Nima Mesgarani

ASSOCIATE PROFESSOR OF ELECTRICAL ENGINEERING

CEPSR 910 (Schapiro)

Nima Mesgarani’s research is focused on exploration of human speech communication. He studies representational and computational characteristics of the human brain areas involved in naturalistic speech communication, and integrating these attributes into novel mathematical models used in machine emulation of speech communication. Better models of the underlying neural mechanisms involved in speech processing can critically impact research in artificial intelligence, neurolinguistics, systems neuroscience, translational medicine, and brain computer interfaces. 

Research Interests

Speech neurophysiology, deep neural network models, brain computer interfaces
Neurophysiology of Speech Communication
The process by which acoustic signals received by a listener are transformed into linguistic and non-linguistic categories is largely unknown. Using the latest advances in invasive and non-invasive human recording techniques, Mesgarani’s group aim to uncover fundamental characteristics of cortical speech processing that enables a listener to extract information from the speech signal. 
 
Neuro-inspired Computational Models
Mesgarani’s research in this area focuses on advancing and refining artificial neural network models used to emulate the cognitive abilities of humans. Furthermore, it is crucial to better understand the representation and transformation performed by these models to identify their limitations and compare them accurately with their biological counterparts, thereby reducing the performance gap between biological and artificial computing.
 
Brain Computer Interfaces
Better understanding of how speech communication occurs in healthy brains enables better therapeutic approaches to help those suffering from speech and language disorders, or those who suffer from peripheral and central auditory pathway disorders. Mesgarani’s research in this area includes better neural decoding algorithms and novel solutions for cognitively controlled hearable devices. 
 
Nima Mesgarani is an associate professor of Electrical Engineering at Columbia University. He received his Ph.D. from University of Maryland where he worked on neuromorphic speech technologies and neurophysiology of auditory cortex. He was a postdoctoral scholar in Center for Language and Speech Processing at Johns Hopkins University and the neurosurgery department of University of California San Francisco before joining Columbia in fall 2013. 

RESEARCH EXPERIENCE

  • 2017, Associate Professor, Department of Electrical Engineering, Columbia University, New York, NY
  • 2013, Assistant Professor, Department of Electrical Engineering, Columbia University, New York, NY
  • 2010-2013, Postdoctoral Scholar, Department of Neurological Surgery, University of California San Francisco, CA
  • 2008-2010, Postdoctoral Scholar, Center for Language and Speech Processing, Johns Hopkins University, MD

HONORS & AWARDS

  • 2016, National Science Foundation (NSF), Faculty Early Career Development Award
  • 2016, Collaborative and Multidisciplinary Pilot Research Award for Basic Science and Clinical/Translational Investigators (CaMPR BASIC), Irving Institute
  • 2015, Pew Charitable Trust, Pew Scholar in the Biomedical Sciences Award
  • 2015, Research Initiatives in Science and Engineering Award (RISE)
  • 2015, Kavli Institute for Brain Science Award          
  • 2005, George Harhalakis Outstanding Systems Engineering Graduate Student Award​

GRANT SUPPORT

  • NSF, NIH, Kavli Institute, Pew Charitable Trust, Starkey

SELECTED PUBLICATIONS

  • Khalighinejad, B.,  Cruzatto da Silva, G., Mesgarani, N.,  Dynamic Encoding of Acoustic Features in Neural Responses to Continuous Speech, (2017), Journal of Neuroscience
  • Chen, Z., Luo, Y., Mesgarani, N., Deep attractor network for single-microphone speech separation,  in Proc. IEEE Int. Conf. Acoust. Speech and Signal Process., 2017
  • Nagamine, T., Chen, Z., Mesgarani, N., (2016), Adaptation of neural networks constrained by prior statistics of node co-activations, In Sixteenth Annual Conference of the International Speech Communication Association, San Francisco,
  • Yang, M., Sheth, S. A., Schevon, C. A., II, G. M. M., & Mesgarani, N. (2015). Speech reconstruction from human auditory cortex with deep neural networks. In Sixteenth Annual Conference of the International Speech Communication Association, Dresden, Germany
  • Mesgarani, N., Cheung, C., Jonson, K., Chang, E. F., (2014), Phonetic feature encoding in human superior temporal gyrus, Science 1245994
  • Mesgarani, N., David, S. V., Fritz, J., Shamma, S., (2014), Mechanisms of noise robust Representation of Speech in Primary Auditory Cortex, Proceedings of the National Academy of Sciences (PNAS), 111.18
  • Mesgarani, N., Chang, E. F., (2012), “Selective cortical representation of attended speaker in multi-talker speech perception”, Nature 485