonist picrotoxin, on early reversals (reversals 1-3) versus late reversals (reversal 5 onwards) on a 2-lever discrimination task in adult male rats. mPFC functional inhibition by muscimol impaired only early reversals, increasing perseveration and impairing lose-shift behavior at reversal 2. In contrast, mPFC disinhibition by picrotoxin impaired late reversals, reducing lose-shift and win-stay behavior. Using chemogenetic mPFC disinhibition (hM4Di-mediated inhibition of GABAergic neurons), we further tested hypothesis (2). Similar to mPFC picrotoxin, chemogenetic mPFC disinhibition impaired late reversals, primarily disrupting win-stay behavior. Our findings suggest that reduced and disinhibited mPFC activity impair distinct aspects of reversal learning.Significance statement Schizophrenia is associated with reduced activation ("hypofrontality") and neural disinhibition (reduced GABAergic inhibition) within the prefrontal cortex (PFC). However, it is not clear if and how these distinct aspects of prefrontal dysfunction contribute to impaired reversal learning, a key feature of the cognitive inflexibility characterizing schizophrenia. Here, we combined bi-directional manipulations of prefrontal GABAergic inhibition with testing of reversal learning in rats. Increasing prefrontal functional inhibition (i.e., reducing prefrontal activation) selectively impaired early reversals, enhancing perseveration and reducing lose-shift behavior, whereas prefrontal disinhibition disrupted late reversals, impairing both lose-shift and win-stay behavior. Our findings suggest that reduced activation and disinhibition of PFC disrupt distinct aspects of reversal learning, by distinct mechanisms.
