Fused deep learning enables 6D single-molecule localization in polarization-resolved microscopy

Area of Science:

• Optical microscopy
• Biophysics
• Materials science

Background:

• Single-molecule orientation localization microscopy (SMOLM) is crucial for understanding molecular transport in complex environments.
• Current SMOLM techniques face challenges in achieving simultaneous high-resolution position and orientation measurements with good photon efficiency.

Purpose of the Study:

• To develop an advanced optical microscope for six-dimensional (6D) localization of dipole emitters.
• To integrate a deep learning approach for enhanced localization accuracy.

Main Methods:

• Utilized a double-helix point spread function in an optical fluorescence microscope.
• Developed a fused deep learning approach for 6D localization.
• Applied the system to localize rhodamine B molecules in polymer films and Nile red molecules in lipid bilayers.

Main Results:

• Achieved median spatial precision of 10 nm and angular precision below 10° for single fluorophore localization.
• Demonstrated successful 6D SMOLM of molecules in polymer films and lipid bilayers.
• Observed ordered molecular orientation in both demonstrated systems.

Conclusions:

• The developed microscope and deep learning method enable simultaneous 6D localization with high precision.
• This technique offers a powerful tool for investigating molecular behavior in challenging biological and material systems.
• The study overcomes limitations in existing SMOLM instrumentation.