All India Institute of Medical Sciences, Bhopal

Social Anxiety Disorder (SAD) is a widespread mental health condition, yet its lack of objective markers hinders timely detection and intervention. While previous research has focused on behavioral and non-verbal markers of SAD in structured activities (e.g., speeches or interviews), these settings fail to replicate in real-world, unstructured social interactions. Identifying non-verbal markers in naturalistic, unstaged environments is essential for developing ubiquitous and non-intrusive monitoring solutions. To address this gap, we present AnxietyFaceTrack, a study leveraging facial video analysis to detect state anxiety in unstaged social settings. A cohort of 91 participants engaged in a social setting with unfamiliar individuals, and their facial videos were recorded using a low-cost smartphone camera. We examined facial features, including eye movements, head position, facial landmarks, and facial action units, and used self-reported survey data to establish ground truth for multiclass (anxious, neutral, non-anxious) and binary (e.g., anxious versus neutral) classifications. Our results demonstrate that a Random Forest classifier trained on the top 20% of features achieved the highest accuracy of 91.0% for multiclass classification and an average accuracy of 92.33% across binary classifications. Notably, head position and facial landmarks yielded the best performance for individual facial regions, achieving 85.0% and 88.0% accuracy, respectively, in multiclass classification, and 89.66% and 91.0% accuracy, respectively, across binary classifications. Post-hoc analysis identified head rotation, facial edge features, and eye landmarks as key contributors to the detection of anxiety. This study introduces a non-intrusive, cost-effective solution that can be seamlessly integrated into everyday smartphones for continuous anxiety monitoring, offering a promising pathway for early detection and intervention.

The genus Burkholderia comprises, clinically significant Burkholderia pseudomallei and the Burkholderia cepacia complex which are causative for melioidosis. Both are Gram-negative aerobic bacilli that resides in either water or soil and cause opportunistic infection. Head and neck involvement is uncommon and often mimics tuberculosis or pyogenic infections, resulting in delayed diagnosis and inappropriate treatment. Here we have performed a descriptive observational study among confirmed cases of Burkholderia of the head and neck. Diagnosis was established by microbiological analysis with Gram stain and culture sensitivity of pus. A Computerized tomography scan of the neck was performed as per the signs and symptoms of localized infection. The cases under study included four patients of Bukholderia-pseudomallei and one with Burkholderia-cepacia. All patients were initially suspected to have common bacterial or tuberculosis infections, and two had failed empirical penicillin therapy before referral. They were all compared on various clinical/presenting parameters, and risk factors. All cases were successfully treated with antibiotics with or without a surgical procedure, demonstrating the significance of early diagnosis, with identification of the etiological agent by culture sensitivity and early management. Diagnosis of Burkholderia is challenging as presentation mimics other chronic illnesses such as tuberculosis, along with delay in laboratory confirmation, or underdiagnosis. Involvement of the head and neck is rare and scarcely reported in existing literature. Hence, this study is an attempt to expand the limited literature on head and neck Burkholderia infections from Central India and emphasizes that localized melioidosis may occur even in immunocompetent individuals. Persistent cervical or facial abscesses that fail conventional antibiotic therapy, particularly in tuberculosis endemic region, should prompt microbiological evaluation for Burkholderia species, so as to avoid treatment failure and to reduce morbidity and mortality of melioidosis.

DNA sequencing has revolutionized biological and biomedical research, offering profound insights into genome organization, function, and variability. From the pioneering Sanger capillary electrophoresis method to the advent of next-generation sequencing, the field has evolved toward unprecedented speed, scalability, and cost decreases over the years. These advancements have enabled diverse applications across genomics, transcriptomics, metagenomics, epigenomics, and precision medicine, powering global initiatives such as the Human Genome Project, the Human Microbiome Project, and the 1000 Genomes Project. Bioinformatics has also advanced in data processing, variant detection, and functional annotation, helping transform raw sequencing data into biologically meaningful insights and knowledge. Although highly advanced, sequencing technologies still encounter challenges, including accuracy trade-offs and the need for efficient management of rapidly increasing volumes of data. Leveraging the genomic revolution, this review explores the shifts toward next-generation phenomics (NGP), an archetype that uses artificial intelligence that integrates multi-omics data with digital phenotyping, the Internet of Things, and real-time analytics. The goal of NGP is to integrate genotypic and phenotypic data to support predictive modeling of health, disease, and environmental interactions. By tracing history, advances in sequencing technologies, and future perspectives on NGP, this article offers a comprehensive overview for researchers and clinicians, highlighting how the integration of omics and digital data will drive the generation of personalized and systems-level biology.

Background Blunt abdominal trauma remains to be a significant factor of surgical admissions in emergencies and causes significant morbidity and mortality. Factors like rapid urbanization, higher numbers of vehicles on roads, and occupation-related injuries have resulted in increased incidences of injuries, especially in developing nations. Knowledge of patterns of injury in a locality is crucial in improving the trauma care system. Objectives To assess the clinical outcomes, therapeutic strategies, organ involvement, mechanisms of injury, and demographic profile of patients who report with blunt abdominal trauma at a tertiary care facility. Methods An observational investigation was carried out on 70 individuals with blunt abdominal trauma at a tertiary care centre. Information related to demographics, mechanism of injury, presence of extra-abdominal injuries, involved organs, treatment methods, and mortality rates was gathered and analyzed in a descriptive manner. Results The sample size for this study comprised 70 individuals. 87.1% of the study sample consisted of males, thus resulting in a sex ratio of 6.8:1. Patients aged between 20 to 29 years comprised the largest age group (40%). Road traffic accidents emerged as the leading cause of injury (50%), with falls from height coming in second (41.4%). Surgery was done on 74.3% of the total patients whereas conservative treatment was applied in 25.7% of the cases. Extra-abdominal injuries were found in 22.9% of the total number of individuals, with chest and pelvic injuries being the most common. Splenic injuries turned out to be the commonest visceral injuries (20%), followed by jejunal (13.7%) and hepatic injuries (11.8%). Conclusion Most blunt abdominal injuries occur in younger male patients and are commonly due to road traffic accidents. Injuries to the spleen were the most frequent abdominal organ injuries identified. Prompt diagnosis, resuscitation, treatment by a multidisciplinary team, and prevention through public health measures remain vital.

This study explores a comparative analysis of base and ensemble classifiers for detecting breast cancer. The classifiers analyze data from fine needle aspiration biopsies to categorize samples as benign or malignant. Artificial intelligence (AI), specifically eXplainable AI (XAI), plays a vital role in this study by providing interpretability to the classification results. Initially, classifier models are built and optimized to ensure accurate predictions. Their performance is then evaluated using metrics like F1 score and recall to identify the most effective model. The best-performing classifier is further analyzed using SHapley Additive exPlanations (SHAP), an XAI method that highlights the significance of each feature and its contribution to the model's decisions. This approach enhances understanding of the classifier's reasoning, which is crucial for medical applications. Finally, a web application is developed to present classifier performance metrics and illustrate how decisions were made, improving transparency and usability.