Category: AI & Technology - Sci-Tech Bulletin

AI-Powered Digital Twins Enable Early Alzheimer’s Diagnosis with 88% Accuracy

Early diagnosis has been made possible by a digital twin model (DADD) that derives personalized Alzheimer’s disease (AD) biomarkers from non-invasive electroencephalographic (EEG) recordings. DADD model could identify early stage alzheimer’s with 79% accuracy while also identifying cerebrospinal fluid biomarkers with 88% accuracy and future cognitive decline with 87% accuracy using only non-invasive data. The DADD model enables scalable, low-cost screening and early intervention for preclinical AD such as subjective cognitive decline (SCD), reducing reliance on invasive tests with strong diagnostic and prognostic value.

AI-Powered Digital Twins Enable Early Alzheimer’s Diagnosis with 88% Accuracy Read More »

LEA, an AI model, can decode the microbial environments all the way from the gut to the ocean

An AI model, termed as LEA (Latent Environment Allocation) has been developed to evaluate the heterogeneity of microbial communities. This study is based on the idea that the microbial ecosystem reflects environments as continuous mixtures. LEA is a machine learning model that predicts the environmental composition of a new sample by placing it on a global microbiome map and comparing it with thousands of existing samples. The model can be applied for environmental monitoring, health diagnostics, and microbiome research, enabling classification of new samples, detection of contamination or dysbiosis, and semantic searching across large datasets without relying on rigid environmental labels.

LEA, an AI model, can decode the microbial environments all the way from the gut to the ocean Read More »

AIRVIC: The web-based AI Tool, automating viral detection in cell cultures

AIRVIC (AI Recognition of Viral CPE) is an AI-powered automated system specifically designed to detect and classify cytopathic effects caused by animal viruses like SARS-CoV-2, BAdV-1, BPIV3, BoAHV-1, and two strains of BoGHV-4 in Vero and MDBK cell lines. Generated using ResNet50-based convolutional neural network, AIRVIC achieved high detection performance with 100% accuracy for BoGHV-4 DN-599 in MDBK cells. AIRVIC demonstrates strong potential for viral isolation and antiviral efficacy testing through its accessible web-based platform, allowing global researchers to leverage its capabilities in viral diagnostics and beyond.

AIRVIC: The web-based AI Tool, automating viral detection in cell cultures Read More »

NanoLuc Biosensor Technology Offering a simple & rapid detection of Nipah virus infection

NanoLuc technology developed using a split NanoLuc biosensor, enables simple mix-and-read detection of Nipah virus antibodies. The biosensor is 98.6 sensitive and 100% specific for detecting IgG antibodies in Nipah virus infected patients. The NanoLuc Biosensor thus offers a fast, sensitive, and scalable approach for serological surveillance and retrospective analysis of the deadliest viral infections

NanoLuc Biosensor Technology Offering a simple & rapid detection of Nipah virus infection Read More »

Your Breath Could Reveal Cancer: How AI Is Changing Early Detection

A breath analyzer based on a hierarchical deep convolutional neural network (HD-CNN) framework can detect and classify cancer subtypes with accuracies of 95.1% and 98.1% for lung and gastric cancers, respectively. The sensor array within the device detects volatile organic compounds present in the breath of cancer patients and converts these signals into digital data for analysis.

Your Breath Could Reveal Cancer: How AI Is Changing Early Detection Read More »