Foundation models (FMs) are large-scale deeplearning models that are developed using large datasets and self-supervised learning methods. These models serve as a base for different downstream tasks, including healthcare. FMs have been adopted with great success across various domains within healthcare. Existing healthcare-based surveys have not yet included all of these domains. Therefore, we provide a detailed survey of FMs in healthcare. We focus on the history, learning strategies, flagship models, applications, and challenges of FMs. We explore how FMs such as the BERT and GPT families are reshaping various healthcare domains, including clinical large language models, medical image analysis, and omics. Furthermore, we provide a detailed taxonomy of healthcare applications facilitated by FMs, such as clinical NLP, medical computer vision, graph learning, and other biology-related tasks. Despite the promising opportunities FMs provide, they also have several associated challenges, which are explained in detail. We also outline open research issues and potential lessons learned to provide researchers and practitioners with insights into the capabilities of FMs in healthcare to advance their deployment and mitigate associated risks.

SUMMARY: Deep brain stimulation is an established therapy for multiple brain disorders, with rapidly expanding potential indications. Neuroimaging has advanced the field of deep brain stimulation through improvements in delineation of anatomy, and, more recently, application of brain connectomics. Older lesion-derived, localizationist theories of these conditions have evolved to newer, network-based “circuitopathies,” aided by the ability to directly assess these brain circuits in vivo through the use of advanced neuroimaging techniques, such as diffusion tractography and fMRI. In this review, we use a combination of ultra-high-field MR imaging and diffusion tractography to highlight relevant anatomy for the currently approved indications for deep brain stimulation in the United States: essential tremor, Parkinson disease, drug-resistant epilepsy, dystonia, and obsessive-compulsive disorder. We also review the literature regarding the use of fMRI and diffusion tractography in understanding the role of deep brain stimulation in these disorders, as well as their potential use in both surgical targeting and device programming.

Traditional targeting methods for thalamic deep brain stimulation (DBS) performed to address tremor have predominantly relied on indirect atlas-based methods that focus on the ventral intermediate nucleus despite known variability in thalamic functional anatomy. Improvements in preoperative targeting may help maximize outcomes and reduce thalamic DBS-related complications. In this study, we evaluated the ability of thalamic parcellation with structural connectivity-based segmentation (SCBS) to predict tremor improvement following thalamic DBS. In this retrospective analysis of 40 patients with essential tremor, hard segmentation of the thalamus was performed by using probabilistic tractography to assess structural connectivity to 7 cortical targets. The volume of tissue activated (VTA) was modeled in each patient on the basis of the DBS settings. The volume of overlap between the VTA and the 7 thalamic segments was determined and correlated with changes in preoperative and postoperative Fahn-Tolosa-Marin Tremor Rating Scale (TRS) scores by using multivariable linear regression models. A significant association was observed between greater VTA in the supplementary motor area (SMA) and premotor cortex (PMC) thalamic segment and greater improvement in TRS score when considering both the raw change (P = .001) and percentage change (P = .011). In contrast, no association was observed between change in TRS score and VTA in the primary motor cortex thalamic segment (P ≥ .19). Our data suggest that greater VTA in the thalamic SMA/PMC segment during thalamic DBS was associated with significant improvement in TRS score in patients with tremor. These findings support the potential role of thalamic SCBS as an independent predictor of tremor improvement in patients who receive thalamic DBS.

Introduction: Essential tremor (ET) is a common movement disorder with an estimated prevalence of 0.9% worldwide. Deep brain stimulation (DBS) is an established therapy for medication refractory and debilitating tremor. With the arrival of next generation technology, the implementation and delivery of DBS has been rapidly evolving. This review will highlight the current applications and constraints for DBS in ET. Areas covered: The mechanism of action, targets for neuromodulation, next generation guidance techniques, symptom specific applications, and long-term efficacy will be reviewed. Expert opinion: The posterior subthalamic area and zona incerta are alternative targets to thalamic DBS in ET. However, they may be associated with additional stimulation induced side effects. Novel stimulation paradigms and segmented electrodes provide innovative approaches to DBS programming and stimulation induced side effects.