DIFFEOMORPHIC REGISTRATION FOR RETINOTOPIC MAPPING VIA QUASICONFORMAL MAPPING

Area of Science:

• Neuroscience
• Vision Science
• Neuroimaging

Background:

• The human visual cortex contains distinct functional regions crucial for vision science research.
• Retinotopic mapping using functional magnetic resonance imaging (fMRI) is a key non-invasive technique for identifying these visual areas.
• fMRI-derived retinotopic maps often lack diffeomorphism due to low signal-noise, hindering precise boundary delineation.

Purpose of the Study:

• To develop a novel registration procedure for generating diffeomorphic retinotopic maps from fMRI data.
• To enhance the accuracy of retinotopic atlases by overcoming limitations of existing registration methods.
• To improve the identification and delineation of visual areas in the human brain.

Main Methods:

• A quasiconformal geometry-based registration model was developed, incorporating unique features of retinotopic mapping.
• The model was solved using efficient numerical methods.
• The proposed method was compared against conventional registration techniques on synthetic and real fMRI datasets.

Main Results:

• The developed registration method demonstrated superior performance compared to popular conventional methods on synthetic data.
• Application to real retinotopic mapping data resulted in significantly reduced registration errors.
• The method successfully generated more accurate and diffeomorphic retinotopic maps.

Conclusions:

• The proposed quasiconformal geometry-based registration method effectively improves the accuracy of retinotopic maps derived from fMRI.
• This advancement facilitates more precise delineation of visual areas in the human brain.
• The developed technique offers a valuable tool for neurophysiology and vision science research.