Boston College researchers developed a generative AI method that removes noise from fMRI brain scans, achieving over 200% improvement compared to prior techniques.
Key Details
- 1Boston College team created DeepCor, an AI-based denoising method for fMRI images.
- 2DeepCor improved performance by over 200% compared to previous methods such as CompCor.
- 3In real fMRI data, DeepCor achieved a 215% improvement in noise removal from face responses and 339% for clarifying synthetic data.
- 4The study was published in Nature Methods on November 28, 2025 (DOI: 10.1038/s41592-025-02967-x).
- 5Generative AI distinguishes between noise patterns in neural and non-neural brain regions.
- 6Accessible deployment and denoising large public datasets are targeted next steps.
Why It Matters
Effective removal of noise from fMRI data facilitates more accurate brain imaging, potentially accelerating neuroscience discoveries and clinical research. This improvement in image quality could set a new standard for preprocessing in neuroimaging and radiology AI workflows.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.
•EurekAlert
KAIST, MIT, Microsoft Develop Efficient AI Image Upsampling for Robotics
KAIST, MIT, and Microsoft have created 'Upsample Anything,' a training-free AI method to restore high-resolution visual data from compressed images with up to 16x improved GPU memory efficiency.