A tendency discrimination device includes a discriminator to objectively discriminate a tendency of a person to be tested, based on brain information obtained by magnetic resonance imaging (MRI). In the generation of the discriminator, a gray matter volume and a diffusion anisotropy degree are calculated for a frontal pole of each of multiple test subjects as a region of interest, and machine learning is performed on the relationship of the information obtained by classifying results of a test for discriminating the tendencies of the multiple test subjects, to gray matter volumes and diffusion anisotropy degrees obtained by MRI for each of the multiple test subjects.
