Artificial Intelligence-Enhanced Transcranial Doppler Ultrasound and Optic Nerve Sheath Diameter as Neuroimaging Biomarkers in Psychiatric and Neurological Disorders: A Systematic Review
preprint
OA: closed
Abstract
Artificial Intelligence (AI) is transforming neuroimaging by significantly enhancing diagnostic speed, accuracy, and reproducibility. Ultrasound-based modalities, particularly Transcranial Doppler (TCD) and Optic Nerve Sheath Diameter (ONSD), offer substantial advantages as portable, economical, and practical alternatives to traditional imaging techniques, especially in resource-limited or emergency settings. This systematic review critically evaluates the integration of AI with TCD and ONSD imaging for diagnosing and prognosticating neurological and psychiatric disorders, emphasizing diagnostic accuracy, methodological rigor, clinical applicability, and preparedness for clinical translation. A total of 70 studies published from 2010 to 2025 were identified from five major databases. CNNs, SVMs, and ensemble methods were the predominant AI techniques, demonstrating excellent diagnostic performance for neurological conditions such as stroke, traumatic brain injury, and intracranial hypertension, with mean area-under-the-curve (AUC) scores frequently exceeding 0.85. In contrast, psychiatric conditions, including depression and schizophrenia, showed moderate predictive accuracy (AUC ~0.70), likely due to the absence of specific ultrasound biomarkers. Crucially, methodological limitations were prevalent, as fewer than 20% of studies employed external validation, and fewer than 10% utilized explainable AI methods—significantly restricting clinical trust, reproducibility, and generalizability. Ethical transparency was also inconsistent, with nearly half of the reviewed studies failing to report ethical approval clearly, and limited availability of open data or code. This review underscores the critical need for standardized imaging protocols, robust external validation, enhanced AI interpretability, and multimodal approaches to facilitate effective and scalable clinical adoption of AI-enhanced ultrasound imaging. Keywords: Artificial Intelligence, Transcranial Doppler, Optic nerve sheath diameter, Traumatic brain injury
My notes (saved in your browser only)
Citation neighborhood (no data yet)
We don't have any in-corpus citations linked to this paper yet. This is a recent paper (2025) — citers typically take a year or two to land, and the OpenAlex reference graph may still be filling in.
Source provenance
- europepmc
- last seen: 2026-05-20T01:45:00.602351+00:00