Figure 1From: Novelty detection in long-term attentional habituation processes using a Bayesian change point algorithmThe following images show the likelihood of change points over all ALR sweeps arranged in matrix. The run-length likelihood is computed recursively by p ( r t , a t | x 1 : t ) = p ( r t , a t , x 1 : t ) ∑ r t ∑ a t p ( r t , a t , x 1 : t ) - 1 . The resulting likelihoods for having change point are plotted, (brighter corresponds to higher likelihoods). There is more variations in the likelihood of change points in 50dB throughout 80-120 ms across the sweeps. For 100dB the likelihood of change point remains the same through all sweeps indicating no habituation is occurring.Back to article page