A preliminary investigation of the deconvolution of auditory evoked potentials using a session jittering paradigm

In the high stimulus rate studies of auditory evoked potentials (AEPs), deconvolution techniques have been developed to unwrap the overlapped responses based on the jittering strategy of stimulus onset asynchronies (SOAs). This study investigates an alternative deconvolution method (multi-rate steady-state averaging deconvolution, MSAD) using a session-jittering strategy where steady-state responses recorded at different SOAs can be adequate to derive the transient-AEP.

A linear transform model was developed to solve the deconvolution problem, and the mathematical properties of the transform matrix were explored by singular value decomposition, which indicates the need for regularization techniques to solve the ill-conditioning of the matrix.

The performance evaluated by both synthetic and experimental data is satisfactory compared with the classic SOA-jittering method commonly known as the continuous loop averaging deconvolution.

Our initial investigations suggest that the MSAD method is promising in terms of SOA insensitivity, sequence robustness and recording flexibility. However, more evaluation is needed to make the method suitable for a general application of the high stimulus rate paradigm.