Ensemble Spontaneous Activity Alterations Detected by CISA Approach

In this paper, we propose a method for detecting alterations in the Ensemble Spontaneous Activity (ESA), a random signal representing the composite spontaneous contribution of the auditory nerve recorded on the round window. The proposed method is based on shape analysis of the ESA amplitude histogram. For this task, we use a recent approach, the Corrected Integral Shape Averaging (CISA). Using this approach, a shape clustering algorithm is proposed to classify healthy and pathological ESA signals generated by a recent ESA model. This model allows a precise simulation of neural mechanisms occurring in the auditory nerve. The obtained results demonstrate that this shape analysis is very sensitive for detecting a small number of fibers with correlated firing, supposed to occur during a particular type of tinnitus. In comparison, the classical spectral index fails in this detection.