Novelty detection in long-term attentional habituation processes using a Bayesian change point algorithm

Habituation is defined as a response decrement resulting from repeated stimulation [1]. This process is very important in filtering a large amount of information in the surrounding environment and focusing our attention on only the features of the environment that are important to us. Here we address the problem of objectively assessing long-term habituation using electroencephalographic auditory late responses (ALRs). We propose using a Bayesian change point algorithm to measure the novelty process and apply it to two experimental data sets of 50dB and 100dB stimuli responses. These correspond to cases of habituation and non-habituation respectively.

We use unprocessed raw data along with a denoised variant of the same data for the analysis. The nonlocal means algorithm is used for 2 diemensional denoising of ALR single sweeps, arranged in matrix, see [2] for detials. Data consists of 900 artefact free sweeps of 410 samples. We analyzed the data at all time instances over all sweeps. This procedure is applied to the unprocessed and processed data set. The Bayesian change point detection algorithm estimates a probability distribution over a run-length parameter. The run-length parameter has two states. It either increases by one time-step, or resets to zero indicating a change point. We calculate an average run-length variation value for all signals, based on changes in the maximum likelihood run-length. This provides a metric for comparing habituation and non-habituation processes.