Profiling oocytes with neural networks from images and mechanical data

Area of Science:

• Biophysics
• Reproductive Biology
• Materials Science

Background:

• Assisted reproductive technologies (ART) success relies on selecting viable oocytes.
• Current oocyte selection methods lack predictive accuracy.
• Mechanical profiling offers a novel approach to assess oocyte quality.

Purpose of the Study:

• To characterize the mechanical properties of the human oocyte's zona pellucida.
• To correlate mechanical properties with oocyte reproductive potential.
• To develop improved methods for oocyte selection in ART.

Main Methods:

• Indentation studies were performed on human oocytes.
• Mechanical properties of the zona pellucida were characterized using a physical model.
• Localized shear-thinning behavior was investigated.

Main Results:

• Excellent fitting was achieved with the physical model during indentation.
• Localized shear-thinning behavior of the zona pellucida was clearly illustrated.
• This shear-thinning behavior has not been previously reported.

Conclusions:

• The mechanical properties of the zona pellucida can be effectively characterized.
• Localized shear-thinning is a novel finding in oocyte mechanics.
• A methodology combining neural networks and optical imaging shows promise for cytoplasmic mechanical property isolation.