Humans use various sequential signals, such as language, music, and mathematics, to convey complex information and facilitate communication. Previous research has identified two fundamental frameworks underlying human sequential signal processing: a structural framework, emphasizing rule-based hierarchical organization (e.g., syntax), and a predictive framework, which focuses on the brain's capacity to anticipate upcoming inputs based on statistical regularities. While the structural approach underscores human-specific abilities, the predictive approach highlights mechanisms shared with non-human animals. We review behavioral and neural evidence across domains, demonstrating overlapping neural substrates, particularly in the inferior frontal gyrus (IFG), involved in structural processing of language, music, and mathematics. Likewise, predictive processing, indexed by ERP components such as N400 and P600, operates across domains to detect violations of expectation. Importantly, we argue that these frameworks are not mutually exclusive: structural knowledge can inform prediction, and predictive processes can, in turn, influence perceived structure. Cross-domain experiments and computational modeling suggest shared cognitive mechanisms, although domain-specific variations remain. We propose that the human brain integrates hierarchical structures with statistical learning to support flexible and generalized sequence processing in humans. Future research should aim to develop unified models across domains, leveraging neuroimaging techniques and large language models.

Neural organoids are self-assembled three-dimensionally shaped aggregates generated from pluripotent stem cells for the purpose of generating brain-like structures. The features of the disease, from molecular to functional levels, can be recapitulated by neural organoids derived from patient induced pluripotent stem cells (iPSCs). These features are not fully reproduced by other culture systems or in vivo models. Neural organoids have been applied to model dementia including Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis/frontotemporal dementia, and they have recapitulated aspects of their complex pathophysiology, including neuronal network dysfunction and accumulation of pathogenic proteins. Although research using neural organoids still faces challenges such as heterogeneity and the absence of non-neural lineage cells, these limitations are being progressively addressed. Recent advances, including the integration of gene-editing technologies and the co-assembly of organoids with specific cell types, have demonstrated the remarkable potential of this approach. This article reviews current research on iPSC-derived neural organoids for dementia, discussing both the technical hurdles and the potential for translational applications.

Induced pluripotent stem cells (iPSCs) are widely used in research because they can be used to create models of diseases with the same genomic background as in patients. iPSC-based screening is recognized as a valuable approach in drug discovery research. Additionally, efforts are underway to develop high-quality models for drug discovery and to better integrate translational research with clinical studies. This review focuses on neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS) and Alzheimer's disease (AD), and provides a broad overview of research using iPSCs, ranging from studies of disease mechanisms to applications in drug discovery. Furthermore, several clinical trials based on iPSC research have been initiated, including those of bosutinib, ropinirole, and ezogabine for ALS, and WVE-004 and BII078 for ALS/FTD. Finally, we also wish to highlight screening studies that incorporate artificial intelligence (AI).

Alzheimer's disease (AD) is the most common progressive neurodegenerative disease. It is characterized by cognitive decline and brain pathology, including amyloid-β (Aβ) plaques, neurofibrillary tangles, and neuroinflammation. While rodent models have contributed to our understanding of a multiplicity of disease mechanisms, their limitations in replicating human age-related neurodegenerative diseases pose challenges for translating research findings to the clinical setting. The common marmoset (Callithrix jacchus), a small non-human primate, has emerged as a promising model for neuroscience and aging research. Age-related Aβ and tau pathologies develop naturally in the marmoset, which also possess a brain network organization, including a default mode network, that closely resembles that of humans. AD-related pathologies have also been experimentally induced in marmosets through injection of materials extracted from post-mortem brain tissue of AD patients. Recent advances in genome-editing technologies have enabled the development of marmoset models carrying familial AD mutations. These models offer new opportunities to investigate early pathological changes in the disease and to evaluate potential therapeutic agents. Taken together, these findings highlight the value of the marmoset as a translational model bridging the gap between rodent studies and human AD research.

Antipsychotic drugs are increasingly recognized to exert therapeutic effects not only through neurotransmitter modulation but also, in part, through epigenetic mechanisms. Risperidone is an atypical antipsychotic drug widely used to treat schizophrenia, yet its in vivo epigenomic effects remain poorly understood. We investigated genome-wide DNA methylation changes induced by chronic risperidone administration in the common marmoset, a non-human primate with high translational relevance. Adult males were treated orally with risperidone for 28 days. DNA methylation was analyzed in four brain regions (frontal cortex, hippocampus, cerebellum, and caudate nucleus) and two peripheral tissues (blood and liver) using a HumanMethylation450K BeadChip adapted to the marmoset genome. Risperidone induced region- and tissue-specific methylation alterations. The brain showed predominant hypermethylation, and the hippocampus had the largest number of differentially methylated probes. Rank-rank hypergeometric overlap analysis revealed partial hypermethylation-hypermethylation concordance between the hippocampus or caudate and peripheral tissues, indicating partially coordinated changes. We also found that risperidone-treated neuroblastoma cells showed methylation patterns closely resembling those of the hippocampus, suggesting that the epigenetic changes are partly conserved in the cell models. These findings offer a framework for understanding the molecular basis of antipsychotic actions and for identifying potential epigenetic markers relevant to clinical effects.

Bilateral vestibular dysfunction disrupts balance and spatial orientation, yet mechanisms of injury and compensation remain incompletely defined. We established a mouse model of bilateral acute inner-ear (labyrinthine) injury with vestibular involvement by sequentially incising the round and oval windows, and evaluate auditory brainstem responses (ABR), vestibular sensory-evoked potentials (VsEP), behavior, and histology across postoperative days 1-28. Threshold elevations and behavioral impairment peaked on day 3, with hair-cell degeneration most prominent early and partial morphological recovery by day 28. These findings delineate the acute injury phase and early stabilization, providing a platform to study vestibular repair and functional recovery.

Tauopathies are a group of neurodegenerative diseases characterized by the aberrant accumulation of tau protein in the brain. While numerous mouse models have been developed to study tauopathies, the majority depend on tau overexpression, which may encompass non-physiological artifacts and limit the translational relevance of findings. In this review, we highlight the development and application of an isogenic panel of MAPT knock-in (KI) mouse lines that carry single or multiple pathogenic mutations within the human MAPT gene. In these models, the endogenous murine Mapt gene was replaced with the humanized MAPT sequence, and tau is expressed under the control of the native murine Mapt promoter. This approach preserves spatiotemporal regulation of tau, providing a more physiological representation of human tauopathies. As such, these mutant MAPT KI models serve as powerful tools for elucidating the pathomechanisms of tauopathies and discovering drugs that aid tau-mediated neurodegeneration.

Feeding behavior is influenced by both metabolic drive and emotional context, yet how acute stress interferes with hunger-driven motivation remains poorly understood. Using a conflict-based open-field feeding paradigm, we examined how 30-min restraint stress alters food-seeking behavior in fasted mice and how pharmacological or physiological interventions modulate this effect. Acute restraint abolished fasting-induced increases in food seeking and intake, an effect that was reversed by intraperitoneal diazepam, an anxiolytic, and by MK-677, a ghrelin receptor agonist that enhances appetite. To identify neural correlates, we quantified c-Fos (neuronal activation) and phosphorylated pyruvate dehydrogenase (pPDH; neuronal inhibition) in the paraventricular hypothalamic nucleus (PVN). Diazepam suppressed restraint-induced c-Fos expression, whereas MK-677 increased pPDH, revealing distinct PVN signatures for anxiolysis and enhanced feeding drive. Notably, refeeding after fasting induced a similar pPDH-dominant pattern and attenuated stress-induced anxiety-related behaviors, indicating that restoration of energy balance exerts intrinsic anti-stress effects through PVN inhibition. Together, these findings reveal that acute stress suppresses hunger-driven food seeking via PVN activation, pharmacological inhibition reverses this suppression, and physiological refeeding promotes stress resilience via PVN-medicated inhibition, highlighting PVN modulation as a shared mechanism linking emotional and metabolic homeostasis.

Tau protein is a microtubule-associated protein that plays a crucial role in maintaining neuronal morphology and axonal transport. While phosphorylation is known to regulate Tau-microtubule interactions, the contribution of specific phosphorylation patterns in situ remains poorly understood due to the complexity of the intracellular environment. In this study, we combined fluorescence recovery after photobleaching (FRAP) in primary cultured rat hippocampal neurons with dephosphorylation-mimetic mutations and computational modeling to analyze the effects of phosphorylation on Tau-microtubule interaction. We particularly focused on the proline-rich region, of which phosphorylation has been studied in physiological and pathological perspectives, and generated a dephosphorylation-mimetic Tau mutant by substituting key phosphorylation sites with alanine residues and compared its microtubule-binding dynamics to those of WT-Tau in FRAP experiments. Experimental data, together with simulation-based parameter exploration, revealed that the overall number of non-phosphorylated sites plays a more dominant role than their specific locations in modulating Tau-microtubule affinity. These findings provide new insights into the post-translational regulation of Tau and establish a computational-experimental framework for interrogating intracellular protein dynamics.

Schizophrenia is characterized by profound semantic impairments that manifest as disrupted language and thought. We provide empirical support for the hypothesis that predictive coding forms a unifying framework for understanding these deficits by reinforcing theoretical ideas with quantitative neuroimaging evidence. According to predictive coding theory, the brain continuously generates predictions about incoming information, and prediction errors drive model updates when expectations diverge from sensory input. This review synthesizes evidence from cognitive neuroscience, computational psychiatry, and neurolinguistics to demonstrate how aberrant prediction error signaling disrupts hierarchical semantic processing in schizophrenia. Behavioral studies have revealed atypical semantic processing in priming and fluency tasks. Electrophysiological studies have shown altered neural responses to semantic incongruence, particularly reduced N400 effects. Furthermore, we have used voxel-wise modeling, graph theory, and topological analysis to demonstrate fundamentally disorganized semantic networks in schizophrenia, characterized by reduced small-worldness, excessive homogenization, and diminished representational variability. These converging findings are consistent with a neurocomputational account wherein semantic deficits reflect disrupted predictive mechanisms. This theoretical framework suggests that miscalibrated precision weighting of prediction errors leads to either over-activation of irrelevant semantic associations or impoverished semantic processing. This perspective offers insights into schizophrenia pathophysiology and guidance for targeted interventions to restore predictive coding function.

Microglia and macrophages (M/M) play significant roles in intracerebral hemorrhage (ICH) injury, but the underlying mechanism is complicated and remains largely unknown. Lipocalin-2 (LCN2) expression elevates in M/M after ICH, yet its role in M/M-induced inflammation after ICH has not been fully elucidated. In the current study, a mouse ICH model was established in LCN2 male (LCN2 cKO) mice and LCN2 male mice. We then aimed to inject LCN2 protein or the Nrf2 inhibitor brusatol groups for mechanism study, and the BV2 and Raw264.7 cell lines were used for in vitro study. LCN2 induces pro-inflammatory phenotype and promotes pro-inflammatory secretion in M/M after blood injury, and M/M LCN2 knockout reduced the pro-inflammatory phenotype after ICH. The Nrf2 protein is activated and nuclear-translocated by LCN2 knockout/downregulation, and the Nrf2 inhibition abrogates the anti-inflammatory effect induced by LCN2 knockout/downregulation. LCN2 knockout/downregulation boosts M/M phagocytosis after ICH, which is partially reversed by Nrf2 inhibition. The LCN2 expression also increases in aged mouse brains and participates in pro-inflammation induction and phagocytosis inhibition after ICH. Our study demonstrates that post-ICH LCN2 expression in M/M induces pro-inflammatory phenotype via inhibiting Nrf2 signaling pathway.

Neurodegenerative diseases are characterized by the gradual deterioration of specific neuronal populations, ultimately resulting in motor, cognitive, or behavioral impairments. Despite the worldwide increase in disease incidence, effective therapies remain unavailable. A common pathological hallmark of neurodegenerative diseases is the accumulation of misfolded protein aggregates, which impair normal cellular function. Accordingly, numerous studies and therapeutic strategies have focused on targeting these toxic aggregates and protein quality control via autophagy, a vital cellular recycling mechanism. Autophagy dysregulation has been implicated in the pathogenesis of several neurodegenerative diseases. Induced pluripotent stem cell (iPSC) technology has emerged as a powerful platform for modeling neurodegenerative diseases, and iPSC-derived models provide human-relevant systems for studying autophagic dysfunction in vitro. In this review, we discuss the key findings of recent studies investigating autophagy in iPSC-based models of neurodegenerative diseases, including Alzheimer's disease, amyotrophic lateral sclerosis, frontotemporal dementia, and other diseases.

The advent of directly induced neurons (iNs) from human somatic cells has revolutionized disease modeling in neurodegeneration. This approach bypasses pluripotent stage during the neuronal cell inducing steps and preserves donor age-related signatures. This review explores the trajectory of iNs, including multiple induction methods, their applications in modeling neurodegenerative diseases, recent innovations such as three-dimensional (3D) culture platforms, and their potential to advance personalized medicine.

Alzheimer's disease and related neurodegenerative diseases (ADRD) represent a major global public health challenge, with their disease mechanisms remain largely unknown, and few treatments are available. Increasing epidemiological evidence underscores the critical role of traumatic brain injury (TBI) in the initiation and progression of ADRD, suggesting shared pathogenic mechanisms between these conditions. While there are still lack of perfect AD models in the field, TBI models may serve as useful and alternative platforms for investigating ADRD. In this review we delineate the definition and epidemiological characteristics of TBI. We further briefly compare the major experimental TBI animal models, outlining their respective strengths and limitations in replicating human neuropathology. Finally, we provide our perspective on potential mechanistic links between TBI and ND, including axonal injury, calcium homeostasis dysregulation, mitochondrial dysfunction, chronic neuroinflammation, blood-brain barrier disruption, and genetic susceptibility. We believe advancing preclinical and translational research on TBI not only enhances our understanding of the pathogenesis in ADRD but also holds promise for developing interventions to mitigate long-term consequences and improving clinical outcomes for many neurodegenerative diseases.

Numerous mouse models of Alzheimer's disease (AD) have developed since the discovery of mutations causing familial AD. These models successfully recapitulate the progressive amyloid pathology over time, thus serving as indispensable tools for improving our understanding of the pathogenesis. However, there is a growing concern about artificial phenotypes in these transgenic mouse models, resulting from overexpression of mutant amyloid precursor protein (APP) under artificial promoters. To address this issue, App knock-in (KI) mice were developed to produce mutated human β-amyloid (Aβ) from the endogenous App locus. Since the first characterisation in 2014, gathering evidence has made significant progress in the phenotype analysis of this mouse model. Here, we provide an update on novel phenotypes observed in App KI mice. In particular, we will highlight how the progression of amyloid pathology is related to neuronal pathology, behavioural phenotype, and microglial response.

This study was intended to verify the potential role of Fluoxetine (Flx) in treating depression associated with opioid-induced constipation (OIC). We established a mouse model of chronic unpredictable mild stress (CUMS) and used loperamide to induce constipation based on the CUMS mice, generating a mouse model of depression associated with OIC (CUMS+OIC). The depressive behavior was evaluated via the open field and sucrose preference tests, while constipation was evaluated using defecation frequency and fecal water content. Liquid chromatography-mass spectrometry (LC-MS)-based metabolomics and bioinformatics analyses were performed. Treatment with Flx alleviated the depressive behavior of mice and inhibited OIC. We identified 153 differential metabolites between the control and the CUMS+OIC groups, among which 51 were downregulated while the other 102 were upregulated. These metabolites were involved in metabolic pathways such as pyrimidine metabolism, purine metabolism, and beta-alanine metabolism. Moreover, 64 differential metabolites between the Flx and the CUMS+OIC groups were involved in nicotinate and nicotinamide metabolism, and prion disease metabolism metabolic pathways. Through cluster analysis, we identified metabolites deregulated by CUMS+OIC and restored by Flx. Conclusively, Flx can improve the behavior and metabolic profile changes of CUMS associated with OIC, providing a basis for treating depression-related constipation.

One main focus in neuroscience is to understand the relationship between decision making and various brain regions. Researchers use machine learning approaches to model the neural circuits of cerebral cortices, cerebellum, and basal ganglia. This review focuses on artificial neural networks (ANNs), particularly recurrent neural networks (RNNs), to model cortical functions for decision making. We first introduce the basic architecture of RNNs and explain how researchers compare the activity and circuits between artificial and biological networks. We also summarize how RNNs model the prefrontal and posterior parietal cortical in tasks involving short-term memory, perceptual decision making, and value-based decision making. We then show our recent challenges to develop a real-cyber hybrid network, that integrates neuronal activity in mice with RNN-based artificial units to better generate continuous-time body movements, compared to conventional RNNs that only use artificial units. The hybrid network tries to develop RNNs which have similar activity to the brain by using real neurons, rather than developing artificial RNNs and comparing their functions with biological brain. We propose that such integrative approaches in neuroscience and AI will further our understanding of both natural and artificial intelligence in the field of neuro-AI.

Tau protein, a central player in Alzheimer's disease (AD), exhibits cytotoxicity upon fibril formation. Understanding the early stages of tau fibrillization is therefore critical for the development of effective therapeutics. Previous work [Rasmussen. et. al, J. Mol. Biol., 2023] reported the rapid formation of Thioflavin T (ThT)-inactive clusters upon mixing tau with anionic polymers, yet the functional role of these clusters remained unclear. Here, we demonstrate that these transient clusters act as obligatory precursors in the fibrillization pathway. Using small-angle X-ray scattering (SAXS) and ThT fluorescence, we show that disrupting the clusters via NaCl addition hinders fibril formation, highlighting their reversible and targetable nature. This behavior is analogous to polymer crystallization, in which disordered chains undergo structural ordering through intermediate precursor states. We propose that similar physical principles underlie the aggregation of other intrinsically disordered proteins such as α-synuclein.

Brachial plexus injury is a neurological injury caused by trauma, and effective treatments remain limited. Understanding its pathology is necessary to develop new therapeutic strategies. In this study, we used a dorsal root avulsion mouse model to determine whether injury-induced alterations in feedback circuitry from the dorsal root ganglion (DRG) to motor neurons contribute to functional recovery. We visualized axons originating from DRG neurons by directly injecting adeno-associated virus encoding green fluorescent protein (AAV-GFP) into the DRG adjacent to the injury site and analyzed the total length of axons on the lateral side of the ventral horn in the spinal cord. Following injury, the fine motor function of the affected forepaw was immediately impaired and then gradually recovered. In parallel, axons originating from an adjacent, uninjured DRG extended into the deafferented spinal segments, possibly contributing to the reinnervation of motor neurons that had lost their original sensory input. Indeed, after spontaneous motor recovery, when we performed an additional dorsal root avulsion originating from the adjacent DRG, functional impairment of the forepaw re-emerged. Our results demonstrate that the plasticity of the adjacent DRG may facilitate the recovery of fine motor function after dorsal root injury.

During wakefulness, cortical neurons fire tonically and asynchronously to constantly support ongoing cognitive functions, whereas they show synchronized cessation of firing, so called 'OFF periods', during NREM sleep. Previous studies reported that cortical neurons start showing local OFF periods even in awake animals as sleep pressure builds up during the course of sleep deprivation. However, spatio-temporal dynamics of OFF periods across the cerebral cortex during prolonged wakefulness remained unknown. In this study, our extracellular recording in free-behaving mice showed that prolonged wakefulness causes neuronal firing lapses in the medial prefrontal cortex (mPFC) but not in motor cortex, suggesting the importance of mPFC in maintaining cognition during wakefulness. We next examined cortical area-specific effects of sustained neural activation on vigilance states. Chemogenetic activation of excitatory neurons in mPFC, but not in motor or sensory cortex, promoted wakefulness. Additionally, chemogenetic inhibition of mPFC reduced time spent in wakefulness, further supporting the causal role of mPFC in wakefulness.

Neural signal processing in the cerebral cortex is often regarded as robust and stereotyped; however, the brain's internal environment undergoes dynamic fluctuations across the day. Whether these diurnal rhythms modulate cortical responsiveness and plasticity remains unclear. Here, we examined diurnal modulation of neural responsiveness and plasticity in the primary visual cortex (V1). Using transgenic rats expressing channelrhodopsin-2 (ChR2), we optically stimulated V1 neurons with brief light pulses and recorded local field potentials (LFPs) over several days. V1 responses to single-pulse stimulation showed clear diurnal variation, with delta- and gamma-band activity modulated in a time-of-day-dependent manner. Administration of the adenosine A1 receptor antagonist DPCPX enhanced neural responses at Zeitgeber time (ZT) 0 (Sunrise) but not at ZT 12 (Sunset). LTP-like potentiation was observed only when train stimulation was applied at Sunrise, indicating that plasticity is also gated by diurnal phase. These findings demonstrate that both excitability and plasticity of V1 circuits are regulated by diurnal factors. Although it remains unclear whether these effects are driven by intrinsic circadian rhythms or light/dark-triggered mechanisms, our results highlight that cortical processing is dynamically modulated across the day, with implications for sensory function, learning, and neuromodulatory regulation.