Cross-frequency coincidence detection in the processing of complex sounds

Thesis (Ph.D.)--Boston UniversityPLEASE NOTE: Boston University Libraries did not receive an Authorization To Manage form for this thesis or dissertation. It is therefore not openly accessible, though it may be available by request. If you are the author or principal advisor of this work and would like to request open access for it, please contact us at open-help@bu.edu. Thank you.Responses of coincidence-detecting neurons are a direct function of the temporal structure of their input patterns. Quantitative studies of coincidence-detection provide insight into how neural processing of temporal information contributes to psychophysical performance. This study explored in detail the response properties of model coincidence-detection cells that receive inputs from auditory-nerve (AN) fibers. It also focused on the role of these model cells in coding of complex sounds related to psychophysical tasks for which temporal cues are believed to be important. Performance of model cells was evaluated quantitatively for different model parameters, including the width of the coincidence window, the number of input AN fibers, the characteristic frequencies (CFs) of the input AN fibers, and mixed strengths of the inputs. Results suggest that model cells with low CFs are very sensitive to the phase relationship of the input AN responses. The response properties of the model cells were also compared with results of physiological studies, and the coincidence-detection model predicts several response properties that were previously believed to be difficult to explain. Models for psychophysical detection and discrimination were designed based on population responses of model coincidence cells. Quantitative predictions of masked detection suggest that the most sensitive model cells for detection are the cells whose input AN responses are out of phase when a tone is added to the noise. The temporal structure in AN responses changes with signal-to-noise ratio and does not change as the overall level changes; thus, this model predicts psychophysical performance better than energy-based models under conditions in which the overall level of the stimulus varies randomly from trial to trial. The comparison of the coincidence-detection model and models based on other cues (e.g. envelope detector and channel theory) and implications for the theory of complex sound processing are also discussed.2031-01-0