Background: The complex pathophysiology of epilepsy hampers the development of effective treatments. Although more than ten kinds of anti-seizures drugs (ASDs) have good effects on seizure control worldwide, about 30% of patients still display pharmacoresistance against ASDs. Neuroinflammation seems to play a crucial role in disease progression. G protein-coupled receptor 120 (GPR120) has been shown to negatively regulate inflammation and apoptosis. However, the role of GPR120 in epilepsy remains unclear. In this study, we aimed to explore the mechanism of GPR120 in epilepsy. Methods: Male adult C57BL/6 mice were intracranially injected with kainic acid (KA) to establish epilepsy model, and the adeno associated virus (AAV) was administered intracranially at 3 weeks before KA injection. VX765 was administered by intragastric administration at 30 min before KA induced and an equal dose administrated twice a day (10 a.m. and 4 p.m.) lasting 7 days until the mice were killed. Western blot analysis, immunofluorescence staining, video monitoring of seizure, LFP recording, Nissl staining were performed. Results: GPR120 was increased in both the hippocampus and cortex in the KA-induced model with temporal lobe epilepsy (TLE), and both were most highly expressed at 7 days after KA injection. Overexpression of GPR120 significantly alleviated epileptic activity, reduced neuronal death after status epilepticus (SE), downregulated the expression of IL-1β, IL-6, IL-18, and pyrin domain-containing protein 3 (NLRP3) inflammasome, whereas knockdown GPR120 showed the opposite effect. The effects of GPR120 knockdown were reversed by VX765 inhibition cysteinyl aspartate specific proteinase-1 (Caspase-1). Conclusion: GPR120 modulates epileptic seizure activity and affects neuronal survival in KA-induced mouse model of temporal lobe epilepsy. Furthermore, GPR120 regulated neuroinflammation in epileptic animals through NLRP3/Caspase-1/IL-1β signaling pathway.

The microbiota-gut-brain (MGB) axis is implicated in Alzheimer’s disease (AD), but evidence for interventional strategies targeting this axis remains limited. In a 24-week, single-blind, randomized controlled trial, 84 individuals with mild cognitive impairment (MCI) or mild AD received either computerized cognitive training (CCT) or treatment as usual (TAU). Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-cog) was the primary outcome. We also assessed functional connectivity (fNIRS), plasma trimethylamine N-oxide (TMAO) levels, and gut microbiota at baseline and 24 weeks. Seventy-four participants completed the study. The CCT group showed significant improvement in ADAS-cog scores compared to controls (Cohen’s d = 1.57 by week 24). Notably, CCT also induced a distinct reorganization of prefrontal functional connectivity and significantly reduced plasma TMAO levels. Microbiome analysis revealed that CCT mitigated the expansion of Ruminococcus torques group (R.torques), which was observed in the control group. Crucially, R.torques was the only genus significantly correlated with improvements in cognition (ADAS-cog, r = 0.407), neuropsychiatric symptoms (NPI, r = 0.395), TMAO reduction (r = 0.443), and functional connectivity changes (r = 0.449). A 24-week CCT program improves cognitive function in MCI and mild AD, potentially through downregulating the R.torques-TMAO pathway within the MGB axis. This pathway represents a promising novel target for multi-domain intervention in AD.

(1) Background: Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that threatens the population health of older adults. However, the mechanisms of the altered metabolism involved in AD pathology are poorly understood. The aim of the study was to identify the potential biomarkers of AD and discover the metabolomic changes produced during the progression of the disease. (2) Methods: Gas chromatography–mass spectrometry (GC–MS) was used to measure the concentrations of the serum metabolites in a cohort of subjects with AD (n = 88) and a cognitively normal control (CN) group (n = 85). The patients were classified as very mild (n = 25), mild (n = 27), moderate (n = 25), and severe (n = 11). The serum metabolic profiles were analyzed using multivariate and univariate approaches. Least absolute shrinkage and selection operator (LASSO) logistic regression was applied to identify the potential biomarkers of AD. Biofunctional enrichment analysis was performed using the Kyoto Encyclopedia of Genes and Genomes. (3) Results: Our results revealed considerable separation between the AD and CN groups. Six metabolites were identified as potential biomarkers of AD (AUC > 0.85), and the diagnostic model of three metabolites could predict the risk of AD with high accuracy (AUC = 0.984). The metabolic enrichment analysis revealed that carbohydrate metabolism deficiency and the disturbance of amino acid, fatty acid, and lipid metabolism were involved in AD progression. Especially, the pathway analysis highlighted that l−glutamate participated in four crucial nervous system pathways (including the GABAergic synapse, the glutamatergic synapse, retrograde endocannabinoid signaling, and the synaptic vesicle cycle). (4) Conclusions: Carbohydrate metabolism deficiency and amino acid dysregulation, fatty acid, and lipid metabolism disorders were pivotal events in AD progression. Our study may provide novel insights into the role of metabolic disorders in AD pathogenesis and identify new markers for AD diagnosis.

Alzheimer's Disease (AD) poses a considerable global health challenge, necessitating early and accurate diagnostics. The Mini-Mental State Examination (MMSE) is widely used for initial screening, but its traditional application often underutilizes the rich multimodal data it generates, such as videos, images, and speech. Integrating these modalities with modern Large Language Models (LLMs) offers untapped potential for improved diagnostics. In this study, we propose a multimodal LLM framework fundamentally reinterprets MMSE data. Instead of relying on conventional, often limited MMSE features, proposed LLM acts as a sophisticated cognitive analyst, directly processing MMSE modalities. This deep multimodal understanding allows for the extraction of novel, high-level features that transcend traditional metrics. These are not merely visual or acoustic signals, but rich semantic representations imbued with cognitive insights gleaned by the LLM. We then construct an interpretable decision tree classifier and derive a succinct rule list, yielding transparent diagnostic pathways readily understandable by clinicians. Finally, framework integrates a counterfactual explanation module to provide individualized "what-if" analyses, illuminating how minimal feature changes could alter model outputs. Our empirical study on real-world clinical data achieves a diagnostic accuracy of approximately 6% percentage points improvements with diagnosing explanation, reinforcing the viability of our framework as a promising, interpretable, and scalable solution for early AD detection.