Meridian Institute

AI systems produce large volumes of logs as they interact with tools and users. Analysing these logs can help understand model capabilities, propensities, and behaviours, or assess whether an evaluation worked as intended. Researchers have started developing methods for log analysis, but a standardised approach is still missing. Here we suggest a pipeline based on current best practices. We illustrate it with concrete code examples in the Inspect Scout library, provide detailed guidance on each step, and highlight common pitfalls. Our framework provides researchers with a foundation for rigorous and reproducible log analysis.

Abstract Many trials are conducted without first assessing which outcomes are important to patients receiving the interventions. The goal of this work was to ask participants of an ongoing exercise program what they believed the most important impacts of the program were to better align outcomes for a future trial with participant priorities. The Rev6 program is a seated, 30-minute exercise class designed for older adults with varying cognitive abilities. The program is offered twice a week by a large assisted living community (ALC) in Rhode Island. We observed three Rev6 sessions and held a focus group with seven Rev6 participants. We also interviewed one ALC staff member responsible for delivering the Rev6 program and two corporate leaders. Three members of the research team analyzed the focus group and interview transcripts to identify themes. Themes related to physical health included staff- and participant-reported impact of the Rev6 program on balance and walking, strength, and pain. Themes related to mental health included staff- and participant-reported impact of the Rev6 program on “feeling better” and increased social connection. With input from ALC staff, researchers developed visual cue cards for each theme and met with nine program participants to vote on the most important theme to evaluate as an outcome measure for a future study. Four residents voted for balance and walking, three residents voted for feeling better, and two residents voted for pain as the most important impact of the program. This framework for evaluation codesign includes the voices of people with dementia.

Abstract Albania, located in the Mediterranean region, possesses a climate highly suitable for the growth of numerous plant species (over 3,000), of which more than 300 are aromatic and medicinal. These plants are widely used in culinary applications, traditional medicine, and as ornamental species. Medicinal plants have been utilized since ancient times due to their health-promoting effects, and they continue to play an important role today, both in traditional therapies and in the formulation of pharmaceutical products and dietary supplements. Owing to their natural origin and minimal side effects compared to synthetic drugs, herbal medicines are increasingly incorporated into long-term treatments and clinical applications. The biological properties of these plants are mainly attributed to their phytochemical constituents, particularly essential oil components. This study presents data on the chemical composition of essential oils from three of the most commonly used aromatic plants in Albania, primarily from the southern region: Salvia officinalis L. (sage), Origanum vulgare L. (oregano), and Laurus nobilis L. (laurel). Wild populations of these species were collected from the Llogara Mountain area (southern Albania) in June 2025. Air-dried plant materials were subjected to hydro-distillation using a Clevenger-type apparatus to obtain the essential oils. The chemical composition was analyzed by gas chromatography equipped with a flame ionization detector (GC/FID) using a VF-1ms capillary column for compound separation. The results showed that α- and β-thujones were the most abundant constituents in S. officinalis oils; carvacrol and thymol were predominant in O. vulgare ; and 1,8-cineole (cineol) was identified as the major component of L. nobilis essential oil. The chemical profiles of these Albanian wild populations were consistent with previously reported data for the Balkan and broader Mediterranean regions.

Pipeline alternating current (AC) corrosion is a function of the induced AC pipeline voltage and is further influenced by pipeline material, pipeline coating systems, and soil characteristics local to the pipeline. When considering modern pipelines, more conductive soils result in greater AC current density at coating defects (i.e., holidays). Soil stratification can further significantly influence both the resultant induced AC pipeline voltage and possible pipeline AC corrosion. This article explores the effects that soil resistivity, layer stratification, and pipeline coating quality have on AC interference, including impacts from induced AC voltage, holiday current densities, and conductive coupling hazards.

In a digitalized world, accessing disease information through online platforms has become essential in influencing health-promoting behaviors among patients with heart failure. Patients must possess sufficient digital health literacy to obtain accurate and reliable information from digital platforms; however, there is no known data about the mediating role of digital health literacy in the relationship between health behaviors and symptom burden among persons with heart failure. To examine the relationship between digital health literacy, health behaviors, and symptom status (frequency, severity, and level of discomfort) in persons with heart failure and the mediating role of digital health literacy in the relationship between health behaviors and symptom status. This cross-sectional, correlational study was conducted between November 2023 and June 2024 with 250 persons diagnosed with heart failure. Data were collected using the Heart Failure Health Behavior Scale, e-Health Literacy Scale, and Heart Failure Symptom Status Scale. Pearson correlation analysis, linear regression, and hierarchical regression analysis with PROCESS Model 4 were used to analyze the continuous variables. Digital health literacy was positively but weakly correlated with health behavior and negatively correlated with symptom status. The mediation analysis revealed that part of the effect of health behaviors on symptom status was mediated by digital health literacy. Health behaviors directly affected symptom status and indirectly affected symptom status through digital health literacy. Strategies developed by nurses to enhance digital health literacy could play a critical role in improving the health of persons with heart failure by reducing symptom burden and promoting health behaviors. Research is needed on the influence of various predictors on symptom burden and assessing whether interventions to enhance digital health literacy can improve health outcomes among those with heart failure.

Herbal remedies have been a mainstay of medicine for thousands of years. This systematic literature review investigated the efficacy and safety of chamomile herbal products among peripartum or postpartum women. Four peer-reviewed databases were searched through June 2024. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol was implemented, and the quality of studies was assessed using Cochrane Risk-of-Bias Assessment tools. A total of 23 studies (16 clinical trials and seven observational studies involving 2,065 women from nine countries) were included in this review. The development of clinical practice recommendations on using chamomile products during pregnancy is not feasible based on the available evidence, indicating the need for randomized, double-blind placebo control studies with larger study populations and consistent study protocols (e.g., type and dosage of chamomile ingested). Limitations of this review include its small sample size and the inclusion of multiple studies by the same research teams; these findings likely reflect insights from the same cohorts of women. Some studies reported clinically significant findings that were not statistically significant. Thus, despite weak evidence supporting the efficacy and safety of chamomile usage during pregnancy, a provider might still share the potential benefits and risks of using chamomile products with patients who use or desire to use chamomile products.

To observe the effect of "Huayu Tongluo"(blood stasis-removing and meridian-collateral dredging) moxibustion on oligodendrocyte-related protein expression and myelin ultrastructure in the corpus callosum tissue of rats with vascular dementia (VD), so as to explore its mechanism underlying improvement of VD. Male SD rats were randomly divided into model group, "Huayu Tongluo" moxibustion group and non-acupoint moxibustion group, with 42 rats in each group. According to the treatment time, the rats were further divided into 1 course (7 d) group (n=12), 2 courses (14 d) group (n=12) and 4 courses (28 d) group (n=18), and the other 18 rats were used as the sham operation group. The VD model was established by using modified 2-vessel occlusion (VO) method. The rats in the "Huayu Tongluo" moxibustion group were treated with suspended moxibustion of "Baihui"(GV20), "Dazhui"(GV14) and "Shenting"(GV24) for 20 min, once a day for 7 d, 14 d or 28 d, respectively. The rats of the non-acupoint moxibustion group were treated with suspended moxibustion at the three locations 2, 3 and 5 cm away from the tail, and the course of treatment was the same as that of "Huayu Tongluo" moxibustion group. Morris water maze test was used to assess the rats' learning-memory ability. A transmission electron microscopy was used to observe the ultrastructure of myelin sheath and the number of myelinated nerve axons, and to calculate the g-ratio (axon diameter/(axon diameter + myelin sheath thickness)) of the myelin sheath thickness. Western blot and immunofluorescence staining were used to separately detect the expressions and fluorescence intensities of oligodendrocyte markers 2', 3'-cyclic nucleotide-3'-phosphodiesterase (CNPase), oligodendrocyte transcription factor 2 (Olig2) and myelin oligodendrocyte glycoprotein (MOG) in the corpus callosum of each group. Compared with the sham operation group, the model group showed prolonged escape latency (P<0.01) and reduced number of platform quadrant crossings (P<0.01) in the 1, 2 and 4 courses of treatment, significantly decreased number of myelinated nerve axons in the corpus callosum (P<0.01), increased g-ratio value of axonal myelin (P<0.01), and decreased relative fluorescence intensities and expression levels of CNPase, Olig2, and MOG proteins in the 1, 2 and 4 courses of intervention (P<0.01). Compared with the model group, the "Huayu Tongluo" moxibustion group showed shortened escape latency (P<0.01) and increased number of platform quadrant crossing after 2 and 4 courses of intervention (P<0.01), significantly increased number of myelinated nerve axons (P<0.01), decreased g-ratio value of axonal myelin, and significantly elevated relative fluorescence intensities and protein expression levels of CNPase, Olig2, and MOG in all the different courses of intervention (P<0.01, P<0.05);and the non-acupoint moxibustion group had an increase in the immunofluorescence intensities of CNPase and MOG, and the expression of CNPase protein in the 4 courses of intervention (P<0.05). Comparison between the two treatment groups showed that the effect of "Huayu Tongluo" moxibustion was significantly superior to that of non-acupoint moxibustion in lowering the g-ratio of axonal myelin (P<0.05), and in up-regulating the number of myelinated nerve axons, and the immunofluorescence intensities and protein expressions of CNPase, Olig2 and MOG in the 3 different courses of intervention (P<0.01, P<0.05). "Huayu Tongluo" moxibustion can improve the learning-memory ability of VD rats, which may be related to its functions in promoting the proliferation and differentiation of oligodendrocyte precursor cells into mature oligodendrocytes, and up-regulating the protein expressions of CNPase, Olig2 and MOG in the corpus callosum. The therapeutic effect was the best in the 4 courses of intervention.

There is widespread of misinformation on futuristic technology and society.To accurately detect such news, the algorithms require up-todate knowledge.The Large Language Models excel in the NLP but cannot retrieve the ongoing events or innovations.For example, GPT and it's variants are restricted till the knowledge of 2021.We introduce a new methodology for the identification of fake news pertaining to futuristic technology and society.Leveraging the power of Google Knowledge, we enhance the capabilities of the GPT-3.5 language model, thereby elevating its performance in the detection of misinformation.The proposed framework exhibits superior efficacy compared to established baselines with the accuracy of 81.04%.Moreover, we propose a novel dataset consisting of fake news in three languages English, Telugu and Tenglish of around 21000 from various sources.