Langchen Fan, Advised by Prof. Laurel Carney

Wednesday, August 9, 2017
12:30 p.m.

Robert B. Goergen Hall, Room 101

People with hearing loss have more difficulty than people with normal hearing in a noisy environment. Therefore, it is important to understand the signal-in-noise detection mechanism. The Tone-in-noise (TIN) detection task is a simplified experimental paradigm designed to explore the mechanism. The task is for the subject to decide if there is a tone presented in the noise. There are diotic and dichotic test conditions. Diotic condition means to present identical tone and noise to the two ears, and dichotic condition means that phase difference is introduced between the two ears. The prevalent theory to explain human detection thresholds of diotic condition is an energy-based model. In the energy model, the auditory periphery is considered a bank of band-pass filters, and the existence of a tone will be determined when the output of the filter is higher, because adding a tone introduces more energy into the input waveform. This model explains many results, but fails when the energy cue is not reliable (i.e., overall sound level varies from trial to trial) – human listeners are less affected than the energy-based model predicts. Therefore, people must use another cue to detect the tone. Envelope, the outline of the waveform amplitude, was proposed to be this cue, but has not been studied in the midbrain. I will study how neurons in the inferior colliculus respond to the envelope change caused by adding a tone to the noise by measuring neural discharge rate and temporal discharge pattern (Aim 1). Out-of-phase tone embedded in identical noise (dichotic condition) is also used a lot in TIN detection. Previous modeling studies have focused on the binaural strategies to explain human dichotic thresholds, but cannot explain human behavior when using pre-selected noise samples. A recently proposed binaural model based on envelope – interaural envelope difference – was shown to explain more variance of human behavior. Therefore, I will study if there is a neural basis for the interaural envelope difference model (Aim 2). Human listeners have much lower detection thresholds for dichotic condition than diotic condition. With data collected in Aim 1 and Aim 2, I will also explore the neural basis of such detection differences (Aim 3).