Camille Noufi on detecting traumatic brain injury from speech
Date:
Fri, 03/15/2019 - 10:30am - 12:00pm
Location:
CCRMA Seminar Room
Event Type:
Hearing Seminar Camille Noufi will be talking about a similar coordination test using the voice. Producing speech requires the coordinated action of many different muscle groups, all tied together with perception and feedback to the motor system. When something goes wrong, it will show up in the voice, where perhaps it can be measured, better than anywhere else. Camille, and her colleagues from Lincoln Labs in Massachusetts, have been studying the vocal patterns produced by children, after an accident and as they heal. They investigate several kinds of vocal measures and evaluate their performance.
Who: Camille Noufi (CCRMA and Lincoln Labs)
What: Vocal Biomarker Assessment Following Severe Pediatric Traumatic Brain Injury: A Longitudinal Prospective Cohort Study
When: Friday, March 15th at 10:30AM
Where: CCRMA Seminar Room, Top Floor of the Knoll at Stanford
Why: Voices, brains, and perception are all tied together
Bring your voice and brain to CCRMA on Friday, and we’ll exercise both at the Hearing Seminar. (We do encourage discussion!)
- Malcolm
Title: Vocal Biomarker Assessment Following Severe Pediatric Traumatic Brain Injury: A Longitudinal Prospective Cohort Study
Abstract: A system of objective instrumental measurement is essential in sensitively, robustly and non-invasively assessing speech changes over time following pediatric traumatic brain injury (TBI). In this paper, we discuss a method for measuring changes in the speech patterns of a cohort of ten pediatric subjects diagnosed with a severe TBI. We apply a diverse set of both well-known and novel speech feature measurements to audio recordings taken monthly over the course of one year following diagnosis. We summarize the longitudinal changes in these features via an acoustic feature trend profile, which allows us to clearly track how different characteristics of speech subsystems are changing in relation to each other. We use the profile to summarize changes occurring to each individual as well as to the entire cohort. Across older children, we find statistically significant (p ≤ 0.01) increases in speaking rate, pitch variation and number of unique phonemes spoken, while we see significant reduction in variability of consonant duration. Across subjects younger children, we find similar reduction in variability of consonant duration and additionally see an increase in articulatory coordination.
Bio: Camille Noufi is a PhD Student at the Center of Computer Research in Music and Acoustics (CCRMA) at Stanford University. She is pursuing the application of expressive voice and musical engagement to assistive and interactive technologies. Her research links fields such as digital signal processing and machine learning to vocal production, perception and cognition. Previously, Camille conducted speech and audio signal processing research at MIT Lincoln Laboratory. She holds a degree in Electrical & Computer Engineering (BSEE-cum laude) and pursued a vocal performance degree at the University of Colorado-Boulder.
FREE
Open to the Public