Robust presurgical functional MRI at 7 T using response consistency.

Area of Science:

• Neuroimaging
• Functional Magnetic Resonance Imaging (fMRI)
• Brain Mapping

Background:

• Functional MRI (fMRI) is crucial for non-invasive presurgical planning, offering repeatability and availability.
• Ultra-high field MRI enhances specificity and sensitivity, improving localization accuracy and reducing scan times.
• fMRI analysis ideally should handle patient task timing deviations and pathology-induced hemodynamic response changes.

Purpose of the Study:

• To evaluate a model-free fMRI analysis method, UNBIASED, for localizing the hand area using ultra-high field MRI.
• To compare the UNBIASED approach with the conventional General Linear Model (GLM) for fMRI data analysis in patients.

Main Methods:

• Applied the UNBIASED model-free analysis, based on fMRI response consistency across runs, to 7 Tesla fMRI data.
• Analyzed data from ten patients with brain tumors and epilepsy performing a hand motor task in a block design.
• Compared UNBIASED results against the General Linear Model (GLM) approach.

Main Results:

• UNBIASED successfully identified and excluded unreliable fMRI runs with minimal or no activation.
• UNBIASED analysis showed reduced motion artifact contamination compared to GLM.
• UNBIASED detected reproducible, time-locked activations missed by GLM, including delayed or transient responses.

Conclusions:

• UNBIASED provides a robust method for generating fMRI activation maps without assumptions on response timing or shape.
• This model-free approach can complement model-based methods in presurgical planning.
• UNBIASED aids surgeons in optimizing surgical access and resection margins, even with altered hemodynamic responses due to pathology.