arXiv subjects

Allison Smith

Publications and source records attributed to Allison Smith.

3 recordsLinked to original sources

arXiv2026

A foundation-model approach to pediatric headache classification from rs-fMRI

Headache is the most common neurological disorder in children and substantially affects quality of life. We investigated whether resting-state functional MRI (rs-fMRI) can support pediatric headache classification using machine learning. We encoded rs-fMRI data using NeuroSTORM, a recent foundation model, and fine-tuned it to distinguish healthy controls from children with headache and subsequently classify headache subtypes. We compared NeuroSTORM with a standard neuroscience approach using functional-connectivity (FC) matrices derived from brain activity as predictors. Using 189 rs-fMRI scans from 110 individuals collected across two visits (prevalence of any headache: 74%), NeuroSTORM achieved an area under the receiver operating characteristic curve (AUROC) of 0.82 (95% CI, 0.82-0.82) and an area under the precision-recall curve (AUPRC) of 0.93 (95% CI, 0.93-0.94) for discriminating headache from non-headache. In contrast, models trained on FC matrices showed lower performance (AUROC, 0.67 [95% CI, 0.67-0.67]; AUPRC, 0.85 [95% CI, 0.85-0.85]). In multiclass classification of healthy controls, chronic migraine, and non-chronic headaches (e.g., post-viral headache, new daily persistent headache, post-traumatic headache), NeuroSTORM achieved a macro-AUROC of 0.69 (95% CI, 0.68-0.69). Results suggest that the approach can distinguish chronic migraine but has difficulty differentiating other headache subtypes from chronic migraine. Overall, under limited-data conditions, NeuroSTORM appears to capture latent rs-fMRI representations that transfer to headache-related tasks without relying on FC features. These findings provide proof of concept for fMRI-based prediction of pediatric headache and highlight potential future utility for subtype identification and individualized treatment strategies.

cs.LG↗

arXiv2023

The Next Generation Arecibo Telescope: A preliminary study

The Next Generation Arecibo Telescope (NGAT) was a concept presented in a white paper Roshi et al. (2021) developed by members of the Arecibo staff and user community immediately after the collapse of the 305 m legacy telescope. A phased array of small parabolic antennas placed on a tiltable plate-like structure forms the basis of the NGAT concept. The phased array would function both as a transmitter and as a receiver. This envisioned state of the art instrument would offer capabilities for three research fields, viz. radio astronomy, planetary and space & atmospheric sciences. The proposed structure could be a single plate or a set of closely spaced segments, and in either case it would have an equivalent collecting area of a parabolic dish of size 300 m. In this study we investigate the feasibility of realizing the structure. Our analysis shows that, although a single structure ~300 m in size is achievable, a scientifically competitive instrument 130 to 175 m in size can be developed in a more cost effective manner. We then present an antenna configuration consisting of one hundred and two 13 m diameter dishes. The diameter of an equivalent collecting area single dish would be ~130 m, and the size of the structure would be ~146 m. The weight of the structure is estimated to be 4300 tons which would be 53% of the weight of the Green Bank Telescope. We refer to this configuration as NGAT-130. We present the performance of the NGAT-130 and show that it surpasses all other radar and single dish facilities. Finally, we briefly discuss its competitiveness for radio astronomy, planetary and space & atmospheric science applications.

astro-ph.IM↗

arXiv2023

A Search for OH 18-cm Emission from Intermediate-Velocity Gas at High Galactic Latitudes

We present search results of 22 high latitude (b > 25 deg.) sightlines for OH 18-cm emission using the 305-m radio telescope at the Arecibo Observatory. These sightlines appear in neutral hydrogen emission at intermediate velocities (V_lsr values ranging from -90 to -20 km/s) and are predicted to have a sufficient molecular composition so as to be detectable in molecular emission. Such objects, known as Intermediate-Velocity Molecular Clouds (IVMCs), have historically been detected through 12CO emission. Recent studies indicate that IVMCs may be widespread in the Galaxy and have important implications for models of the interstellar medium and star formation. However, we report non-detections of OH emission toward the 22 sightlines and provide stringent upper limits on the OH column density. Using available HI and Av data in combination with existing state-of-the-art PDR models, we estimate H2 column densities and find that they are more than an order of magnitude lower than the predicted values. We also find that the hydrogen volume density of these clouds is less than roughly 25 per cubic centimeter. In addition, we discuss the known IVMCs with previous 12CO detections in the context of the PDR models. Our analysis of these clouds indicates that the structure of molecular material in IVMCs is morphologically clumpy. These results motivate the need for future sensitive, on-the-fly searches (rather than targeted searches) for CO emission from IVMCs with of order roughly 1' resolution. High angular resolution (1') HI and Av data will also be helpful to better constrain the structure and composition of IVMCs.

astro-ph.GA↗