Michael Brown
Villanova University, PA·Psychology·Early-career
Spoken words rarely reach our ears cleanly. Background noise, different voices, different language backgrounds and varied pronunciations all blur the signal.
- Fields
Psychology
Neuroscience
Computer Science
- Topics
spoken language comprehension
speech perception
spoken word recognition
- Grants
- 1 known grant, through Nov 2026
What they work on
- Identify which sound information listeners use to tell spoken words apart by recording brain responses in the first few hundred milliseconds after hearing speech.
- Test how a listener's expectation of an upcoming word, built from sentence context, changes the brain's early responses to speech sounds.
- Build machine-learning neural network models that recognize spoken words in context, trained in a way that mimics how children learn language.
- Compare the neural network models' behavior with human brain data to test whether the models capture how people understand speech in different contexts.
More about the research
People still understand each other, largely by using context, such as the other words in a sentence, to work out what they heard. This project asks how listeners combine the raw sound of speech with expectations from context in real time. It also asks whether computer models can learn to do the same thing.
Why it matters
Understanding how people pull meaning from noisy, variable speech explains how humans communicate in real-world settings. The work may also show how speech-recognition software could become more reliable by processing language more like human listeners do. Examples include using context to resolve unclear words and handling different voices and pronunciations.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Susannah Levi
New York University, NY·Psychology
Test whether familiarizing listeners with a specific talker's voice makes words and phrases spoken by that talker easier to remember than those from unfamiliar talkers, even when the words are clearly understood.
Adeen Flinker
New York University, NY·Neuroscience
Map the integration periods, the time windows over which neurons combine sound, across primary and non-primary auditory cortex, using direct cortical recordings in neurosurgical patients.
David Harwath
University of Texas at Austin, TX·Computer Science·Early-career
Develop attention-based neural network segmentation layers that discover units in raw speech audio, applied hierarchically to recover structure at multiple levels such as phones, words and phrases.
The school
Villanova University: 4 labs on PI Pages, $33M a year on research, 32 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.