According to the Felder and Silverman Learning Styles Model (FSM), students have learning preferences regarding how information is obtained, processed, perceived and understood. The Index of Learning Styles (ILS) is an online questionnaire created by Felder and Soloman to classify students according to their learning styles. With a priori knowledge of students' learning styles, one might hypothesize that the instructor could adapt his/her class to support, and even improve, students' learning. Still, one question that remains open is whether it is possible to individualize the FSM, i.e., if students of determined learning styles perform different (better or worse) on questions mapped to different styles. In this work, we assessed the correlation between students' performance and their learning styles in an Introductory Physics course. We designed a Learning Styles Classification Method (LSCM) and implemented it online (LSQuiz) to predict individual students' performance on pre-class questionnaires (N = 63). We found that, in general, the ability of the ILS to predict individual student performance on pre-class questionnaires was not better than random, with no significant correlation between students' performance and their learning styles, indicated by a Pearson's correlation coefficient of 0.54. Nevertheless, when independently analyzing the learning styles dimensions, we have identified heterogeneous data between the dimensions, with a greater correlation in the Sequential-Global dimension, with a 0.76 coefficient, followed by a coefficient of 0.50 in the Visual-Verbal and 0.35 in the Sensory Intuitive. We found, however, that the results related to the Sequential-Global dimension are not supported by the internal consistency of that dimension in the ILS (Cronbach's alpha of 0.30). We conclude suggesting that the adoption of customized learning practices in the Visual-Verbal dimension have potential and could be the focus of further studies.