Kalman filter based estimation of auditory steady state response parameters

Auditory steady state responses (ASSRs) are brain responses to modulated or repetitive stimuli that can be captured in the EEG recording. ASSRs can be used as an objective measure to clinically determine frequency specific hearing thresholds, to quantify the sensitivity of the auditory system to modulation, and have been related to speech intelligibility. However, the detection of ASSRs is difficult due to the low signal to noise ratio of the responses. Moreover, minimizing measurement time is important for clinical applications. Traditionally ASSRs are analyzed using discrete Fourier transform (DFT) based methods. We present a Kalman filter based ASSR analysis procedure and illustrate several benefits over traditional DFT based methods. We show on a data set of 320 measurements that the proposed method reaches valid amplitude estimates significantly faster than the state of the art DFT method. Further, we provide two extensions to the proposed method. First, we demonstrate information can be incorporated from multiple recording electrodes by extending the system model. Secondly, we extend the model to incorporate artifacts from cochlear implant (CI) stimulation and demonstrate electrically evoked auditory steady state responses (EASSRs) can be accurately measured.