Ph.D. Dissertation Defense: Peng Zan

Thursday, October 24, 2019
11:00 a.m.
AVW 2328
Maria Hoo
301 405 3681
mch@umd.edu

ANNOUNCEMENT:  Ph.D. Dissertation Defense


Name: Peng Zan

Committee:
Professor Jonathan Z. Simon, Chair/Advisor
Professor Steve Marcus
Professor Shihab Shamma
Professor Daniel A. Butts
Professor Samira Anderson, Dean’s Representative

Date/time: Thursday, Oct. 24, 2019 at 11:00am-2:00pm  

Location:  AVW 2328 

Title: Decoding Auditory Brain Responses with Mutual Information and the Effects of Aging 


Abstract:

The ability to segregate and understand speech in complex listening scenarios is an inherent property of the human brain. However, this ability deteriorates as the brain ages. The underlying age-associated alteration of neural mechanisms remains unclear. Understanding neural representations, in midbrain and cortex, of acoustics and speech along the auditory pathway holds the key to investigate the effects of aging on the auditory system. However, non-linearities in neural processing may conceal important internal mechanisms from standard linear methodology. This thesis develops a novel non-linear approach based on measures from information theory, and explores the non-linear representation of speech in both midbrain and cortex, which reveals a change in mechanism in auditory processing, with aging, along the auditory pathway from subcortical to cortical structures. First, midbrain activity was recorded from both younger and older listeners with normal hearing, using non-invasive electroencephalography. By estimating mutual information between stimulus and response both in amplitude and phase, we identify, for older listeners, a decrease in the earlier response of the midbrain despite there being an over-representation in the low-frequency response of the cortex. When processing speech in noise, the midbrain of older listeners extracts less information from speech than that of younger listeners, but it benefits when the background switches from meaningful, e.g., English speech, to meaningless, e.g. Dutch speech. Next, cortical activity was non-invasively recorded by magnetoencephalography. Results demonstrate that age-associated over-representation is widespread and is seen at various latencies. The specific response component with ~200 ms latency shows a relative information enhancement of attended over unattended speech, which depends on the signal to noise ratio (SNR), but in opposite directions for older listeners and younger listeners (respectively decreasing vs. increasing with worsening SNR). Critically, for the same late component in older listeners, the mutual information is larger in amplitude, and is negatively correlated with performance in a visual behavioral inhibition task. This negative correlation suggests that this late over-representation may be associated with a loss of behavioral inhibition control. Further, results show that this stronger response of late latency may play a critical role in determining speech intelligibility for older listeners. Finally, an analysis of cortical frequency-following response (FFR) on the same MEG-recorded responses shows robust cortical contribution in high-Gamma band (60-100 Hz). Contrary to the over-represented speech envelope in low-frequency (1-8 Hz) response, older listeners’ cortex loses its over-representation in cortical FFR compared to younger listeners. We also demonstrate that for poor SNR, both younger and older listeners are unable to sustain the foreground representation at a stable information level. However, younger listeners are adequate to maintain background FFR to a stable level with strengthening background sound level while older listeners fail to. This may suggest a reduced ability to segregate sources using auditory selective attention, with aging. Overall, this dissertation expands current understanding of auditory processes of speech in noise in both auditory subcortical and cortical systems, and the effects of aging. 

Audience: Graduate  Faculty 

 

November 2019

SU MO TU WE TH FR SA
27 28 29 30 31 1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
1 2 3 4 5 6 7
Submit an Event