Untrained Convolutional Neural Networks Match Pretrained Models for MRI Feature Extraction
Researchers have shown that an untrained convolutional neural network (un-CNN) can extract features from structural MRI scans with predictive performance comparable to or exceeding state-of-the-art pretrained AI models. The un-CNN architecture incorporates multi-channel inputs, hierarchical encoding, and covariance pooling applied to three MRI datasets across three different downstream tasks. The finding challenges the assumption that large, computationally expensive pretrained models are necessary for neuroimaging analysis, with potential implications for reproducibility and data privacy.
A study posted to bioRxiv demonstrates that an untrained convolutional neural network, termed un-CNN, can serve as an effective feature extractor for structural MRI without any prior training on data. Across three structural MRI datasets and three downstream tasks, un-CNN matched or outperformed pretrained foundation models that are currently considered state of the art. The architecture extends a classical 3D CNN design with multi-channel inputs, a hierarchical encoder featuring multi-scale feature aggregation, and covariance pooling. Because the model requires no training, it avoids several practical drawbacks of large pretrained models, including high computational and memory demands, the logistical burden of distributing large model weights, and risks of inadvertent data leakage. The authors argue this approach also improves reproducibility, since results do not depend on opaque training procedures or proprietary datasets used to build foundation models.
What's missing
The study has not yet undergone peer review, as it is a preprint. Key limitations not addressed in the abstract include: whether performance advantages hold across a broader range of MRI modalities or clinical tasks beyond the three tested; the sensitivity of results to architectural hyperparameter choices in the untrained network; and whether the covariance pooling step introduces its own computational overhead that partially offsets the efficiency gains claimed.
What different sources said
- bioRxivCenter
Untrained Convolutional Neural Networks as Feature Extractors for Structural MRI
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.