Meeting Banner
Abstract #0616

Disease-Specific Brain Function Representation Generation for Diagnosis Using Large Language Models

Mengjun Liu1, Lichi Zhang1,2, and Qian Wang3,4
1School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China, 2National Engineering Research Center of Advanced Magnetic Resonance Technologies for Diagnosis and Therapy (NERC-AMRT), Shanghai Jiao Tong University, Shanghai, China, 3School of Biomedical Engineering & State Key Laboratory of Advanced Medical Materials and Devices, ShanghaiTech University, Shanghai, China, 4Shanghai Clinical Research and Trial Center, Shanghai, China

Synopsis

Keywords: Diagnosis/Prediction, Brain

Motivation: Functional connectivity (FC) based on rs-fMRI is a classical brain-connectomic measure. However, FC is handcraft feature that insufficently describes complex brain activity and neglects disease specificity.

Goal(s): We aim to learn function representation directly from rs-fMRI and integrate disease-specific brain function abnormalities in the learning process assisted by large language model (LLM).

Approach: We employ an encoder-decoder architecture and introduce neuroscience knowledge from existing literature through an LLM to guide the generation of function representation. The resulting function representations are used for disease diagnosis to verify their effectiveness.

Results: The generated function representations improve the diagnostic performance, verifying the effectiveness of our method.

.

Impact: Our findings indicate that general and disease-specific brain function representations guided with LLM improve diagnostic accuracy. Additionally, the framework’s adaptability across different diseases positions it as a versatile tool in neuroimaging research, with potential applications in studying various disorders.

How to access this content:

For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.

After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.

After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.

Click here for more information on becoming a member.

Keywords