The Bivariate Global Spectral Beta Test: A New Objective Response Detector Applied to the EEG During Photic Stimulation

The objective response detection (ORD) techniques such as Global Spectral F-Test (GSFT) are mathematical methods used to detect potentials evoked to rhythmic stimulation that are embedded in the ongoing electroencephalogram (EEG). The ORDs are based on the detector sampling distribution under the null hypothesis of no response to assess whether a given response is embedded in the electroencephalogram background (EEG). The performance of these detectors is strongly affected on both the signal-to-noise ratio (SNR) and the length of the electroencephalogram (EEG) signal. In this context, multivariate detectors are developed to increase the performance of the detection technique without increasing data recording time by using the information from more signals simultaneously. This work proposes three new bivariate detectors based on the normalization of the Global Spectral F-Test. The critical values ​​and detection probabilities for the new techniques were obtained theoretically and using a Monte Carlo simulation. The bivariate new detectors were next applied to the EEG from 10 subjects during intermittent, photic stimulation leading to an increase of up to 32.69% in mean detection rate in comparison with the univariate Global Spectral Beta-Test. The higher detection rate obtained with the proposed techniques without the need of increasing the number of data segments allows evoked responses to be detected faster, which might be useful in clinical practice.