The method of time-frequency analysis are the ways that mainly are used to analyze electroencephalography (EEG) signals. In our previous studies, we proved that watching 2D and 3D of a movie causes different effects in EEG signals. In this paper, we used dataset of previous study and mean of five EEG frequency bands as features. To obtain these frequency bands, a frequency analysis and two time-frequency analysis are used. Aim of this paper is that which one of these methods can be provide better features. These methods are Fast Fourier Transform (FFT), Discrete Wavelet Transform (DWT) and Wigner-Ville Distribution (WVD). To compare these methods Partial Least Squares Regression (PLSR) as a powerful classifier was implemented. The results show that PLSR could be classified Wigner-Ville Distribution's features better than other method's features.