Advancing Thalamic Nuclei Segmentation: The Impact of Compressed Sensing on MRI Processing

Area of Science:

• Neuroimaging
• Radiology
• Structural MRI

Background:

• Thalamic nuclei characterization is vital for neurological disorders like Parkinson's disease and schizophrenia.
• Severe head motion in patients challenges in vivo mapping of thalamic nuclei using MRI.
• Compressed sensing (CS) and fast segmentation tools accelerate MRI acquisition and processing.

Purpose of the Study:

• To evaluate thalamic nuclei segmentation using various CS-accelerated MPRAGE sequences.
• To assess the impact of CS acceleration on segmentation robustness and variability.
• To compare segmentation results initialized with FastSurfer and FreeSurfer.

Main Methods:

• Six MPRAGE variants with varying CS acceleration (1-9 min) were tested.
• Thalamic nuclei segmentation was initialized using FastSurfer and FreeSurfer.
• Volume variability and sequence effects were analyzed for whole thalamus and major nuclei.

Main Results:

• Minimal sequence effects and no systematic bias were observed across CS-accelerated sequences.
• CS-accelerated sequences showed lower volume variability than non-CS sequences.
• FastSurfer and FreeSurfer initializations yielded highly comparable thalamic nuclei segmentations.

Conclusions:

• High-quality thalamic nuclei segmentation is achievable with CS T1-weighted image acceleration on a clinical 3T MRI.
• CS-accelerated sequences and rapid segmentation methods are suitable for in vivo thalamic nuclei studies.
• These advancements benefit research in both healthy individuals and clinical populations with neurological conditions.