Fuzzy logic based modeling of auditory brainstem responses to spoken syllable /da/

The complexities of processing and effects of individual factors on the encoding of sounds is a feature of brainstem processing. No statistical method, except fuzzy logic modeling, is effective in explaining these complex relationships. The aim of the current study was to explore the behavior of the brainstem in response to complex auditory stimuli using a fuzzy based model. Auditory brainstem responses to synthetic /da/ stimuli were recorded for 35 normal subjects (17 female) with a mean age of 22.7 years. A fuzzy based model with a Singleton fuzzifier, Mamdani product inference engine, and a center average defuzzifier were used. By providing recorded responses to create a rule base, the behavior of the brainstem was modeled. It appears that, using the training data, the current fuzzy model robustly anticipated and modeled the behavior of speech-evoked responses in the test data set. A mean error of 16.84% (variance: 1%) was observed for 5 randomly chosen training and testing data sets. Auditory brainstem responses to complex sounds (here; spoken syllable /da/) can be modeled using a fuzzy rule base. By correctly extracting these rules from the responses of subjects with normal hearing, a model was established that could anticipate the signals.