The present invention relates to a method based on concepts of nonlinear optics, self-focusing, and self-defocusing effects under third-order nonlinearity to study protein aggregation in real-time. For nonlinearity monitoring, a laser beam is used as the excitation source for the sample. The beam passing through the sample is detected by a photodiode detector at the far field. A processing unit is used for data acquisition and data analysis. For unfolded monomeric states of a protein, the sign of the nonlinear refractive index (n2) is consistently negative, whereas at the onset of aggregation, corresponding to the formation of early-stage oligomers, a critical transition occurs in which n2 changes sign from negative to positive. The proposed method is applicable throughout the entire aggregation time window to determine the exact protein conformation in real-time via a unique magnitude of n2. This method also identifies the toxic species by a sign shift in n2.  It is completely chemical-free, making it safer and environmentally friendly. Additionally, the simplicity and sensitivity allow for precise identification of disease markers without the need for complex equipment or extensive sample preparation.