Enhancing password-based user authentication: implicit motor sequence learning in the sisl task

The serial interception sequence learning (SISL) task is used in a user authentication system based on implicit motor sequence learning (Bojinov, Sanchez, Reber, Boneh & Lincoln, 2014). This study aims to investigate the influence of training length on performance of the SISL task. Participants were distributed across three conditions: a training phase of 480 trials, 960 trials and 1,440 trials. Experiment 1 comprised two phases: a training phase and a test phase. In the former, participants substantially trained on a fixed sequence, in the latter, sequence knowledge was tested, using both the trained sequence and new, random sequences. Results showed that participants performed significantly better on the trained sequence at both experiments and that training length did not significantly influence SISL task performance. From this, it can be concluded that a SISL-based authentication system could be successful in practice. This is supported by a performed security analysis. However, future research should explore the robustness of the SISL-based authentication system.