Author

Chenfei Ye

View on Semantic Scholar

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Open access Sep 2026

Unveiling the Distinctive Brain Functional Dynamics Between Parkinson's Disease and Progressive Supranuclear Palsy

ABSTRACT Background Dynamic analysis of resting‐state fMRI (rs‐fMRI) offers a novel approach to differentiate Parkinson's disease (PD) from progressive supranuclear palsy (PSP) by capturing temporal features of brain network activity, which may shed light on the mechanisms underlying non‐motor symptoms. Objectives To characterize differences in brain dynamics between PD and PSP using dynamic brain metrics, evaluate their exploratory discriminative performance, and investigate associations with non‐motor symptoms. Methods Sixty‐nine healthy controls, 82 PD patients, and 29 PSP patients underwent standardized clinical assessment and rs‐fMRI. Hidden Markov models extracted temporal features including fraction occurrence (FO), dwell time, and transition probability. These metrics were used for group comparisons, correlation analyses, and machine learning classification. Results PD and PSP showed opposite trends in fraction occurrence of unimodal network‐dominant states. Compared to PD and controls, PSP exhibited prolonged duration in the dorsal attention and limbic network (DAN&LIM) state and increased occurrence in the dorsal attention and frontoparietal control network (DAN&FPCN) state. Reduced unimodal state occupancy correlated with cognitive decline in both groups, while increased DAN&LIM and unimodal persistence linked to worse mood and sleep disturbances in PSP. Machine learning with these metrics achieved moderate accuracy in differentiating PD from PSP. Conclusions Temporal features of brain network dynamics may provide candidate imaging markers for distinguishing PD and PSP while offering mechanistic insights into non‐motor symptomatology. However, their diagnostic applicability requires validation in independent external multicenter cohorts.

Chen-Fei Ye, Jia-Hui Shi, Ji-Liang Fang et al. · CNS Neuroscience & Therapeut... · 0 citations

Preprint Jul 2026

UniMedSeg: Unified In-Context Learning for Multi-Paradigm 2D/3D Medical Image Segmentation

Medical image segmentation foundation models are expected to generalize across diverse clinical scenarios, yet existing universal methods remain fragmented by prompt paradigms and spatial dimensions. Visual in-context learning, interactive segmentation, and language-guided segmentation are typically handled by paradigm-specific models, while 2D and 3D images are also modeled separately. Such isolation prevents heterogeneous annotations and data from being jointly absorbed by a single scalable model and limits cross-paradigm knowledge transfer. To address this bottleneck, we propose UniMedSeg, a Transformer-centric universal segmentation framework that maps visual examples, geometric interactions, language instructions, and 2D/3D images into a shared sequence space, enabling heterogeneous medical supervision to be jointly learned through a unified in-context interface without prompt- or dimension-specific branches. To overcome the long-sequence memory bottleneck caused by visual contexts, we introduce Decoupled Split Attention, which reduces attention complexity to linear while preserving hardware-friendly computation and focused context-target interaction. Extensively trained and evaluated on a large corpus curated from 27 public datasets, UniMedSeg achieves state-of-the-art performance across visual in-context, interactive, and language-guided segmentation without task-specific fine-tuning, demonstrating strong generalization on diverse held-out tasks. The code and model weights are publicly available at https://github.com/Lii1228/UniMedSeg

Yunzhou Li, Jiesi Hu, Yanwu Yang et al. · 0 citations