Analysis of auditory middle latency responses at low and high stimulus rates during sleep

Date of Award




Degree Name

Doctor of Philosophy (Ph.D.)


Biomedical Engineering

First Committee Member

Ozcan Ozdamar, Committee Chair


Auditory evoked responses (AEPs) during sleep indicate the degree of activation and responsiveness of auditory information processing centers during various wakefulness levels. Although AEPs obtained at conventional slow rates during sleep have been described, detailed studies with continuous recordings are still missing. High rate recordings especially at 40 Hz are also suspected to provide more information on brain response generators and their generation mechanisms and processing during sleep. This work introduces new methods to generate, acquire, process and analyze auditory middle latency responses at high stimulation rates during sleep. Overlapped Auditory Evoked Potential (AEP) transient signals are deconvolved using the Continuous Loop Averaging Deconvolution (CLAD) method via frequency domain deconvolution of overlapped responses. The high stimulation rate is obtained using a stimulation sequence specially designed to reduce noise gain in the deconvolution process. Simultaneous acquisition of Auditory Brainstem Response (ABR) and Middle Latency Response (MLR) is obtained at an average stimulation rate of 5 and 40 Hz, using 10 and 100 second EEG recordings. Continuous AEPs were obtained during a short time afternoon sleep session that included awake, stages I-IV and REM sleep stages. To reduce MLR component variability and improve the peak component measurement, a new AEP filtering method was developed. The filtering method implements an optimized zero phase shift spectral high pass filtering combined with time-frequency block wavelet denoising. ABR and MLR peak components were detected and measured automatically using wavelet detail decompositions of the filtered evoked responses. AEP waveform analysis and comparisons between stimulation at 5 Hz and 40 Hz demonstrate no changes in ABR waves but clear and consistent differences in Na, Pa, Nb and Pb components, indicating different auditory response mechanisms. Deconvolved responses at 40 Hz also confirm previous observations on the reduction and disappearance of the Pb component during stage III and IV, obtained with standard averaging methods and stimulation schemes. Reductions on Pb amplitude during sleep showed a strong relation with the corresponding sleep stage. The developed method helps correlating the sleep EEG, indicative of different arousal and consciousness levels, with corresponding responses to auditory stimulation. Waveform differences in middle latency evoked responses at low and high rates need to be further explored.


Health Sciences, Audiology; Engineering, Biomedical; Psychology, Cognitive

Link to Full Text