To detect neural activities using a low-field MRI, we focused on spin-lock techniques, which are potentially effective tools for low-field functional MRI (fMRI). First, we estimated the magnetic field strength generated from neural activities using an equivalent current dipole model. Subsequently, we compared two spin-lock techniques by simulation and phantom experiments. It was found that the lowest detectable field strength was several tens pT in the simulation, whereas the experimental one was 2.34 nT. In future work, we plan to improve SNR using noise-reduction processing to realize the low-field fMRI.